JP2008196506A - Engine knock control device and engine ignition timing control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device for avoiding knock in the way of prohibiting the repetition of delaying an ignition timing to avoid knock, followed by advancing operation. <P>SOLUTION: The knock control device comprises a knock detecting means 47 for detecting whether knock is actually generated or not in a combustion chamber 5, a knock generation timing detecting means 47 for detecting a knock generation timing in the combustion chamber, a comparing means 31 for comparing a detected value for the knock generation timing with a predicted value for the knock generation timing, a knock correlation parameter estimated value calculating means 31 for calculating an estimated value for a knock correlation parameter in accordance with the comparison result and the knock detection result, a knock generation timing predicted value calculating means 31 for calculating a predicted value for the knock generation timing in accordance with the estimated value for the knock correlation parameter, a knock limit ignition timing calculating means 31 for calculating a knock limit ignition timing in accordance with the predicted value for the knock generation timing, and ignition executing means 31, 14 for performing spark ignition at the knock limit ignition timing. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an engine (internal combustion engine) knock control device and an engine ignition timing control device.

Paying attention to the fact that there are two fuels on the market, high octane fuel (octane number 98) and low octane fuel (octane number 91), when the fuel cap is opened, first set the base ignition timing for high octane fuel. If knocking occurs, it is determined that the fuel is a high-octane fuel, and if knocking does not occur, it is determined that the fuel is a low-octane fuel. When it is determined that the fuel is used, the operation is continued as it is, and when it is determined that the fuel is a low octane fuel, the operation is switched to the base ignition timing for the low octane fuel (see Patent Document 1).
Japanese Patent Laid-Open No. 5-280454

  By the way, in the overseas market, there are markets where various octane number fuels are used and the octane number is not known in advance. In the case where the technology disclosed in Patent Document 1 is applied to the fuel sold in such a market and the base ignition timing for the low octane fuel is set, the local octane number of the fuel is determined for the low octane fuel. Knock occurs when the octane number of the fuel used for matching the base ignition timing is smaller.

  In this case, in the technique of Patent Document 1 as well, the knock control by the knock sensor is performed as in the prior art. Therefore, when the knock sensor detects that the knock has occurred, the base ignition timing is stepped by the first predetermined value. After that, the base ignition timing is gradually advanced by the second predetermined value, and then the knock sensor detects that the knock has occurred again by the advance of the ignition timing in this operation. Then, the above operation is repeated.

  Thus, even if knocking occurs when the octane number of the local fuel is smaller than the octane number of the fuel used for matching the base ignition timing for the low-octane fuel, according to the technique of Patent Document 1, in order to avoid knocking The ignition timing delay and the subsequent advance operation are repeated, and knocking can be avoided by this operation. However, fuel consumption deteriorates and output decreases due to the ignition timing delay for avoiding knocking. In order to avoid such fuel consumption deterioration and output decrease, a base ignition timing calculation map is prepared for each of a plurality of different octane numbers from the maximum octane number to the minimum octane number. The ROM capacity for storing the map for calculating the base ignition timing becomes large.

  The above octane number is a parameter correlated with knock when gasoline is used as fuel, and in the case of a mixed fuel of gasoline and alcohol, the alcohol concentration in the mixed fuel is a knock correlation parameter. There is a market in the overseas market that the alcohol concentration in the mixed fuel is not known in advance. The technology of Patent Document 1 is applied to the mixed fuel sold in such a market as it is, and the alcohol concentration in the mixed fuel is high. When the base ignition timing for mixed fuel is set, and the alcohol concentration of the mixed fuel in the field is higher than the alcohol concentration of the mixed fuel used for matching the base ignition timing for the mixed fuel with high alcohol concentration Knocks. On the other hand, if the ignition timing delay operation and the subsequent advance operation to avoid knocking by the knock sensor are executed, the fuel consumption deteriorates and the output decreases. In order to avoid this, a map for calculating the base ignition timing is used. Is provided for each of a plurality of different alcohol concentrations from the lowest alcohol concentration to the highest alcohol concentration, the ROM capacity for storing the base ignition timing calculation map for each alcohol concentration increases.

  On the other hand, the compression ratio is also a knock correlation parameter. When using a fuel whose octane number is determined in advance, the compression ratio is determined by the engine specification, so the base ignition timing is matched so that knock does not occur at the compression ratio determined by the engine specification. . However, knocking occurs when the actual compression ratio becomes higher than the compression ratio of the engine specification for some reason. At this time, if the delay of the ignition timing for avoiding the knock and the subsequent advance operation are repeated based on the knock sensor, the fuel consumption deteriorates and the output decreases.

  Therefore, the present invention feeds back the knock detection result by the knock sensor to the knock correlation parameter instead of the ignition timing and converges the knock correlation parameter at an early stage, thereby retarding the ignition timing for avoiding knock as in the conventional device. It is an object of the present invention to provide a device that avoids knocking by a method that does not repeat the following and advance angle operations.

  The present invention detects whether or not knock has actually occurred in the combustion chamber, detects the knock generation timing in the combustion chamber, and detects the knock generation timing detection value (θknkreal) and the knock generation timing prediction value (θknkest). And calculating an estimated value of the knock correlation parameter based on the comparison result and the knock detection result, calculating the knock occurrence timing predicted value (θknest) based on the knock correlation parameter estimated value, and generating the knock A knock limit ignition timing (KNOCKcal) is calculated based on the predicted timing value (θknest), and spark ignition is performed at the knock limit ignition timing.

  The present invention also detects whether or not knock has actually occurred in the combustion chamber, detects the knock intensity in the combustion chamber, and compares this knock intensity detection value (KICREAL) with the knock intensity estimation value (KICET). Then, an estimated value of the knock correlation parameter is calculated based on the comparison result and the knock detection result, and a predicted knock generation timing (θknest) in the combustion chamber is calculated based on the knock correlation parameter estimated value. A knock limit ignition timing (KNOCKcal) is calculated based on the predicted generation timing value (θknquest), and spark ignition is performed at this knock limit ignition timing.

  Further, the present invention detects whether or not knock is actually generated in the combustion chamber, detects the knock generation timing in the combustion chamber, detects the knock intensity in the combustion chamber, detects the knock detection result, the knock generation timing A compression ratio estimated value (CMPEST2) is calculated based on the detected value (θknkreal) and the knock intensity detected value (KICREAL), and a knock correlation other than the compression ratio is calculated based on the knock detection result and the knock occurrence timing detected value (θknkreal). An estimated value of the parameter is calculated, a knock generation timing prediction value (θknkest) in the combustion chamber is calculated based on the knock correlation parameter, and a knock limit ignition timing (KNOCKcal) is calculated based on the knock generation timing prediction value (θknkest). And the spark ignition is performed at the knock limit ignition timing.

  The present invention also calculates a laminar combustion velocity (SL1, SL2) that is a combustion velocity in a laminar flow state of the combustion gas, calculates a volume corresponding to the combustion gas volume in the combustion chamber, and reaches a predetermined crank angle. The combustion mass ratio (BR1, BR2) of the gas combusted in the combustion chamber is calculated, the reaction probability (RPROBA) indicating the ease of combustion of the combustion gas under a predetermined operating condition is calculated, and these laminar combustion speeds ( SL1, SL2), a combustion gas volume equivalent volume, a combustion mass ratio (BR1, BR2), and a reaction probability (RPROBA) to calculate a combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle, and this combustion period Based on (BURN1, BURN2), the basic ignition timing (MBTCAL) from which MBT is obtained is calculated, and the combustion period is set to the initial combustion period (BURN1) and the main combustion. The engine knock control device is divided into periods (BURN2), and the initial combustion period (BURN1) is calculated using the volume (V0) of the combustion start timing in the combustion chamber as the volume corresponding to the combustion gas volume in the combustion chamber. , It is detected whether or not knock has actually occurred in the combustion chamber, the knock generation timing in the combustion chamber is detected, and this knock generation timing detection value (θknkreal) is compared with the knock generation timing prediction value (θknkest). Then, an estimated value of the compression ratio (CMPEST), which is a knock correlation parameter, is calculated based on the comparison result and the knock detection result, and the volume of the combustion chamber at the combustion start timing is calculated based on the estimated compression ratio (CMPEST). It is configured to calculate (V0).

  The present invention also calculates a laminar combustion velocity (SL1, SL2) that is a combustion velocity in a laminar flow state of the combustion gas, calculates a volume corresponding to the combustion gas volume in the combustion chamber, and reaches a predetermined crank angle. The combustion mass ratio (BR1, BR2) of the gas combusted in the combustion chamber is calculated, the reaction probability (RPROBA) indicating the ease of combustion of the combustion gas under a predetermined operating condition is calculated, and these laminar combustion speeds ( SL1), a combustion gas volume equivalent volume, a combustion mass ratio (BR1, BR2), and a reaction probability (RPROBA), a combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated, and this combustion period (BURN1 , BURN2) to calculate the basic ignition timing (MBTCAL) at which MBT is obtained, and the combustion period to the initial combustion period (BURN1) and the main combustion period ( In the engine knock control device, the initial combustion period (BURN1) is calculated using the volume (V0) of the combustion start timing in the combustion chamber as the volume corresponding to the combustion gas volume in the combustion chamber. It is detected whether or not knock has actually occurred in the combustion chamber, the knock strength in the combustion chamber is detected, the knock strength estimated value (KICEST) is compared with the knock strength detected value (KICREAL), and the comparison result Then, an estimated value of the compression ratio (CMPEST), which is a knock correlation parameter, is calculated based on the knock detection result, and a volume (V0) of the combustion chamber at the combustion start timing is calculated based on the estimated compression ratio (CMPEST). To be configured.

  Further, the present invention detects whether or not knock has actually occurred in the combustion chamber, detects the knock generation timing in the combustion chamber, and detects the knock generation timing detection value (θknkreal) and the knock generation timing prediction value (θknest). And an estimated value (CMPEST) of the compression ratio is calculated based on the comparison result and the knock detection result, and the volume (V0) of the combustion chamber at the combustion start timing is calculated based on the estimated compression ratio (CMPEST). ) And a combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated based on the volume (V0) at the combustion start timing, and MBT is obtained based on the combustion period (BURN1, BURN2). The basic ignition timing (MBTCAL) is calculated and spark ignition is performed at this basic ignition timing (MBTCAL).

  The present invention also detects whether or not knock has actually occurred in the combustion chamber, detects the knock intensity in the combustion chamber, and compares this knock intensity detection value (KICREAL) with the knock intensity estimation value (KICET). Then, an estimated value (CMPEST) of the compression ratio is calculated based on the comparison result and the knock detection result, and the volume (V0) of the combustion chamber at the combustion start timing is calculated based on the estimated compression ratio (CMPEST). The combustion period (BURN1, BURN2) from the start of combustion to the predetermined crank angle is calculated based on the volume (V0) at the combustion start timing, and the basic ignition for obtaining MBT based on the combustion period (BURN1, BURN2). A timing (MBTCAL) is calculated, and the spark ignition is performed at the basic ignition timing (MBTCAL).

  According to the present invention, the estimated value of the knock correlation parameter is calculated based on the knock detection result (the knock detection result is fed back to the knock correlation parameter), and the predicted knock occurrence time is calculated based on the knock correlation parameter estimated value In addition, since the knock limit ignition timing is calculated based on the predicted knock generation timing, when using fuel sold in a market where the octane number of the fuel and the alcohol concentration in the mixed fuel are not known in advance, In addition, when using a fuel with a predetermined octane number or a mixed fuel with a predetermined alcohol concentration, if the actual compression ratio becomes higher than the compression ratio of the engine specifications for some reason, In order to avoid knocks like conventional devices that feed back the knock detection results to the ignition timing It is not be repeated and the operation of the retard and advance the subsequent fire season. For this reason, since the knock limit ignition timing can be followed even during a transition such as acceleration or deceleration, fuel consumption deterioration and output reduction can be prevented.

  Further, according to the present invention, based on the comparison result between the knock occurrence time detection value (θknkreal) and the knock occurrence time prediction value (θknkest) or between the knock intensity detection value (KICREAL) and the knock intensity estimation value (KICET). Since the knock correlation parameter estimated value is calculated based on the comparison result, the convergence of the knock correlation parameter estimated value can be accelerated, and the drivability is improved. Further, in order to increase the accuracy in detecting the knock occurrence time, it is necessary to shorten the sampling period. However, if the knock intensity is used, the sampling frequency of the knock detection means can be lowered. The system can be configured at low cost without degrading performance.

  Further, according to the present invention, it is detected whether or not knock is actually generated in the combustion chamber, the knock generation timing in the combustion chamber is detected, the knock generation timing detection value (θknkreal) and the knock generation timing prediction value ( θknest), a compression ratio estimated value (CMPEST) is calculated based on the comparison result and the knock detection result, and the volume of the combustion chamber at the combustion start timing is calculated based on the compression ratio estimated value (CMPEST). (V0) is calculated, a combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated based on the volume (V0) at the combustion start timing, and MBT is calculated based on the combustion period (BURN1, BURN2). The basic ignition timing (MBTCAL) obtained is calculated and spark ignition is performed at this basic ignition timing (MBTCAL), or combustion is performed. It is detected whether knock has actually occurred in the room, the knock intensity in the combustion chamber is detected, this knock intensity detection value (KICREAL) is compared with the knock intensity estimation value (KICEST), and the comparison result and An estimated value (CMPEST) of the compression ratio is calculated based on the knock detection result, a volume (V0) at the combustion start timing of the combustion chamber is calculated based on the estimated compression ratio (CMPEST), and this combustion start timing is calculated. The combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated based on the volume (V0) at, and the basic ignition timing (MBTCAL) at which MBT is obtained is calculated based on the combustion periods (BURN1, BURN2). However, since spark ignition is performed at this basic ignition timing (MBTCAL), the fuel or alcohol concentration with a predetermined octane number is determined. Even when the actual compression ratio becomes higher than the compression ratio of the engine specification for some reason in the case of using a mixed fuel with a fixed value, the basics for obtaining the MBT while accelerating the convergence of the estimated compression ratio (CMPEST) Ignition timing can be given accurately.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a schematic diagram for explaining the system of the present invention.

  The air is stored in the intake collector 2 and then introduced into the combustion chamber 5 of each cylinder via the intake manifold 3. Fuel (gasoline) is injected and supplied from a fuel injector 21 arranged in the intake port 4 of each cylinder. The fuel injected into the air is vaporized and mixed with the air to form a gas (air mixture) and flows into the combustion chamber 5. This air-fuel mixture is confined in the combustion chamber 5 when the intake valve 15 is closed, and is compressed by the rise of the piston 6.

  In order to ignite this compressed air-fuel mixture with a high-pressure spark, an ignition device 11 of an electronic power distribution system is provided in which an ignition coil with a built-in power transistor is arranged in each cylinder. That is, the ignition device 11 includes an ignition coil 13 that stores electrical energy from the battery, a power transistor that energizes and shuts off the primary side of the ignition coil 13, and a primary current of the ignition coil 13 that is provided on the ceiling of the combustion chamber 5. It includes a spark plug 14 that receives a high voltage generated on the secondary side of the ignition coil 13 due to interruption of the spark coil 13 and performs spark discharge.

  When a spark is blown off by the spark plug 14 slightly before the compression top dead center and the compressed mixture is ignited, the flame spreads and then explosively burns, and the gas pressure by this combustion works to push down the piston 6. This work is taken out as the rotational force of the crankshaft 7. The combusted gas (exhaust gas) is discharged into the exhaust passage 8 when the exhaust valve 16 is opened.

A three-way catalyst 9 is provided in the exhaust passage 8. When the air-fuel ratio of the exhaust gas is in a narrow range (window) centered on the stoichiometric air-fuel ratio, the three-way catalyst 9 can efficiently remove harmful three components such as HC, CO, and NOx contained in the exhaust gas simultaneously. Since the air-fuel ratio is the ratio of the intake air amount and the fuel amount, the intake air amount introduced into the combustion chamber 5 per one cycle of the engine (crank angle 720 ° section in a four-cycle engine) and the fuel injector 21 The engine controller 31 uses the intake air flow rate signal from the air flow meter 32 and the fuel from the fuel injector 21 based on the signals from the crank angle sensors (33, 34) so that the ratio to the fuel injection amount becomes the stoichiometric air-fuel ratio. The injection amount is determined, and the air-fuel ratio is feedback controlled based on a signal from an O ₂ sensor 35 provided upstream of the three-way catalyst 9.

  A so-called electronically controlled throttle 22 in which a throttle valve 23 is driven by a throttle motor 24 is provided upstream of the intake collector 2. Since the torque required by the driver appears in the amount of depression of the accelerator pedal 41 (accelerator opening), the engine controller 31 determines a target torque based on a signal from the accelerator sensor 42, and a target air for realizing this target torque. The amount is determined, and the opening degree of the throttle valve 23 is controlled via the throttle motor 24 so as to obtain this target air amount.

  Cam sprockets and crank sprockets are attached to the front portions of the intake valve camshaft 25, the exhaust valve camshaft 26, and the crankshaft 7, respectively, and a timing chain (not shown) is hung around these sprockets so that the camshaft 25 and 26 are driven by the crankshaft 7 of the engine, and are interposed between the cam sprocket and the intake valve camshaft 25 to continuously adjust the phase of the intake valve cam with a constant operating angle. A variable intake valve timing control mechanism (hereinafter referred to as “intake VTC mechanism”) 27 that can be controlled, and a cam sprocket and an exhaust valve camshaft 26 are interposed between the camshaft 26 and the exhaust valve cam. Variable exhaust valve timing control mechanism (hereinafter referred to as “exhaust”) that can control the phase continuously. Provided with that.) 28 and TC mechanism ". When the opening / closing timing of the intake valve 15 and the opening / closing timing of the exhaust valve 16 are changed, the amount of the inert gas remaining in the combustion chamber 5 changes. As the amount of the inert gas in the combustion chamber 5 increases, the pumping loss decreases and the fuel consumption improves. Therefore, the target intake valve closing timing and the target exhaust gas indicate how much inert gas should remain in the combustion chamber 5 depending on the operating conditions. The valve closing timing is determined in advance, and the engine controller 31 determines the target intake valve closing timing and the target exhaust valve closing timing from the operating conditions (engine load and rotation speed) at that time, so that these target values can be obtained. The intake valve closing timing and the exhaust valve closing timing are controlled via the actuators of the intake VTC mechanism 27 and the exhaust VTC mechanism 28.

  An intake air temperature signal from the intake air temperature sensor 43, an intake air pressure signal from the intake air pressure sensor 44, an exhaust gas temperature signal from the exhaust air temperature sensor 45, and an exhaust gas pressure signal from the exhaust air pressure sensor 46 are output from the water temperature sensor 37. The engine controller 31 that is input together with the coolant temperature signal controls the ignition timing that is the primary current cutoff timing of the spark plug 14 via the power transistor 13.

  In this case, the ignition timing when knock does not occur is the basic ignition timing MBTCAL according to the operating conditions. However, knocking may occur in the combustion chamber 5 in a high load low rotation speed region of the engine. If this occurs, the durability of the engine decreases, so the engine controller 31 performs knock control.

  In this case, in the general knock control, when the knock sensor detects that the knock has occurred, the base ignition timing is greatly retarded stepwise by the first predetermined value, and thereafter the base ignition timing is set to the second value. The operation is gradually advanced by a predetermined value, and the above operation is repeated when it is detected again by the knock sensor by the advance of the ignition timing in this operation. In this embodiment, the knock operation is repeated. The result of knock detection by the sensor 47 (knock detection means) is fed back to the estimated value OCTEST of the fuel octane number (knock correlation parameter) instead of the ignition timing, thereby retarding the ignition timing for avoiding knock as in the conventional device. And avoiding knocking in a way that does not repeat the following advance angle operation. That is, the estimated octane number OCTEST is calculated based on the knock detection result by the knock sensor 47, and the self-ignition timing θknk (knock occurrence time) in the combustion chamber 5 is predicted based on the estimated octane number OCTEST. Knock limit ignition timing KNOCKcal is calculated based on the timing θknk. When knocking occurs, the knock limit ignition timing KNOCKcal has a value retarded from the basic ignition timing MBTCAL, and spark ignition is performed using the knock limit ignition timing KNOCKcal as the ignition timing.

  FIG. 2 is a flowchart showing the overall flow of ignition timing control. This flow shows a flow of operations, not a flow executed at regular intervals.

  In steps 1 and 2, the basic ignition timing MBTcal [degBTDC] and the knock limit ignition timing KNOCKcal [degBTDC] are calculated.

  Here, the calculation of the basic ignition timing MBTcal will be described first. First, the ignition timing control based on the combustion analysis will be outlined (the basic concept is described in Japanese Patent Laid-Open No. 2003-148236).

  As shown in FIG. 3, the crank angle at which the combustion pressure of the air-fuel mixture reaches the maximum value Pmax when the air-fuel mixture is ignited with MBT (minimum advance angle value at which maximum torque is obtained) is defined as a reference crank angle θPMAX [degATDC]. The reference crank angle θPMAX is substantially constant regardless of the combustion method, and is generally in the range of 12 to 15 degrees after compression top dead center, and at most 10 to 20 degrees after compression top dead center.

  FIG. 4 shows changes in the combustion mass ratio BR (combustion gas mass ratio) obtained by the combustion analysis in the combustion chamber in the spark ignition engine. The combustion mass ratio BR representing the ratio of the combustion mass to the fuel supplied to the combustion chamber is 0% at the time of ignition, and reaches 100% by complete combustion. Experiments have confirmed that the combustion mass ratio at the reference crank angle θPMAX is constant and about 60%.

  The combustion period corresponding to the change allowance until the combustion mass ratio BR reaches about 60% corresponding to the reference crank angle θPMAX from 0% is a period in which there is almost no change in both the combustion mass ratio and the combustion pressure immediately after the start of combustion. It is divided into an initial combustion period and a main combustion period in which the combustion mass ratio and the combustion pressure increase rapidly. The initial combustion period is a stage from the start of combustion until flame nuclei are formed, and the flame nuclei are formed when the combustion mass ratio changes from 0% to 2% to 10%. During this initial combustion period, the rate of increase in combustion pressure and combustion temperature is small, and the initial combustion period is long with respect to changes in the combustion mass ratio. The length of the initial combustion period is susceptible to changes in temperature and pressure in the combustion chamber.

  On the other hand, in the main combustion period, the flame propagates from the flame kernel to the outside, and the flame speed (that is, the combustion speed) increases rapidly. Therefore, the change in the combustion mass ratio during the main combustion period is larger than the change in the combustion mass ratio during the initial combustion period.

  In the engine controller 31, the period until the combustion mass ratio reaches 2% (changes) is set as the initial combustion period BURN1 [deg], and the period from the end of the initial combustion period BURN1 to the reference crank angle θPMAX (combustion chamber volume ratio) In other words, the main combustion period BURN2 [deg] is distinguished from 2% to about 60%. Then, a combustion period BURN [deg] that is the sum of the initial combustion period BURN1 and the main combustion period BURN2 is calculated, a reference crank angle θPMAX [degATDC] is subtracted from the combustion period BURN, and an ignition dead time equivalent crank described later is further calculated. The crank angle position to which the angle IGNDEAD [deg] is added is set as the basic ignition timing MBTCAL [degBTDC], which is the ignition timing at which MBT is obtained.

  The pressure and temperature in the combustion chamber 5 during the initial combustion period in which flame nuclei are formed are almost equivalent to the pressure and temperature at the time of ignition, but the ignition timing is calculated from this, but accurate ignition is performed from the beginning. The time cannot be set. Therefore, as shown in FIG. 13, the previous value of the basic ignition timing is calculated as the previous combustion start timing MBTCYCL [degBTDC] (step 44), and this value is used to calculate the initial combustion period as shown in FIG. (Step 162), the calculation of the initial combustion period is cyclically repeated to obtain a highly accurate result without time delay.

  Next, the calculation of the basic ignition timing MBTCAL executed by the engine controller 31 will be described in detail with reference to the following flowchart.

  FIG. 5 is for calculating various physical quantities necessary for calculating the ignition timing, and is executed at regular time intervals (for example, every 10 msec).

  First, in step 11, the intake valve closing timing IVC [degBTDC], the collector internal temperature TCOL [K] detected by the temperature sensor 43, the collector internal pressure PCOL [Pa] detected by the pressure sensor 44, and the temperature sensor 45 are detected. Exhaust temperature TEXH [K], internal inert gas rate MRESFR [%], cooling water temperature TWK [K] detected by temperature sensor 37, target equivalent ratio TFBYA, engine rotational speed NRPM [rpm] detected by crank angle sensor ], Dead ignition time DEADTIME [μsec] is read.

  Here, the crank angle sensor includes a position sensor 33 for detecting the position of the crankshaft 7 and a phase sensor 34 for detecting the position of the intake camshaft 25. The engine is based on signals from the two sensors 33 and 34. The rotational speed NRPM [rpm] is calculated.

  The intake valve closing timing IVC is known from the command value given to the intake VTC mechanism 27. Alternatively, the actual intake valve closing timing may be detected by the phase sensor 34.

  The internal inert gas ratio MRESFR is a value obtained by dividing the amount of inert gas remaining in the combustion chamber by the total amount of gas in the combustion chamber, and the calculation thereof will be described later. The ignition dead time DEADTIME is a constant value.

  The target equivalent ratio TFBYA is calculated in a fuel injection amount calculation flow (not shown). The target equivalent ratio TFBYA is an unnamed number, and is a value represented by the following expression when the theoretical air-fuel ratio is 14.7.

TFBYA = 14.7 / target air-fuel ratio (1)
For example, from equation (1), when the target air-fuel ratio is the stoichiometric air-fuel ratio, TFBYA = 1.0, and when the target air-fuel ratio is a lean value such as 22.0, TFBYA is a positive value less than 1.0. is there.

In step 12, the volume of the combustion chamber 5 at the intake valve closing timing IVC (that is, the volume at the compression start timing) VIVC [m ³ ] is calculated. The volume VIVC of the combustion chamber 5 when the intake valve is closed is determined by the stroke position of the piston 6. The stroke position of the piston 6 is determined by the crank angle position of the engine.

  Referring to FIG. 6, consider a case where the rotation center 72 of the crankshaft 71 of the engine is offset from the center axis 73 of the cylinder. Assume that the connecting rod 74, the joint point 75 between the connecting rod 74 and the crankshaft 71, and the piston pin 76 that connects the connecting rod 74 and the piston are in the relationship shown in the figure. At this time, the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve can be expressed by the following equations (2) to (6).

VIVC = f1 (θivc) = Vc + (π / 4) D ² · Hivc (2)
Vc = (π / 4) D ² · Hx / (ε−1) (3)
^{Hivc = {(CND + ST 2} /2) - (CRoff-PISoff) 2} 1/2
− {(ST / 2) · cos (θivc + θoff)}
+ (CND ² −X ² ) ^1/2 (4)
X = (ST / 2) · sin (θivc + θoff) −CRoff + PISoff (5)
θoff = arcsin {(CRoff−PISoff) / (CND · (ST / 2))} (6)
Where Vc: gap volume [m ³ ],
ε: compression ratio,
D: cylinder bore diameter [m],
ST: Full piston stroke [m],
Hivc: Piston pin 76 at the intake valve closing timing
Distance from TDC [m],
Hx: difference between the maximum value and the minimum value of the distance from the TDC of the piston pin 76 [m],
CND: length of connecting rod 74 [m],
CRoff: Offset distance of the nodal point 75 from the cylinder center axis 73
[M],
PISoff: offset distance [m] of the crankshaft rotation center 72 from the cylinder center axis 73,
θivc: Intake valve closing timing crank angle [degATDC],
θoff: an angle [deg] between a line connecting the piston pin 76 and the crankshaft rotation center 72 and a vertical line in TDC,
X: horizontal distance [m] between the nodal point 75 and the piston pin 76,
As described above, the crank angle θivc at the time of closing the intake valve is known because it is determined by the command signal from the engine controller 31 to the intake VTC mechanism 27. By substituting the crank angle θivc (= IVC) at this time into the equations (2) to (6), the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve can be calculated. Therefore, in practice, the volume VIVC of the combustion chamber 5 at the intake valve closing timing is set by a table using the intake valve closing timing IVC as a parameter. When the intake VTC mechanism 27 is not provided, a constant value can be given.

  In step 13, the temperature (that is, the compression start timing temperature) TINI [K] of the combustion chamber 5 at the intake valve closing timing IVC is calculated. The temperature of the gas flowing into the combustion chamber 5 is a temperature of a gas in which the fresh air flowing into the combustion chamber 5 and the inert gas remaining in the combustion chamber 5 are mixed. The temperature of the fresh air flowing into the combustion chamber 5 is Since the temperature of the inert gas equal to the fresh air temperature TCOL in the intake collector 2 and remaining in the combustion chamber 5 can be approximated by the exhaust temperature TEXH in the vicinity of the exhaust port, the temperature of the combustion chamber 5 at the intake valve closing timing IVC. TINI is the following from the fresh air temperature TCOL in the intake collector 2, the exhaust gas temperature TEXH, and the internal inert gas ratio MRESFR that is the ratio of the inert gas remaining in the combustion chamber 5 at the timing when the intake valve closing timing IVC is reached. It can be obtained by an expression.

TINI = TEXH × MRESFR + TCOL × (1−MRESFR) (7)
In step 14, the pressure (that is, compression start timing pressure) PINI [Pa] at the intake valve closing timing IVC of the combustion chamber 5 is calculated. That is, the collector internal pressure PCOL at the timing when the intake valve closing timing IVC is reached is taken in as the pressure PINI at the intake valve closing timing IVC.

  In step 15, a reaction probability RPROBA [%] representing the flammability of the air-fuel mixture in the combustion chamber 5 is calculated. The reaction probability RPROBA is a dimensionless value and depends on the three parameters of the residual inert gas ratio MRESFR, the cooling water temperature TWK [K], and the target equivalent ratio TFBYA, and can be expressed by the following equation.

RPROBA = f3 (MRESFR, TWK, TFBYA) (8)
More specifically, the maximum value of the reaction probability obtained by the combination of the three parameters MRESFR, TWK, and TFBYA is set to 100%, the relationship between these parameters and the reaction probability RPROBA is experimentally obtained, and the obtained reaction probability RPROBA is obtained. Are stored in advance in the memory of the engine controller 31 as a table corresponding to the parameters. In step 15, the reaction probability RPROBA is obtained by searching this table according to the parameters.

  Specifically, a table of water temperature correction coefficients having characteristics as shown in FIG. 7 according to the cooling water temperature TWK, a table of internal inert gas rate correction coefficients (not shown) set similarly, and a target equivalent ratio A table of equivalence ratio correction coefficients having characteristics as shown in FIG. 8 according to TFBYA is stored in the memory in advance. The maximum value of each correction coefficient is 1.0, and the reaction probability RPROBA is calculated by multiplying the product of the three types of correction coefficients by the maximum value of 100% of the reaction probability.

  Explaining each table, the water temperature correction coefficient shown in FIG. 7 becomes larger as the cooling water temperature TWK is higher, and becomes 1.0 when the cooling water temperature TWK is 80 ° C. or higher. The equivalence ratio correction coefficient shown in FIG. 8 is the maximum value of 1.0 when the target equivalence ratio TFBYA is 1.0, that is, the stoichiometric air-fuel ratio. The ratio correction factor decreases. Although the internal inert gas rate correction coefficient is not shown, it is 1.0 when the internal inert gas rate MRESFR is zero.

  In step 16, a reference crank angle θPMAX [degATDC] is calculated. As described above, the reference crank angle θPMAX does not fluctuate very much, but it still tends to advance as the engine speed NRPM increases. Therefore, the reference crank angle θPMAX can be expressed by the following equation as a function of the engine speed NRPM. it can.

θPMAX = f4 (NRPM) (9)
Specifically, the reference crank angle θPMAX is obtained by searching a table of characteristics shown in FIG. 9 stored in advance in the memory of the engine controller 31 from the engine speed NRPM. In order to facilitate calculation, the reference crank angle θPMAX can be regarded as constant.

  Finally, in step 17, the ignition dead time equivalent crank angle IGNDEAD [deg] is calculated. The ignition dead time equivalent crank angle IGNDEAD is a crank angle section from the timing at which the engine controller 31 outputs a signal for cutting off the primary current of the ignition coil 13 until the ignition plug 14 actually ignites, and can be expressed by the following equation. .

IGNDEAD = f5 (DEADTIME, NRPM) (10)
Here, the ignition dead time DEADTIME is set to 200 μsec. Equation (10) is for calculating the ignition dead time equivalent crank angle IGNDEAD that is the crank angle corresponding to the ignition dead time DEADTIME from the engine speed NRPM.

  FIG. 10 is for calculating the initial combustion period BURN1 [deg], and FIG. 12 is for calculating the main combustion period BURN2 [deg], which is executed at regular intervals (for example, every 10 msec). 10 and 12 are executed following FIG. Either of FIGS. 10 and 12 may be executed first.

First, referring to FIG. 10, in step 161, the previous combustion start timing MBTCYCL [degBTDC], the volume VIVC [m ³ ] at the intake valve closing timing of the combustion chamber 5 calculated in step 12 of FIG. 5, and the step of FIG. The temperature TINI [K] at the intake valve closing timing of the combustion chamber 5 calculated in 13, the pressure PINI [Pa] at the intake valve closing timing of the combustion chamber 5 calculated in Step 14 of FIG. 5, and the engine speed NRPM [Rpm], the reaction probability RPROBA [%] calculated in step 15 of FIG. 5 is read.

  Here, the previous combustion start timing MBTCYCL is a value one cycle before [degBTDC] of the basic ignition timing MBTCAL, and the calculation thereof will be described later with reference to FIG.

In step 162, the volume V0 [m ³ ] at the combustion start timing of the combustion chamber 5 is calculated. As described above, the ignition timing (combustion start timing) here is not the basic ignition timing MBTCAL calculated in the current cycle but a value one cycle before the basic ignition timing. That is, the volume V0 at the combustion start timing of the combustion chamber 5 is calculated from MBTCYCL, which is a value one cycle before the basic ignition timing, by the following equation.

V0 = f6 (MBTCYCL) (11)
Specifically, the volume V0 of MBTCYCL in the combustion chamber 5 is calculated from the stroke position of the piston 6 at the previous combustion start timing MBTCYCL and the bore diameter of the combustion chamber 5. In step 12 of FIG. 5, the volume VIVC of the combustion chamber 5 at the intake valve closing timing IVC is obtained by searching a table of intake valve closing timing volumes using the intake valve closing timing as a parameter. Here, MBTCYCL is set as a parameter. The volume V0 of the combustion chamber 5 at the previous combustion start time MBTCYCL may be obtained by searching the table of the previous combustion start time volume.

  In step 163, an effective compression ratio Ec at the combustion start timing is calculated. The effective compression ratio Ec is a dimensionless value, and is a value obtained by dividing the volume V0 of the combustion chamber 5 at the combustion start timing by the volume VIVC of the combustion chamber 5 at the intake valve closing timing, as shown in the following equation.

Ec = f7 (V0, VIVC)
= V0 / VIVC (12)
In step 164, the temperature increase rate TCOMP in the combustion chamber 5 from the intake valve closing timing IVC to the combustion start timing is calculated based on the effective compression ratio Ec as shown in the following equation.

TCOMP = f8 (Ec) = Ec ^ (κ−1) (13)
Where κ: specific heat ratio,
Equation (13) is an equation for the rate of temperature rise of the adiabatic compressed gas. Note that “^” on the right side of the equation (13) represents power calculation. This symbol is also used in the formula described later.

  κ is a value obtained by dividing the constant pressure specific heat of the gas adiabatically compressed by the constant volume specific heat. If the gas adiabatically compressed is air, κ = 1.4, and this value may be used simply. However, the calculation accuracy can be further improved by experimentally determining the value of κ for the air-fuel mixture.

  FIG. 11 illustrates equation (13). Therefore, it is possible to obtain the temperature increase rate TCOMP by storing a table of such characteristics in advance in the memory of the engine controller 31 and searching the table based on the effective compression ratio Ec.

In step 165, the temperature T0 [K] at the combustion start timing of the combustion chamber 5 is multiplied by the temperature TINI at the intake valve closing timing of the combustion chamber 5 by the temperature increase rate TCOMP, that is, T0 = TINI × TCOMP (14)
It is calculated by the following formula.

  Steps 166 and 167 are the same as steps 164 and 165. That is, at step 166, the pressure increase rate PCOMP in the combustion chamber 5 from the intake valve closing timing IVC to the combustion start timing is calculated based on the effective compression ratio Ec as shown in the following equation.

PCOMP = f9 (Ec) = Ec ^ κ (41)
Where κ: specific heat ratio,
The equation (41) is also an equation for the rate of increase in pressure of gas that is adiabatically compressed in the same manner as the equation (13). “^” On the right side of the equation (41) represents power calculation as in the equation (13).

  κ is the same as the value used in the above equation (13). If the gas to be adiabatically compressed is air, κ = 1.4, and this value may be used simply. However, the calculation accuracy can be further improved by obtaining the value of κ from the composition and temperature of the air-fuel mixture.

  It is also possible to obtain a pressure increase rate PCOMP by storing a table having the same characteristics as in FIG. 11 in advance in the memory of the engine controller 31 and searching the table based on the effective compression ratio Ec.

In step 167, the pressure P0 [Pa] at the combustion start timing of the combustion chamber 5 is multiplied by the pressure PINI at the intake valve closing timing of the combustion chamber 5 by the pressure increase rate PCOMP, that is, P0 = PINI × PCOMP (42)
It is calculated by the following formula.

  In step 168, the laminar combustion speed SL1 [m / sec] in the initial combustion period is calculated by the following equation (known).

SL1 = f10 (T0, P0)
= SLstd × (T0 / Tstd) ^2.18 × (P0 / Pstd) ^−0.16 (15)
Where Tstd: reference temperature [K],
Pstd: Reference pressure [Pa]
SLstd: Reference laminar combustion at reference temperature Tstd and reference pressure Pstd
Speed [m / sec],
T0: temperature [K] at the combustion start timing of the combustion chamber 5;
P0: pressure [Pa] at the combustion start timing of the combustion chamber 5;
The laminar combustion velocity (laminar flame velocity) is the flame propagation velocity in the absence of gas flow, and is independent of the compression velocity in the combustion chamber 5 and the intake air flow velocity in the combustion chamber 5. 5 is known to be a function of the temperature and pressure of 5. Therefore, the laminar combustion speed in the initial combustion period is a function of the combustion start temperature T0 and the combustion start pressure P0, and as described later, the main combustion period. Is a function of the compression top dead center temperature TTDC and the compression top dead center pressure PTDC. This is because the laminar combustion speed generally varies depending on the engine load, the inert gas ratio in the combustion chamber 5, the intake valve closing timing, the specific heat ratio, and the intake air temperature. Since it is a factor that affects T and pressure P, it is assumed that the laminar combustion speed can be finally defined by the temperature T and pressure P in the combustion chamber 5.

  In the above equation (15), the reference temperature Tstd, the reference pressure Pstd, and the reference laminar combustion speed SLstd are values determined in advance by experiments.

Under a pressure of 2 bar or more, which is a normal pressure in the combustion chamber 5, the pressure term (P0 / Pstd) ^−0.16 in the equation (15) becomes a small value. Accordingly, it is also possible to define the reference laminar combustion speed SLstd only by the reference temperature Tstd with the pressure term (P0 / Pstd) ^−0.16 being a constant value.

  Accordingly, when the reference temperature Tstd is 550 [K], the reference laminar combustion rate SLstd is 1.0 [m / sec], and the pressure term is 0.7, the temperature T0 and the laminar combustion rate at the combustion start timing The relationship with SL1 can be approximately defined by the following equation.

SL1 = f11 (T0)
= 1.0 × 0.7 × (T0 / 550) ^2.18 (16)
In step 169, the gas flow turbulence intensity ST1 in the initial combustion period is calculated. The turbulence intensity ST1 of the gas flow is a dimensionless value, and depends on the flow rate of fresh air flowing into the combustion chamber 5 and the penetration of injected fuel from the fuel injector 21.

  The flow rate of fresh air flowing into the combustion chamber 5 depends on the shape of the intake passage, the operating state of the intake valve 15, and the shape of the intake port 4 where the intake valve 15 is provided. The penetration of the injected fuel depends on the injection pressure of the fuel injector 21, the fuel injection period, and the combustion injection timing.

  Finally, the turbulence strength ST1 of the gas flow during the initial combustion period can be expressed by the following equation as a function of the engine speed NRPM.

ST1 = f12 (NRPM) = C1 × NRPM (17)
Where C1: constant,
It is also possible to obtain the turbulence intensity ST1 from a table using the rotational speed NRPM as a parameter.

  In step 170, the gas combustion speed FLAME1 [m / sec] in the initial combustion period is calculated from the laminar combustion speed S1 and the turbulence intensity ST1 by the following equation.

FLAME1 = SL1 × ST1 (18)
If there is gas turbulence in the combustion chamber 5, the gas combustion speed changes. Equation (18) takes into account the contribution (influence) to the combustion speed associated with this gas turbulence.

  In step 171, the initial combustion period BURN1 [deg] is calculated by the following equation.

BURN1 = {(NRPM × 6) × BR1 × V0}
/ (RPROBA × AF1 × FLAME1) (19)
However, AF1: Reaction area (fixed value) of flame kernel [m ² ],
The equation (19) and the equation (22) to be described later are derived from the following basic equation that the combustion period is obtained by dividing the mass of the combustion gas by the combustion speed. The equations (19) and (22) The numerator and denominator on the right-hand side do not immediately indicate the mass of combustion gas or the burning rate.

Combustion period [sec] = total mass in cylinder [g] / (unburned gas density [g / m ³ ]
× Flame surface area [m ² ] × Flame speed [m / sec])
... (Supplement 1)
Since the unburned gas density in the right side denominator of (Supplement 1) is a value obtained by dividing the unburned gas mass [g] by the unburned gas volume [m ³ ], the charging efficiency ITAC corresponding to the mass as in the conventional device. It cannot be said that the unburned gas density can be calculated accurately with the function of only. Therefore, experimental formulas shown in the above formula (19) and formula (22) to be described later are obtained for the first time by introducing a predetermined approximation to the formula (complement 1) while checking the experimental results.

  Here, BR1 on the right side of the equation (19) is a change amount of the combustion mass ratio from the combustion start timing to the end timing of the initial combustion period BURN1, and here BR1 = 2% is set. (NRPM × 6) on the right side of the equation (19) is a process for converting the unit from rpm to crank angle (deg). The reaction area AF1 of the flame kernel is set experimentally.

  Further, it can be assumed that the combustion chamber volume does not change during the initial combustion period. Therefore, when calculating the initial combustion period BURN1, the combustion chamber volume V0 at the start of combustion, which is the first combustion chamber volume, is employed.

Next, the flow of FIG. 12 is followed. In step 181, similarly to step 161 of FIG. 10, the volume VIVC [m ³ ] at the intake valve closing timing of the combustion chamber 5 calculated in step 12 of FIG. The temperature TINI [K] at the intake valve closing timing of the combustion chamber 5 calculated in step 13 of FIG. 5, the pressure PINI [Pa] at the intake valve closing timing of the combustion chamber 5 calculated in step 14 of FIG. The speed NRPM [rpm], the reaction probability RPROBA [%] calculated in step 15 of FIG. 5 is read, and the cylinder fresh air amount MACYL [g], the target equivalent ratio TFBYA, the internal inert gas amount MRES [g], The external inert gas amount MEGR [g] is read.

  Here, although an external EGR device is not shown in FIG. 1, as far as FIG. 12 is concerned, an explanation will be given on the premise of an engine equipped with an external EGR device. In this case, the external inert gas amount MEGR may be calculated using, for example, a known method (see Japanese Patent Laid-Open No. 10-141150). It should be noted that when an engine that does not include an external EGR device as in the present embodiment shown in FIG. 1 is used, it is sufficient to handle the external inert gas amount MEGR = 0. The calculation of the cylinder fresh air amount MACYL and the internal inert gas amount MRES will be described later with reference to FIG.

Steps 182 and 183 are the same as steps 163 and 164 in FIG. That is, in step 182, the effective compression ratio Ec at the compression top dead center time. 2 is calculated. Effective compression ratio Ec 2 is also a dimensionless value like the effective compression ratio Ec of the above equation (12), and the volume VTDC at the time of compression top dead center of the combustion chamber 5 is calculated at the closing timing of the intake valve of the combustion chamber 5 as shown in the following equation. The value divided by the volume VIVC.

Ec 2 = f13 (VTDC, VIVC) = VTDC / VIVC
... (43)
In equation (43), the volume VTDC at the time of compression top dead center of the combustion chamber 5 is constant regardless of the operating conditions, and may be stored in the memory of the engine controller 31 in advance.

In step 183, the temperature increase rate TCOMP due to adiabatic compression in the combustion chamber 5 during the period from the intake valve closing timing IVC to the compression top dead center. 2 is an effective compression ratio Ec as shown in the following equation: 2 is calculated.

TCOMP 2 = f14 (Ec 2)
= Ec 2 ^ (κ-1) (44)
Where κ: specific heat ratio,
A table having the same characteristics as in FIG. 11 is stored in advance in the memory of the engine controller 31 and the effective compression ratio Ec. By searching the table from 2, the temperature rise rate TCOMP 2 can also be obtained.

  In step 184, the total gas mass MGAS [g] in the combustion chamber 5 is calculated from the cylinder fresh air amount MACYL, the target equivalent ratio TFBYA, the internal inert gas amount MRES, and the external inert gas amount MEGR by the following equation.

MGAS = MACYL × (1 + TFBYA / 14.7) + MRES + MEGR
... (45)
“1” in parentheses on the right side of the equation (45) is a fresh air, and “TFBYA / 14.7” is a fuel.

  In step 185, the total gas mass MGAS of the combustion chamber 5, the cylinder fresh air amount MACYL, and the target equivalent ratio TFBYA are used to calculate the temperature increase amount (combustion increase temperature) TBURN [K] due to the combustion of the air-fuel mixture using the following equation. .

TBURN = {MACYL × TFBYA / 14.7 × BRk × Q}
/ (Cv × MGAS) (46)
Where Q is the constant calorific value of the fuel,
BRk: Combustion mass ratio of fuel in cylinder,
Cv: constant volume specific heat,
The numerator on the right side of the equation (46) means the total heat generated by the fuel in the cylinder [J], and the denominator means the temperature increase rate per unit generated heat [J / K]. That is, the equation (46) is an approximate equation applied to the thermodynamic formula.

  Here, the combustion mass ratio BRk of the in-cylinder fuel is adapted in advance through experiments or the like. For example, 60% / 2 = 30% is set. In this embodiment, since the combustion mass ratio BR reaches about 60% as the combustion period, the intermediate 30% is set as BRk.

  Since the constant calorific value Q of fuel varies depending on the type of fuel, it is obtained in advance by experiments or the like according to the type of fuel. The constant volume specific heat Cv is a value of 2 to 3, and the representative value is adapted beforehand by an experiment or the like. However, the calculation accuracy can be further improved by obtaining the value of the constant volume specific heat Cv from the composition and temperature of the air-fuel mixture.

In step 186, the temperature TTDC [K] at the compression top dead center of the combustion chamber 5 is changed from the temperature TINI at the intake valve closing timing of the combustion chamber 5 to the temperature increase rate TCOMP to the compression top dead center. Multiply by 2 and add the combustion rise temperature TBURN to the multiplication value, that is, the following equation is used.

TTDC = TINI × TCOMP 2 + TBURN
... (47)
In step 187, the temperature TTDC and the volume VTDC at the compression top dead center of the combustion chamber 5 and the pressure PINI, the volume VIVC and the temperature TINI at the closing timing of the intake valve of the combustion chamber 5 are The pressure PTDC [K] is calculated.

PTDC = PINI × VIVC × TTDC / (VTDC × TINI)
... (48)
Equation (48) is obtained using the equation of state. That is, the following equation of state is established using the pressure, volume and temperature (PINI, VIVC, TINI) at the intake valve closing timing.

PINI x VIVC = n · R · TINI (Supplement 2)
Where n is the number of moles
R: gas constant,
Since the volume is almost equal in the vicinity of the compression top dead center, the following equation of state is established using the pressure, volume and temperature (PTDC, VTDC, TTDC) at the compression top dead center.

PTDC × VTDC = n · R · TTDC (Supplement 3)
When n · R is eliminated from both (complement 3) and (complement 2) and PTPT is solved, the above equation (48) is obtained.

  In step 188, similarly to step 168 of FIG. 10, a laminar combustion speed SL2 [m / sec] in the main combustion period is calculated by the following equation (known).

SL2 = f15 (TTDC, PTDC)
= SLstd × (TTDC / Tstd) ^2.18 × (PTDC / Pstd) ^−0.16
... (49)
Where Tstd: reference temperature [K],
Pstd: Reference pressure [Pa]
SLstd: reference laminar burning velocity at reference temperature Tstd and reference pressure Pstd
[M / sec],
TTDC: temperature [K] at the compression top dead center of the combustion chamber 5;
PTDC: pressure [Pa] at the compression top dead center of the combustion chamber 5
The explanation of the equation (49) is the same as the equation (16). That is, the reference temperature Tstd, the reference pressure Pstd, and the reference laminar combustion speed SLstd in the equation (49) are values determined in advance by experiments. Under a pressure of 2 bar or more, which is a normal pressure in the combustion chamber 5, the pressure term (PTDC / Pstd) ^−0.16 in the equation (49) becomes a small value. Accordingly, it is also possible to define the reference laminar combustion speed SLstd only by the reference temperature Tstd with the pressure term (PTDC / Pstd) ^−0.16 being a constant value. Therefore, when the reference temperature Tstd is 550 [K], the reference laminar combustion speed SLstd is 1.0 [m / sec], and the pressure term is 0.7, the temperature TTDC at the compression top dead center and the laminar combustion The relationship with the speed SL2 can be approximately defined by the following equation.

SL2 = f16 (TTDC)
= 1.0 × 0.7 × (TTDC / 550) ^2.18
... (50)
In step 189, the turbulence intensity ST2 of the gas flow during the main combustion period is calculated. The turbulence intensity ST2 of the gas flow can be expressed by the following equation as a function of the engine speed NRPM, similarly to the turbulence intensity ST1 of the gas flow during the initial combustion period.

ST2 = f17 (NRPM) = C2 × NRPM (20)
Where C2 is a constant,
It is also possible to obtain the turbulence intensity ST2 from a table using the rotation speed as a parameter.

  In step 190, from the laminar combustion speed SL2 [m / sec] and the turbulence intensity ST2 of the gas flow in the main combustion period, the combustion speed FLAME2 [m / sec] in the main combustion period is calculated by the following equation.

FLAME2 = SL2 × ST2 (21)
However, SL2: Laminar burning velocity [m / sec],
Equation (21) considers the contribution to the combustion speed associated with gas turbulence, as in Equation (18).

  In step 191, the main combustion period BURN2 [deg] is calculated by the following equation similar to the equation (19).

BURN2 = {(NRPM × 6) × (BR2 × VTDC)}
/ (RPROBA × AF2 × FLAME2) (22)
However, AF2: reaction area [m ² ] of the flame kernel,
Here, BR2 on the right side of the equation (22) is a change amount of the combustion mass ratio from the start timing to the end timing of the main combustion period. Since the combustion mass ratio BR becomes 2% at the end of the initial combustion period, and then the main combustion period starts and the combustion mass ratio BR reaches 60% and the main combustion period ends, BR2 = 60 % -2% = 58% is set. AF2 is an average reaction area in the growth process of the flame kernel, and is set to a fixed value experimentally determined in advance, like AF1 in the equation (19).

  During the main combustion period, the combustion chamber volume changes with the compression top dead center interposed therebetween. That is, it can be considered that the compression top dead center position exists at the approximate center between the start timing of the main combustion period and the end timing of the main combustion period. In addition, the combustion chamber volume does not change much in the vicinity of the compression top dead center even if the crank angle changes. Therefore, the combustion chamber volume in the main combustion period is represented by the combustion chamber volume VTDC at the compression top dead center.

  FIG. 13 is for calculating the basic ignition timing MBTCAL [degBTDC], and is executed at regular intervals (for example, every 10 msec). This is executed following the flow that is executed later in FIGS.

  In step 41, the initial combustion period BURN1 calculated in step 171 in FIG. 10, the main combustion period BURN2 calculated in step 191 in FIG. 12, and the ignition timing dead time equivalent crank calculated in step 17 in FIG. The angle IGNDEAD, the reference crank angle θPMAX calculated in step 16 of FIG. 5 is read.

  In step 42, the sum of the initial combustion period BURN1 and the main combustion period BURN2 is calculated as the combustion period BURN [deg].

  In step 43, the basic ignition timing MBTCAL [degBTDC] is calculated by the following equation.

MBTCAL = BURN−θPMAX + IGNDEAD (23)
In step 44, the value obtained by subtracting the ignition dead time equivalent crank angle IGNDEAD from the basic ignition timing MBTCAL is calculated as the previous combustion start timing MBTCYCL [degBTDC].

  Assuming that the basic ignition timing MBTCAL calculated in step 43 is used as the ignition timing command value for the current cycle, the previous combustion start timing MBTCYCL calculated in step 44 is calculated until the ignition timing for the next cycle is reached. Used in step 162.

  FIG. 14 is for calculating the internal inert gas ratio MRESFR in the combustion chamber 5 and is executed at regular intervals (for example, every 10 msec). This flow is executed prior to the flow of FIG.

  In step 51, the output of the air flow meter 32 and the target equivalent ratio TFBYA are read. In step 52, based on the output of the air flow meter 32, a new air amount (cylinder fresh air amount) MACYL flowing into the combustion chamber 5 is calculated. As a method for calculating the cylinder fresh air amount MACYL, a known method may be used (see JP 2001-50091 A).

  In step 53, an internal inert gas amount MRES in the combustion chamber 5 is calculated. The calculation of the internal inert gas amount MRES will be described with reference to the flow of FIG.

  In FIG. 15 (subroutine of step 53 in FIG. 14), in step 61, an inert gas amount MRESCYL at the exhaust valve closing timing EVC in the combustion chamber 5 is calculated. The calculation of the inert gas amount MRESCYL will be further described with reference to the flowchart of FIG.

  In FIG. 16 (subroutine of step 61 in FIG. 15), in step 71, the exhaust valve closing timing EVC [degBTDC], the exhaust temperature TEXH [K] detected by the temperature sensor 45, the exhaust pressure PEXH [kPa] detected by the pressure sensor 46 ].

  Here, just as the intake valve closing timing IVC is known from the command value given to the intake VTC mechanism 27, the exhaust valve closing timing EVC is also known from the command value given to the exhaust VTC mechanism 28.

  In step 72, the volume VEVC of the combustion chamber 5 at the exhaust valve closing timing EVC is calculated. This may be obtained by searching a table using the exhaust valve closing timing as a parameter, similarly to the volume VIVC at the intake valve closing timing IVC. That is, when the exhaust valve VTC mechanism 28 is provided, the volume VEVC at the exhaust valve closing timing EVC of the combustion chamber 5 may be obtained by searching the table shown in FIG. 23 from the exhaust valve closing timing EVC. When the exhaust VTC mechanism 28 is not provided, a constant value can be given.

  Although not shown, when a mechanism for changing the compression ratio is provided, the combustion chamber volume VEVC at the exhaust valve closing timing corresponding to the amount of change in the compression ratio is obtained from the table. When a mechanism for changing the compression ratio in addition to the exhaust VTC mechanism 28 is provided, the combustion chamber volume at the exhaust valve closing timing is obtained by searching a map corresponding to the exhaust valve closing timing and the compression ratio change amount.

  In step 73, the gas constant REX of the inert gas in the combustion chamber 5 is obtained by searching the table shown in FIG. 24 from the target equivalent ratio TFBYA. As shown in FIG. 24, the gas constant REX of the inert gas is the smallest when the target equivalent ratio TFBYA is 1.0, that is, the stoichiometric air-fuel ratio, and becomes larger whether it is larger or smaller.

  In step 74, the temperature TEVC of the combustion chamber 5 at the exhaust valve closing timing EVC is estimated based on the exhaust temperature TEXH. For simplicity, the exhaust temperature TEXH may be set as TEVC as it is. Note that the temperature TEVC of the combustion chamber 5 at the closing timing of the exhaust valve changes depending on the amount of heat corresponding to the fuel injection amount of the injector 21. Therefore, if such characteristics are taken into consideration, the calculation accuracy of TEVC is improved.

  In step 75, the pressure PEVC at the exhaust valve closing timing EVC of the combustion chamber 5 is calculated based on the exhaust pressure PEXH. Simply, the exhaust pressure PEXH may be set to PEVC.

  In step 76, from the volume VEVC at the exhaust valve closing timing EVC of the combustion chamber 5, the temperature TEVC at the exhaust valve closing timing EVC, the pressure PEVC at the exhaust valve closing timing EVC, and the inert gas gas constant REX, the exhaust valve of the combustion chamber 5 is obtained. The inert gas amount MRESCYL at the closing timing EVC is calculated by the following equation.

MRESCYL = (PEVC × VEVC) / (REX × TEVC) (24)
When the calculation of the inert gas amount MRESCYL at the exhaust valve closing timing EVC of the combustion chamber 5 is completed in this way, the processing returns to FIG. 15, and in step 62, the overlap of the intake and exhaust valves 15 and 16 (“O / L” in the figure). The amount of inactive gas MRESOL during the overlap is calculated, which is the amount of inert gas that is blown back from the exhaust side to the intake side.

  The calculation of the inert gas amount MRESOL will be described with reference to the flowchart of FIG.

  In FIG. 17 (subroutine of step 62 in FIG. 15), in step 81, the intake valve opening timing IVO [degBTDC], the exhaust valve closing timing EVC [degBTDC], and the exhaust valve of the combustion chamber 5 calculated in step 74 of FIG. The temperature TEVC at the closing timing EVC is read.

  Here, since the intake valve opening timing IVO is a timing earlier than the intake valve closing timing IVC by the opening angle of the intake valve 15, the opening angle of the intake valve 15 (which is known in advance) from the intake valve closing timing IVC. Can be sought.

  In step 82, the intake valve exhaust timing VTCOL [deg] is calculated from the intake valve opening timing IVO and the exhaust valve closing timing EVC by the following equation.

VTCOL = IVO + EVC (25)
For example, the intake valve opening timing IVO is at the intake top dead center position when the intake VTC mechanism 27 actuator is not energized, and the intake valve opening timing is advanced from the intake top dead center when the intake VTC mechanism 27 actuator is energized. The exhaust valve closing timing EVC is at the exhaust top dead center when the exhaust VTC mechanism 28 actuator is not energized, and the exhaust valve closing timing EVC is at the exhaust top dead center when the exhaust valve VTC mechanism 28 actuator is energized. In the case of more advanced characteristics, the sum of IVO and EVC becomes the overlap amount VTCOL of the intake and exhaust valves.

  In step 83, the accumulated effective area ASUMOL during the overlap is calculated by searching the table shown in FIG. 25 from the overlap amount VTCOL of the intake and exhaust valves. As shown in FIG. 25, the integrated effective area ASUMOL during the overlap is a value that increases as the overlap amount VTCOL of the intake and exhaust valves increases.

  Here, FIG. 26 is an explanatory diagram of the integrated effective area ASUMOL during the overlap of the intake and exhaust valves, where the horizontal axis indicates the crank angle, and the vertical axis indicates the respective opening areas of the intake valve 12 and the exhaust valve 15. Yes. The effective opening area at any time during the overlap is the smaller of the exhaust valve opening area and the intake valve opening area. The integrated effective area ASUMOL in the entire period during the overlap is an integral value (hatched portion in the figure) during the period in which the intake valve 15 and the exhaust valve 16 are open.

  By calculating the accumulated effective area ASUMOL during the overlap in this way, the overlap amount between the intake valve 15 and the exhaust valve 16 can be approximated as one orifice (outflow hole), and the state of the exhaust system and The gas flow rate passing through the virtual orifice can be simply calculated from the state of the intake system.

  In step 84, the specific heat ratio SHEATR of the inert gas remaining in the combustion chamber 5 is obtained by searching the map shown in FIG. 27 from the target equivalent ratio TFBYA and the temperature TEVC at the exhaust valve closing timing EVC of the combustion chamber 5. calculate. As shown in FIG. 27, the specific heat ratio SHEATR of the inert gas remaining in the combustion chamber is the smallest when the target equivalent ratio TFBYA is in the vicinity of 1.0, and it is larger or smaller than that. Further, under the condition where the target equivalent ratio TFBYA is constant, the temperature TEVC at the exhaust valve closing timing EVC of the combustion chamber 5 becomes smaller as the temperature becomes higher.

  In step 85, a supercharging determination flag TBCRG and a choke determination flag CHOKE are set. The setting of the supercharging determination flag TBCRG and the choke determination flag CHOKE will be described with reference to the flowchart of FIG.

  In FIG. 18 (subroutine of step 85 in FIG. 17), in step 101, the intake pressure PIN detected by the intake pressure sensor 44 and the pressure PEVC at the exhaust valve closing timing EVC of the combustion chamber 5 calculated in step 75 of FIG. Is read.

  In step 102, an intake exhaust pressure ratio PINBYEX is calculated from the intake pressure PIN and the pressure PEVC at the exhaust valve closing timing EVC of the combustion chamber 5 by the following equation.

PINBYEX = PIN / PEVC (26)
This intake / exhaust pressure ratio PINBYEX is an unknown number, and 1 is compared with this in step 103. When the intake / exhaust pressure ratio PINBYEX is 1 or less, it is determined that there is no supercharging, and the routine proceeds to step 104 where the supercharging determination flag TBCRG (initially set to zero) = 0.

  If the intake / exhaust pressure ratio PINBYEX is greater than 1, it is determined that there is supercharging, and the routine proceeds to step 105 where the supercharging determination flag TBCRG = 1.

  In Step 106, the specific heat ratio MIXAIRSHR of the mixture is obtained by searching the table shown in FIG. 28 from the target equivalent ratio TFBYA read in Step 51 of FIG. Replace with SHEATR. As shown in FIG. 28, the specific heat ratio MIXAIRSHR of the air-fuel mixture is a value that increases as the target equivalent ratio TFBYA decreases.

  In steps 106 and 107, the specific heat ratio SHEATR of the inert gas is replaced with the specific heat ratio MIXAIRSHR of the air-fuel mixture in consideration of supercharging such as turbocharging or inertial supercharging. That is, during supercharging, the gas flow during the overlap of the intake / exhaust valve is directed (blows through) from the intake system to the exhaust system. In this case, the specific heat ratio of the gas passing through the virtual orifice is the specific heat of the inert gas. By changing the ratio to the specific heat ratio of the air-fuel mixture, the amount of gas blown through is accurately estimated, and the amount of internal inert gas is accurately calculated.

  In Step 108, based on the specific heat ratio SHEATR of the inert gas calculated in Step 84 of FIG. 17 or Steps 106 and 107 of FIG. 18, the minimum and maximum choke determination thresholds SLCHOKE and SLCHOKEH are calculated by the following equations. To do.

SLCHOKER = {2 / (SHEATR + 1)}
^ {SHEATR / (SHEATR-1)} (27a)
SLCHOKEH = {− 2 / (SHEATR + 1)}
^ {-SHEATR / (SHEATR-1)} (27b)
These choke determination threshold values SLCHOKE and SLCHOKEH calculate the limit value for choking.

  If it is difficult to calculate each power of (27a) right side and (27b) right side in step 108, the calculation results of equations (27a) and (27b) are used as the minimum choke determination threshold value SLCHOKEL table and maximum choke determination. The threshold value SLCHOKEH table may be stored in advance in the memory of the engine controller 31 and obtained by searching the table from the specific heat ratio SHEATR of the inert gas.

  In step 109, it is determined whether or not the intake / exhaust pressure ratio PINBYEX is within the range of not less than the minimum choke determination threshold value SLCHOKEEL and not more than the maximum choke determination threshold value SLCHOKEH, that is, not in the choke state. If the intake / exhaust pressure ratio PINBYEX is within the range, it is determined that there is no choke, and the routine proceeds to step 110 where the choke determination flag CHOKE (initially set to zero) = 0.

  If the intake / exhaust pressure ratio P1NBYEX is not within the range, it is determined that choke is present, and the routine proceeds to step 111 where the choke determination flag CHOKE = 1.

  When the setting of the supercharging determination flag and the choke determination flag is thus completed, the process returns to FIG. 17 and the following four cases are performed in steps 86 to 88.

<1> When supercharging determination flag TBCRG = 0 and choke determination flag CHOKE = 0 <2> When supercharging determination flag TBCRG = 0 and choke determination flag CHOKE = 0 <3> Supercharging determination flag TBCRG = 0 and choke When determination flag CHOKE = 1 <4> When supercharging determination flag TBCRG = 1 and choke determination flag CHOKE = 0 When the above <1>, the routine proceeds to step 89 and overlaps when there is no supercharging and no choke If the average blown back inert gas flow rate MRESOLtmp1 is <2>, the process proceeds to step 90, and the blown back inert gas flow rate MRESOLtmp2 during the overlap when there is no supercharging and choke is present is step 91 when <3>. Go on to overlap when there is supercharging and no choke The average blowback inert gas flow rate MRESOLtmp3 in the case of <4> is advanced to step 92 to calculate the blowback inert gas flow rate MRESOLtmp4 with supercharging and with choke, respectively. Move to active gas flow rate MRESOLtmp.

Here, calculation of the inert gas flow rate MRESOLtmp1 during the overlap when there is no supercharging and no choke will be described with reference to the flow of FIG. 19. In step 121 of FIG. 19 (subroutine of step 89 in FIG. 17), step 73 in FIG. , 75, the gas constant REX of the inert gas, and the pressure PEVC when the exhaust valve of the combustion chamber 5 is closed are read.

  In step 122, based on the gas constant REX of the inert gas and the temperature TEVC at the exhaust valve closing timing of the combustion chamber 5 read in step 81 of FIG. 17, a density term MRSOLD used in a gas flow rate calculation formula described later. Is calculated by the following equation.

MRSOLD = SQRT {1 / (REX × TEVC)} (28)
Here, “SQRT” on the right side of equation (28) is a function that calculates the square root of the value in the parenthesis on the right.

  If the square root of the density term MRSOLD is difficult to calculate, the calculation result of equation (28) is stored in advance in the memory of the engine controller 31 as a map, and the gas constant REX and the temperature of the combustion chamber 5 at the exhaust valve closing timing are stored. You may obtain | require by searching the map from TEVC.

  In step 123, based on the specific heat ratio SHEATR of the inert gas calculated in step 84 of FIG. 17 and the intake / exhaust pressure ratio PINBYEX calculated in step 102 of FIG. The pressure difference term MRSOLP used is calculated by the following equation.

MRSOLP = SQRT [SHEATR / (SHEATR-1)
× {PTNBYEX ^ (2 / SHEATR)
-PTNBYEX ^ ((SHEATR + 1) / SHEATR)}] (29)
In step 124, from the density term MRSOLD, the pressure difference term MRSOLP, and the pressure PEVC at the exhaust valve closing timing of the combustion chamber 5, the blow-back inert gas flow rate MRESOLtmp1 during overlap without supercharging and without choke is expressed by the following equation. In step 125, the calculated value is transferred to the blown back inert gas flow rate MRESOLtmp.

MRESOLtmp1 = 1.4 × PEVC × MRSOLD × MRSOLP (30)
Next, the calculation of the blown back inert gas flow rate when there is no supercharging and when there is a choke will be described with reference to the flow of FIG. 20. In FIG. 20 (subroutine of step 90 of FIG. 17), in steps 131 and 132, steps 121 and 122 of FIG. In the same manner as described above, the gas constant REX of the inert gas and the pressure PEVC at the time of closing the exhaust valve of the combustion chamber 5 are read, and the density term MRSOLD is calculated from the above equation (28).

  In step 133, the choke pressure difference term MRSOLPC is calculated by the following equation based on the specific heat ratio SHEATR of the inert gas calculated in step 84 of FIG.

MRSOLPC = SQRT [SHEATR × {2 / (SHEATR + 1)} ^ {(SHEATR + 1) / [SHEATR-1)}] (31)
If it is difficult to calculate the power and square root of equation (31), the calculation result of equation (31) is stored in advance in the memory of the engine controller 31 as a table of choke pressure difference term MRSOLPC. Alternatively, it may be obtained by searching the table from the specific heat ratio SHEATR of the inert gas.

  In step 134, from the density term MRSOLD, the choke pressure difference term MRSOLPC, and the pressure PEVC when the combustion chamber 5 is closed, the inert gas flow rate MRESOLtmp2 blown back during overlap without supercharging and with choke is calculated. In step 135, the calculated value is transferred to the blown back inert gas flow rate MRESOLtmp.

MRESOLtmp2 = PEVC × MRSOLD × MRSOLPC (32)
Next, calculation of the blowback gas flow rate with supercharging and without choke will be described with reference to the flow of FIG. 21. In FIG. 21 (subroutine of step 91 in FIG. 17), in step 141, the intake pressure PIN detected by the intake pressure sensor 44 is explained. Is read.

  In step 142, the supercharging pressure difference term MRSOLPT is calculated from the specific heat ratio SHEATR of the inert gas calculated in steps 106 and 107 in FIG. 18 and the intake / exhaust pressure ratio PINBYEX calculated in step 102 in FIG. 18. Is calculated by the following equation.

MRSOLPT = SQRT [SHEATR / (SHEATR-1)
× {PINBYEX ^ (-2 / SHEATR)
-PINBYEX ^ (-(SHEATR + 1) / SHEATR)}] (33)
If the power calculation and the square root calculation of Expression (33) are difficult, the calculation result of Expression (33) is stored in advance in the memory of the engine controller 31 as a map of the supercharging pressure difference term MRSOLPT, You may obtain | require by searching the map from the specific heat ratio SHEATR of an inert gas, and the intake-exhaust pressure ratio PINBYEX.

  In step 143, based on the pressure difference term at the time of supercharging MRSOLPT and the intake pressure PIN, the inactive gas flow rate MRESOLtmp3 during the overlap with supercharging and without choke is calculated by the following equation, and the calculated value is calculated in step 143. At 144, the flow returns to the blown back inert gas flow rate MRESOLtmp.

MRESOLtmp3 = −0.152 × PIN × MRSOLPT (34)
Here, by setting the blown back inert gas flow rate MRESOLtmp3 in the equation (34) to a negative value, the gas flow rate of the air-fuel mixture blown from the intake system to the exhaust system during the overlap can be expressed.

Next, the calculation of the flow rate of the inert gas blown back during overlap with supercharging and with choke will be described with reference to the flow of FIG. 22. In FIG. 22 (subroutine of step 92 in FIG. 17), in steps 151 and 152, the step of FIG. Similarly to 141, the intake pressure PIN detected by the intake pressure sensor 44 is read, and the choke pressure difference term MRSOLPC is calculated by the aforementioned equation (31) as in step 132 of FIG.

  In step 153, based on the choke pressure difference term MRSOLPC and the intake pressure PIN, the overlap blow-back gas flow rate MRESOLtmp4 with supercharging and with choke is calculated by the following equation, and the calculated value is exceeded in step 154. Transfer to blown inert gas flow rate MRESOLtmp during lap.

MRESOLtmp4 = −0.108 × PIN × MRSOLPC (35)
Here, the blown back inert gas flow rate MRESOLtmp4 of the equation (35) can also represent a gas flow rate of the air-fuel mixture blown from the intake side to the exhaust side during the overlap, similarly to MRESOLtmp3.

  In this way, when the calculation of the blown back inert gas flow rate MRESOLtmp during the overlap divided according to the combination of the presence or absence of supercharging and the presence or absence of choke is completed, the process returns to FIG. From the inert gas flow rate MRESOLtmp and the integrated effective area ASUMOL during the overlap period, the blown back inert gas amount MRESOL during the overlap is calculated by the following equation.

MRESOL = (MRESOLtmP × ASUMOL × 60)
/ (NRPM × 360) (36)
When the calculation of the blown back inert gas amount MRESSOL is completed in this way, the process returns to FIG. 15, and in step 63, the inert gas amount MRESCYL at the exhaust valve closing timing EVC in the combustion chamber 5 and the blowback during overlap are returned. The internal inert gas amount MRES is calculated by adding the gas amount MRESOL, that is, the following equation.

MRES = MRESCYL + MRESSOL (37)
As described above, the flow rate of the inert gas blown back during overlap (MRESOLtmp3, MRESOLtmp4) becomes negative when there is supercharging, and therefore, the amount of blown inert gas MRESOL during overlap of the above equation (36) also becomes negative. According to the equation (37), the internal inert gas amount is reduced by the amount of the blown back inert gas amount MRESOL during the overlap.

  When the calculation of the internal inert gas amount MRES is completed in this way, the flow returns to FIG. 14. In step 54, using this internal inert gas amount MRES and the target equivalent ratio TFBYA, the internal inert gas rate MRESFR is calculated by the following equation. (Ratio of the amount of internal inert gas to the total amount of gas in the combustion chamber 5) is calculated.

MRESFR = MRES
/{MRES+MACYL×(1+TFBYA/14.7)} (38)
This completes the calculation of the internal inert gas ratio MRESFR.

  As described above, according to the present embodiment, the internal inert gas amount MRES is configured by the inert gas amount MRESCYL when the exhaust valve of the combustion chamber 5 is closed and the blowback gas amount MRESOL during the overlap of the intake and exhaust valves. In this case, the temperature TEV and pressure PEVC at the closing timing of the exhaust valve in the combustion chamber 5 are calculated (steps 74 and 75 in FIG. 16), and these temperature TEVC, pressure PEVC and inert gas are calculated. Since the inert gas amount MRESCYL at the exhaust valve closing timing of the combustion chamber 5 is calculated from the state constant (Equation (24) above) based on the gas constant REX (see step 76 of FIG. 16), Even during transient operation in which the amount of state (PEVC, VEVC, TEVC) inside the combustion chamber 5 changes every moment, the combustion chamber 5 can be accurately obtained regardless of the operating conditions. The inert gas amount MRESCYL in the exhaust valve closing timing can be calculated (estimated).

  In addition, based on the temperature TEVC and pressure PEVC of the combustion chamber 5 when the exhaust valve is closed, the gas constant REX and specific heat ratio SHEATR of the inert gas, and the intake pressure PIN, the blow-back inert gas flow rate (MRESOLtmp1, MRESOLtmp2) during the overlap is calculated. This was calculated (see FIGS. 19 and 20), and this gas flow rate was multiplied by the integrated effective area ASUMOL during the overlap to calculate the amount of blown back gas MRESOL during the overlap (see step 93 in FIG. 17). Therefore, the overlapped blow-back gas amount MRESOL can be calculated (estimated) with high accuracy.

  As described above, if both the inert gas amount MRESCYL and the overlapped blow-back gas amount MRESSOL in the combustion chamber 5 at the closing timing of the exhaust valve can be calculated (estimated) with high accuracy, the internal inert gas amount MRES, which is the sum of them, can also be accurately calculated. The internal inert gas ratio MRESFR calculated based on the internal inert gas amount MRES that can be estimated with high accuracy can be calculated (estimated). By making use of the temperature TINI at the intake valve closing timing IVC (see step 13 in FIG. 5), the temperature TINI at the intake valve closing timing IVC in the combustion chamber 5 can be accurately calculated. In addition, the engine can be appropriately controlled by making use of the internal inert gas amount MRES that can be accurately estimated for the fuel injection amount, valve opening / closing timing (overlap amount), and the like. .

  Further, since the gas constant REX of the inert gas and the specific heat ratio SHEATR of the inert gas are values corresponding to the target equivalent ratio TFBYA (see FIGS. 24 and 27), the operation is performed at an air-fuel ratio that deviates from the theoretical air-fuel ratio. (E.g. during lean operation where the air / fuel ratio is leaner than the stoichiometric air / fuel ratio, in order to stabilize the engine's originally unstable state, such as during cold start, the engine is on the rich side of the stoichiometric air / fuel ratio. Immediately after starting the engine operating at the air-fuel ratio, since the same large output is required, the exhaust valve closing timing of the combustion chamber 5 is also applied during full load operation where the air-fuel ratio is richer than the stoichiometric air-fuel ratio. Inert gas amount MRESCYL, overlapped blow-back gas amount MRESSOL, total internal inert gas amount MRES, and internal inert gas ratio MRESFR based on this are accurately calculated. It can be.

  Further, since the accumulated effective area ASUMOL during the overlap period is defined as the area of the virtual orifice, and it is assumed that the exhaust gas blows through the virtual orifice from the combustion chamber 5 to the intake system, the amount of the blown back inert gas amount MRESOL during the overlap is assumed. Calculation is simplified.

  This completes the description of the calculation of the internal inert gas ratio MRESFR in the combustion chamber 5, and next, the calculation of the knock limit ignition timing KNOCKcal will be described.

  Here, the concept of the newly created knock control will be described first. FIG. 29 shows the pressure history in the combustion chamber 5 when knocking occurs. When the average pressure from which the high frequency component is removed is redrawn, it can be seen that the pressure in the combustion chamber 5 increases at a stretch at the self-ignition timing θknk (knock generation timing). The pressure increase due to the knock is considered to occur as a result of the unburned air-fuel mixture in the combustion chamber 5 burning at an equal volume, and the pressure increase dP is calculated by a thermodynamic calculation formula as follows.

  Assuming that all the unburned gas of the unburned fuel amount MUB burns by equal volume combustion, the calorific value Q is given by the following equation from thermodynamics.

Q = CF # × MUB (Supplement 4)
Where CF #: lower heating value of fuel,
On the other hand, the gas temperature in the combustion chamber 5 rises due to the calorific value Q. Therefore, if this temperature rise is ΔT, the following equation is established.

Q = Cv × M × ΔT (Supplement 5)
Where M is the mass of all gases in the combustion chamber 5,
Cv: constant volume specific heat of burned gas,
If both equations of (Supplement 4) and (Supplement 5) are equal, and solving for the temperature rise ΔT, the following equation is obtained.

ΔT = (CF # × MUB) / (Cv × M) (Supplement 6)
Differentiate both sides of the gas equation of state PV = nRT (however, V is constant because of constant volume change).

V × dP = dn × R × T + n × R × dT (Supplement 7)
Here, since the change in the number of moles n is small in a state where knocking occurs, the following equation is obtained with dn = 0 on the right side of Equation (7).

dP = (n × R / V) × dT (Supplement 8)
If the temperature rise dT (= ΔT) is eliminated from both equations (Auxiliary 8) and (Auxiliary 6) and the pressure rise dP is arranged, the following equation is finally obtained.

dP = n * R * CF # * MUB / (V * Cv * M) (Supplement 9)
That is, (Supplement 9) is obtained by calculating the unburned fuel amount MUB, the volume V of the combustion chamber 5 at the time of self-ignition, the constant volume specific heat Cv of the burned gas, the mass M of all the gases in the combustion chamber 5, and the combustion chamber 5 It is shown that if the total number of moles n of all the gases is known, the pressure increase DP can be obtained by a calculation formula.

  In this case, the self-ignition timing of the combustion chamber 5 can be obtained using a known method. In the known method, the temperature and pressure in the combustion chamber 5 are calculated for each unit crank angle, 1 / τ with respect to the temperature and pressure is obtained from FIG. 30A or 30B, and the value obtained by integrating 1 / τ is 1. Is set to the self-ignition timing θknk. Here, τ in FIG. 30A or 30B is the time until the fuel in the combustion chamber 5 reaches self-ignition. When the self-ignition timing θknk is obtained, the volume Vknk at the self-ignition time of the combustion chamber 5 is mechanically obtained from the self-ignition timing θknk.

  In the first embodiment, when gasoline is used as the fuel, the octane number estimated value OCTEST of this fuel is calculated, and it is necessary to calculate 1 / τ for the fuel of this octane number estimated value OCTEST. For this reason, the fuel with the estimated octane number OCTEST is 1 / τ based on the octane number 100 (maximum octane number) shown in FIG. 30A and 1 / τ based on the octane number 80 (minimum octane number) fuel shown in FIG. 30B. 1 / τ is calculated (described later).

  On the other hand, if the self-ignition timing θknk is known, the combustion mass ratio BRknk at the time of self-ignition can be obtained from FIG. 31, and the unburned fuel amount is calculated from the combustion mass ratio BRknk at the time of self-ignition and the fuel amount QINJ by the following formula. MUB can be determined. That is, since BRknk is already combusted in the fuel amount QINJ, the remaining 1-BRknk is not combusted yet.

MUB = QINJ × (1-BRknk) (Supplement 10)
However, in FIG. 31, in order to simplify the calculation, the combustion start delay period, the initial combustion period, and the main combustion period are divided into three, and the characteristics of each period are approximated by a straight line.

  In this case, the characteristics shown in FIG. 31 do not change even if the engine load or the rotational speed is different. Therefore, if the engine specifications are determined, the characteristics are uniquely determined.

  Next, the constant volume specific heat Cv of burned gas can also be calculated by a thermodynamic calculation formula as follows. That is, since the definition of the constant pressure specific heat Cp is Cp = (∂E / ∂T) p, when this equation is integrated, the following equation is obtained.

∫dE = Cp × ∫dT (Supplement 11)
∴E = Cp × T (Supplement 12)
From the formula (Supplement 12), the constant pressure specific heat Cp is given by the following formula.

Cp = E / T (Supplement 13)
On the other hand, Cp−Cv = R is established when the pressure is constant and the ideal gas is changed. Therefore, when the constant pressure specific heat Cp is eliminated from this equation and (Equation 13) and the constant volume specific heat Cv is arranged, the following equation is finally obtained. .

Cv = E / TR (Supplement 14)
E: Enthalpy,
T: Average temperature at the time of self-ignition in the combustion chamber 5,
The mass M of all gases in the combustion chamber 5 of the above (Supplement 9) can be calculated by the following equation.

M = MRES + MACYL + QINJ (Supplement 15)
Where MRES: amount of internal inert gas,
MACYL: Cylinder fresh air volume,
QINJ: Fuel amount,
In this way, the unburned fuel amount MUB, the constant volume specific heat Cv of the burned gas, and the mass M of all the gases in the combustion chamber 5 are also calculated by the formulas of (Supplement 10), (Supplement 14), and (Supplement 15), respectively. It can be seen that it can be obtained. The remaining unknowns are the total number of moles n of all the gases in the combustion chamber 5 of the above (Supplement 9), the enthalpy E of the (Supplement 14) and the average temperature T (= TE) at the self-ignition timing in the combustion chamber 5. It is.

  Here, the total number of moles n of all the gases in the combustion chamber 5 and the number of moles of each component gas can be obtained by calculation using the basic equation of combustion, and the moles of each component gas. Using the number and the empirical formula, the enthalpy E of the formula (Supplement 14) can be calculated. The average temperature TE at the self-ignition timing in the combustion chamber 5 can also be obtained by calculation using a thermodynamic equation.

  In this way, the pressure increase dP due to knocking is configured so that it can be obtained almost by calculation formula without relying on a table or map, thereby greatly reducing the man-hours and time required for creating a table or map. Can be reduced.

  After that, the pressure increase dP obtained in this way is related to the knock, and this dP is converted into a knock strength estimated value.

  Next, calculation of the knock limit ignition timing KNOCKcal will be described in detail with reference to the following flowchart.

  32 and 33 (subroutine of step 2 in FIG. 2) are for calculating the knock limit ignition timing KNOCKcal, and are executed when the crank angle reaches a predetermined timing (for example, MBTCAL). In the following description, there is a part where the physical quantity already obtained in the above-described flow is obtained in duplicate.

  32, in step 201, the cylinder fresh air amount MACYL [g] and the internal inert gas amount MRES [g], the fuel amount QINJ [g] calculated in steps 52 and 53 in FIG. 14, and the step 171 in FIG. In addition to the initial combustion period BURN1 [deg] calculated in step 4, the main combustion period BURN2 [deg] calculated in step 34 of FIG. 12, and the basic ignition timing MBTCAL [degBTDC] calculated in step 191 of FIG. , The ignition dead time equivalent value IGNDEAD [deg] calculated in step 17 of FIG. 5, the collector internal temperature TCOL [K] detected by the temperature sensor 43, the exhaust gas temperature TEXH [K] detected by the temperature sensor 45, A collector internal pressure PCOL [Pa] detected by the pressure sensor 44 is read. The fuel amount QINJ [g] may be obtained in proportion to the fuel injection pulse width TI [ms].

  In step 202, the cylinder fresh air amount MACYL [g] is moved to WIDRY [g], and the internal inert gas amount MRES [g] is moved to MASSZ [g]. These WIDRY and MASSZ are introduced for use only in calculating the knock strength index KNKI, WIDRY is the cylinder fresh air amount, and MASSZ is the internal inert gas amount.

  In step 203, a value obtained by subtracting the basic ignition timing MBTCAL [degBTDC] and the ignition dead time equivalent crank angle IGNDEAD [deg] (that is, the crank angle at the start of combustion) is set to the crank angle θ [degBTDC].

  In step 204, the compression start temperature TC0 [K] of the combustion chamber 5 is calculated by the following equation.

TC0 = {(WIDRY + QINJ) × TCOL + MASSZ × TEXH}
/ (WIDRY + QINJ + MASSZ) (60)
Here, the specific heat of the inert gas and fresh air is made equal to simplify the equation.

  In step 205, the compression start pressure PC0 [Pa] of the combustion chamber 5 is calculated. For this purpose, the collector internal pressure PCOL at the intake valve closing timing IVC detected by the pressure sensor 44 may be set to PC0.

  Steps 206 to 208 are parts for calculating 1 / τ of the estimated octane number OCTEST for the fuel. If a 1 / τ map is provided for each of a plurality of different octane numbers from the maximum octane number to the minimum octane number, the ROM capacity becomes too large. Therefore, here, the 1 / τ map for the fuel having the maximum octane number (for example, 100) and the minimum octane number Only 1 / τ map for (for example, 80) fuels, and 1 / τ for fuels with an octane number (octane number estimated value OCTEST) between the maximum octane number and the minimum octane number are the fuels of those octane number 100 fuels. It is calculated by interpolating between 1 / τ and 1 / τ for fuel with an octane number of 80.

  Specifically, in steps 206 and 207, by searching the maps having the contents shown in FIGS. 30A and 30B from the compression start temperature TC0 and the compression start pressure PC0, 1 / τ and octane number 80 for the fuel having the octane number 100 are obtained. 1 / τ is calculated for each fuel. Each 1 / τ is a value that increases as the temperature and pressure increase as shown in FIGS. 30A and 30B. Further, if the temperature and pressure are the same, 1 / τ for the fuel with an octane number of 100 tends to be greater than 1 / τ for the fuel with an octane number of 80. Therefore, in step 208, 1 / τ in the fuel of the estimated octane number OCTEST is calculated by the following equation (interpolation calculation equation).

1 / τEST = 1 / τ80 + (OCTEST-80)
× (1 / τ100-1 / τ80) / (100-80)
... (Supplement 17)
Where 1 / τEST: 1 / τ in the fuel of the estimated octane number OCTEST
1 / τ100: 1 / τ with 100-octane fuel
1 / τ80: 1 / τ with fuel of octane number 80
Here, calculation of the estimated octane value OCTEST will be described later with reference to FIG.

  In step 209, 1 / τ of the estimated octane number OCTEST with fuel is added to the SUM. SUM represents an integrated value of 1 / τ. The initial value of the integrated value SUM is zero.

  In step 210, the integrated value SUM is compared with 1. When the integrated value SUM is less than 1, it is not the timing of self-ignition, so the routine proceeds to step 211 where the current crank angle θ is compared with the predetermined value const01. Here, a crank angle position (for example, 90 degATDC) at which knocking after ignition no longer occurs is set as the predetermined value const01. When the current crank angle θ does not exceed the predetermined value const01, the routine proceeds to step 212, and the crank angle is advanced by a predetermined angle const02 (for example, 1 deg).

  In step 213, the instantaneous compression ratio εθ in the combustion chamber 5 is calculated. This instantaneous compression ratio εθ is the reciprocal of the value obtained by dividing the clearance volume Vc of the combustion chamber 5 by the volume of the combustion chamber 5 at the current crank angle θ. Since the volume of the combustion chamber 5 at the current crank angle θ is determined by the stroke position of the piston 6, that is, the crank angle of the engine, a table with the crank angle θ as a parameter is created in advance and this table is searched from the current crank angle θ. What is necessary is just to ask for.

  In step 214, the combustion mass ratio BR at the current crank angle θ is calculated. For this purpose, first, a crank angle Θ [degATDC] for obtaining a combustion mass ratio is calculated from the current crank angle θ.

  In this case, the crank angle Θ as a variable is a value obtained by taking the compression top dead center TDC as a reference zero and adding the retard side to the plus and the advance side minus to the crank angle Θ [degATDC]. When used, the combustion mass ratio BR becomes the following linear expression.

Combustion delay period;
BR = 0 (61)
Initial combustion period;
BR = SS1 × (Θ + MBTCAL−IGNEAD) (62)
Main combustion period;
BR = 0.02 + SS2 × (Θ + MBTCAL-IGNDEAD-BURN1)
... (63)
However, SS1: 0.02 / BURN1,
SS2: 0.58 / BURN2,
Accordingly, the combustion mass ratio is calculated by the equation (61) when the calculated crank angle Θ is in the combustion delay period, by the equation (62) when in the initial combustion period, and by the equation (63) when in the main combustion period.

  In steps 215 and 216, an average temperature TC [K] and an average pressure PC [Pa] when the fuel in the combustion chamber 5 burns are calculated by the following equations.

TC = TC0 × εθ ^ 0 ・ 35
+ CF # × QINJ × BR / (MASSZ + WIDRY + QINJ)
... (64)
PC = PC0 × εθ ^ 1.35 × TC / TC0 / εθ ^ 0.35 (65)
Where εθ is the instantaneous compression ratio,
CF #: Lower heating value of fuel,
The equations (64) and (65) are equations when it is assumed that the gas is adiabatically compressed in the combustion chamber 5 and combusted with a constant volume change. That is, the first term on the right side of equation (64) is the temperature after adiabatic compression, PC0 × εθ ^ 1.35 on the right side of equation (65) is the pressure after adiabatic compression, and the second term on the right side of equation (64). TC / TC0 / εθ ^ 0.35 on the right side of equation (65) represents the rate of pressure increase due to combustion at constant volume change, where the term is the temperature rise due to combustion at constant volume change.

  In step 217, the temperature Tub of the unburned mixture in the combustion chamber 5 is calculated by the following equation.

Tub = TC0 × εθ ^ 0 · 35 × (PC / PC0 / εθ ^ 1.35)
^ (0.35 / 1.35) (66)
The equation (66) is an equation when it is assumed that the gas is adiabatically compressed in the combustion chamber 5 and is burned by a reversible adiabatic change, unlike the equation (64). That is, TC0 × εθ ^ 0.35 on the right side of equation (66) represents the temperature after adiabatic compression, and (PC / PC0 / εθ ^ 1.35) ^ (0.35 / 1.35) on the right side of equation (66). Represents the rate of temperature rise due to combustion in a reversible adiabatic change.

  It is assumed that the pressure of the unburned mixture is equal to the average pressure PC in the above equation (65).

  Here, the difference between the average temperature TC of the equation (64) and the temperature Tub of the unburned mixture of the equation (66) is as follows. That is, the average temperature TC in the equation (64) is a temperature when it is assumed that the heat generated in the combustion chamber 5 raises the temperature of all the gases in the combustion chamber 5. On the other hand, the unburned gas mixture temperature Tub of the equation (66) is in a state where the gas is separated into the burned gas and the unburned gas in the combustion chamber 5, and the heat generated in the combustion chamber 5. Is a temperature when it is assumed that only the burned gas is heated. Then, a sudden pressure increase occurs due to self-ignition of the unburned mixture, and knocking occurs.

  Thereafter, the process returns to Step 206. In Steps 206 and 207, instead of the initial combustion start temperature TC0 and the combustion start pressure PC0, the unburned mixture temperature Tub obtained in Steps 216 and 217 is changed. By searching maps containing the contents of FIG. 30A and FIG. 30B from the unburned mixture pressure (= PC), 1 / τ for an octane number 100 fuel and 1 / τ for an octane number 80 fuel are calculated. In step 208, 1 / τ in the fuel of the estimated octane number OCTEST is calculated based on these two 1 / τ using the above-mentioned (Supplement 17), and 1 / τ in the fuel of the calculated octane number estimated value OCTEST is calculated. In step 209, τ is integrated into the integrated value SUM. Then, the summed SUM and 1 are compared in step 210, and the current crank angle θ and the predetermined value const01 are compared in step 211. When the integrated value SUM is less than 1 and the crank angle θ does not exceed the predetermined value const02, the operations of steps 212 to 217 are performed to calculate the combustion chamber average pressure PC and the unburned mixture temperature Tub, and the step is again performed. The operations 206 to 217 are repeated.

  Thus, every time the crank angle θ is advanced by the predetermined value const02, the combustion chamber average pressure PC and the unburned mixture temperature Tub are recalculated to calculate 1 / τ in the fuel of the estimated octane number OCTEST, and this is used as the integrated value SUM. By integrating, the integrated value SUM in step 209 gradually increases toward 1.

  When the integrated value SUM of 1 / τ eventually becomes 1 or more, it is determined that the self-ignition timing (knock occurrence time) has come, and the routine proceeds from step 210 to step 218 in FIG. Move to the predicted value θknquest [degBTDC].

  In step 219 of FIG. 33, the combustion mass ratio BRknk at the auto-ignition timing predicted value θknkest is calculated. This is because when the predicted self-ignition timing θknkest is in the initial combustion period, the predicted self-ignition timing θknkest is converted to the crank angle Θ based on the compression top dead center TDC, and the converted crank angle Θ is When the predicted auto-ignition timing value θknkest is in the main combustion period, the auto-ignition timing predicted value θknkest is converted into the crank angle Θ based on the compression top dead center TDC, and the converted crank The angle Θ can be obtained by substituting each into the above equation (63).

  In step 220, the average temperature TE at the time of self-ignition of the combustion chamber 5 is calculated. This may be obtained by calculating the average temperature TC of the combustion chamber 5 obtained by substituting 1.0 into the combustion mass ratio BR on the right side of the equation (64) as the self-ignition average temperature TE of the combustion chamber 5.

  In step 221, the volume Vknest at the predicted self-ignition timing θknkest of the combustion chamber 5 is calculated. The volume Vknkest of the combustion chamber 5 at the predicted self-ignition timing θknkest is determined by the stroke position of the piston 6, that is, the crank angle of the engine, similarly to the volume at the current crank angle θ of the combustion chamber 5. This table may be obtained by creating a table in advance and searching this table from the predicted self-ignition timing value θknquest.

  In step 222, the unburned fuel amount MUB1 [g] at the time of self-ignition is calculated from the fuel amount QINJ [g] and the combustion mass ratio BRknest at the self-ignition timing predicted value θknkest by the following equation.

MUB1 = QINJ × (1-BRknkest) (67)
Expression (67) is an expression obtained by replacing MUB → MUB1 and BRknk → BRknquest in the above (complement 10).

  In step 223, the total gas mole number MLALL is calculated. This will be described with reference to the flowchart of FIG.

  34 (subroutine of step 223 in FIG. 33), in step 241, the internal inert gas amount MASSZ [g], the cylinder fresh air amount WIDRY [g], and the fuel amount QINJ [g] calculated in step 202 of FIG. In step 242, the internal inert gas ratio RTOEGR in the combustion chamber 5 is calculated by the following equation.

RTOEGR = MASSZ / (MASSZ + WIDRY + QINJ) (68)
In step 243, the number of moles of each gas component when all the fuel in the combustion chamber 5 is burned (that is, BR = 1) is calculated. However, the gas components are limited to O ₂ , N ₂ , CO ₂ , CO, and H ₂ O other than the fuel. Further, the fuel composition of gasoline is approximated by C ₇ H ₁₄ .

First, the number of moles WEDRY [mol] of the total exhaust gas generated after the fuel amount QINJ [g] is combusted and the respective gas components such as O ₂ , N ₂ , CO ₂ , CO, and H ₂ O in the exhaust gas The number of moles of XEO2 [mol], XEN2 [mol], XECO2 [mol], XECO [mol], and XEH2O [mol] is calculated as follows.

Total exhaust gas; WEDRY = MIDRY # × WlDRY-QlNJ
/ (B # × AC # + A # × AH) × (A # / 4) (69.1)
Oxygen; XEO2 = {MIDRY # × WlDRY × 0.21-QINJ
/ (B # × AC # + A # × AH) × (B # + A # / 4)}
/ WEDRY (69.2)
Carbon dioxide; XECO2 = {QINJ / (B # × AC # + A # × AH #) × B #}
/ WEDRY (69.3)
Carbon monoxide; XECO = 0 (69.4)
Nitrogen: XEN2 = 1-XEO2-XECO2-XECO (69.5)
Water; XEH2O = {MIDRY # × WIDRY × 15/745
+ QINJ / (B # × AC # + A # × AH #)
× A # / 2} / WEDRY (69.6)
However, MIDRY #: Number of moles of fresh gas per gram
AH #: molar mass of hydrogen,
AC #: molar mass of carbon,
A #, B #: constant,
Here, since the composition of gasoline is approximated by C ₇ H ₁₄ , the constant A # is 14 and the constant B # is 7.

  Next, the number of moles of each gas component at the initial stage of the combustion cycle WGAS [mol], WEGR [mol], WO2 [mol], WN2 [mol], WCO2 [mol], WCO [mol], WH2O [mol] Calculate as follows.

Fuel: WGAS = QINJ / (B # × AC # + A # × AH #)
... (70.1)
Inert gas; WEGR = MIDRY # × WIDRY × RTOEGR (70.2)
Oxygen: WO2 = MIDRY # × WIDRY × 0.21 + WEGR × XEO2
... (70.3)
Nitrogen; WN2 = MIDRY # × WIDRY × 0.89 + WEGR × XEN2
... (70.4)
Carbon dioxide; WCO2 = WEGR × XECO2 (70.5)
Carbon monoxide; WCO = WEGR × XECO (70.6)
Water; WH2O = MIDRY # × WIDRY × 15/745
+ WEGR × XEH2O (70.7)
Next, the number of moles MGLAS [mol], MLO2 [mol], MLN2 [mol], MLCO2 [mol], MLCO [mol], MLH2O [mol] of each gas component when all are combusted (that is, BR = 1). Calculate as follows.

Fuel; MLGAS = WGAS-QINJ / (B # × AC # + A # × AH #)
... (71.1)
Oxygen; MLO2 = WO2- (B # + A # / 4) × QlNJ
/ (B # × AC # + A # × AH #) (71.2)
Nitrogen; MLN2 = WN2 (71.3)
Carbon dioxide; MLCO2 = WCO2 + B # × QINJ
/ (B # × AC # + A # × AH #) (71.4)
Carbon monoxide; MLCO = WCO (71.5)
Water; MLH2O = WH2O + A # / 2 × QINJ
/ (B # × AC # + A # × AH #) (71.6)
This completes the calculation of the number of moles of each gas component when all the fuel in the combustion chamber 5 has been burned (that is, BR = 1), and thus the process proceeds to step 244 to calculate the sum of the number of moles of each gas component. As the total number of gas moles MLALL when all of the fuel is burned, that is, the total number of gas moles MLALL is calculated by the following equation.

MLALL = MLGAS + MLO2 + MLN2 + MLCO2 + MLCO + MLH2O
... (71.7)
When the calculation of the total gas mole number MLALL is completed in this way, the process returns to step 224 in FIG. 33 to calculate the gas enthalpy (enthalpy of self-ignited fuel gas) E [cal / mol]. The calculation of the gas enthalpy will be described with reference to the flow of FIG. 35 (subroutine of step 224 in FIG. 33), in step 251, the self-ignition average temperature TE of the combustion chamber 5 calculated in step 220 of FIG. 33, and each gas calculated in steps 243 and 244 of FIG. The number of moles of components MLGAS, MLO2, MLN2, MLCO2, MLCO, MLH2O, and the total number of moles of gas MLALL are read.

In step 252, enthalpy EO2, EN2, ECO2, ECO, EH2O of each gas component is calculated from the average temperature TE during self-ignition. What is necessary is just to calculate the enthalpy of each gas component using the following Mizutani empirical formula (refer to internal combustion engine vol.11 No.125p79).
(1) When TE <1200K E = A0 # + 1000 × (A1 # × (TE / 1000)
+ A2 # / 2 × (TE / 1000) ^ 2
+ A3 # / 3 × (TE / 1000) ^ 3
+ A4 # / 4 × (TE / 1000) ^ 4
+ A5 # / 5 × (TE / 1000) ^ 5) + HDL # (72.1)
(2) When TE> 1200K E = B0 # + 1000 × (B1 # × (TE / 1000)
+ B2 # × LN (TE / 1000)
-B3 # / (TE / 1000)
-B4 # / 2 / (TE / 1000) ^ 2
-B5 # / 3 / (TE / 1000) ^ 3) + HDL # (72.2)
However, A0 # to A5 #, B0 to B5 #, and HDL # are adapted values obtained by experiments,
In step 253, the enthalpy EG of the fuel is calculated by the following equation.

EG = B # / AC # × ECO2 + A # / AH # × EH20 / 2
+ (B # / AC # + A # / AH # / 4) × EO2 (72.3)
In step 254, the average enthalpy E of each gas component is calculated by the following equation, the processing in FIG. 35 is terminated, and the processing returns to step 225 in FIG.

E = (MLGAS × EG + MLO2 × EO2 + MLN2 × EN2
+ MLCO2 × ECO2 + MLCO × ECO + MLH2O × EH2O)
/ MLALL (72.4)
In step 225 of FIG. 33, the constant volume specific heat Cv [J / K · g] of the burned gas is calculated by the following equation using the gas enthalpy E and the average temperature TE at the time of self-ignition of the combustion chamber 5.

Cv = E / TE-R # (73)
Where R #: universal gas constant,
Expression (73) is an expression obtained by replacing T → TE and R → R # in the above (complement 14).

  In step 226, the pressure increase due to self-ignition, that is, the pressure increase DP1 [Pa] due to knocking is calculated by the following equation.

DP1 = (MLALL × R # × CF # × MUB1)
/ {Vknest × Cv × (MASSZ + QINJ + WIDRY)}
... (74)
Where CF #: lower heating value of fuel,
As shown in FIG. 29, the pressure increase DP1 in equation (74) is such that the pressure in the combustion chamber 5 increases stepwise due to the occurrence of knocking, and this pressure increase is obtained by a calculation formula.

  Expression (74) is an expression obtained by replacing dP → DP1, n → MLALL, R → R #, V → Vknest, M → MASSZ + WIDRY + QINJ in the above (Appendix 9).

  In step 227, the knock magnitude estimated value basic value KIC0 is calculated by the following equation.

KIC0 = correlation coefficient 1 × DP1 (75)
Here, the correlation coefficient 1 on the right side of equation (75) is a coefficient for representing the correlation with the knock intensity. In this case, the knock magnitude estimated value basic value KIC0 increases as the pressure increase DP1 due to knock increases.

  In step 228, a rotational speed correction coefficient KN is calculated from the engine rotational speed NRPM by searching a table having the contents shown in FIG. 36, and in step 229, this rotational speed correction coefficient KN is set to the knock magnitude estimated value basic value KIC0. The multiplied value is set as a knock strength estimated value KICEEST, that is, a knock strength estimated value KICEEST is calculated by the following equation.

KICEST = KIC0 × KN (76)
Here, the rotational speed correction coefficient KN is for reflecting the difference in the knock strength estimation value because the driver feels pressure vibration due to knock more strongly when the engine rotational speed NRPM is lower than when the rotational speed is high. is there. That is, as shown in FIG. 36, the value of KN is 1.0 when the reference rotational speed NRPM0 is 1.0 and exceeds 1.0 in a lower rotational speed range, and conversely in a higher rotational speed range than the reference rotational speed NRPM0. The value is less than 1.0. Actual values are adapted by experiment.

  In step 230, the knock retard amount KNRT [deg] is calculated by the following equation.

KNRT = KICEST-trace knock strength (77)
Here, the trace knock intensity of the equation (77) is a knock intensity when a light knock occurs as is well known. The trace knock intensity is obtained by searching a table having the contents shown in FIG. 37 from the engine speed NRPM.

  In step 231, the value obtained by subtracting the knock retard amount KNRT from the basic ignition timing MBTCAL is set as the knock limit ignition timing KNOCKcal, that is, the knock limit ignition timing KNOCKcal [degBTDC] is calculated by the following equation.

KNOCKcal = MBTCAL-KNRT (78)
On the other hand, the integrated value SUM of 1 / τ may not reach 1, and at this time, the current crank angle θ eventually exceeds the predetermined value const01 in step 211 of FIG. At this time, the process proceeds from step 211 in FIG. 32 to step 232 in FIG. 33, and after the knock retard amount KNRT = 0, the operation in step 231 is executed.

  After that, the process waits until the crank angle reaches the basic ignition timing MBTCAL of the next combustion cycle, and the processes of FIGS. 32 and 33 are executed again. Thus, the knock limit ignition timing KNOCKcal is obtained for each combustion cycle.

  When the calculation of the knock limit ignition timing KNOCKcal is completed in this manner, the process returns to step 3 in FIG. Is selected as the ignition timing minimum value PADV, and in step 4, a value obtained by adding various corrections based on the water temperature or the like is set as an ignition timing command value QADV [degBTDC]. After completion of engine warm-up, there is no correction due to the water temperature or the like, so the ignition timing command value QADV becomes equal to the ignition timing minimum value PADV.

  The ignition timing command value QADV set in this way is transferred to the ignition register in step 5, and an ignition signal for cutting off the primary current from the engine controller 31 at the timing when the actual crank angle coincides with the ignition timing command value QADV is generated by the ignition coil 13. Is output.

  Next, calculation of the estimated octane number OCTEST of the fuel during operation will be described with reference to the flow of FIG. Since the estimation of the octane number is performed while detecting whether or not knocking has occurred based on a signal from the knock sensor 47, the flow in FIG. 38 is executed immediately after ignition for each ignition. Here, in order to execute each ignition, it may be executed at a timing at which a predetermined crank angle has elapsed from the input of a reference position signal generated by a signal from the crank angle sensor (33, 34).

  In FIG. 38, in step 261, the knock sensor 47 is used to check whether knock has occurred. For example, since the frequency component of the knock detected by the knock sensor 47 (knock detection means, knock generation timing detection means) changes as shown in FIG. 39, a predetermined threshold value is determined in advance according to conformity, and this threshold is set. When crossing the value, it is determined that knocking has occurred, and the routine proceeds to step 262, where the actual self-ignition timing detection value is read. In order to obtain the actual self-ignition timing detection value, the timing crossing the threshold in FIG. 39 may be measured as the actual self-ignition timing. In FIG. 39, the horizontal axis represents the time from a predetermined crank angle position, and the vertical axis represents the absolute value of the amplitude of the frequency component for knocking. Here, the measured timing [s] is multiplied by a conversion coefficient to be converted into a crank angle unit, and the converted value is read as a self-ignition timing detection value θknkreal [degBTDC] (knock occurrence timing detection value).

  Steps 263 to 266 in FIG. 38 are portions for converging the estimated octane number OCTEST without causing knock so that the ignition timing difference Δθ, which is the difference between the detected self-ignition timing value and the predicted self-ignition timing value, falls within the allowable range. It is. First, at step 263, the ignition timing difference Δθ is calculated from the self-ignition timing predicted value θknkest calculated at step 218 of FIG. 33 and the self-ignition timing detection value θknkreal read at step 262 of FIG.

Δθ = θknkreal−θknkest (79)
Here, the positive ignition timing difference Δθ indicates that the predicted self-ignition timing θknkest is on the more retarded side than the detected self-ignition timing θknkreal. On the contrary, there is no possibility that the self-ignition timing detection value θknkreal is on the retard side of the predicted self-ignition timing value θknkest. However, since there is no problem even if the two are not distinguished in terms of computation, the processing is performed without distinguishing both in order not to complicate the computation.

  In step 264, the absolute value of this ignition timing difference Δθ is compared with a predetermined value (for example, 1 deg). The predetermined value defines the allowable range. If the absolute value of the ignition timing difference Δθ is less than the predetermined value, the ignition timing difference Δθ is within the allowable range and knocking has occurred due to a factor other than the error in the estimated octane number. It is determined that the current process is terminated.

  When the absolute value of the ignition timing difference Δθ is greater than or equal to a predetermined value, the routine proceeds to step 265. As described above, when the predicted self-ignition timing value θknkest is on the more retarded side than the detected self-ignition timing value θknkreal (Δθ is positive), it is determined that the estimated octane number OCTEST is too large than the actual octane number, and the estimated octane number OCTEST Is reduced by a value obtained by multiplying the first predetermined value const03 by the ignition timing difference Δθ. That is, the estimated octane value OCTEST is updated by the following equation.

OCTEST (new) = OCTEST (old) −const03 × Δθ (80)
However, OCTEST (new): Estimated value of octane number after update,
OCTEST (old): Estimated octane value before update,
const03: update rate to the side to be reduced (positive anonymous number),
Here, the second term on the right side of equation (80) defines the amount of update per one estimated octane value. The reason why the ignition timing difference Δθ is introduced into the update amount per time is to accelerate the convergence of the estimated octane number OCTEST. That is, when the estimated octane number OCTEST is larger than the actual octane number and close to the actual octane number, the autoignition timing predicted value θknkest does not deviate so much from the autoignition timing detection value θknkreal, but the estimated octane number OCTEST is actually When the value is larger than the actual octane number and deviated from the actual octane number, the self-ignition timing predicted value θknkest is greatly deviated from the self-ignition timing detection value θknkreal to the retard side. Therefore, when the autoignition timing predicted value θknkest is greatly deviated from the autoignition timing detection value θknkreal (that is, Δθ is large), the renewal amount per time is increased accordingly, and the estimated octane number OCTEST is It is intended to accelerate convergence.

  In step 266 of FIG. 38, the auto-ignition timing predicted value θknquest is again calculated using the updated octane number estimated value OCTEST. This self-ignition timing predicted value θknkest is calculated for the second time (the self-ignition timing predicted value θknkest is calculated for the first time in step 218 in FIG. 33), and this second self-ignition timing predicted value θknkest is calculated for the first time. It is overwritten on the predicted ignition timing value θknquest.

  The calculation of the second and subsequent self-ignition timing predicted values θknquest will be described with reference to the flowcharts of FIGS. 40 and 41. The calculation process of the predicted self-ignition timing θknquest for the second and subsequent times is based on the calculation process of the predicted self-ignition timing in FIGS. 32 and 33. Accordingly, in FIGS. 40 and 41, the same parts as those in FIGS. Are given the same step number.

  32 and FIG. 33 are only the steps 281 and 282 of FIG. That is, when the estimated octane value OCTEST updated in step 265 of FIG. 38 is used in step 208 of FIG. 40, when the integrated value SUM of 1 / τ becomes 1 or more in step 210 of FIG. Step 281 is entered, and the crank angle θ at that time is entered into the autoignition timing predicted value θknquest [degBTDC] to calculate the second autoignition timing predicted value θknest.

  40 and 41, the self-ignition timing predicted value θknkest calculated for the second time is closer to the self-ignition timing detection value θknkreal than the self-ignition timing predicted value θknkest calculated for the first time in step 218 of FIG. Therefore, returning to step 263 of FIG. 38, the ignition timing difference Δθ is calculated again by using the self-ignition timing predicted value θknest calculated for the second time, and in step 264, this ignition timing difference Δθ is compared with a predetermined value. When the ignition timing difference Δθ is greater than or equal to the predetermined value, steps 265, 266, 263, and 264 are repeated. When the ignition timing difference Δθ calculated using the self-ignition timing predicted value θknkest calculated several times in due to this repetition has fallen below the predetermined value, the process proceeds from step 264 in FIG. 38 to “END”. The process ends.

  As described above, when knocking is detected, the operation of repeatedly updating the estimated octane number OCTEST until the ignition timing difference Δθ falls below a predetermined value (the loop operation of steps 263 to 266 in FIG. 38) reaches the next combustion cycle. By the end of the combustion cycle, the estimated octane number OCTEST has converged when the combustion cycle in which knocking occurs is completed.

  On the other hand, when the crank angle θ exceeds the predetermined value const01 without the integrated value SUM of 1 / τ reaching 1 in FIG. 40, the routine proceeds from step 211 in FIG. 40 to step 282 in FIG. 41, and the predetermined value const06 is predicted as the auto-ignition timing. After moving to the value θknest, the current process is terminated. As the predetermined value const06, a value equal to or larger than the predetermined value const01 in step 211 in FIG. 32 (a value on the retard side) is used.

  Thus, in this embodiment (first embodiment), when it is detected that a knock has occurred, the octane number is estimated until the ignition timing difference Δθ is within an allowable range in the combustion cycle in which the knock is detected. The value OCTEST is reduced.

  On the other hand, if no knock is detected in FIG. 38, the routine proceeds from step 261 to step 267, where the ignition timing minimum value PADV [degBTDC] calculated in step 3 of FIG. 2 and the basic ignition calculated in step 1 of FIG. Compare the timing MBTCAL [degBTDC]. When the ignition timing minimum value PADV coincides with the basic ignition timing MBTCAL, the octane number estimated value OCTEST and the actual octane number coincide with each other. Therefore, since it is not necessary to change the octane number estimated value, the current process is terminated.

  On the other hand, when the minimum ignition timing value PADV does not coincide with the basic ignition timing MBTCAL, it is determined that the estimated octane number OCTEST does not coincide with the actual octane number, and as a result, the ignition timing is retarded. Then, the counter value count is compared with the predetermined value const04. The initial value of the counter value count is zero. Since the counter value count is initially less than the predetermined value const04 after proceeding to step 268, the process proceeds to step 269 and the counter value count is incremented by one. That is, the counter value count is incremented by 1 each time the flow of FIG. 38 is executed once when no knock is detected, so the counter value count eventually becomes equal to or greater than the predetermined value const04. At this time, the routine proceeds from step 268 to step 270, where the octane number estimated value OCTEST is increased by the second predetermined value const05. That is, the estimated octane value OCTEST is updated by the following equation.

OCTEST (new) = OCTEST (old) + const05 (81)
However, OCTEST (new): Estimated value of octane number after update,
OCTEST (old): Estimated octane value before update,
const05: Update amount to the side to be increased,
Since the estimated octane value OCTEST is updated every time the counter value count reaches the predetermined value const04, the counter value count is reset to zero in step 271.

  FIG. 42 shows the movement of the ignition timing, the counter value count, and the estimated octane number OCTEST as a model. In FIG. 42, the alternate long and short dash line indicates the case of the first embodiment (the present invention), and the solid line indicates the case of Comparative Example 1.

Here, “Comparative Example 1” does not include steps 262, 263, 264, and 266 in FIG. 38, and instead of the above equation (80) in step 265,
OCTEST (new) = OCTEST (old) -const03
... (Supplement 18)
However, OCTEST (new): Estimated value of octane number after update,
OCTEST (old): Estimated octane value before update,
const03: update amount to the side to be reduced,
In this case, the estimated octane value OCTEST is updated by the following formula, and after this operation, the process proceeds to “END” and the current process is terminated. In the case of the comparative example 1, as shown in FIG. 42, when knocking is detected by the knock sensor 47 at timing t01, it is determined that the estimated octane number OCTEST is greater than the actual octane number, and the estimated octane number OCTEST is the predetermined value const03. Only be reduced step by step. As a result, if knock does not occur, every time the counter value count reaches the predetermined value const04, the estimated octane value OCTEST increases by the second predetermined value const05. Then, after the ignition timing minimum value PADV coincides with MBTCAL at the timing of t02, the update of the octane number estimated value OCTEST is stopped and the value at that time is held. If knocking occurs again at the subsequent timing t03, the above operation is repeated.

  Thus, according to the comparative example 1, since the octane number estimated value OCTEST is updated by the predetermined value const03 every time a knock is actually detected, the octane number estimated value OCTEST converges when no knock is detected. It will be.

  Further, according to Comparative Example 1, when the self-ignition timing predicted value θknkest is slightly deviated to the retarded side from the actual self-ignition timing, the self-ignition timing predicted value θknkest is retarded from the actual self-ignition timing. Therefore, in each case, the updated amount per one time (= const03) of the estimated octane number is the same, so that the autoignition timing predicted value θknkest is The convergence of the estimated octane number OCTEST is delayed when the ignition timing deviates far from the ignition timing.

  On the other hand, according to the first embodiment, when a knock is detected at the timing t01, the self-ignition timing at that time is detected, and the self-ignition timing detection value θknkreal and the self-ignition timing predicted value θknkest are A value obtained by multiplying the first predetermined value const03 by the ignition timing difference Δθ, which is the difference, is an update amount per one time of the estimated octane number, so the estimated octane number OCTEST is larger than the actual octane number and the estimated octane number OCTEST is If the predicted self-ignition timing θknkest deviates far from the detected self-ignition timing θknkreal due to a large deviation from the actual octane number, the estimated octane number OCTEST is larger than the actual octane number and the estimated octane number OCTEST is not much different from the actual octane number Update amount per single estimated octane value is greater than if a more ignition timing predicted value θknkest is not a little only deviated to the retard side of ignition timing detected value Shitaknkreal.

In addition, since the update of the octane number estimated value OCTEST is repeated many times until the ignition timing difference Δθ falls within the allowable range, the octane number estimated value OCTEST converges in the combustion cycle in which knocking is detected. Therefore, the convergence value of the estimated octane value OCTEST is larger than that in the case of Comparative Example 1 (see the dashed line in the lower part of FIG. 42).
As described above, according to the first embodiment, when knock is detected, the octane number estimated value OCTEST is larger as the octane number estimated value OCTEST is larger than the actual octane number and the octane number estimated value OCTEST is larger than the actual octane number. Since it converges quickly and in the combustion cycle in which knocking is detected, the second knocking is not detected after the ignition timing minimum value PADV coincides with the basic ignition timing MBTCAL (after t02).

  The octane number estimated value OCTEST calculated in accordance with FIG. 38 in this way is used to calculate 1 / τ in the fuel of the octane number estimated value OCTEST in step 208 of FIG.

  Here, the effect of this embodiment is demonstrated.

  According to the first embodiment (the invention described in claim 1), when gasoline is used as the fuel, the knock detection result by the knock sensor 47 is fed back to the octane number of the fuel instead of the ignition timing (see FIG. 38). . That is, as indicated by a one-dot chain line at the bottom of FIG. 42, the estimated octane value OCTEST decreases stepwise when it is detected that knocking has occurred, and thereafter, the second predetermined value const05 is incremented by a predetermined cycle. It is getting bigger gradually.

  Thus, according to the first embodiment (the invention described in claim 1), the estimated octane number OCTEST is calculated based on the knock detection result by the knock sensor 47 (see steps 261, 265, and 270 in FIG. 38). Based on the estimated octane value OCTEST, a self-ignition timing predicted value θknest (knock occurrence timing predicted value) in the combustion chamber 5 is calculated (see steps 206 to 210 in FIG. 32 and step 218 in FIG. 33). Since the knock limit ignition timing KNOCKcal is calculated based on the predicted ignition timing value θknquest (see steps 219 to 231 in FIG. 33), even when using a fuel sold in a market where the octane number of the fuel cannot be known in advance, Regardless of the driving state, the knock detection result by the knock sensor 47 is fed back to the ignition timing. Unlike the conventional apparatus, the ignition timing delay for avoiding knocking and the subsequent advance operation are not repeated. For this reason, since the knock limit ignition timing can be followed even during a transition such as acceleration or deceleration, fuel consumption deterioration and output reduction can be prevented.

  Further, when the estimated octane number OCTEST is updated (calculated), as shown in the above equation (80), the self-ignition timing detection value θknkreal (knock generation timing detection value) and the self-ignition timing prediction value θknkest (knock generation timing prediction value) The ignition timing difference [Delta] [theta] (= [theta] knkreal- [theta] knkest) is also taken into consideration. In other words, when the estimated octane number OCTEST is larger than the actual octane number and deviates from the actual octane number (that is, the predicted self-ignition timing value θknkest is deviated larger than the detected self-ignition timing value θknkreal), the estimated octane number Update of the estimated octane number per time than when the value OCTEST is larger than the actual octane number and close to the actual octane number (that is, the predicted autoignition timing value θknkest is not much shifted to the retarded side from the detected autoignition timing value θknkreal) The octane number estimated value OCTEST is larger than the actual octane number and larger than the actual octane number, and the octane number estimated value OCTEST is larger than the actual octane number and close to the actual octane number. Price estimate OCTEST to converge quickly.

  Thus, according to the first embodiment (the invention described in claim 1), the octane number is estimated even based on the ignition timing difference Δθ (comparison result of the self-ignition timing detection value θknkreal and the self-ignition timing predicted value θknkest). Since the value OCTEST (knock correlation parameter estimated value) is calculated (see step 265 in FIG. 38), the convergence of the octane number estimated value OCTEST can be accelerated, and the drivability is improved.

  In the first embodiment, as shown in FIGS. 5, 10, 12, and 13, laminar combustion velocities SL1 and SL2, combustion gas volume equivalent volume (V0, VTDC), combustion mass ratio (BR1, BR2) The combustion period from the start of combustion to the predetermined crank angle (BURN1, BURN2) is calculated based on the reaction probability RPROBA, and the basic ignition timing MBTCAL is calculated based on the combustion period (BURN1, BURN2). Instead of calculating the basic ignition timing MBTCAL, a base ignition timing map may be provided. In this case, the knock limit ignition timing calculation means is a pressure increase amount estimation means for estimating the pressure increase amount DP1 due to knocking of the combustion chamber 5 based on the self-ignition timing predicted value θknest (knock occurrence timing predicted value) and the operating conditions. (See steps 219 to 226 in FIG. 33), knock strength estimated value calculation means (see steps 227 to 229 in FIG. 33) for calculating a knock strength estimated value KIC based on the pressure increase DP1, and this knock strength estimation Knock retard amount calculating means (see step 230 in FIG. 33) for calculating the knock retard amount KNRT based on the value KIC, and the knock limit ignition timing obtained by correcting the basic ignition timing MBTCAL to the retard side by the knock retard amount KNRT. Knock limit ignition timing setting means for setting as KNOCKcal (see step 231 in FIG. 33); Even if the base ignition timing as the basic ignition timing is given as a map by including it (the invention according to claim 2), a map of the base ignition timing for each of a plurality of different octane numbers from the maximum octane number to the minimum octane number Therefore, the ROM capacity does not increase.

  According to the first embodiment (the invention according to claim 3), in response to the fact that the octane number has the greatest influence on the knock when gasoline is used as the fuel, the estimated octane number OCTEST is obtained from the knock detection result and Since the calculation is based on the ignition timing difference Δθ (comparison result of the knock occurrence timing detection value and the knock occurrence timing prediction value) (see steps 261, 262, 263, and 265 in FIG. 38), the gasoline whose octane number is not previously known is calculated. Even when the fuel is used, it is possible to accurately calculate the predicted self-ignition timing θknkes (predicted knock occurrence timing).

  According to the first embodiment (the invention described in claim 4), when the octane number estimated value calculating means detects that a knock has occurred, the self-ignition timing detection value θknkreal (knock) in the combustion cycle in which the knock has been detected. The estimated octane number OCTEST is updated to a smaller value until the ignition timing difference Δθ (occurrence timing difference) between the occurrence timing detection value) and the predicted self-ignition timing value θknest (predicted knock occurrence timing) falls within the allowable range (see FIG. 38 (see steps 263 to 266 in the loop operation), and thereafter, the octane number estimated value OCTEST is updated at a constant cycle so as to increase by the second predetermined value const05 (see steps 261 and 267 to 271 in FIG. 38). Therefore, the estimated octane value OCTEST can be converged in the combustion cycle in which knocking has been detected.

  According to the first embodiment (the invention described in claim 5), the update (see step 270 in FIG. 38) is performed so that the estimated octane number OCTEST is increased by the second predetermined value const05 (the side where knocking occurs). Since the basic ignition timing MBTCAL is executed only when knocking occurs, that is, when the ignition timing minimum value PADV is retarded from the basic ignition timing MBTCAL (step 267 in FIG. 38), the estimated octane number OCTEST is erroneously set. There is no renewal.

  According to the first embodiment (the invention described in claim 9), the basic ignition timing calculation means calculates the laminar combustion velocity (SL1, SL2) that is the combustion velocity in the laminar flow state of the combustion gas. Combustion rate calculation means (see step 168 in FIG. 10 and step 188 in FIG. 12) and combustion gas volume equivalent volume calculation means (FIG. 10) for calculating the volume (V0, VTDC) corresponding to the combustion gas volume in the combustion chamber 5. Step 162 of FIG. 12 and 182 of FIG. 12) and combustion mass ratio calculating means for calculating the combustion mass ratio (BR1, BR2) of the gas combusted in the combustion chamber 5 up to a predetermined crank angle (Step 171 of FIG. 10, Step 191 in FIG. 12), reaction probability calculation means for calculating the reaction probability RPROBA indicating the ease of combustion of the combustion gas under a predetermined operating condition (see Step 15 in FIG. 5), and these laminar flows Based on the burning rate (SL1, SL2), the combustion gas volume equivalent volume (V0, VTDC), the combustion mass ratio (BR1, BR2), and the reaction probability RPROBA, the combustion period (BURN1, BURN2) from the start of combustion to the predetermined crank angle is calculated. Combustion period calculating means for calculating (see step 171 in FIG. 10 and step 191 in FIG. 12) and basic ignition timing calculating means for calculating a basic ignition timing MBTCAL from which MBT is obtained based on the combustion periods (BURN1, BURN2) ( 13), the knock limit ignition timing KNOCKcal, which is a value obtained by correcting the basic ignition timing MBTCAL to the retard side, is calculated based on the combustion analysis. The optimum knock limit ignition timing KNOCKcal can be calculated regardless of the conditions.

  The flow of FIGS. 43 and 45 is the second embodiment, FIG. 43 replaces FIG. 32 of the first embodiment, and FIG. 45 replaces FIG. 38 of the first embodiment. 43, the same step number is assigned to the same part as FIG. 32, and the same step number is assigned to the same part as FIG. 38 in FIG. FIG. 33 is a flow common to the first and second embodiments.

  A mixed fuel of gasoline and alcohol (alcohol-containing fuel) may be used. In this case, when the base ignition timing is set, the alcohol concentration in the mixed fuel is checked, and the base ignition timing is matched so that knocking does not occur when the mixed fuel having the checked alcohol concentration is used.

  However, for example, in the overseas market, when the alcohol concentration in the mixed fuel is different from the mixed fuel used for matching the base ignition timing, for example, when the alcohol concentration in the mixed fuel is lower than the alcohol concentration of the mixed fuel used for matching, knocking occurs. If the ignition timing delay and the subsequent advance operation are repeated, even if knocking can be avoided due to the ignition timing retardation, fuel consumption deteriorates and output decreases.

  The second embodiment is applied to a case where a mixed fuel of alcohol and gasoline is used as a fuel to be used. Based on a knock detection result by the knock sensor 47, an estimated value ALLCEST of the alcohol concentration (knock correlation parameter) in the mixed fuel. Is calculated based on the estimated alcohol concentration value ALLCEST, and a predicted ignition timing (knock occurrence timing predicted value) in the combustion chamber 5 is calculated. The knock limit ignition timing KNOCKcal is calculated based on the predicted self-ignition timing value θknest. Is calculated.

  The portion different from the first embodiment will be mainly described. In FIG. 43, steps 291 to 293 are portions for calculating 1 / τ for the mixed fuel of the alcohol concentration estimated value ALCEST. If a map of 1 / τ is provided for each of a plurality of different alcohol concentrations from the lowest alcohol concentration to the highest alcohol concentration, the ROM capacity becomes too large, and therefore, 1 for the fuel with the lowest alcohol concentration (for example, 0%). Only the map of / τ and the map of 1 / τ for the fuel with the highest alcohol concentration (for example, 85%), and the alcohol concentration between the lowest alcohol concentration and the highest alcohol concentration (alcohol concentration estimated value) 1 / τ for the mixed fuel of (ALCEST) is calculated by interpolating between 1 / τ for the fuel having the alcohol concentration of 0% and 1 / τ for the fuel having the alcohol concentration of 85%.

  Specifically, in steps 291 and 292, the first time is obtained by searching a map having the contents shown in FIGS. 44A and 44B from the compression start temperature TC0 and the compression start pressure PC0, respectively. / Τ, 1 / τ for a mixed fuel with an alcohol concentration of 85% is calculated. Each 1 / τ is a value that increases as the temperature and pressure increase as shown in FIGS. 44A and 44B. Further, if the temperature and pressure are the same, 1 / τ in the mixed fuel having the alcohol concentration of 0% tends to be larger than 1 / τ in the mixed fuel having the alcohol concentration of 85%. Therefore, in step 293, 1 / τ in the mixed fuel of the alcohol concentration estimated value ALCEST is calculated by the following equation (interpolation calculation equation).

1 / τEST = 1 / τ85 + (85-ALCTEST)
× (1 / τ0−1 / τ85) / (85-0)
... (Supplement 19)
However, 1 / τEST: 1 / τ in the mixed fuel of the alcohol concentration estimated value ALCEST,
1 / τ0: 1 / τ with a mixed fuel with an alcohol concentration of 0%,
1 / τ85: 1 / τ with a mixed fuel with an alcohol concentration of 85%,
Here, the calculation of the alcohol concentration estimated value ALCEST will be described later.

  In step 209 of FIG. 43, 1 / τ of the alcohol concentration estimated value ALCEST in the mixed fuel is added to the integrated value SUM.

  Next, in FIG. 45, when it is detected at step 261 that knocking has occurred by knock sensor 47, it is determined that the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration, and after operations of steps 262 and 263 are performed, step 264 is performed. Then, the absolute value of the ignition timing difference Δθ is compared with a predetermined value. When the absolute value of the ignition timing difference Δθ is greater than or equal to a predetermined value, the routine proceeds to step 301, where the alcohol concentration estimated value ALCEST is increased by a value obtained by multiplying the first predetermined value const13 by the ignition timing difference Δθ. That is, the alcohol concentration estimated value ALEST is updated by the following equation.

ALLCEST (new) = ALCEST (old) + const13 × Δθ (82)
However, ALCEST (new): Estimated alcohol concentration after update,
ALCEST (old): Estimated alcohol concentration before update,
const13: update rate to the higher side (positive unnamed number),
Here, the second term of the equation (82) defines the update amount per one time of the estimated alcohol concentration value. The reason why the ignition timing difference Δθ is introduced into the renewal amount per time is to accelerate the convergence of the alcohol concentration estimated value ALCEST. That is, when the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration and close to the actual alcohol concentration, the autoignition timing predicted value θknkest is not much shifted to the retard side from the autoignition timing detected value θknkreal. When the ALCEST is lower than the actual alcohol concentration and larger than the actual alcohol concentration, the auto-ignition timing predicted value θknkest is largely shifted to the retard side from the auto-ignition timing detection value θknkreal. Therefore, when the predicted self-ignition timing θknkest is greatly shifted to the retarded side from the self-ignition timing detection value θknkreal (Δθ is large), the alcohol concentration estimated value per update amount is increased accordingly and the alcohol The concentration estimated value ALCEST is converged earlier.

  In step 302, the auto-ignition timing predicted value θknquest is calculated again using the alcohol concentration estimated value ALLCEST updated in step 301. This self-ignition timing predicted value θknkest is calculated for the second time (the self-ignition timing predicted value θknkest is calculated for the first time in step 218 in FIG. 33), and this second self-ignition timing predicted value θknkest is calculated for the first time. It is overwritten on the predicted ignition timing value θknquest.

  The calculation of the second and subsequent self-ignition timing prediction values θknquest will be described with reference to the flowcharts of FIGS. 46 and 47. The calculation process of the predicted self-ignition timing θknquest for the second and subsequent times is based on the calculation process of the predicted self-ignition timing in FIGS. 43 and 33. Accordingly, in FIGS. 46 and 47, the same parts as those in FIGS. 43 and 33 are used. Are given the same step number.

  43 and FIG. 33 are only the steps 281 and 311 in FIG. That is, when the estimated alcohol concentration value ALLCEST updated in step 301 in FIG. 45 is used in step 293 in FIG. 46, when the integrated value SUM of 1 / τ becomes 1 or more in step 210 in FIG. Proceeding to step 281 of 47, the crank angle θ at that time is entered into the predicted self-ignition timing θknquest [degBTDC] to calculate the second predicted self-ignition timing θknest.

  The self-ignition timing predicted value θknkest calculated in the second time in this way is closer to the self-ignition timing detection value θknkreal than the self-ignition timing predicted value θknkest calculated in the first time in step 218 in FIG. Returning to step 263, the ignition timing difference Δθ is calculated again using the self-ignition timing predicted value θknkest calculated for the second time. In step 264, the ignition timing difference Δθ is compared with a predetermined value, and the ignition timing difference Δθ is calculated. When it is equal to or greater than the predetermined value, steps 301, 302, 263, and 264 are repeated. When the ignition timing difference Δθ calculated using the self-ignition timing predicted value θknkest calculated several times in due to this repetition has fallen below the predetermined value, the process proceeds from step 264 in FIG. The process ends.

  In this way, when knocking is detected, the operation of repeatedly updating the alcohol concentration estimated value ALCEST until the ignition timing difference Δ falls below a predetermined value (loop operation of steps 263, 264, 301, 302 in FIG. 45) is as follows. Therefore, when the combustion cycle in which knocking has been detected is completed, the alcohol concentration estimated value ALLCEST has converged.

  On the other hand, when the crank angle θ exceeds the predetermined value const11 without the integrated value SUM of 1 / τ reaching 1 in FIG. 46, the process proceeds from step 294 in FIG. 46 to step 311 in FIG. 47 to predict the predetermined ignition timing const16. After moving to the value θknest, the current process is terminated. As the predetermined value const16, a value (a value on the retard side) that is equal to or larger than the predetermined value const11 in step 294 of FIG.

  Thus, in the second embodiment, when it is detected that knocking has occurred, the alcohol concentration estimated value ALCEST is increased until the ignition timing difference Δθ falls within the allowable range in the combustion cycle in which the knocking is detected. is doing.

  On the other hand, when knocking is not detected in FIG. 45, when the ignition timing minimum value PADV does not coincide with the basic ignition timing MBTCAL and the counter value count is equal to or greater than the predetermined value const14, steps 261, 267, and 303 are followed by step 304. Then, the alcohol concentration estimated value ALCEST is lowered by the second predetermined value const15. That is, the alcohol concentration estimated value ALEST is updated by the following equation.

ALLCEST (new) = ALCEST (old) -const15 (83)
However, ALCEST (new): Estimated alcohol concentration after update,
ALCEST (old): Estimated alcohol concentration before update,
const15: update amount to the lower side,
Also, in steps 294 and 295 in FIG. 43 and step 303 in FIG. 45, the predetermined values const11, const12, and const14 are set to values different from those in the first embodiment. These predetermined values const11, const12, and const14 are matched by a preliminary experiment or the like in advance. However, the predetermined values const11 and const12 may be the same as the predetermined values const01 and const02 of the first embodiment.

  FIG. 48 shows the movement of the ignition timing, the counter value count, and the alcohol concentration estimated value ALCEST in the second embodiment as a model. In FIG. 48, the alternate long and short dash line indicates the case of the second embodiment, and the solid line indicates the case of Comparative Example 2.

Here, “Comparative Example 2” does not include Steps 262, 263, 264, and 302 in FIG. 45. In Step 301, instead of the above equation (82),
ALLCEST (new) = OCTEST (old) + const13
... (Supplement 20)
However, ALCEST (new): Estimated alcohol concentration after update,
ALCEST (old): Estimated alcohol concentration before update,
const13: update amount to the side to be increased,
This is a case where the alcohol concentration estimated value ALEST is updated by the following formula, and after this operation, the flow proceeds to “END” to end the current process. In the case of Comparative Example 2, when knocking occurs at the timing of t11 as shown in FIG. 48, it is determined that the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration, and the alcohol concentration estimated value ALCEST is stepwise. It increases by a predetermined value const13 of 1. As a result, if knocking does not occur, every time the counter value count reaches the predetermined value const14, the alcohol concentration estimated value ALCEST decreases by the second predetermined value const15. Then, after the ignition timing minimum value PADV coincides with MBTCAL at the timing of t12, the update of the alcohol concentration estimated value ALCEST is stopped and the value at that time is held. If knocking occurs again at the timing of t13 thereafter, the above operation is repeated.

  As described above, according to the second comparative example, the alcohol concentration estimated value ALCEST is updated by the predetermined value const13 every time a knock is detected, so that the alcohol concentration estimated value ALTEST converges when no knock is detected. It will be.

  Further, according to Comparative Example 2, when the predicted self-ignition timing θknkest is slightly out of the actual self-ignition timing, and when the predicted self-ignition timing θknkes is retarded from the actual self-ignition timing. Therefore, in each case, the renewal amount per time (= const13) of the alcohol concentration estimated value is the same, so that the autoignition timing predicted value θknest is the actual value. The convergence of the estimated alcohol concentration value ALLCEST is delayed when it is greatly deviated from the self ignition timing to the retard side.

  On the other hand, according to the second embodiment, when knock is detected at the timing of t11, the self-ignition timing at that time is detected, and the self-ignition timing detection value θknkreal and the self-ignition timing predicted value θknkest are A value obtained by multiplying the first predetermined value const13 by the ignition timing difference Δθ, which is a difference, is an update amount per one time of the alcohol concentration estimated value, so that the alcohol concentration estimated value ALLCEST is lower than the actual alcohol concentration and the alcohol concentration When the estimated ignition time value ALCEST deviates significantly from the actual alcohol concentration, and the predicted self-ignition timing value θknkest deviates far from the detected self-ignition timing value θknkreal, the estimated alcohol concentration value ALLCEST is Lower than the concentration and the alcohol concentration estimate ALSEC is the actual alcohol Updating amount per one time of the alcohol concentration estimated value than if the ignition timing predicted value θknkest does not deviate little retarded than the self-ignition timing detected value θknkreal increases by not less deviated from degrees.

In addition, since the update of the alcohol concentration estimated value ALCEST is repeated many times until the ignition timing difference Δθ falls within the allowable range, the alcohol concentration estimated value ALLCEST converges in the combustion cycle in which the knock is detected. Therefore, the convergence value of the alcohol concentration estimated value ALCEST is larger than that in the case of Comparative Example 2 (see the dashed line in the lower part of FIG. 48).
As described above, according to the second embodiment, when a knock is detected, the alcohol concentration estimated value ALLCEST is lower than the actual alcohol concentration and the alcohol concentration estimated value ALLCEST is larger than the actual alcohol concentration. Since the concentration estimated value ALCEST converges quickly and converges in the combustion cycle in which the knock is detected, the second knock is detected after the ignition timing minimum value PADV coincides with the basic ignition timing MBTCAL (after t12). There is no.

  The estimated alcohol concentration value ALEST calculated in FIG. 45 in this way is used to calculate 1 / τ in the mixed fuel of the estimated alcohol concentration value ALLCEST in step 293 of FIG.

  According to the second embodiment, when the fuel is a mixed fuel of gasoline and alcohol, the knock detection result by the knock sensor 47 is fed back to the alcohol concentration in the mixed fuel instead of the ignition timing (see FIG. 45). That is, as indicated by the one-dot chain line at the bottom of FIG. 48, the alcohol concentration estimated value ALCEST increases stepwise when it is detected that knocking has occurred, and thereafter the second predetermined value const15 is incremented by a predetermined period. It is getting smaller gradually.

  Thus, according to the second embodiment (the invention described in claim 1), the alcohol concentration estimated value ALLCEST is calculated based on the knock detection result by the knock sensor 47 (see steps 261, 301, and 304 in FIG. 45). ) Based on the estimated alcohol concentration value ALLCEST, the auto-ignition timing predicted value θknest (knock occurrence timing predicted value) in the combustion chamber 5 is calculated (steps 291 to 293, 209, 210 in FIG. 43, steps in FIG. 33). 218), the knock limit ignition timing KNOCKcal is calculated on the basis of the predicted self-ignition timing value θknquest (see steps 219 to 231 in FIG. 33), so that the alcohol concentration in the mixed fuel is not sold in advance. Even when using a mixed fuel, the knock detection result is input to the ignition timing regardless of the operating conditions. Unlike the conventional device that performs back-back, the ignition timing delay for avoiding knock and the subsequent advance operation are not repeated. For this reason, since the knock limit ignition timing can be followed even in a transition as in the first embodiment, it is possible to prevent fuel consumption deterioration and output reduction.

  Further, when the alcohol concentration estimated value ALCEST is updated (calculated), as shown in the above equation (82), the self-ignition timing detection value θknkreal (knock generation timing detection value) and the self-ignition timing prediction value θknkest (knock generation timing prediction value) ) Ignition timing difference Δθ (= θknkreal−θknkest) is also taken into consideration. That is, when the alcohol concentration estimated value ALCTEST is lower than the actual alcohol concentration and greatly deviated from the actual alcohol concentration (that is, the self-ignition timing predicted value θknkest is greatly deviated from the self-ignition timing detection value θknkreal to the retard side). The alcohol concentration estimation is performed when the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration and close to the actual alcohol concentration (that is, the predicted self-ignition timing value θknkest is not significantly shifted to the retarded side from the detected self-ignition timing value θknkreal). The amount of update per value increases, and the estimated alcohol concentration value ALLCEST is lower than the actual alcohol concentration when the estimated alcohol concentration value ALLCEST is lower than the actual alcohol concentration and larger than the actual alcohol concentration. And the alcohol concentration estimated value ALCEST converges more quickly than when the actual close to the alcohol concentration.

  Thus, according to the second embodiment (the invention described in claim 1), the alcohol concentration is also based on the ignition timing difference Δθ (comparison result of the auto-ignition timing detection value θknkreal and the auto-ignition timing prediction value θknkest). Since the estimated value ALCEST (knock correlation parameter estimated value) is calculated (see step 301 in FIG. 45), the convergence of the alcohol concentration estimated value ALEST can be accelerated, and the drivability is improved.

  According to the second embodiment (the invention according to claim 6), in the mixed fuel of gasoline and alcohol, the alcohol concentration estimated value ALCEST is knocked in response to the influence of the alcohol concentration in the mixed fuel on the knocking. Since it is calculated based on the detection result and the ignition timing difference Δθ (comparison result of the detected knock timing and the predicted knock timing) (see steps 261, 262, 263, and 301 in FIG. 45), the alcohol concentration is known in advance. Even when an alcohol-containing fuel that cannot be obtained is used as a fuel to be used, it is possible to accurately calculate the auto-ignition timing predicted value θknest (knock occurrence timing predicted value).

  According to the second embodiment (the invention described in claim 7), when the alcohol concentration estimated value calculation means detects that a knock has occurred, the self-ignition timing detection value θknkreal (in the combustion cycle in which the knock is detected) The estimated alcohol concentration value ALCEST is updated to a higher value until the ignition timing difference Δθ (occurrence timing difference) between the knock occurrence timing detection value) and the auto-ignition timing predicted value θknkes (predicted knock occurrence timing) falls within the allowable range. (Refer to the loop operation in steps 263, 264, 301, and 302 in FIG. 45), and thereafter, the alcohol concentration estimated value ALCEST is updated at a constant cycle to the lower side by the second predetermined value const15 (step 261 in FIG. 45, 267, 303, 269, 304, and 271), the arc is detected in the combustion cycle in which knock is detected. It is possible to converge the Le density estimate Alcest.

  According to the second embodiment (the invention described in claim 8), the alcohol concentration estimated value ALCEST is updated to a side that is lowered by a second predetermined value const15 (a side where knocking occurs) (see step 304 in FIG. 45). ) Is executed under the condition that knocking occurs at the basic ignition timing MBTCAL (step 267 in FIG. 45), so that the alcohol concentration estimated value ALLCEST is not erroneously updated.

  The flow of FIGS. 49 and 50 is the third embodiment, FIG. 49 replaces FIG. 32 of the first embodiment, and FIG. 50 replaces FIG. 38 of the first embodiment. 49, the same step number is assigned to the same part as FIG. 32, and the same step number is assigned to the same part in FIG. FIG. 33 is a flow common to the first, second, and third embodiments.

  Both the octane number of the fuel described in the first embodiment and the alcohol concentration in the mixed fuel described in the second embodiment are parameters having a correlation with knock, but the parameters having a correlation with knock are limited to these. There is no compression ratio. When using a fuel with a predetermined octane number, the compression ratio is determined in advance by the engine specifications, so the base ignition timing must be matched so that knock does not occur even at the compression ratio determined by the engine specifications. Become. For this reason, knocking occurs when the actual compression ratio becomes higher than the compression ratio of the engine specification for some reason. On the other hand, the ignition timing retarded to avoid knocking and the subsequent advance operation are repeated. Doing so will result in poor fuel consumption and reduced output.

  As shown in FIGS. 5, 10, 12, and 13, the third embodiment has a laminar combustion velocity (SL 1, SL 2), a combustion gas volume equivalent volume (V 0, VTDC), and a combustion mass ratio (BR 1, BR 2). ) And the reaction probability RPROBA, a combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated, and a basic ignition timing MBTCAL from which MBT is obtained is calculated based on the combustion period (BURN1, BURN2). The combustion period is divided into an initial combustion period BURN1 and a main combustion period BURN2, and among these, the initial combustion period BURN1 corresponds to the volume V0 at the compression start timing of the combustion chamber 5 corresponding to the combustion gas volume in the combustion chamber 5. Fuel that uses gasoline with a predetermined octane number (e.g. 80), assuming that the volume is calculated In this case, a compression ratio estimated value CMPEST, which is a knock correlation parameter, is calculated based on the knock detection result from the knock sensor 47, and the volume V0 of the combustion chamber 5 at the compression start timing is calculated based on the compression ratio estimated value CMPEST. To do.

  The difference from the first embodiment will be mainly described. In FIG. 49, in step 321, the map is searched for from the compression start temperature TC0 and the compression start pressure PC0 for the first time, and will be described later. Regardless of the estimated compression ratio CMPEST, 1 / τ for the fuel having an octane number of 80 is calculated, and the calculated 1 / τ for the fuel having an octane number of 80 is integrated to the integrated value SUM in step 209.

  Next, in FIG. 50, when it is detected at step 261 that knocking has occurred by the knock sensor 47, it is determined that the compression ratio estimated value CMPALCEST is smaller than the actual compression ratio and the operations of steps 262 and 263 are performed. In step 264, the absolute value of the ignition timing difference Δθ is compared with a predetermined value. When the absolute value of the ignition timing difference Δθ is greater than or equal to a predetermined value, the routine proceeds to step 331, where the compression ratio estimated value CMPEST is increased by a value obtained by multiplying the first predetermined value const23 by the ignition timing difference Δθ. That is, the compression ratio estimated value CMPALCEST is updated by the following equation.

CMPEST (new) = CMPEST (old) + const23 × Δθ (84)
Where CMPEST (new): updated compression ratio estimate,
CMPEST (old): estimated compression ratio before update,
const23: update rate to the side to be increased (positive unnamed number),
Here, the second term of the equation (82) defines the amount of update per one time of the estimated compression ratio. The reason why the ignition timing difference Δθ is introduced into the update amount per time is to accelerate the convergence of the compression ratio estimated value CMPEST. That is, when the compression ratio estimated value CMPEST is smaller than the actual compression ratio and close to the actual compression ratio, the self-ignition timing predicted value θknkest does not deviate so much from the self-ignition timing detected value θknkreal, but the compression ratio estimated value CMPEST. Is smaller than the actual compression ratio and larger than the actual compression ratio, the self-ignition timing predicted value θknkest is greatly deviated from the self-ignition timing detection value θknkreal to the retard side. Therefore, when the predicted self-ignition timing value θknkest is greatly deviated from the self-ignition timing detected value θknkreal (Δθ is larger), the compression ratio estimation value is updated by increasing the amount of updating per time accordingly. The convergence of the ratio estimated value CMPEST is accelerated.

  In step 332, the auto-ignition timing predicted value θknest is calculated again using the updated compression ratio estimated value CMPEST. This self-ignition timing predicted value θknkest is calculated for the second time (the self-ignition timing predicted value θknkest is calculated for the first time in step 218 in FIG. 33), and this second self-ignition timing predicted value θknkest is calculated for the first time. It is overwritten on the predicted ignition timing value θknquest.

  Calculation of the second and subsequent self-ignition timing predicted values θknquest will be described with reference to the flowcharts of FIGS. 51 and 52. The calculation process of the predicted self-ignition timing θknquest for the second and subsequent times is based on the calculation process of the predicted self-ignition timing in FIGS. 49 and 33. Accordingly, in FIGS. 51 and 52, the same parts as those in FIGS. 49 and 33 are used. Are given the same step number.

  49 and FIG. 33 are only the steps 281 and 341 in FIG. That is, when the integrated value SUM of 1 / τ becomes 1 or more in step 210 in FIG. 51, the process proceeds to step 281 in FIG. 52, and the crank angle θ at that time is set to the autoignition timing predicted value θknkes [degBTDC]. The predicted self-ignition time θknkes for the second time is calculated.

  Since the self-ignition timing predicted value θknkest calculated in the second time in this way is closer to the self-ignition timing detection value θknkreal than the self-ignition timing predicted value θknkest calculated in step 218 in FIG. Returning to step 263, the ignition timing difference Δθ is calculated again using the self-ignition timing predicted value θknkest calculated for the second time. In step 264, the ignition timing difference Δθ is compared with a predetermined value, and the ignition timing difference Δθ is calculated. When it is equal to or greater than the predetermined value, steps 331, 332, 263, and 264 are repeated. When the ignition timing difference Δθ calculated using the self-ignition timing predicted value θknkest calculated several times in due to this repetition falls below a predetermined value, the process proceeds from step 264 in FIG. 50 to “END”. The process ends.

  As described above, when the knock is detected, the operation of repeatedly updating the compression ratio estimated value CMPEST until the ignition timing difference Δ falls below the predetermined value (the loop operation of steps 263, 264, 331, and 332 in FIG. 50) is as follows. Therefore, the compression ratio estimated value CMPEST has converged when the combustion cycle in which knocking has been detected is completed.

  On the other hand, when the crank angle θ exceeds the predetermined value const21 without the integrated value SUM of 1 / τ reaching 1 in FIG. 51, the routine proceeds from step 322 in FIG. 51 to step 341 in FIG. 52, and the predetermined value const26 is predicted as the auto-ignition timing. After moving to the value θknest, the current process is terminated. As the predetermined value const26, a value equal to or larger than the predetermined value const21 in step 322 in FIG.

  Thus, in the third embodiment, when it is detected that knocking has occurred, the compression ratio estimated value CMPEST is increased until the ignition timing difference Δθ falls within the allowable range in the combustion cycle in which the knocking is detected. ing.

  On the other hand, when knocking is not detected in FIG. 50, when ignition timing minimum value PADV does not coincide with basic ignition timing MBTCAL and counter value count becomes equal to or greater than predetermined value const24, steps 261, 267, and 333 are followed by step 334. Then, the compression ratio estimated value CMPEST is decreased by the second predetermined value const25. That is, the compression ratio estimated value CMPEST is updated by the following equation.

CMPEST (new) = CMPEST (old) −const25 (85)
Where CMPEST (new): updated compression ratio estimate,
CMPEST (old): estimated compression ratio before update,
const25: update amount to the side to be reduced,
Further, in steps 322 and 323 in FIG. 49 and in step 333 in FIG. 50, the predetermined values const21, const22, and const24 are set to values different from those in the first embodiment. These predetermined values const21, const22, and const24 are matched by a preliminary experiment or the like in advance. Among these, the predetermined values const21 and const22 may be the same as the predetermined values const01 and const02 of the first embodiment.

  In the third embodiment, the volume VIVC of the combustion chamber 5 at the intake valve closing timing and the volume V0 of the combustion chamber 5 at the combustion start timing (MBTCYCL) are calculated based on the compression ratio estimated value CMPEST calculated in this way. To do. This will be described with reference to the flowcharts of FIGS.

  The flow of FIGS. 53 and 54 replaces FIGS. 5 and 10 of the first embodiment. 53, the same step number is assigned to the same part as in FIG. 5, and the same step number in FIG. 54 is assigned to the same part as in FIG.

The difference from the first embodiment will be mainly described. In FIG. 53, in step 351, the gap volume Vc [m ³ ] is calculated by the following equation.

Vc = {1 / (CMPEST-1)} × (π / 4) D ² · Hx (86)
Where CMPEST: estimated compression ratio,
D: cylinder bore diameter [m],
Hx: difference between the maximum value and the minimum value of the distance from the TDC of the piston pin 76 [m]
,
Here, the equation (86) replaces the above equation (3) of the first embodiment. In the first embodiment, the compression ratio ε in the equation (3) is assumed to be constant, but in the third embodiment, the compression ratio estimated value CMPEST as a variable value is used.

  Using the clearance volume Vc thus determined, in step 352, the volume VIVC of the combustion chamber 5 at the intake valve closing timing is calculated by the following equation.

VIVC = Vc + (π / 4) D ² · Hivc (87)
D: cylinder bore diameter [m],
Hivc: Distance of piston pin 76 from TDC at intake valve closing timing
[M],
The expression (87) is the same as the expression (2) in the first embodiment.

Next, in FIG. 54, in step 361, the clearance volume Vc [m ³ ] is calculated by the above equation (86), and this clearance volume Vc is used, and in step 362, in the combustion start timing (MBTCYCL) of the combustion chamber 5 by the following equation. The volume V0 is calculated.

V0 = Vc + (π / 4) D ² · Hmbtcycl (88)
Where D: cylinder bore diameter [m]
Hmbtcycl: Piston pin at the start of combustion (MBTCYCL)
Distance from 76 TDCs [m],
According to the third embodiment (the invention described in claim 10), when using a fuel having a predetermined octane number, in this case, a fuel having an octane number of 80, the knock detection result by the knock sensor 47 is compressed instead of the ignition timing. The ratio is fed back (see FIG. 50). Therefore, the compression ratio estimated value CMPEST moves in the same manner as the alcohol concentration estimated value ALEST shown by the one-dot chain line at the bottom of FIG. That is, when knocking occurs, it is determined that the compression ratio estimated value CMPEST is lower than the actual compression ratio, and the compression ratio estimated value CMPEST is increased stepwise. As a result, if knocking does not occur, every time the counter value count reaches the predetermined value const24, the compression ratio estimated value CMPEST decreases by a second predetermined value const25 at a constant cycle. Then, after the ignition timing minimum value PADV coincides with MBTCAL, the update of the compression ratio estimated value CMPEST is stopped and the value at that time is held. Based on the compression ratio estimated value CMPEST that changes in this way, the volume VIVC of the combustion chamber 5 at the intake valve closing timing and the volume V0 of the combustion chamber 5 at the combustion start timing (MBTCYCL) are calculated.

  Thus, according to the third embodiment (the invention described in claim 10), the compression ratio estimated value CMPEST is calculated based on the knock detection result by the knock sensor 47 (see steps 261, 331, and 334 in FIG. 50). ) Since the volume V0 at the combustion start timing of the combustion chamber 5 used for calculating the initial combustion period BURN1 is calculated based on the compression ratio estimated value CMPEST (see steps 361 and 362 in FIG. 54), the fuel having an octane number of 80 Even when the actual compression ratio becomes higher than the compression ratio of the engine specification for some reason when using (a fuel with a predetermined octane number), the knock detection result is fed back to the ignition timing regardless of the operating state. Like the conventional device, the ignition timing delay to avoid knocking and the subsequent advance operation are repeated. No. For this reason, as in the first and second embodiments, the knock limit ignition timing can be followed even in a transient state, so that fuel consumption deterioration and output reduction can be prevented.

  Further, according to the third embodiment (the invention described in claim 45), it is detected whether or not knock is actually generated in the combustion chamber, and the knock generation timing in the combustion chamber is detected, and this knock generation timing is detected. The detected value θknkreal is compared with the knock occurrence timing predicted value θknkest, the compression ratio estimated value CMPEST is calculated based on the comparison result and the knock detection result, and the combustion chamber combustion is calculated based on the compression ratio estimated value CMPEST. A volume V0 at the start time is calculated, a combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated based on the volume V0 at the start time of combustion, and MBT is calculated based on the combustion period (BURN1, BURN2). Is calculated at the basic ignition timing MBTCAL, and spark ignition is performed at this basic ignition timing MBTCAL. When using a fuel with a fixed Kutan number or a mixed fuel with a fixed alcohol concentration, the compression ratio estimated value CMPEST will converge even if the actual compression ratio becomes higher than the compression ratio of the engine specification for some reason. The basic ignition timing with which MBT can be obtained can be given with high accuracy while being advanced.

  Further, when the compression ratio estimated value CMPEST is updated (calculated), as shown in the above equation (84), the self-ignition timing detection value θknkreal (knock occurrence timing detection value) and the self-ignition timing prediction value θknkest (knock occurrence timing prediction value). ) Ignition timing difference Δθ (= θknkreal−θknkest) is also taken into consideration. That is, when the self-ignition timing predicted value θknkest is greatly deviated from the self-ignition timing detection value θknkreal (Δθ is large), the self-ignition timing prediction value θknkest is much less retarded than the self-ignition timing detection value θknkreal. The amount of update per one compression ratio estimation value is larger than when there is no deviation (Δθ is small), and the self-ignition timing predicted value θknkest is larger than the self-ignition timing detection value θknkreal on the retarded side. The compression ratio estimated value CMPEST converges faster than when the self-ignition timing predicted value θknkest is not shifted far from the self-ignition timing detected value θknkreal.

  Thus, according to the third embodiment (the invention described in claim 10), the compression ratio is also based on the ignition timing difference Δθ (comparison result between the self-ignition timing detection value θknkreal and the self-ignition timing prediction value θknkest). Since the estimated value CMPEST (knock correlation parameter estimated value) is calculated (see step 331 in FIG. 50), the convergence of the compression ratio estimated value CMPEST can be accelerated, and the drivability is improved.

  According to the third embodiment (the invention described in claims 12 and 47), when the compression ratio estimated value calculating means detects that a knock has occurred, the auto-ignition timing detection value in the combustion cycle in which the knock has been detected. The compression ratio estimated value CMPEST is increased until the ignition timing difference Δθ (occurrence timing difference) between θknkreal (knock occurrence timing detection value) and the self-ignition timing predicted value θknest (knock occurrence timing prediction value) falls within an allowable range. (Refer to the loop operation of steps 263, 264, 331, and 332 in FIG. 50), and thereafter, the compression ratio estimated value CMPEST is updated at a constant cycle so as to decrease by the second predetermined value const25 (step in FIG. 50). 261, 267, 333, 269, 334, and 271), the compression ratio estimated value CMP in the combustion cycle in which knocking is detected. It is possible to converge the ST.

  According to the third embodiment (the inventions described in claims 13 and 48), the compression ratio estimated value CMPEST is updated to the side where the second predetermined value const25 is decreased (the side where knocking occurs) (step 334 in FIG. 50). (See) is executed under the condition that knocking occurs at the basic ignition timing MBTCAL (step 267 in FIG. 50), so that the compression ratio estimated value CMPEST is not erroneously updated.

  According to the third embodiment (the invention described in claims 14 and 49), as shown in FIG. 54, the intake valve is based on the volume VIVC of the combustion chamber 5 at the intake valve closing timing and the volume V0 at the combustion start timing. An effective compression ratio Ec from the closing timing IVC to the combustion start timing is calculated (see step 163 in FIG. 54), and combustion starts in the combustion chamber 5 from the temperature TINI at the intake valve closing timing of the combustion chamber 5 and the effective compression ratio Ec. The temperature T0 at the timing is calculated from the pressure PINI at the intake valve closing timing of the combustion chamber 5 and the effective compression ratio Ec, respectively, and the pressure P0 at the combustion start timing of the combustion chamber 5 is calculated (see steps 164 to 167 in FIG. 54). A laminar combustion speed SL1 used for calculating the initial combustion period BURN1 is calculated based on the temperature T0 and the pressure P0 at the combustion start timing of the combustion chamber 5 (step 168 in FIG. 54). In this case, since the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve is calculated based on the estimated compression ratio CMPEST (steps 351 and 352 in FIG. 53), the fuel having a predetermined octane number is used. Even when the actual compression ratio becomes higher than the scheduled compression ratio for some reason, the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve can be calculated with high accuracy.

  In the third embodiment, the self-ignition timing predicted value θknkest (knock occurrence timing predicted value) has been described based on a characteristic that represents the distribution of the reciprocal of the time until the fuel in the combustion chamber reaches self-ignition. Instead of the self-ignition timing predicted value θknkest, the self-ignition timing detection value θknkreal (knock occurrence timing detection value) may be used (the invention according to claims 11 and 46).

  The flow of FIG. 55, FIG. 59, and FIG. 61 is the fourth, fifth, and sixth embodiments, respectively, FIG. 55 is FIG. 38 of the first embodiment, FIG. 61 replaces FIG. 50 of the third embodiment. 55, the same step number is assigned to the same portion as FIG. 38, the same step number is assigned to the same portion as FIG. 45 in FIG. 59, and the same step number is assigned to the same portion as FIG. ing.

  In the first to third embodiments, when it is detected that knocking has occurred, the difference between the self-ignition timing detection value θknkreal and the self-ignition timing prediction value θknkest in the combustion cycle in which knocking is detected. The knock correlation parameter was changed until the ignition timing difference Δθ was within the allowable range (the octane number estimated value OCTEST was decreased for the first embodiment, the alcohol concentration estimated value ALLCEST was increased for the second embodiment, In the third embodiment, the compression ratio estimated value CMPEST is increased), and in the fourth to sixth embodiments, when it is detected that knocking has occurred, the combustion cycle in which knocking is detected is detected. , The knock strength difference ΔKIC, which is the difference between the knock strength estimated value KICEEST and the knock strength detected value KICREAL, is an allowable range. The knock correlation parameters (the estimated octane value for the fourth embodiment, the estimated alcohol concentration for the fifth embodiment, and the estimated compression ratio for the sixth embodiment) are changed until they fall within the range.

  In the fourth embodiment shown in FIG. 55, the parts different from the first embodiment shown in FIG. 38 are steps 371 to 375, and in the fifth embodiment shown in FIG. 59, the parts different from the second embodiment shown in FIG. In the sixth embodiment shown in FIG. 61, steps 371 to 373, 391, and 392 are different from the third embodiment shown in FIG. That is, the actual knock strength is read in step 371 common to the fourth, fifth, and sixth embodiments. In order to obtain the actual knock intensity, the area above the threshold value in FIG. 56 showing the characteristics of the frequency component for the knock may be obtained by integration. The integral value thus obtained is multiplied by the correlation coefficient 2 to be converted into a knock intensity, and the converted value is read as a knock intensity detected value KICREAL. The characteristic diagram of FIG. 56 is the same as the characteristic diagram of FIG.

  In the fourth embodiment, the loop operation in steps 372 to 375 in FIG. 55 is a part for converging the octane number estimated value OCTEST without causing knock so that the knock magnitude difference ΔKIC falls within the allowable range. In the fifth embodiment, FIG. The loop operation of steps 372, 373, 381, and 382 of FIG. 61 converges the alcohol concentration estimated value ALCEST without causing knock so that the knock intensity difference ΔKIC falls within the allowable range. The loop operation of steps 372, 373, 391, and 392 is a part for converging the compression ratio estimated value CMPEST without causing knock so that the knock magnitude difference ΔKIC falls within the allowable range. First, at step 372 common to the fourth, fifth, and sixth embodiments, the knock magnitude difference ΔKIC is calculated by the following equation.

ΔKIC = KICEST−KICREAL (89)
Similarly, in step 373 common to the fourth, fifth, and sixth embodiments, the absolute value of the knock intensity difference ΔKIC is compared with a predetermined value const6. The predetermined value const6 defines an allowable range. If the absolute value of the knock intensity difference ΔKIC is less than the predetermined value const6, the knock intensity difference ΔKIC is within the allowable range, and the estimated octane number, alcohol concentration estimated value, and compression ratio estimated value. It is determined that knocking has occurred due to factors other than these errors, and the current process is terminated.

  When the absolute value of the knock magnitude difference ΔKIC is greater than or equal to a predetermined value, the process proceeds to step 374 in FIG. 55 in the fourth embodiment, to step 381 in FIG. 59 in the fifth embodiment, and to FIG. 61 in the sixth embodiment. Proceed to step 391. That is, when the knock magnitude estimated value KICETEST is larger than the knock magnitude detected value KICREAL, it means that the knock magnitude estimated value is too large. In the fourth embodiment, the octane number estimated value OCTEST is set to the first octane number estimated value OCTEST in step 374 of FIG. Is reduced by a value obtained by multiplying the predetermined value const03 by the knock intensity difference ΔKIC. That is, the estimated octane value OCTEST is updated by the following equation.

OCTEST (new) = OCTEST (old) −const03 × ΔKIC (90)
However, OCTEST (new): Estimated value of octane number after update,
OCTEST (old): Estimated octane value before update,
const03: update rate to the side to be reduced (positive anonymous number),
Similarly, in the fifth embodiment, in step 381 of FIG. 59, the alcohol concentration estimated value ALCEST is increased by a value obtained by multiplying the first predetermined value const13 by the knock intensity difference ΔKIC. That is, the alcohol concentration estimated value ALEST is updated by the following equation.

ALLCEST (new) = ALCEST (old) + const13 × ΔKIC (91)
However, ALCEST (new): Estimated alcohol concentration after update,
ALCEST (old): Estimated alcohol concentration before update,
const13: update rate (positive unnamed number) to the increase side,
In the sixth embodiment, the compression ratio estimated value CMPEST is increased by a value obtained by multiplying the first predetermined value const23 by the knock magnitude difference ΔKIC in step 391 of FIG. That is, the compression ratio estimated value CMPEST is updated by the following equation.

CMPEST (new) = CMPEST (old) + const23 × ΔKIC (92)
Where CMPEST (new): updated compression ratio estimate,
CMPEST (old): estimated compression ratio before update,
const23: update rate (positive unnamed number) to the larger side,
Here, the second term on the right side of the equation (90) defines the update amount per one time of the estimated octane number. The reason why the knock intensity difference ΔKIC is introduced into the update amount per time is to accelerate the convergence of the estimated octane value OCTEST. That is, when the estimated octane number OCTEST is larger than the actual octane number and close to the actual octane number, the knock magnitude estimated value KICEST does not deviate much to the side larger than the knock magnitude detected value KICREAL, but the octane number estimated value OCTEST is larger than the actual octane number and When the deviation is larger than the actual octane number, the knock magnitude estimated value KICEEST is greatly displaced to the side larger than the knock magnitude detected value KICREAL. Therefore, when the knock magnitude estimated value KICETEST is greatly deviated to the larger side than the knock magnitude detected value KICREAL (that is, ΔKIC is large), the update amount per one is increased accordingly, and the convergence of the octane number estimated value OCTEST is accelerated. It is what I did. For the same reason, the first predetermined value const13 is multiplied by the knock intensity difference ΔKIC in the equation (91) and the first predetermined value const23 in the equation (92).

  In the fourth embodiment, in step 375 of FIG. 55, the estimated octane number OCTEST updated in the immediately preceding step 374 is used. In the fifth embodiment, in step 382 of FIG. 59, the estimated alcohol concentration updated in the immediately preceding step 381. In the sixth embodiment, in step 392 in FIG. 61, the knock strength estimated value KICEEST is calculated again using the compression ratio estimated value CMPEST updated in the immediately preceding step 391 in the sixth embodiment. Each calculation of these knock strength estimation values KICEEST is the second time (the calculation of the knock strength estimation value KICEST in step 229 in FIG. 33 is the first calculation), and this second knock strength estimation value KICEEST is calculated as the first knock strength estimation. Overwrite the value KICEEST.

  Regarding the calculation of the knock magnitude estimated value KICEEST for the second and subsequent times, FIG. 57 and FIG. 58 (subroutine of step 375 in FIG. 55) for the fourth embodiment, and FIG. 60 and FIG. 58 (FIG. 59) for the fifth embodiment. The sixth embodiment will be described with reference to the flowchart in FIG. 62 and FIG. 58 (subroutine in step 392 in FIG. 61). The calculation process of the knock intensity estimated value KICEEST for the second and subsequent times is the calculation process of the knock intensity estimated value in FIGS. 32 and 33 of the first embodiment in the fourth embodiment, and the second embodiment in the fifth embodiment. 43 and 33 in FIG. 43, the calculation processing of the knock strength estimation value in FIG. 49 and FIG. 33 of the third embodiment is diverted in the sixth embodiment. In FIGS. 57 and 58 of the embodiment, the same parts as FIGS. 32 and 33 of the first embodiment, and in FIGS. 60 and 58 of the fifth embodiment, the same parts as FIGS. 43 and 33 of the second embodiment, 62 and 58 of the sixth embodiment, the same step numbers are assigned to the same parts as those of FIGS. 49 and 33 of the third embodiment.

  The difference from the calculation process of the knock strength estimated value KICEEST in each of the first, second, and third embodiments is that steps 230, 231 shown in FIG. 33 common to the first, second, and third embodiments are shown in FIG. 232 is not shown in FIG. 58 common to the fourth, fifth, and sixth embodiments, and a step 381 is newly added in FIG. 58 common to the fourth, fifth, and sixth embodiments. Is a point. Accordingly, when the estimated octane number OCTEST updated in step 374 of FIG. 55 in the fourth embodiment is used in step 208 of FIG. 57, the integrated value SUM of 1 / τ becomes 1 or more in step 210 of FIG. When the alcohol concentration estimated value ALEST updated in step 381 in FIG. 59 in the fifth embodiment is used in step 293 in FIG. 60, the integrated value SUM of 1 / τ is 1 or more in step 210 in FIG. FIG. 6 is a diagram common to the fourth, fifth, and sixth embodiments when the integrated value SUM of 1 / τ becomes 1 or more in step 210 of FIG. 62 in the sixth embodiment. Proceeding to step 229 at 58, the knock magnitude estimated value KICEST is calculated.

  In this way, in the fourth embodiment, the knock magnitude estimated value KICEST calculated for the second time is closer to the knock magnitude detected value KICREAL than the knock magnitude estimated value KICEEST calculated for the first time in step 229 of FIG. 55 returns to step 372 in FIG. 55, step 372 in FIG. 59 in the fifth embodiment, and step 372 in FIG. 61 in the sixth embodiment, and the knock strength estimated value KICEEST calculated for the second time is used again. The difference ΔKIC is calculated, and the knock intensity difference ΔKIC is compared with a predetermined value in step 373 of FIGS. 55, 59 and 61. If the knock intensity difference ΔKIC is equal to or larger than the predetermined value, the step of FIG. 55 is performed in the fourth embodiment. 374, 375, 372, and 373, in the fifth embodiment, steps 381, 382, 372, and 373 in FIG. , Also in the sixth embodiment repeating steps 391,392,372,373 of Figure 61. When the knock magnitude difference ΔKIC calculated using the knock magnitude estimated value KICEEST calculated several times by repeating this process falls below a predetermined value, the process proceeds from step 373 in FIGS. 55, 59, and 61 to “END”. Exit this process.

  As described above, when knock is detected in the fourth embodiment, an operation of repeatedly updating the estimated octane number OCTEST until the knock magnitude difference ΔKIC falls below a predetermined value (loop operation in steps 372 to 375 in FIG. 55), fifth When knocking is detected for the embodiment, an operation of repeatedly updating the alcohol concentration value estimated value ALCEST until the knock intensity difference ΔKIC falls below a predetermined value (loop operation of steps 372, 373, 381, and 382 in FIG. 59), and When knock is detected for the sixth embodiment, the operation of repeatedly updating the compression ratio estimated value CMPEST until the knock intensity difference ΔKIC falls below a predetermined value (loop operation of steps 372, 373, 391, and 392 in FIG. 61) is as follows. With plenty of time before the combustion cycle reaches If the combustion cycle in which knocking is detected is completed, the octane number estimated value OCTEST for the fourth embodiment, the alcohol concentration estimated value ALLCEST for the fifth embodiment, and the compression ratio estimated value CMPEST for the sixth embodiment. Have converged.

  On the other hand, in FIG. 57 of the fourth embodiment, when the crank angle θ exceeds the predetermined value const01 without the integrated value SUM of 1 / τ reaching 1, the process proceeds from Step 211 of FIG. 57 to 1 in FIG. 60 of the fifth embodiment. When the crank angle θ exceeds the predetermined value const11 without the integrated value SUM of / τ reaching 1, the integrated value SUM of 1 / τ reaches 1 from step 294 of FIG. 60 and in FIG. 62 of the sixth embodiment. When the crank angle θ exceeds the predetermined value const21, the routine proceeds from step 322 in FIG. 60 to step 381 in FIG. 58 that is common to the fourth, fifth, and sixth embodiments, and the knock magnitude estimated value KICEST = 0. After that, the current process is terminated.

  When it is detected that knocking has occurred in this way, the estimated octane number OCTEST is decreased in the fourth embodiment until the knock intensity difference ΔKIC falls within the allowable range in the combustion cycle in which the knocking is detected. In the fifth embodiment, the alcohol concentration estimated value ALLCEST is increased, and in the sixth embodiment, the compression ratio estimated value CMPST is increased.

  According to the fourth, fifth, and sixth embodiments (inventions of claims 17 and 26), the estimated values of knock correlation parameters such as octane number, alcohol concentration, and compression ratio are calculated based on the knock detection result ( The knock detection result is fed back to the knock correlation parameter (steps 261, 374, and 270 in FIG. 55 for the fourth embodiment, steps 261, 381, and 304 in FIG. 59 for the fifth embodiment, and FIG. 61 for the sixth embodiment in FIG. 61). Steps 261, 391, and 334), and based on the knock correlation parameter estimated value, a self-ignition timing predicted value θknest (knock occurrence timing predicted value) in the combustion chamber 5 is calculated (step of FIG. 32 for the fourth embodiment). 206 to 210, step 218 in FIG. 33, steps 291 to 293 and 20 in FIG. 43 for the fifth embodiment. , 210, step 218 in FIG. 33, and steps 321, 209, and 210 in FIG. 49 for the sixth embodiment, and step 218 in FIG. 33), the knock limit ignition timing KNOCKcal is calculated based on the predicted self-ignition timing θknkest. (Refer to steps 219 to 231 in FIG. 33) As described above, even when using a fuel sold in a market where the octane number of the fuel or the alcohol concentration in the mixed fuel cannot be known in advance, the octane number is determined in advance. Even when the actual compression ratio is higher than the engine-specific compression ratio for some reason when using the existing fuel, knocking is performed as in the conventional device that feeds back the knock detection result to the ignition timing regardless of the operating condition. The ignition timing delay for avoidance and the subsequent advance operation are repeated. No. For this reason, since the knock limit ignition timing can be followed even during a transition such as acceleration or deceleration, fuel consumption deterioration and output reduction can be prevented.

  In addition, when updating (calculating) the estimated octane number OCTEST, alcohol concentration estimated value ALTEST, and compression ratio estimated value CMPEST when knocking is detected, the above equations (90), (91), and (92) are shown. As described above, the knock magnitude difference ΔKIC (= KICEST−KICREAL) between the knock magnitude estimated value KICEEST and the knock magnitude detected value KICREAL is also taken into consideration. That is, for the fourth embodiment, when the estimated octane number OCTEST is larger than the actual octane number and deviates from the actual octane number (that is, the estimated knock intensity value KICETEST is greatly deviated more than the detected knock intensity value KICREAL), The amount of update per octane number estimated value when the estimated octane number OCTEST is larger than the actual octane number and close to the actual octane number (that is, the knock strength estimated value KICEST is not shifted so much to the side larger than the knock strength detected value KICREAL). When the estimated octane number OCTEST is larger than the actual octane number and deviates from the actual octane number, the estimated octane number OCTEST is larger than the actual octane number and closer to the actual octane number. Octane estimate OCTEST to converge faster than the time. In the fifth embodiment, when the alcohol concentration estimated value ALCTEST is lower than the actual alcohol concentration and deviates more than the actual alcohol concentration (that is, the knock strength estimated value KICEEST is greatly deviated to the side larger than the knock strength detected value KICREAL). When the estimated alcohol concentration value ALCEST is lower than the actual alcohol concentration and close to the actual alcohol concentration (that is, the estimated knock strength value KICEST is not much shifted to the side larger than the detected knock strength value KICREAL), When the renewal amount per time becomes large and the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration and deviated more than the actual alcohol concentration, the alcohol concentration estimated value ALTEST is the actual alcohol concentration. Alcohol concentration estimated value ALCEST converges more quickly than when close to and actual alcohol concentration low Ri. In the sixth embodiment, when the knock magnitude estimated value KICEEST is greatly shifted to the side larger than the knock magnitude detected value KICREAL (Δθ is large), the knock magnitude estimated value KICEEST is not shifted much to the side larger than the knock magnitude detected value KICREAL. The amount of update per one compression ratio estimated value is larger than when (Δθ is small), and the knock strength estimated value KICEEST is greater when the knock strength estimated value KICEEST is shifted farther than the knock strength detected value KICREAL. The compression ratio estimated value CMPEST converges faster than when it is not so much shifted to the side larger than the knock magnitude detection value KICREAL.

  As described above, according to the fourth, fifth, and sixth embodiments (the inventions according to claims 17 and 26), the knock magnitude difference ΔKIC (comparison result between the knock magnitude estimated value KICEEST and the knock magnitude detected value KICREAL) , The knock correlation parameter estimated values such as the estimated octane number OCTEST, the alcohol concentration estimated value ALLCEST, and the compression ratio estimated value CMPEST are calculated (see step 374 in FIG. 55, 381 in FIG. 59, and step 391 in FIG. 61). Therefore, the convergence of the knock correlation parameter estimation value can be accelerated, and the drivability is improved.

  Further, in order to improve accuracy in detecting the self-ignition timing (knock occurrence timing) using the knock sensor 47, it is necessary to shorten the sampling cycle. However, if the knock intensity is used, the sampling frequency of the knock sensor 47 can be lowered. Therefore, according to the fourth, fifth, and sixth embodiments using the knock strength (the inventions according to claims 17 and 26), the system can be configured at low cost without deteriorating the performance.

  In the fourth embodiment, as shown in FIG. 5, FIG. 10, FIG. 12, and FIG. 13, laminar combustion velocities SL1, SL2, combustion gas volume equivalent volume (V0, VTDC), and combustion mass ratio (BR1, BR2) The combustion period from the start of combustion to the predetermined crank angle (BURN1, BURN2) is calculated based on the reaction probability RPROBA, and the basic ignition timing MBTCAL is calculated based on the combustion period (BURN1, BURN2). Instead of calculating the basic ignition timing MBTCAL, a base ignition timing map may be provided. In this case, the knock limit ignition timing calculation means is a pressure increase amount estimation means for estimating the pressure increase amount DP1 due to knocking of the combustion chamber 5 based on the self-ignition timing predicted value θknest (knock occurrence timing predicted value) and the operating conditions. (See steps 219 to 226 in FIG. 33), knock strength estimated value calculation means (see steps 227 to 229 in FIG. 33) for calculating a knock strength estimated value KIC based on the pressure increase DP1, and this knock strength estimation Knock retard amount calculating means (see step 230 in FIG. 33) for calculating the knock retard amount KNRT based on the value KIC, and the knock limit ignition timing obtained by correcting the basic ignition timing MBTCAL to the retard side by the knock retard amount KNRT. Knock limit ignition timing setting means for setting as KNOCKcal (see step 231 in FIG. 33); Even if the base ignition timing as the basic ignition timing is given as a map by including the configuration (the invention according to claim 18), a map of the base ignition timing for each of a plurality of different octane numbers from the maximum octane number to the minimum octane number Therefore, the ROM capacity does not increase.

  According to the fourth embodiment (the invention described in claim 19), in response to the greatest influence of the octane number on the knock when gasoline is used as the fuel, the estimated octane value OCTEST is obtained from the knock detection result and the knock. Since the calculation is based on the magnitude ΔKIC (comparison result of the knock magnitude detected value KICREAL and the knock magnitude estimated value KICEEST) (see steps 261, 371, 372, and 374 in FIG. 55), gasoline whose octane number is not known in advance is used as the fuel. Even when the automatic ignition timing is predicted, the self-ignition timing predicted value θknkest (knock occurrence timing predicted value) can be accurately calculated.

  Further, according to the fourth embodiment (the invention described in claim 20), when the octane number estimated value calculating means detects that a knock has occurred, in the combustion cycle in which the knock has been detected, The estimated octane number OCTEST is updated to a smaller value until the knock magnitude difference ΔKIC with the knock magnitude detected value KICREAL falls within the allowable range (refer to the loop operation in steps 372 to 375 in FIG. 55 in particular), and thereafter the estimated octane number OCTEST. Is updated at a constant cycle to the side where the second predetermined value const05 is increased (see steps 261, 267 to 271 in FIG. 55), so that the estimated octane number OCTEST is converged in the combustion cycle in which the knock is detected. be able to.

  According to the fourth embodiment (the invention described in claim 21), the update (see step 270 in FIG. 55) is performed so that the estimated octane value OCTEST is increased by the second predetermined value const05 (the side where knocking occurs). Since the basic ignition timing MBTCAL is executed only when knocking occurs, that is, when the ignition timing minimum value PADV is retarded from the basic ignition timing MBTCAL (step 267 in FIG. 38), the estimated octane number OCTEST is erroneously set. There is no renewal.

  According to the fourth embodiment (the invention described in claim 25), the basic ignition timing calculation means calculates the laminar combustion velocity (SL1, SL2) that is the combustion velocity in the laminar state of the combustion gas. Combustion rate calculation means (see step 168 in FIG. 10 and step 188 in FIG. 12) and combustion gas volume equivalent volume calculation means (FIG. 10) for calculating the volume (V0, VTDC) corresponding to the combustion gas volume in the combustion chamber 5. Step 162 of FIG. 12 and 182 of FIG. 12) and combustion mass ratio calculating means for calculating the combustion mass ratio (BR1, BR2) of the gas combusted in the combustion chamber 5 up to a predetermined crank angle (Step 171 of FIG. 10, Step 191 in FIG. 12), reaction probability calculation means for calculating the reaction probability RPROBA indicating the ease of combustion of the combustion gas under a predetermined operating condition (see Step 15 in FIG. 5), and these layers Based on the combustion speed (SL1, SL2), the combustion gas volume equivalent volume (V0, VTDC), the combustion mass ratio (BR1, BR2), and the reaction probability RPROBA, the combustion period (BURN1, BURN2) from the start of combustion to the predetermined crank angle is calculated. Combustion period calculating means for calculating (see step 171 in FIG. 10 and step 191 in FIG. 12) and basic ignition timing calculating means for calculating a basic ignition timing MBTCAL from which MBT is obtained based on the combustion periods (BURN1, BURN2) ( 13), the knock limit ignition timing KNOCKcal, which is a value obtained by correcting the basic ignition timing MBTCAL to the retard side, is calculated based on the combustion analysis. The optimum knock limit ignition timing KNOCKcal can be calculated regardless of the conditions.

  According to the fifth embodiment (the invention described in claim 22), in the mixed fuel of gasoline and alcohol, the alcohol concentration estimated value ALTEST is knocked in response to the influence of the alcohol concentration in the mixed fuel on the knocking. Since the calculation is based on the detection result and the knock intensity difference ΔKIC (comparison result of the knock intensity estimated value KICEEST and the knock intensity detected value KICreal) (see steps 261, 371, 372, 381 in FIG. 59), the alcohol concentration can be known in advance. Even when a non-alcohol-containing fuel is used as the fuel to be used, it is possible to accurately calculate the autoignition timing predicted value θknkest (knock generation timing predicted value).

  According to the fifth embodiment (the invention described in claim 23), when the alcohol concentration estimated value calculating means detects that a knock has occurred, the knock intensity estimated value KICEST and the knock in the combustion cycle in which the knock has been detected. The alcohol concentration estimated value ALCEST is updated to a higher side until the knock intensity difference ΔKIC with the intensity detection value KICREAL is within the allowable range (see particularly the loop operation of steps 372, 373, 381, and 382 in FIG. 59), and thereafter Since it is a means for updating the alcohol concentration estimated value ALCEST at a constant cycle to the lower side by the second predetermined value const15 (see steps 261, 267, 303, 269, 304, 271 in FIG. 59), knock is detected. It is possible to converge the estimated alcohol concentration value ALLCEST in the combustion cycle.

  According to the fifth embodiment (the invention described in claim 24), the alcohol concentration estimated value ALCEST is updated to the side where the second predetermined value const15 is lowered (the side where knocking occurs) (see step 304 in FIG. 59). ) Is executed under the condition that knocking occurs at the basic ignition timing MBTCAL (step 267 in FIG. 59), the alcohol concentration estimated value ALLCEST is not erroneously updated.

  According to the sixth embodiment (the invention described in claim 52), it is detected whether or not knock has actually occurred in the combustion chamber, the knock strength in the combustion chamber is detected, and this knock strength detection value KICREAL The knock strength estimated value KICEEST is compared, and the compression ratio estimated value CMPEST is calculated based on the comparison result and the knock detection result. The volume V0 of the combustion chamber at the combustion start timing is calculated based on the compression ratio estimated value CMPEST. Is calculated, the combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated based on the volume V0 at the combustion start timing, and the basic ignition with which MBT is obtained based on the combustion period (BURN1, BURN2) Since the timing MBTCAL is calculated and spark ignition is performed at this basic ignition timing MBTCAL, the octane number is determined in advance. Even when the actual compression ratio becomes higher than the compression ratio of the engine specification for some reason when using fuel or mixed fuel with a fixed alcohol concentration, the MBT can be obtained while accelerating the convergence of the compression ratio estimated value CMPEST. The basic ignition timing can be given with high accuracy.

  According to the sixth embodiment (the invention described in claims 28 and 55), when the compression ratio estimated value calculation means detects that a knock has occurred, the knock intensity estimated value KICEST in the combustion cycle in which the knock is detected. And the compression ratio estimated value CMPEST is updated to a larger side until the knock magnitude difference ΔKIC between the detected value and the knock magnitude detected value KICREAL falls within the allowable range (refer to the loop operation of steps 372, 373, 391, and 392 in FIG. 61), After that, since the compression ratio estimated value CMPEST is a means for updating the compression ratio estimated value CMPEST at a constant cycle to the lower side by the second predetermined value const25 (see steps 261, 267, 333, 269, 334, 271 in FIG. 61), The compression ratio estimated value CMPEST can be converged by the detected combustion cycle.

  According to the sixth embodiment (the invention described in claims 29 and 55), the compression ratio estimated value CMPEST is updated to the side where the second predetermined value const25 is decreased (the side where knocking occurs) (step 334 in FIG. 61). (See FIG. 61) is executed under the condition that knocking occurs at the basic ignition timing MBTCAL (step 267 in FIG. 61), so that the compression ratio estimated value CMPEST is not erroneously updated.

  According to the sixth embodiment (the invention described in claims 30 and 56), as shown in FIG. 54, the intake valve is based on the volume VIVC of the combustion chamber 5 at the intake valve closing timing and the volume V0 at the combustion start timing. An effective compression ratio Ec from the closing timing IVC to the combustion start timing is calculated (see step 163 in FIG. 54), and combustion starts in the combustion chamber 5 from the temperature TINI at the intake valve closing timing of the combustion chamber 5 and the effective compression ratio Ec. The temperature T0 at the timing is calculated from the pressure PINI at the intake valve closing timing of the combustion chamber 5 and the effective compression ratio Ec, respectively, and the pressure P0 at the combustion start timing of the combustion chamber 5 is calculated (see steps 164 to 167 in FIG. 54). A laminar combustion speed SL1 used for calculating the initial combustion period BURN1 is calculated based on the temperature T0 and the pressure P0 at the combustion start timing of the combustion chamber 5 (step 168 in FIG. 54). In this case, since the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve is calculated based on the estimated compression ratio CMPEST (steps 351 and 352 in FIG. 53), the fuel having a predetermined octane number is used. Even when the actual compression ratio becomes higher than the scheduled compression ratio for some reason, the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve can be calculated with high accuracy.

  In the sixth embodiment, the self-ignition timing predicted value θknkest (knock occurrence timing predicted value) has been described based on the case where it is calculated based on the characteristics representing the distribution of the reciprocal of the time until the fuel in the combustion chamber reaches self-ignition. The self-ignition timing detection value θknkreal (knock occurrence timing detection value) may be used instead of the self-ignition timing prediction value θknquest (the inventions according to claims 27 and 53).

  As described above, the fourth, sixth, and seventh embodiments also have the same operational effects as those of the first, second, and third embodiments.

  The flow of FIG. 63 is the seventh embodiment, the flow of FIG. 68 is the eighth embodiment, FIG. 63 is FIG. 38 of the first embodiment or FIG. 55 of the fourth embodiment, and FIG. 68 is the second embodiment. 45 or FIG. 59 of the fifth embodiment, respectively. 63, the same parts as those in FIGS. 38 and 55 are given the same step numbers, and in FIG. 68, the same parts as those in FIGS. 45 and 59 are given the same step numbers.

  In the first embodiment, when it is detected that knocking has occurred, the ignition timing difference Δθ, which is the difference between the self-ignition timing detection value θknkreal and the predicted self-ignition timing value θknest in the combustion cycle in which the knock is detected, Although the estimated octane number OCTEST is reduced until it falls within the allowable range, in the seventh embodiment, when it is detected that a knock has occurred, the compression ratio is estimated in addition to this in the combustion cycle in which the knock is detected. The value is also calculated. Similarly, in the second embodiment, when it is detected that a knock has occurred, the ignition that is the difference between the self-ignition timing detection value θknkreal and the self-ignition timing prediction value θknkest in the combustion cycle in which the knock is detected. The alcohol concentration estimated value ALCEST is increased until the time difference Δθ falls within the allowable range. In the eighth embodiment, when it is detected that a knock has occurred, the estimated alcohol concentration is determined in the combustion cycle in which the knock is detected. In addition, an estimated compression ratio is calculated.

  The seventh embodiment shown in FIG. 63 differs from the first embodiment shown in FIG. 38 or the fourth embodiment shown in FIG. 55 in steps 401 and 402, and in the eighth embodiment shown in FIG. Steps 401 and 461 are mainly different from the second embodiment shown in FIG. 5 or the fifth embodiment shown in FIG. That is, in step 401 common to the seventh and eighth embodiments, the compression ratio estimated value CMPEST2 is calculated. The calculation of the compression ratio estimated value CMPEST2 will be described with reference to the flow of FIG. 64 common to the seventh and eighth embodiments.

  Here, since the calculation method of the compression ratio estimated value in the seventh and eighth embodiments is different from the calculation method of the compression ratio estimated value in the third embodiment, the compression ratio calculated in the seventh and eighth embodiments. The estimated value is distinguished from the compression ratio estimated value “CMPEST” calculated in the third embodiment as “CMPEST2”.

  In FIG. 64 (subroutine of step 401 in FIGS. 63 and 68) common to the seventh and eighth embodiments, in step 411, the cylinder fresh air amount MACYL [g] calculated in steps 52 and 53 in FIG. The internal inert gas amount MRES [g] and the fuel amount QINJ [g] are read. The fuel amount QINJ [g] may be obtained in proportion to the fuel injection pulse width TI [ms].

  In step 412, the cylinder fresh air amount MACYL [g] is moved to WIDRY [g], and the internal inert gas amount MRES [g] is moved to MASSZ [g]. These WIDRY and MASSZ are introduced for use only in calculating the knock strength index KNKI, WIDRY is the cylinder fresh air amount, and MASSZ is the internal inert gas amount.

  In step 413, using the self-ignition timing detection value θknkreal obtained in step 262 of FIGS. 63 and 68, the combustion mass ratio BRknreal at the self-ignition timing detection value θknkreal is calculated. This is because, when the self-ignition timing detection value θknkreal is in the initial combustion period, the self-ignition timing detection value θknkreal is converted to the crank angle Θ based on the compression top dead center TDC, and the converted crank angle Θ is When the self-ignition timing detection value θknkreal is in the main combustion period, the self-ignition timing detection value θknkreal is converted into the crank angle Θ based on the compression top dead center TDC, and the converted crank The angle Θ can be obtained by substituting each into the above equation (63).

  In step 414, an average temperature TE at the time of self-ignition of the combustion chamber 5 is calculated. This may be obtained by calculating the average temperature TC of the combustion chamber 5 obtained by substituting 1.0 into the combustion mass ratio BR on the right side of the equation (64) as the self-ignition average temperature TE of the combustion chamber 5.

  In step 415, the unburned fuel amount MUB2 [g] at the time of self-ignition is calculated from the fuel amount QINJ [g] and the combustion mass ratio BRkreal at the self-ignition timing detection value θknkreal by the following equation.

MUB2 = QINJ × (1-BRknkreal) (93)
Expression (93) is an expression obtained by replacing MUB → MUB2 and BRknk → BRknreal in the above (complement 10).

  In steps 416 to 418, the total number of gas moles MLALL, the gas enthalpy E, and the specific heat Cv of burned gas are calculated. Since these calculations are the same as steps 223, 224, and 225 in FIG. 33, description thereof is omitted.

  In step 419, the pressure increase due to self-ignition, that is, the pressure increase DP2 [Pa] due to knock is calculated by the following equation using the knock magnitude detection value KICREAL obtained in step 371 of FIGS.

DP2 = KICREAL / correlation coefficient 3 (94)
Expression (94) is the same as the above expression (75). That is, the above expression (75) converts the pressure increase into a knock intensity, whereas the expression (94) converts the knock intensity into a pressure increase. Therefore, the correlation coefficient 3 on the right side of the equation (94) is a coefficient for representing the correlation with the knock intensity. For simplicity, the correlation coefficient 3 may be the same as the correlation coefficient 1 used in step 227 of FIG.

  In step 420, the volume Vknk2 at the self-ignition timing detection value θknkreal of the combustion chamber 5 is calculated by the following equation.

Vknk2 = (MLALL × R # × CF # × MUB2)
/ {DP2 × Cv × (MASSZ + QINJ + WIDRY)}
... (95)
Where CF #: lower heating value of fuel,
Equation (95) is the same as equation (74) above. That is, the above equation (74) is an equation arranged for the pressure increase, and the equation (95) is an equation arranged for the combustion chamber volume.

  In step 421, the clearance volume Vc2 is calculated by the following equation using the volume Vknk2 at the self-ignition timing detected value θknkreal of the combustion chamber 5.

Vc2 = Vknk2- (π / 4) D ² · Hknkreal (96)
Where D: cylinder bore diameter [m]
Hknkreal: distance [m] from the TDC of the piston pin 76 in the self-ignition timing detection value,
Equation (96) is the same as equation (2) above. That is, the second term on the right side of the equation (96) is the volume (not including the gap area) at the auto-ignition timing detection value θknkreal of the combustion chamber 5, so that the gap volume is obtained by subtracting this volume from Vknk2 that includes the gap volume. Can be obtained. D and Hknkreal are known from the engine specifications.

  In step 422, the compression ratio estimated value CMPEST2 is calculated by the following equation using the gap volume Vc2.

CMPEST2 = (π / 4) D ² · Hx / Vc2 + 1 (97)
Where D: cylinder bore diameter [m]
Hx: difference between the maximum value and the minimum value of the distance from the TDC of the piston pin 76 [m]
,
Equation (97) is the same as equation (86) above. That is, the equation (86) is an equation arranged for the clearance volume, and the equation (97) is an equation arranged for the compression ratio. D and Hx are also known from the engine specifications.

  When the calculation of the compression ratio estimated value CMPEST2 is thus completed, the process returns to FIG. 63 in the seventh embodiment, and the octane number estimated value OCTEST is returned in step 402. In the eighth embodiment, the process returns to FIG. 68 and the alcohol concentration is estimated in step 461. The value ALCEST is calculated. The calculation of the estimated octane number OCTEST of the seventh embodiment will be described with reference to the flow of FIG. 65, and the calculation of the alcohol concentration estimated value ALTEST of the eighth embodiment will be described with reference to the flow of FIG.

  When it is detected that a knock has occurred, in the combustion cycle in which the knock is detected, in the first embodiment, the ignition timing difference Δθ is less than a predetermined value, and in the fourth embodiment, the knock intensity difference ΔKIC is less than a predetermined value. The estimated octane number OCTEST is repeatedly updated until it has settled, and the estimated octane number OCTEST is converged. In the combustion cycle in which the knock is detected when the occurrence of the knock is detected, the seventh embodiment is the fuel in the combustion chamber 5. Is the time until auto-ignition, and the reciprocal (1 / τ) of time τ, which differs depending on the pressure and temperature in the combustion chamber 5, extends from the combustion start timing to the knock generation timing detection value θknkreal for each predetermined crank angle. Accumulate and repeat the estimated octane number OCTEST until the absolute value of the difference between the accumulated value SUM and 1 falls within the allowable range. Update those to converge the estimated octane value OCTEST.

  Similarly, when it is detected that knocking has occurred, in the combustion cycle in which the knocking is detected, until the ignition timing difference Δθ is less than a predetermined value in the second embodiment, and in the fifth embodiment, the knock intensity difference ΔKIC is The alcohol concentration estimated value ALTEST is updated repeatedly until it falls below the predetermined value, and the alcohol concentration estimated value ALLCEST is converged. However, in the combustion cycle in which the knock is detected when it is detected that the knock has occurred, the eighth embodiment The time taken for the fuel in the combustion chamber 5 to self-ignite and the reciprocal (1 / τ) of the time that varies depending on the pressure and temperature in the combustion chamber 5 is knocked from the combustion start timing at every predetermined crank angle. Accumulated over the detected value θknkreal, the alcohol concentration estimated value until the absolute value of the difference between the accumulated value SUM and 1 falls within the allowable range The ALLCEST is repeatedly updated to converge the alcohol concentration estimated value ALLCEST.

  Here, as a method of calculating the estimated octane value OCTEST in the seventh embodiment, the calculation method of the autoignition timing predicted value in FIGS. 40 and 41 of the first embodiment or in FIGS. 57 and 58 of the fourth embodiment. As a calculation method of the predicted self-ignition timing, and as a calculation method of the alcohol concentration estimated value ALCEST in the eighth embodiment, the calculation method of the predicted self-ignition timing in FIG. 46 and FIG. 47 of the second embodiment or the fifth embodiment. The calculation method of the auto-ignition timing predicted value in the above-described FIGS. 60 and 58 is used.

  More specifically, in FIG. 65 of the seventh embodiment (subroutine of step 402 in FIG. 63), the same step numbers are assigned to the same parts as those in FIG. 40 of the first embodiment or FIG. 57 of the fourth embodiment. In FIG. 69 of the eighth embodiment (subroutine of step 461 in FIG. 68), the same step numbers are assigned to the same parts as in FIG. 46 of the second embodiment or FIG. 60 of the fifth embodiment.

  FIG. 65 of the seventh embodiment differs from FIG. 40 of the first embodiment or FIG. 57 of the fourth embodiment in that steps 431, 432, 433 and FIG. 69 of the eighth embodiment are different from FIG. 46 of the second embodiment. Alternatively, steps 431, 432, and 471 are different from FIG. 60 of the fifth embodiment.

  The difference of the seventh embodiment from the first and fourth embodiments will be mainly described. In FIG. 65, first, in step 431, the crank angle θ and the auto-ignition timing detection value θknkreal obtained in step 262 of FIG. And compare. Before the crank angle θ reaches the self-ignition timing detection value θknkreal, the routine proceeds to step 212 in FIG. 65, and the operations of steps 213 to 217, 206 to 209, and 431 in FIG. 65 are performed while the crank angle θ is advanced by a predetermined angle const02. repeat. When the crank angle θ eventually reaches the self-ignition timing detection value θknkreal, the routine proceeds from step 431 in FIG. 65 to step 432, where the absolute value of the difference between 1 / τ integrated value SUM and 1 in the fuel of the estimated octane number OCTEST is calculated. Is compared with a predetermined value ε (positive value). The predetermined value ε defines an allowable range. If the absolute value of the difference between 1 and 1 / τ integrated value SUM is less than the predetermined value ε, the process proceeds to “END” and the current process is terminated.

  On the other hand, if the absolute value of the difference between the integrated value SUM of 1 and 1 / τ is greater than or equal to the predetermined value ε, it is determined that the estimated octane number OCTEST is too large, and the routine proceeds to step 433 to decrease the estimated octane number OCTEST. That is, the estimated octane value OCTEST is updated by the following equation.

OCTEST (new) = OCTEST (old) -const03 × (1-SUM)
... (98)
However, OCTEST (new): Estimated value of octane number after update,
OCTEST (old): Estimated octane value before update,
const03: update rate to the side to be reduced (positive anonymous number),
Here, the second term on the right side of the equation (98) defines the update amount per one time of the estimated octane number. The reason why the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is introduced in order to accelerate the convergence of the estimated octane value OCTEST. That is, when the estimated octane number OCTEST is larger than the actual octane number and close to the actual octane number, the value obtained by subtracting the 1 / τ integrated value SUM is relatively small, but the estimated octane number OCTEST is the actual octane number. When the deviation is larger and larger than the actual octane number, the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is relatively large. Therefore, when the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is relatively large, the update amount per one time of the estimated octane number is increased accordingly, so that the convergence of the estimated octane number OCTEST is accelerated. Is.

  Thereafter, returning to step 203 in FIG. 65, the operation from step 203 is repeatedly executed. When the crank angle θ reaches the self-ignition timing detection value θknkreal, the process proceeds from step 431 to step 432 in FIG. If the absolute value of the difference between the accumulated values SUM1 is equal to or greater than the predetermined value ε, it is determined that the octane number estimated value OCTEST is still too large, and the octane number estimated value OCTEST is decreased again in step 433, and the process returns to step 203 in FIG. When the above operation is repeated again, the absolute value of the difference between the integrated value SUM of 1 and 1 / τ eventually becomes less than the predetermined value ε. At this time, the process proceeds to “END” and the process of FIG. As a result, the octane number estimated value when the absolute value of the difference between the integrated value SUM of 1 and 1 / τ is less than the predetermined value ε becomes the convergence value of the octane number estimated value.

  Thus, when it is detected that knocking has occurred, the integrated value SUM of 1 / τ from the combustion start timing (MBTCAL + IGNEAD) to the self-ignition timing detection value θknkreal in the combustion cycle in which the knock is detected The estimated octane number OCTEST is repeatedly updated until the absolute value of the difference from 1 falls within the allowable range to converge the estimated octane number OCTEST.

  Next, the differences of the eighth embodiment from the second and fifth embodiments will be mainly described. In FIG. 69, first, in step 431, the crank angle θ and the self-ignition timing obtained in step 262 of FIG. The detected value θknkreal is compared. Before the crank angle θ reaches the self-ignition timing detection value θknkreal, the routine proceeds to step 295 in FIG. 69, and the crank angle θ is advanced by a predetermined angle const12 by steps 213 to 217, 291, 292, 293, 209 in FIG. The operation of 431 is repeated. When the crank angle θ eventually reaches the auto-ignition timing detection value θknkreal, the routine proceeds from step 431 in FIG. 69 to step 432, where the absolute difference between the 1 / τ integrated value SUM and 1 in the fuel mixture of the alcohol concentration estimated value ALCEST is absolute. The value is compared with a predetermined value ε (positive value). If the absolute value of the difference between the integrated value SUM of 1 and 1 / τ is less than the predetermined value ε, the process proceeds to “END” and the current process is terminated.

  On the other hand, if the absolute value of the difference between the integrated value SUM of 1 and 1 / τ is greater than or equal to the predetermined value ε, it is determined that the alcohol concentration estimated value ALCEST is too small, and the routine proceeds to step 471 where the alcohol concentration estimated value ALSEC is increased. That is, the alcohol concentration estimated value ALEST is updated by the following equation.

ALLCEST (new) = ALCEST (old) + const13 × (1-SUM)
... (99)
However, ALCEST (new): Estimated alcohol concentration after update,
ALCEST (old): Estimated alcohol concentration before update,
const13: update rate (positive unnamed number) to the increase side,
Here, the second term on the right side of the equation (99) defines the update amount per one time of the alcohol concentration estimated value. The reason why the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is introduced to the update amount per time is to accelerate the convergence of the estimated alcohol concentration value ALSEC. That is, when the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration and close to the actual alcohol concentration, the value obtained by subtracting the integrated value SUM from 1 to 1 / τ is relatively small, but the alcohol concentration estimated value ALTEST is Is lower than the actual alcohol concentration and larger than the actual alcohol concentration, the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is relatively large. Therefore, when the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is relatively large, the update amount per one time of the alcohol concentration estimated value is increased accordingly, so that the convergence of the alcohol concentration estimated value ALTEST is accelerated. It is a thing.

  Thereafter, returning to step 203 in FIG. 69, the operation from step 203 is repeatedly executed. When the crank angle θ reaches the self-ignition timing detection value θknkreal, the process proceeds from step 431 to step 432 in FIG. If the absolute value of the difference between the integrated values SUM is greater than or equal to the predetermined value ε, it is determined that the alcohol concentration estimated value ALLCEST is still too small. In step 471, the alcohol concentration estimated value ALLCEST is increased again, and the process returns to step 203. When the above operation is repeated again, the absolute value of the difference between the integrated value SUM of 1 and 1 / τ eventually becomes less than the predetermined value ε. At this time, the process proceeds to “END” and the process of FIG. 69 is terminated. As a result, the alcohol concentration estimated value when the absolute value of the difference between the integrated value SUM of 1 and 1 / τ is less than the predetermined value ε becomes the convergence value of the alcohol concentration estimated value.

  Thus, when it is detected that knocking has occurred, the integrated value SUM of 1 / τ from the combustion start timing (MBTCAL + IGNEAD) to the self-ignition timing detection value θknkreal in the combustion cycle in which the knock is detected The alcohol concentration estimated value ALLCEST is repeatedly updated until the absolute value of the difference from 1 falls within the allowable range to converge the alcohol concentration estimated value ALLCEST.

  The compression ratio estimated value CMPEST2 calculated by FIG. 64 common to the seventh and eighth embodiments as described above is similar to that of the third embodiment in the seventh and eighth embodiments, as shown in FIGS. 66, the volume VIVC of the combustion chamber 5 at the intake valve closing timing and the volume V0 of the combustion chamber 5 at the combustion start timing (MBTCYCL) are used (steps 441 and 352 in FIG. 66, step 451 in FIG. 67). 362).

  In FIGS. 66 and 67 common to the seventh and eighth embodiments, the same step numbers are assigned to the same portions as those in FIGS. 53 and 54 of the third embodiment, and the description thereof is omitted.

  According to the seventh and eighth embodiments (the invention described in claim 33), the estimated values of the knock correlation parameters other than the compression ratio based on the knock detection result (the octane number estimated value OCTEST, eighth for the seventh embodiment) Alcohol concentration estimated value (ALCEST) is calculated for the embodiment (the knock detection result is fed back to the knock correlation parameter) (steps 261, 402, and 270 in FIG. 63 for the seventh embodiment, and step 261 in FIG. 68 for the eighth embodiment). , 461, 304), a predicted self-ignition timing θknest (predicted knock occurrence timing) in the combustion chamber 5 is calculated based on the knock correlation parameter estimated value (step 206 to FIG. 32 in FIG. 32 for the seventh embodiment). 210, step 218 in FIG. 33, steps 291 to 293 in FIG. 43 for the eighth embodiment, 09, 210, step 218 in FIG. 33), and knock limit ignition timing KNOCKcal is calculated based on the predicted self-ignition timing value θknkest (see steps 219 to 231 in FIG. 33 for the seventh and eighth embodiments). Therefore, even when using fuel sold in a market where the octane number of the fuel and the alcohol concentration in the mixed fuel are not known in advance, the knock detection result is fed back to the ignition timing regardless of the operating state. As described above, the ignition timing delay for avoiding knocking and the subsequent advance operation are not repeated. For this reason, since the knock limit ignition timing can be followed even during a transition such as acceleration or deceleration, fuel consumption deterioration and output reduction can be prevented.

  In addition, updating of the estimated value of the knock correlation parameter m other than the compression ratio (the octane number estimated value OCTEST for the seventh embodiment and the alcohol concentration estimated value ALTEST for the eighth embodiment) when the occurrence of knocking is detected ( In the calculation, the value obtained by subtracting the integrated value SUM of 1 / τ from 1 as shown in the above formulas (98) and (99) is also taken into consideration. That is, when the estimated octane number OCTEST is larger than the actual octane number and is larger than the actual octane number (that is, the value obtained by subtracting the integrated value SUM of 1 / τ is relatively large) in the seventh embodiment, the octane number The update amount per one time of the estimated octane number is larger than when the estimated value OCTEST is larger than the actual octane number and close to the actual octane number (that is, the value obtained by subtracting 1 / τ integrated value SUM is relatively small). Therefore, when the estimated octane number OCTEST is larger than the actual octane number and deviates from the actual octane number, the estimated octane number OCTEST converges faster than when the estimated octane number OCTEST is larger than the actual octane number and close to the actual octane number. . When the estimated alcohol concentration value ALCTEST is lower than the actual alcohol concentration and larger than the actual alcohol concentration (that is, the value obtained by subtracting the integrated value SUM of 1 / τ from 1 is relatively large) in the eighth embodiment, When the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration and close to the actual alcohol concentration (that is, the value obtained by subtracting the integrated value SUM of 1 / τ is relatively small), the alcohol concentration estimated value per time When the alcohol concentration estimated value ALCEST is lower than the actual alcohol concentration and deviated more than the actual alcohol concentration, the alcohol concentration estimated value ALTEST is lower than the actual alcohol concentration and becomes the actual alcohol concentration. Alcohol concentration estimated value ALC than near The EST converges quickly.

  As described above, according to the seventh and eighth embodiments (the invention described in claim 33), knock correlation parameter estimation values other than the compression ratio even based on the self-ignition timing detection value θknkreal (knock occurrence timing detection value). (The octane number estimated value OCTEST for the seventh embodiment and the alcohol concentration estimated value ALCEST for the eighth embodiment) are calculated (see step 433 in FIG. 65 for the seventh embodiment and step 471 in FIG. 69 for the seventh embodiment). The convergence of the estimated values of the knock correlation parameters other than the compression ratio can be accelerated, and the drivability is improved.

  In the seventh and eighth embodiments, as shown in FIG. 66, FIG. 67, FIG. 12, and FIG. 13, laminar combustion speeds SL1, SL2, combustion gas volume equivalent volume (V0, VTDC), combustion mass ratio ( (BR1, BR2) and the reaction probability RPROBA, the combustion period (BURN1, BURN2) from the start of combustion to a predetermined crank angle is calculated, and the basic ignition timing MBTCAL is calculated based on the combustion period (BURN1, BURN2). As described above, instead of calculating the basic ignition timing MBTCAL, a base ignition timing map may be provided. In this case, the knock limit ignition timing calculation means is a pressure increase amount estimation means for estimating the pressure increase amount DP1 due to knocking of the combustion chamber 5 based on the self-ignition timing predicted value θknest (knock occurrence timing predicted value) and the operating conditions. (See steps 219 to 226 in FIG. 33), knock strength estimated value calculation means (see steps 227 to 229 in FIG. 33) for calculating a knock strength estimated value KIC based on the pressure increase DP1, and this knock strength estimation Knock retard amount calculating means (see step 230 in FIG. 33) for calculating the knock retard amount KNRT based on the value KIC, and the knock limit ignition timing obtained by correcting the basic ignition timing MBTCAL to the retard side by the knock retard amount KNRT. Knock limit ignition timing setting means for setting as KNOCKcal (see step 231 in FIG. 33); By including it (the invention according to claim 34), even if the base ignition timing as the basic ignition timing is given by a map, the map of the base ignition timing for each of a plurality of different octane numbers from the maximum octane number to the minimum octane number Therefore, the ROM capacity does not increase.

  According to the seventh embodiment (the invention described in claim 35), in response to the fact that the octane number has the largest influence on knock when gasoline is used as fuel, the estimated octane number OCTEST is obtained from the knock detection result and Since it is calculated based on the self-ignition timing detection value θknkreal (knock occurrence timing detection value) (see steps 261, 262, 402 and FIG. 65 in FIG. 63), this is the case where gasoline whose octane number is not known in advance is used as the fuel. However, the self-ignition timing predicted value θknest (knock occurrence timing predicted value) can be accurately calculated.

  According to the seventh embodiment (the invention described in claim 36), when the estimated octane number calculating means detects that a knock has occurred, the fuel in the combustion chamber 5 is not automatically detected in the combustion cycle in which the knock has been detected. The reciprocal (1 / τ) of the time until ignition is varied depending on the pressure and temperature in the combustion chamber 5 from the combustion start timing (MBTCAL + IGNEAD) to the knock generation timing detection value θknkreal for each predetermined crank angle const02. Until the absolute value of the difference between the integrated value SUM and 1 falls within the allowable range, and the octane number estimated value OCTEST is updated to a smaller value (particularly, the loop operation of steps 203 to 209 and 431 to 433 in FIG. 65) After that, the estimated octane value is updated at a constant cycle so as to increase by the second predetermined value const05. (Refer to steps 261, 267 to 271 in FIG. 63) Since it is a means, the estimated octane number OCTEST can be converged in the combustion cycle in which knocking is detected.

  According to the seventh embodiment (the invention described in claim 37), the update (see step 270 in FIG. 63) is performed so that the estimated octane value OCTEST is increased by the second predetermined value const05. The basic ignition timing MBTCAL is executed only when knocking occurs, that is, only when the minimum ignition timing value PADV is retarded from the basic ignition timing MBTCAL (step 267 in FIG. 63), so that the estimated octane number OCTEST is erroneously set. There is no renewal.

  According to the eighth embodiment (the invention described in claim 38), in the mixed fuel of gasoline and alcohol, the alcohol concentration estimated value ALCEST is knocked in response to the influence of the alcohol concentration in the mixed fuel on the knocking. Since it is calculated based on the detection result and the self-ignition timing detection value θknkreal (knock occurrence timing detection value) (see steps 261, 262, and 461 in FIG. 68), an alcohol-containing fuel whose alcohol concentration cannot be known in advance is used as the fuel. Even in this case, it is possible to accurately calculate the auto-ignition timing predicted value θknkest (knock occurrence timing predicted value).

According to the eighth embodiment (the invention described in claim 39), when the alcohol concentration estimated value calculating means detects that a knock has occurred, the fuel in the combustion chamber 5 is discharged in the combustion cycle in which the knock has been detected. The reciprocal number (1 / τ) of the time until self-ignition, which differs depending on the pressure and temperature in the combustion chamber 5, is detected from the combustion start time (MBTCAL + IGNEAD) for each predetermined crank angle const12, and the detected value θknkrear Until the absolute value of the difference between the integrated value SUM and 1 falls within the allowable range, and the alcohol concentration estimated value ALCEST is updated to a higher side (particularly, steps 203 to 205, 291 to 293 in FIG. 69). 209, 431, 432, and 471 (see the loop operation)), and thereafter the alcohol concentration estimated value is set to the second predetermined value const Since it is a means of updating at a constant cycle to the lower side by 15 (see steps 261, 267, 303, 269, 304, 271 in FIG. 68), the alcohol concentration estimated value ALSEC is converged in the combustion cycle in which knocking is detected. Can be made.

  According to the eighth embodiment (the invention described in claim 40), the alcohol concentration estimated value ALCEST is updated to the side where the second predetermined value const15 is lowered (the side where knocking occurs) (see step 304 in FIG. 68). ) Is executed under the condition that knocking occurs at the basic ignition timing MBTCAL (step 267 in FIG. 68), so that the alcohol concentration estimated value ALLCEST is not erroneously updated.

  According to the seventh and eighth embodiments (the invention described in claim 41), the basic ignition timing calculation means calculates the laminar combustion velocity (SL1, SL2) which is the combustion velocity in the laminar state of the combustion gas. A laminar combustion velocity calculating means for calculating (see step 168 in FIG. 67 and step 188 in FIG. 12), and a volume corresponding to the combustion gas volume in the combustion chamber 5 (V0, VTDC) is calculated. Means (see step 162 in FIG. 67, 182 in FIG. 12) and combustion mass ratio calculating means for calculating the combustion mass ratio (BR1, BR2) of the gas combusted in the combustion chamber 5 up to a predetermined crank angle (FIG. 67). Step 171 of FIG. 12, step 191 of FIG. 12), and reaction probability calculation means (see step 15 of FIG. 66) for calculating the reaction probability RPROBA indicating the ease of combustion of the combustion gas under predetermined operating conditions. Based on these laminar combustion velocity (SL1, SL2), combustion gas volume equivalent volume (V0, VTDC), combustion mass ratio (BR1, BR2) and reaction probability RPROBA, the combustion period (BURN1, BURN2) is calculated (see step 171 in FIG. 67, step 191 in FIG. 12), and basic ignition timing MBTCAL is calculated based on this combustion period (BURN1, BURN2). Calculation means (see steps 41 to 43 in FIG. 13), the knock limit ignition timing KNOCKcal, which is a value obtained by correcting the basic ignition timing MBTCAL to the retard side, is calculated based on the combustion analysis. As a result, the optimum knock limit ignition timing KNOCKcal can be calculated regardless of the operating conditions. .

  According to the seventh and eighth embodiments (the invention described in claim 42), the compression ratio estimated value CMPEST2 is calculated based on the knock detection result by the knock sensor 47 (step 261 in FIG. 63 for the seventh embodiment). , 401, FIG. 64, and the eighth embodiment (see steps 261, 401, FIG. 64 in FIG. 68), and the compression start timing of the combustion chamber 5 used for calculating the initial combustion period BURN1 based on the compression ratio estimated value CMPEST2 Is calculated (see steps 451 and 362 in FIG. 67 common to the seventh and eighth embodiments). Therefore, when a fuel having an octane number of 80 (a fuel having a predetermined octane number) is used for some reason, Even when the actual compression ratio is higher than the compression ratio of the engine specification, the knock detection result is It is never repeated and the operation of the retard and advance to subsequent ignition timing for knock avoidance as in the conventional apparatus for feedback. For this reason, as in the first and second embodiments, the knock limit ignition timing can be followed even in a transient state, so that fuel consumption deterioration and output reduction can be prevented.

  According to the seventh and eighth embodiments (the invention described in claim 43), as shown in FIG. 67, based on the volume VIVC of the combustion chamber 5 at the intake valve closing timing and the volume V0 at the combustion start timing. An effective compression ratio Ec from the intake valve closing timing IVC to the combustion start timing is calculated (see step 163 in FIG. 67), and the temperature TINI at the intake valve closing timing of the combustion chamber 5 and this effective compression ratio Ec The temperature T0 at the combustion start timing is calculated from the pressure PINI at the intake valve closing timing of the combustion chamber 5 and the effective compression ratio Ec, respectively, and the pressure P0 at the combustion start timing of the combustion chamber 5 is calculated (see steps 164 to 167 in FIG. 67). ), A laminar combustion speed SL1 used for calculating the initial combustion period BURN1 is calculated based on the temperature T0 and the pressure P0 at the combustion start timing of the combustion chamber 5 (step 16 in FIG. 67). In this case, since the volume VIVC of the combustion chamber 5 at the intake valve closing timing is calculated based on the compression ratio estimated value CMPEST2 (steps 441 and 352 in FIG. 66), the fuel whose octane number is determined in advance is used. Even when the actual compression ratio becomes higher than the scheduled compression ratio for some reason, the volume VIVC of the combustion chamber 5 at the closing timing of the intake valve can be calculated with high accuracy.

  In the first aspect of the present invention, the function of the comparison means is the step 263 of FIG. 38 and the function of the knock correlation parameter estimation value calculation means is the step 261, 265, 270 of FIG. The functions are performed by steps 206, 207, 208, 209, 210 of FIG. 32, and 218 of FIG. 33, and the function of the knock limit ignition timing calculation means is performed by steps 219 to 231 of FIG.

  In the invention according to claim 10, the function of the laminar combustion velocity calculation means is the step 168 in FIG. 54 and step 188 in FIG. 12, and the function of the combustion gas volume equivalent volume calculation means is the step 362 in FIG. 54, the function of the combustion mass ratio calculating means is step 171 of FIG. 54, step 191 of FIG. 12, the function of the reaction probability calculating means is step 15 of FIG. 53, and the function of the basic ignition timing calculating means is FIG. In steps 41, 42 and 43, the function of the comparison means is in step 263 in FIG. 50, the function of the compression ratio estimated value calculation means is in steps 261, 331 and 334 in FIG. 54 steps 361 and 362, respectively.

  According to the seventeenth aspect of the present invention, the function of the comparison means is the step 372 of FIG. 55, and the function of the knock correlation parameter estimated value calculation means is the step 261, 374, 270 of FIG. The functions are performed by steps 206, 207, 208, 209, 210 of FIG. 32, and 218 of FIG. 33, and the function of the knock limit ignition timing calculation means is performed by steps 219 to 231 of FIG.

  In the invention described in claim 26, the function of the laminar combustion velocity calculating means is step 168 of FIG. 54 and step 188 of FIG. 12, and the function of the combustion gas volume equivalent volume calculating means is step 362 of FIG. 54, the function of the combustion mass ratio calculating means is step 171 of FIG. 54, step 191 of FIG. 12, the function of the reaction probability calculating means is step 15 of FIG. 53, and the function of the basic ignition timing calculating means is FIG. Steps 41, 42 and 43 indicate that the function of the comparison means is in accordance with step 372 in FIG. 61, and that the function of the compression ratio estimated value calculation means is in accordance with steps 261, 391 and 334 in FIG. 54 steps 361 and 362, respectively.

  In the invention according to claim 33, the function of the compression ratio estimated value calculating means is the steps 261, 262, 371, 401 of FIGS. 63 and 68, and the function of the knock correlation parameter estimated value calculating means is the step 261 of FIG. 262, 402 or step 261 in Fig. 68. By 262, 461, the function of the knock occurrence time predicted value calculation means is step 206, 207, 208, 209, 210 in Fig. 32, 218 in Fig. 33, or step 291 in Fig. 43. 293, 209, 210, and 218 in FIG. 33, the function of the knock limit ignition timing calculation means is performed by steps 219 to 231 in FIG. 33, respectively.

  In the invention according to claim 45, the function of the knock occurrence timing detecting means is in step 262 of FIG. 50, the function of the comparing means is in step 263 of FIG. 50, and the function of the compression ratio estimated value calculating means is in step 261 of FIG. 331, 224, the function of the combustion start timing volume calculating means is based on steps 361 and 362 in FIG. 54, and the function of the combustion period calculating means is based on steps 171 in FIG. 54 and step 191 in FIG. The function is performed by steps 42 and 43 in FIG.

  In the invention according to claim 52, the function of the comparison means is based on step 372 in FIG. 55, and the function of the compression ratio estimated value calculation means is based on steps 261, 374, and 270 in FIG. 54, the function of the combustion period calculation means is performed by step 171 of FIG. 54, step 191 of FIG. 12, and the function of the basic ignition timing calculation means is performed by steps 42 and 43 of FIG. Yes.

The engine control system figure of one Embodiment. The flowchart for demonstrating ignition timing control. The pressure change figure of a combustion chamber. The characteristic view explaining the change of a combustion mass ratio. The flowchart for demonstrating calculation of the physical quantity of 1st, 2nd, 4th, 5th embodiment. Diagram explaining the positional relationship between the crankshaft of the engine and the connecting rod. The characteristic diagram of a water temperature correction coefficient. The characteristic view of an equivalence ratio correction coefficient. The characteristic figure of a standard crank angle. The flowchart for demonstrating calculation of the initial combustion period of 1st, 2nd, 4th, 5th embodiment. The characteristic figure of a temperature rise rate. The flowchart for demonstrating calculation of the main combustion period. The flowchart for demonstrating calculation of basic ignition timing. The flowchart for demonstrating calculation of an internal inert gas rate. The flowchart for demonstrating calculation of an internal inert gas amount. The flowchart for demonstrating calculation of the amount of inert gas at the time of EVC. The flowchart for demonstrating calculation of the amount of inert gas blown back during overlap. The flowchart for demonstrating the setting of a supercharging determination flag and a choke determination flag. The flowchart for demonstrating calculation of the inactive gas flow rate during the overlap when there is no supercharging and there is no choke. The flowchart for demonstrating the calculation of the inactive gas flow rate during the overlap at the time of supercharging without a choke. The flowchart for demonstrating the calculation of the inactive gas flow rate during the overlap at the time of supercharging and no choke. The flowchart for demonstrating calculation of the inactive gas flow rate during the overlap at the time of supercharging and a choke. The characteristic figure of the combustion chamber volume in the exhaust valve closing timing. The characteristic figure of the gas constant of an inert gas. The characteristic figure of the integrated effective area during overlap. Explanatory drawing of the integrated effective area during overlap. The characteristic figure of the specific heat ratio of an inert gas. The characteristic figure of the specific heat ratio of air-fuel | gaseous mixture. The characteristic view which shows the pressure history in the combustion chamber at the time of knock generation. The characteristic diagram of 1 / τ with the fuel of the octane number 100 of the first, fourth, and seventh embodiments. The characteristic figure of 1 / τ with the fuel of the octane number 80 of the first, third, fourth, and seventh embodiments. The characteristic view which shows the change of the combustion mass ratio at the time of approximating with a straight line. The flowchart for demonstrating calculation of the knock limit ignition timing of 1st, 4th, 7th embodiment. The flowchart for demonstrating calculation of the knock limit ignition timing of 1st, 2nd, 3rd, 4th, 5th, 6th, 7th, 8th embodiment. The flowchart for demonstrating calculation of a total gas mole number. The flowchart for demonstrating calculation of gas enthalpy. The characteristic diagram of a rotational speed correction coefficient. Trace knock strength characteristic chart. The flowchart for demonstrating calculation of the octane number estimated value of 1st Embodiment. The characteristic figure of the frequency component for the knock of 1st, 2nd, 3rd, 7th, 8th embodiment. The flowchart for demonstrating calculation of the self-ignition timing predicted value of 1st Embodiment. The flowchart for demonstrating calculation of the self-ignition timing predicted value of 1st Embodiment. The wave form diagram which shows the motion of the octane number estimated value at the time of knock detection of 1st Embodiment. The flowchart for demonstrating calculation of the knock limit ignition timing of 2nd, 5th, 8th embodiment. FIG. 9 is a characteristic diagram of 1 / τ with a mixed fuel having an alcohol concentration of 0% according to the second, fifth, and eighth embodiments. The characteristic diagram of 1 / τ in the mixed fuel of 85% alcohol concentration of the second, fifth and eighth embodiments. The flowchart for demonstrating calculation of the alcohol concentration estimated value of 2nd, 5th, 8th embodiment. The flowchart for demonstrating calculation of the self-ignition time estimated value of 2nd Embodiment. The flowchart for demonstrating calculation of the self-ignition time estimated value of 2nd Embodiment. The wave form diagram which shows the motion of the alcohol concentration estimated value at the time of knock detection of 2nd Embodiment. The flowchart for demonstrating calculation of the knock limit ignition timing of 3rd, 6th embodiment. The flowchart for demonstrating calculation of the compression ratio estimated value of 3rd Embodiment. The flowchart for demonstrating calculation of the self-ignition time estimated value of 3rd Embodiment. The flowchart for demonstrating calculation of the self-ignition time estimated value of 3rd Embodiment. 10 is a flowchart for explaining calculation of physical quantities according to third and sixth embodiments. The flowchart for demonstrating calculation of the initial combustion period of 3rd, 6th embodiment. The flowchart for demonstrating calculation of the octane number estimated value of 4th Embodiment. The characteristic figure of the frequency component for knock of 4th, 5th, 6th, 7th, 8th embodiment. The flowchart for demonstrating calculation of the knock strength estimated value of 4th Embodiment. The flowchart for demonstrating calculation of the knock strength estimated value of 4th, 5th, 6th embodiment. The flowchart for demonstrating calculation of the alcohol concentration estimated value of 5th Embodiment. The flowchart for demonstrating calculation of the knock strength estimated value of 5th Embodiment. The flowchart for demonstrating calculation of the compression ratio estimated value of 6th Embodiment. The flowchart for demonstrating calculation of the knock strength estimated value of 6th Embodiment. The flowchart for demonstrating calculation of the compression ratio estimated value and octane number estimated value of 7th Embodiment. The flowchart for demonstrating calculation of the compression ratio estimated value of 7th, 8th embodiment. The flowchart for demonstrating calculation of the octane number estimated value of 7th Embodiment. The flowchart for demonstrating calculation of the physical quantity of 7th, 8th embodiment. The flowchart for demonstrating calculation of the initial combustion period of 7th, 8th embodiment. The flowchart for demonstrating calculation of the compression ratio estimated value and alcohol concentration estimated value of 8th Embodiment. The flowchart for demonstrating calculation of the alcohol concentration estimated value of 8th Embodiment.

Explanation of symbols

1 Engine 5 Combustion chamber 11 Ignition device (spark ignition means)
21 Fuel Injector 31 Engine Controller 47 Knock Sensor (Knock Detection Means, Knock Strength Detection Means, Knock Generation Timing Detection Means)

Claims (58)

Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock generation timing detection means for detecting the knock generation timing in the combustion chamber;
Comparison means for comparing the knock occurrence time detection value and the knock occurrence time prediction value;
Knock correlation parameter estimated value calculating means for calculating an estimated value of the knock correlation parameter based on the comparison result and the knock detection result;
A knock occurrence time predicted value calculating means for calculating the knock occurrence time predicted value based on the knock correlation parameter estimated value;
Knock limit ignition timing calculation means for calculating the knock limit ignition timing based on the knock generation timing prediction value;
An engine knock control device comprising: ignition executing means for performing spark ignition at the knock limit ignition timing.   The knock limit ignition timing calculation means calculates pressure increase amount estimation means for estimating a pressure increase amount due to knocking of the combustion chamber based on the knock generation timing prediction value and operation conditions, and calculates basic ignition timing according to the operation conditions And a knock limit ignition timing setting means for setting, as a knock limit ignition timing, a value obtained by correcting the basic ignition timing to the retard side based on the amount of pressure increase in the fuel chamber. The engine knock control apparatus according to claim 1.   The said knock correlation parameter estimated value calculation means is an octane number estimated value calculation means for calculating an estimated octane number based on the comparison result and the knock detection result when gasoline is used as fuel. Engine knock control device.   The estimated octane value calculating means detects that a knock has occurred, and the difference in the occurrence timing between the knock occurrence timing detection value and the knock occurrence timing prediction value is within an allowable range in the combustion cycle in which the knock is detected. 4. The engine according to claim 3, wherein the estimated octane number is updated to a smaller side until the octane number estimated value is increased by a second predetermined value thereafter at a constant cycle. Knock control device.   The update to make the octane number estimated value larger by a second predetermined value is executed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device.   When the fuel is a mixed fuel of gasoline and alcohol, the knock correlation parameter estimated value calculating means is an alcohol concentration estimated value calculating means for calculating an estimated alcohol concentration value in the mixed fuel based on the comparison result and the knock detection result. The knock control apparatus for an engine according to claim 2.   When the alcohol concentration estimated value calculation means detects that a knock has occurred, the difference in the generation timing between the knock generation timing detection value and the knock generation timing prediction value is within an allowable range in the combustion cycle in which the knock is detected. The alcohol concentration estimated value is updated to a higher side until it falls within the range, and thereafter the alcohol concentration estimated value is updated to a lower side by a second predetermined value at a constant cycle. Engine knock control device.   The update to lower the estimated alcohol concentration value by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device.   The basic ignition timing calculating means includes a laminar combustion speed calculating means for calculating a laminar combustion speed that is a combustion speed in a laminar flow state of the combustion gas, and a combustion gas for calculating a volume corresponding to the combustion gas volume in the combustion chamber. A volume equivalent volume calculating means, a combustion mass ratio calculating means for calculating a combustion mass ratio of the gas combusted in the combustion chamber up to a predetermined crank angle, and a reaction indicating the ease of combustion of the combustion gas under a predetermined operating condition A reaction probability calculating means for calculating a probability, a combustion period calculating means for calculating a combustion period from the start of combustion to a predetermined crank angle based on the laminar combustion velocity, the combustion gas volume equivalent volume, the combustion mass ratio, and the reaction probability; The engine knock control apparatus according to claim 2, further comprising basic ignition timing calculation means for calculating a basic ignition timing at which MBT is obtained based on the combustion period. Laminar combustion speed calculating means for calculating a laminar combustion speed that is a combustion speed in a laminar state of the combustion gas;
A combustion gas volume equivalent volume calculating means for calculating a volume corresponding to the combustion gas volume in the combustion chamber;
Combustion mass ratio calculating means for calculating the combustion mass ratio of the gas combusted in the combustion chamber by a predetermined crank angle;
Reaction probability calculation means for calculating a reaction probability indicating the ease of combustion of combustion gas under a predetermined operating condition;
A combustion period calculating means for calculating a combustion period from the start of combustion to a predetermined crank angle based on these laminar combustion speed, combustion gas volume equivalent volume, combustion mass ratio and reaction probability;
A basic ignition timing calculation means for calculating a basic ignition timing for obtaining MBT based on the combustion period;
Dividing the combustion period into an initial combustion period and a main combustion period;
Among these, in the engine knock control device for calculating the initial combustion period using the volume of the combustion start timing of the combustion chamber as the combustion gas volume equivalent volume in the combustion chamber,
Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock generation timing detection means for detecting the knock generation timing in the combustion chamber;
Comparison means for comparing the knock occurrence time detection value and the knock occurrence time prediction value;
A compression ratio estimated value calculating means for calculating an estimated value of the compression ratio which is a knock correlation parameter based on the comparison result and the knock detection result;
An engine knock control device comprising: a combustion start time volume calculating unit that calculates a volume of the combustion chamber at a combustion start time based on the estimated compression ratio. A knock occurrence time predicted value calculating means for calculating the knock occurrence time predicted value;
Knock limit ignition timing calculation means for calculating a knock limit ignition timing based on the knock generation timing predicted value or the knock generation timing detection value;
The engine knock control device according to claim 10, further comprising ignition execution means for performing spark ignition at a retarded value of the knock limit ignition timing and the basic ignition timing.   When the compression ratio estimated value calculation means detects that a knock has occurred, the difference in the generation timing between the knock generation timing detection value and the knock generation timing prediction value is within an allowable range in the combustion cycle in which the knock is detected. The compression ratio estimated value is updated to a larger side until it falls within, and thereafter, the compression ratio estimated value is updated to a lower side by a second predetermined value at a constant cycle. Engine knock control device.   The update to reduce the compression ratio estimated value to the lower side by a second predetermined value is executed when the knock limit ignition timing is on the retard side with respect to the basic ignition timing. Engine knock control device. Calculate the effective compression ratio from the intake valve closing timing to the combustion starting timing based on the volume at the intake valve closing timing of the combustion chamber and the volume at the combustion starting timing,
The temperature at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the temperature at the combustion chamber combustion start timing. The pressure at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the pressure at the combustion chamber combustion start timing. Respectively,
When calculating the laminar combustion speed used for calculation of the initial combustion period based on the temperature and pressure at the combustion start timing of these combustion chambers,
The engine knock control device according to claim 10, wherein a volume of the combustion chamber at an intake valve closing timing is calculated based on the estimated compression ratio.   12. The knock generation time predicted value calculation means calculates the knock generation time predicted value based on a characteristic representing a distribution of a reciprocal of time until fuel in a combustion chamber reaches self-ignition. Engine knock control device.   11. The engine knock control apparatus according to claim 1, wherein the comparison result is a value obtained by subtracting the predicted knock occurrence timing from the detected knock occurrence timing. Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock strength detection means for detecting the knock strength in the combustion chamber;
A comparison means for comparing the knock strength detection value with the knock strength estimation value;
Knock correlation parameter estimated value calculating means for calculating an estimated value of the knock correlation parameter based on the comparison result and the knock detection result;
A knock generation time predicted value calculating means for calculating a knock generation time predicted value in the combustion chamber based on the knock correlation parameter estimated value;
Knock limit ignition timing calculation means for calculating the knock limit ignition timing based on the knock generation timing prediction value;
An engine knock control device comprising: ignition executing means for performing spark ignition at the knock limit ignition timing.   The knock limit ignition timing calculation means calculates pressure increase amount estimation means for estimating a pressure increase amount due to knocking of the combustion chamber based on the knock generation timing prediction value and operation conditions, and calculates basic ignition timing according to the operation conditions And a knock limit ignition timing setting means for setting, as a knock limit ignition timing, a value obtained by correcting the basic ignition timing to the retard side based on the amount of pressure increase in the fuel chamber. The engine knock control apparatus according to claim 17.   The said knock correlation parameter estimated value calculation means is an octane number estimated value calculation means for calculating an estimated octane number based on the comparison result and the knock detection result when gasoline is used as fuel. Engine knock control device.   The octane number estimated value calculating means detects the occurrence of knock when the knock intensity difference between the knock intensity estimated value and the knock intensity detected value falls within an allowable range in the combustion cycle in which the knock is detected. 20. The engine knock control according to claim 19, wherein the estimated value is updated to a smaller side, and thereafter the octane number estimated value is updated to a larger predetermined second value at a constant cycle. apparatus.   The update to increase the octane number estimated value by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device.   When the fuel is a mixed fuel of gasoline and alcohol, the knock correlation parameter estimated value calculating means is an alcohol concentration estimated value calculating means for calculating an estimated alcohol concentration value in the mixed fuel based on the comparison result and the knock detection result. The engine knock control apparatus according to claim 18, wherein   When the alcohol concentration estimated value calculation means detects that a knock has occurred, until the knock intensity difference between the knock intensity estimated value and the knock intensity detected value falls within an allowable range in the combustion cycle in which the knock is detected. 23. The engine according to claim 22, wherein the engine is a means for updating the alcohol concentration estimated value to a higher side and thereafter updating the alcohol concentration estimated value to a lower side by a second predetermined value at a constant cycle. Knock control device.   The update to reduce the estimated alcohol concentration value to a lower side by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device.   The basic ignition timing calculating means includes a laminar combustion speed calculating means for calculating a laminar combustion speed that is a combustion speed in a laminar flow state of the combustion gas, and a combustion gas for calculating a volume corresponding to the combustion gas volume in the combustion chamber. A volume equivalent volume calculating means, a combustion mass ratio calculating means for calculating a combustion mass ratio of the gas combusted in the combustion chamber up to a predetermined crank angle, and a reaction indicating the ease of combustion of the combustion gas under a predetermined operating condition A reaction probability calculating means for calculating a probability, a combustion period calculating means for calculating a combustion period from the start of combustion to a predetermined crank angle based on the laminar combustion velocity, the combustion gas volume equivalent volume, the combustion mass ratio, and the reaction probability; 19. The engine knock control apparatus according to claim 18, further comprising basic ignition timing calculation means for calculating a basic ignition timing at which MBT is obtained based on the combustion period. Laminar combustion speed calculating means for calculating a laminar combustion speed that is a combustion speed in a laminar state of the combustion gas;
A combustion gas volume equivalent volume calculating means for calculating a volume corresponding to the combustion gas volume in the combustion chamber;
Combustion mass ratio calculating means for calculating the combustion mass ratio of the gas combusted in the combustion chamber by a predetermined crank angle;
Reaction probability calculation means for calculating a reaction probability indicating the ease of combustion of combustion gas under a predetermined operating condition;
A combustion period calculating means for calculating a combustion period from the start of combustion to a predetermined crank angle based on these laminar combustion speed, combustion gas volume equivalent volume, combustion mass ratio and reaction probability;
A basic ignition timing calculation means for calculating a basic ignition timing for obtaining MBT based on the combustion period;
Dividing the combustion period into an initial combustion period and a main combustion period;
Among these, in the engine knock control device for calculating the initial combustion period using the volume of the combustion start timing of the combustion chamber as the combustion gas volume equivalent volume in the combustion chamber,
Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock strength detection means for detecting the knock strength in the combustion chamber;
A comparison means for comparing the estimated knock strength value with the detected knock strength value;
A compression ratio estimated value calculating means for calculating an estimated value of the compression ratio which is a knock correlation parameter based on the comparison result and the knock detection result;
An engine knock control device comprising: a combustion start time volume calculating unit that calculates a volume of the combustion chamber at a combustion start time based on the estimated compression ratio. Knock limit ignition timing calculation means for calculating the knock limit ignition timing based on the knock generation timing prediction value or the knock generation timing detection value;
27. The engine knock control apparatus according to claim 26, further comprising ignition execution means for performing spark ignition at a retarded value of the knock limit ignition timing and the basic ignition timing.   The compression ratio estimated value calculating means detects the occurrence of knock when the compression ratio is reduced until the knock intensity difference between the knock intensity estimated value and the knock intensity detected value is within an allowable range in the combustion cycle in which the knock is detected. 28. The engine knock according to claim 27, wherein the estimated value is updated to a larger side, and thereafter the compression ratio estimated value is updated to a lower side by a second predetermined value at a constant cycle. Control device.   29. The updating of the compression ratio estimated value to be decreased by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device. Calculate the effective compression ratio from the intake valve closing timing to the combustion starting timing based on the volume at the intake valve closing timing of the combustion chamber and the volume at the combustion starting timing,
The temperature at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the temperature at the combustion chamber combustion start timing. The pressure at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the pressure at the combustion chamber combustion start timing. Respectively,
When calculating the laminar combustion speed used for calculation of the initial combustion period based on the temperature and pressure at the combustion start timing of these combustion chambers,
27. The engine knock control apparatus according to claim 26, wherein a volume of the combustion chamber at the closing timing of the intake valve is calculated based on the estimated compression ratio.   28. The knock generation time prediction value calculation means calculates the knock generation time prediction value based on a characteristic representing a distribution of a reciprocal of time until fuel in a combustion chamber reaches self-ignition. Engine knock control device.   27. The engine knock control apparatus according to claim 17 or 26, wherein the comparison result is a value obtained by subtracting the estimated knock magnitude from the detected knock magnitude. Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock generation timing detection means for detecting the knock generation timing in the combustion chamber;
Knock strength detection means for detecting the knock strength in the combustion chamber;
A compression ratio estimated value calculating means for calculating a compression ratio estimated value based on the knock detection result, the knock occurrence timing detected value and the knock intensity detected value;
A knock correlation parameter estimated value calculating means for calculating an estimated value of a knock correlation parameter other than the compression ratio based on the knock detection result and the knock occurrence timing detected value;
A knock generation time prediction value calculating means for calculating a knock generation time prediction value in the combustion chamber based on the knock correlation parameter;
Knock limit ignition timing calculation means for calculating the knock limit ignition timing based on the knock generation timing prediction value;
An engine knock control device comprising: ignition executing means for performing spark ignition at the knock limit ignition timing.   The knock limit ignition timing calculation means calculates pressure increase amount estimation means for estimating a pressure increase amount due to knocking of the combustion chamber based on the knock generation timing prediction value and operation conditions, and calculates basic ignition timing according to the operation conditions And a knock limit ignition timing setting means for setting, as a knock limit ignition timing, a value obtained by correcting the basic ignition timing to the retard side based on the amount of pressure increase in the fuel chamber. The engine knock control apparatus according to claim 33.   The said knock correlation parameter estimated value calculation means is an octane number estimated value calculation means which calculates an octane number estimated value based on the said knock detection result and a knock generation | occurrence | production time detection value when using gasoline as a fuel. The engine knock control device described.   The octane number estimated value calculation means is the time until the fuel in the combustion chamber reaches self-ignition in the combustion cycle in which the knock is detected when it is detected that the knock has occurred, and differs depending on the pressure and temperature in the combustion chamber. The reciprocal of time is integrated from the combustion start timing to the knock generation timing detection value at every predetermined crank angle, and the estimated octane number is decreased until the absolute value of the difference between the integrated value and 1 falls within the allowable range. 36. The engine knock control apparatus according to claim 35, wherein the engine knock control unit updates the octane number estimated value at a constant cycle to the side where the estimated octane number increases by a second predetermined value thereafter.   The update of the octane number estimated value to be increased by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device.   In the case of a mixed fuel of gasoline and alcohol, the knock correlation parameter estimated value calculating means is an alcohol concentration estimated value calculating means for calculating an alcohol concentration estimated value in the mixed fuel based on the knock detection result and the knock generation timing detected value. The engine knock control apparatus according to claim 34, wherein the engine knock control apparatus is provided.   The alcohol concentration estimated value calculating means is a time until the fuel in the combustion chamber reaches self-ignition in the combustion cycle in which the knock is detected when it is detected that the knock has occurred, and differs depending on the pressure and temperature in the combustion chamber. The reciprocal of the time to be accumulated is accumulated for every predetermined crank angle from the combustion start timing to the knock occurrence timing detection value, and the alcohol concentration estimated value is increased until the absolute value of the difference between the accumulated value and 1 falls within the allowable range. 39. The engine knock control apparatus according to claim 38, wherein the engine knock control unit updates the estimated alcohol concentration value by a second predetermined value and then updates the estimated alcohol concentration value at a constant cycle.   40. The updating of the alcohol concentration estimated value to be decreased by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine knock control device.   The basic ignition timing calculating means includes a laminar combustion speed calculating means for calculating a laminar combustion speed that is a combustion speed in a laminar flow state of the combustion gas, and a combustion gas for calculating a volume corresponding to the combustion gas volume in the combustion chamber. A volume equivalent volume calculating means, a combustion mass ratio calculating means for calculating a combustion mass ratio of the gas combusted in the combustion chamber up to a predetermined crank angle, and a reaction indicating the ease of combustion of the combustion gas under a predetermined operating condition A reaction probability calculating means for calculating a probability, a combustion period calculating means for calculating a combustion period from the start of combustion to a predetermined crank angle based on the laminar combustion velocity, the combustion gas volume equivalent volume, the combustion mass ratio, and the reaction probability; 35. The engine knock control apparatus according to claim 34, further comprising basic ignition timing calculation means for calculating a basic ignition timing at which MBT is obtained based on the combustion period. Dividing the combustion period into an initial combustion period and a main combustion period;
Of these, an initial combustion period calculating means for calculating the initial combustion period using the volume at the compression start timing of the combustion chamber as a volume corresponding to the combustion gas volume in the combustion chamber;
42. The engine knock control device according to claim 41, further comprising: a compression start time volume calculating means for calculating a volume of the combustion chamber at a compression start time based on the estimated compression ratio value. Calculate the effective compression ratio from the intake valve closing timing to the combustion starting timing based on the volume at the intake valve closing timing of the combustion chamber and the volume at the combustion starting timing,
The temperature at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the temperature at the combustion chamber combustion start timing. The pressure at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the pressure at the combustion chamber combustion start timing. Respectively,
When calculating the laminar combustion speed used for calculation of the initial combustion period based on the temperature and pressure at the combustion start timing of these combustion chambers,
43. The engine knock control apparatus according to claim 42, wherein a volume of the combustion chamber at the closing timing of the intake valve is calculated based on the estimated compression ratio.   The compression ratio estimated value calculating means is based on the means for calculating the combustion mass ratio in the knock generation timing detection value, the means for calculating the unburned fuel amount in the knock generation timing detection value, and the knock intensity detection value. Based on the means for calculating the amount of pressure increase due to knock, the combustion mass ratio in the detected value of knock occurrence time, the amount of unburned fuel in the detected value of knock occurrence time, and the amount of pressure increase due to knock, the volume at the detected value of knock occurrence time 34. The method according to claim 33, further comprising means for calculating, means for calculating a clearance volume based on the volume at the knock occurrence time detection value, and means for calculating an estimated compression ratio based on the clearance volume. The engine knock control device described. Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock generation timing detection means for detecting the knock generation timing in the combustion chamber;
Comparison means for comparing the knock occurrence time detection value and the knock occurrence time prediction value;
A compression ratio estimated value calculating means for calculating an estimated value of the compression ratio based on the comparison result and the knock detection result;
Combustion start time volume calculating means for calculating the volume of the combustion chamber at the combustion start time based on the compression ratio estimated value;
Combustion period calculation means for calculating a combustion period from the start of combustion to a predetermined crank angle based on the volume at the start of combustion;
Basic ignition timing calculating means for calculating a basic ignition timing for obtaining MBT based on the combustion period;
An ignition timing control device for an engine, comprising: ignition executing means for performing spark ignition at the basic ignition timing. A knock occurrence time predicted value calculating means for calculating the knock occurrence time predicted value;
Knock limit ignition timing calculation means for calculating a knock limit ignition timing based on the knock generation timing predicted value or the knock generation timing detection value;
46. The engine ignition timing control device according to claim 45, further comprising ignition execution means for performing spark ignition at a retarded value of the knock limit ignition timing and the basic ignition timing.   When the compression ratio estimated value calculation means detects that a knock has occurred, the difference in the generation timing between the knock generation timing detection value and the knock generation timing prediction value is within an allowable range in the combustion cycle in which the knock is detected. 47. The unit according to claim 46, wherein the compression ratio estimation value is updated to a larger side until it falls within, and thereafter the compression ratio estimation value is updated at a constant period to a lower side by a second predetermined value. Engine ignition timing control device.   The update to reduce the compression ratio estimated value to the lower side by a second predetermined value is executed when the knock limit ignition timing is on the retard side with respect to the basic ignition timing. Engine ignition timing control device. Calculate the effective compression ratio from the intake valve closing timing to the combustion starting timing based on the volume at the intake valve closing timing of the combustion chamber and the volume at the combustion starting timing,
The temperature at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the temperature at the combustion chamber combustion start timing. The pressure at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the pressure at the combustion chamber combustion start timing. Respectively,
When calculating the laminar combustion speed used for calculation of the initial combustion period based on the temperature and pressure at the combustion start timing of these combustion chambers,
46. The engine ignition timing control device according to claim 45, wherein a volume of the combustion chamber at the closing timing of the intake valve is calculated based on the estimated value of the compression ratio.   47. The knock generation timing predicted value calculation means calculates the knock generation timing predicted value based on a characteristic representing a distribution of a reciprocal of time until fuel in a combustion chamber reaches self-ignition. Engine ignition timing control device.   46. The engine ignition timing control apparatus according to claim 45, wherein the comparison result is a value obtained by subtracting the knock occurrence timing predicted value from the knock occurrence timing detection value. Knock detecting means for detecting whether or not knock actually occurs in the combustion chamber;
Knock strength detection means for detecting the knock strength in the combustion chamber;
A comparison means for comparing the knock strength detection value with the knock strength estimation value;
A compression ratio estimated value calculating means for calculating an estimated value of the compression ratio based on the comparison result and the knock detection result;
Combustion start time volume calculating means for calculating the volume of the combustion chamber at the combustion start time based on the compression ratio estimated value;
Combustion period calculation means for calculating a combustion period from the start of combustion to a predetermined crank angle based on the volume at the start of combustion;
Basic ignition timing calculating means for calculating a basic ignition timing for obtaining MBT based on the combustion period;
An ignition timing control device for an engine, comprising: ignition executing means for performing spark ignition at the basic ignition timing. Knock limit ignition timing calculation means for calculating the knock limit ignition timing based on the knock generation timing prediction value or the knock generation timing detection value;
53. The engine ignition timing control apparatus according to claim 52, further comprising: ignition execution means for performing spark ignition at a retarded side value of the knock limit ignition timing and the basic ignition timing.   The compression ratio estimated value calculating means detects the occurrence of knock when the compression ratio is reduced until the knock intensity difference between the knock intensity estimated value and the knock intensity detected value is within an allowable range in the combustion cycle in which the knock is detected. 54. The engine ignition according to claim 53, wherein the estimated value is updated to a larger side, and thereafter, the compression ratio estimated value is updated at a constant cycle to a lower side by a second predetermined value. Timing control device.   55. The updating of the compression ratio estimated value to be lowered by a second predetermined value is performed when the knock limit ignition timing is retarded from the basic ignition timing. Engine ignition timing control device. Calculate the effective compression ratio from the intake valve closing timing to the combustion starting timing based on the volume at the intake valve closing timing of the combustion chamber and the volume at the combustion starting timing,
The temperature at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the temperature at the combustion chamber combustion start timing. The pressure at the combustion chamber intake valve closing timing and the effective compression ratio are used to determine the pressure at the combustion chamber combustion start timing. Respectively,
When calculating the laminar combustion speed used for calculation of the initial combustion period based on the temperature and pressure at the combustion start timing of these combustion chambers,
53. The engine ignition timing control device according to claim 52, wherein a volume of the combustion chamber at an intake valve closing timing is calculated based on the estimated compression ratio.   54. The knock generation time predicted value calculation means calculates the knock generation time predicted value based on a characteristic representing a distribution of a reciprocal of time until fuel in a combustion chamber reaches self-ignition. Engine ignition timing control device.   53. The engine ignition timing control apparatus according to claim 52, wherein the comparison result is a value obtained by subtracting the knock magnitude estimated value from the knock magnitude detected value.

Images