JP2010169038A - Device for determining variation in air-fuel ratio among cylinders of multiple cylinder internal combustion engine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To promptly and certainly determine occurrence of "variation in an air-fuel ratio among cylinders" caused by abnormality of a fuel injection valve, in a multiple cylinder internal combustion engine including a dual injection system. <P>SOLUTION: An integration value (sum value) of deviation between the output value of an air-fuel ratio sensor downstream of a catalyst and a value equivalent to a theoretical air-fuel ratio is individually calculated with respect to each of injection modes (P, D, PD modes). Air-fuel ratio feedback control is performed based on the sum value corresponding to the present injection mode and the output value of the air-fuel ratio sensor upstream of the catalyst. During occurrence of "the variation in the air-fuel ratio AF", by using such a phenomenon that the sum value corresponding to the present injection mode is increased in a fuel increasing direction by "rich deviation" of the output value of the air-fuel ratio sensor upstream of the catalyst, occurrence of "the variation in the air-fuel ratio AF" is detected when the sum value corresponding to the present injection mode reaches a guard value Guard. During the P(D) mode, when the sum value SUMp(SUMd) for the P(D) mode is a value in an increasing direction and "SUMp(SUMd)-SUMd(SUMp)>A" is satisfied, the injection mode is fixed to the P(D) mode. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention is applied to a multi-cylinder internal combustion engine, and feeds back an air-fuel ratio based on output values of an upstream air-fuel ratio sensor and a downstream air-fuel ratio sensor respectively disposed upstream and downstream of a catalyst disposed in an exhaust passage. The present invention relates to an apparatus that controls and determines (detects) variations in air-fuel ratio among cylinders. Hereinafter, the “internal combustion engine” may be simply referred to as “engine”.

  Conventionally, as this type of apparatus, for example, one disclosed in Patent Document 1 is known. In this apparatus, a catalyst is disposed in a collective exhaust passage formed by a collection of exhaust passages extending from each cylinder, and an upstream air-fuel ratio sensor and a downstream air-fuel ratio sensor are disposed in upstream and downstream collective exhaust passages of the catalyst. Are arranged respectively. The deviation between the output value of the downstream air-fuel ratio sensor and the target value corresponding to the target air-fuel ratio (theoretical air-fuel ratio) is proportional / integrated / differentiated (PID processing) to calculate the feedback correction amount. Based on a value obtained by correcting the output value of the upstream air-fuel ratio sensor with this feedback correction amount, the port is set so that the air-fuel ratio of the exhaust gas (mixed exhaust gas) passing through the collective exhaust passage becomes the target air-fuel ratio (theoretical air-fuel ratio). The amount of fuel injected from the injection valve is feedback controlled. The port injection valve is a fuel injection valve that injects fuel in an intake passage (intake port) upstream of the intake valve.

  In general, the difference between the intake air flow rate measured by the air flow meter used to determine the amount of fuel injected from the fuel injection valve and the actual air flow rate (error of the air flow meter), and the injection instruction to the fuel injection valve The difference between the commanded fuel injection amount and the amount of actually injected fuel (error of the fuel injection valve) or the like (hereinafter, these are collectively referred to as “error of the fuel injection amount”) may occur.

  The feedback correction amount includes the value of the integral term (I term), that is, the value obtained by multiplying the deviation integral value that is updated by integrating the “deviation” by the feedback gain. As a result, even if the “fuel injection amount error” occurs, the “fuel injection amount error” can be compensated by the deviation integral value (integral term) by executing the feedback control described above. The fuel ratio can be matched and converged with the target air-fuel ratio. In other words, the value of the deviation integral value (integral term) can be a value representing the magnitude of the “error in fuel injection amount”.

  In the apparatus described in Patent Document 1, a “learning process” of a deviation integral value having such a character is executed. Specifically, every time a predetermined timing arrives, the steady component of the deviation integral value (specifically, the value obtained by low-pass filtering the deviation integral value) becomes the learning value (the steady component of the deviation integral value). Is obtained as an update value for updating. The updated value is added to the learned value stored in the backup RAM (SRAM) at that time to update the learned value. Then, the updated value is subtracted from the deviation integrated value at that time.

  In this way, every time a predetermined timing arrives, the steady component of the deviation integral value becomes the learned value without changing the sum of the deviation integral value and the learned value (hereinafter referred to as “sum total value”). It will be transferred. That is, the total value functions as a substantial deviation integral value in the feedback correction amount. In other words, in the apparatus described in Patent Document 1, the air-fuel ratio is feedback-controlled based on the feedback correction amount based on the total value (including the total value).

  By the way, in a multi-cylinder internal combustion engine, variation in the injection amount from the fuel injection valve between cylinders, variation in the maximum lift amount of the intake valve, variation in distribution of the EGR gas amount recirculated to the intake system by the EGR mechanism for each cylinder. Etc. may occur. When the characteristic variation between the cylinders occurs, the air-fuel ratio variation (inter-cylinder air-fuel ratio variation) may occur between the cylinders. When the air-fuel ratio variation between cylinders occurs, even if the air-fuel ratio of the mixed exhaust gas coincides with the stoichiometric air-fuel ratio, the air-fuel ratio and the cylinder (rich cylinder) in which the air-fuel ratio becomes rich (rather than the stoichiometric air-fuel ratio) There may be a cylinder (lean cylinder) that is leaner (than the theoretical air-fuel ratio).

  Generally, hydrogen is discharged from the rich cylinder. As the richness of the air-fuel ratio of the rich cylinder increases, the hydrogen concentration of the exhaust gas discharged from the rich cylinder increases (see FIG. 5 described later). On the other hand, when the air-fuel ratio of the mixed exhaust gas coincides with the stoichiometric air-fuel ratio, the richer the air-fuel ratio of the rich cylinder, the greater the degree of variation in the inter-cylinder air-fuel ratio. Therefore, the greater the degree of variation in the air-fuel ratio between the cylinders, the higher the hydrogen concentration of the exhaust gas discharged from the rich cylinder (and hence the hydrogen concentration of the mixed exhaust gas) (see FIG. 7 described later).

  In general, an oxygen concentration sensor (particularly, a limiting current oxygen concentration sensor) is used as the upstream air-fuel ratio sensor. When the exhaust gas contains hydrogen, it is known that the higher the hydrogen concentration of the exhaust gas, the more the output of the oxygen concentration sensor shows an air-fuel ratio that is shifted to a richer side than the true air-fuel ratio of the exhaust gas. This phenomenon is called “rich shift”. This “rich shift” means that hydrogen molecules that are smaller than oxygen molecules enter the detection unit of the oxygen concentration sensor before the oxygen molecules, making it difficult for the oxygen molecules to ionize within the detection unit. It is thought to be caused by.

  From the above, when the air-fuel ratio of the mixed exhaust gas matches the stoichiometric air-fuel ratio, the larger the degree of variation in the air-fuel ratio between the cylinders, the higher the output of the upstream air-fuel ratio sensor is due to the “rich deviation”. The air-fuel ratio deviated to the rich side with respect to the true air-fuel ratio (that is, the stoichiometric air-fuel ratio) of the mixed exhaust gas is shown.

  Therefore, during the feedback control described above, if a “rich deviation” occurs in the upstream air-fuel ratio sensor output, the fuel injection amount is reduced by the “rich deviation”, and the air-fuel ratio of the mixed exhaust gas becomes the “rich deviation”. An attempt is made to adjust the air / fuel ratio leaner than the stoichiometric air / fuel ratio. For this reason, the exhaust gas whose air-fuel ratio is leaner than the stoichiometric air-fuel ratio flows into the catalyst, and as a result, the state where the air-fuel ratio of the exhaust gas reaching the downstream air-fuel ratio sensor becomes leaner than the stoichiometric air-fuel ratio continues.

  Thus, when the state in which the air-fuel ratio of the exhaust gas reaching the downstream air-fuel ratio sensor is leaner than the stoichiometric air-fuel ratio continues, the above-described deviation integral value (or learning value) increases the fuel injection amount. It gradually increases in the (increase direction). In other words, during the above-described feedback control, when “inter-cylinder air-fuel ratio variation” occurs, the above-described deviation integral value (or learning value) gradually increases in the increasing direction.

  Based on such knowledge, the present applicant, in Japanese Patent Application No. 2007-192474 (Non-Patent Document), when the above-mentioned deviation integral value (or learning value) reaches a predetermined guard value in the increasing direction, It has already been proposed to determine that the “inter-air-fuel variation” has occurred.

  By the way, in recent years, a fuel injection valve (hereinafter referred to as “in-cylinder injection valve”) that directly injects fuel in a combustion chamber is provided as a main fuel injection valve in order to achieve improvement in combustion efficiency and reduction in fuel consumption. At the same time, an internal combustion engine having a fuel injection valve that also follows the port injection valve has been developed in order to ensure startability at the time of engine startup (see, for example, Patent Document 2). Hereinafter, a system including two fuel injection valves, i.e., the in-cylinder injection valve and the port injection valve, is referred to as a “dual injection system”.

  In such a dual injection system, normally, a port injection mode in which fuel is injected only from the port injection valve, a cylinder injection mode in which fuel is injected only from the cylinder injection valve, and fuel from the port injection valve and cylinder injection valve. The two injection modes in which the fuel is injected are selectively used according to the operating state of the engine (see FIG. 9 described later). The present inventor has described the above-described “when the deviation integral value (or learning value) reaches a predetermined guard value in the increasing direction” for the multi-cylinder internal combustion engine having such a dual injection system. It has been found that a new problem occurs when the “apparatus that determines that“ air-fuel variation ”occurs” is applied.

JP 2005-113729 A JP 2004-60474 A

  Hereinafter, the new problem will be specifically described. Generally, among a plurality of port injection valves and a plurality of in-cylinder injection valves provided in a multi-cylinder internal combustion engine equipped with a dual injection system, a plurality of port injection valves (a part thereof) and a plurality of in-cylinder injection valves ( It is rare that an abnormality (including the occurrence of a misfire such as a valve clogging) occurs at the same time. For example, it is assumed that an abnormality has occurred only in (a part of) a plurality of port injection valves. It is assumed that “in-cylinder air-fuel ratio variation” due to other components does not occur. Further, it is assumed that one type of deviation integral value (or learning value) is calculated and updated regardless of which injection mode is selected.

  In this case, in the port injection mode, a “rich deviation” occurs in the upstream air-fuel ratio sensor due to the occurrence of “inter-cylinder air-fuel ratio variation”. Therefore, based on the upstream air-fuel ratio sensor output value, the air-fuel ratio of the mixed exhaust gas tends to be adjusted to a leaner air-fuel ratio than the stoichiometric air-fuel ratio. As a result, when the port injection mode continues, the deviation integral value (or learning value) gradually increases in the increasing direction. In contrast, in the in-cylinder injection mode, no “rich deviation” occurs in the upstream air-fuel ratio sensor because “inter-cylinder air-fuel ratio variation” does not occur. Therefore, the air-fuel ratio of the mixed exhaust gas tends to be adjusted to the stoichiometric air-fuel ratio itself based on the upstream air-fuel ratio sensor output value. As a result, when the in-cylinder injection mode continues, the deviation integral value (or learning value) converges toward zero.

  Therefore, when the operating state in which the port injection mode is selected continues for a relatively long time, the deviation integral value (or learning value) can reach a predetermined guard value in the increasing direction. It can be determined that "variation" has occurred. On the other hand, when the operating state of the engine frequently changes, the selected injection mode can be frequently switched between the port injection mode and the in-cylinder injection mode (and both injection modes). In such a case, even if the deviation integrated value (or learning value) increases in the increasing direction in the port injection mode, the deviation integrated value (or learning value) is set to zero in the subsequent in-cylinder injection mode. It decreases (see FIG. 10A described later).

  For this reason, when the injection mode is frequently switched between the port injection mode and the in-cylinder injection mode, the deviation integral value (or learning value) cannot reach the predetermined guard value in the increasing direction (or the guard value). As a result, there is a problem that it cannot be determined that the “inter-cylinder air-fuel ratio variation” has occurred (or it takes a long time to make a determination).

  The present invention has been made to cope with such a problem, and an object of the present invention is to generate `` inter-cylinder air-fuel ratio variation '' caused by abnormality of a fuel injection valve in a multi-cylinder internal combustion engine equipped with a dual injection system. An object of the present invention is to provide an inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine that can make an early and reliable determination.

  An inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine equipped with a dual injection system according to the present invention has a catalyst (for example, a three-way catalyst) in a collective exhaust passage formed by collecting exhaust passages extending from each cylinder. And an upstream air-fuel ratio sensor and a downstream air-fuel ratio sensor upstream and downstream of the catalyst. Here, in particular, as the upstream air-fuel ratio sensor, an oxygen concentration sensor (for example, a limiting current oxygen concentration sensor) that can generate the above-described “rich shift” when hydrogen is contained in the exhaust gas is used. As the downstream air-fuel ratio sensor, for example, an electromotive force type oxygen concentration sensor can be used.

  The determination apparatus according to the present invention includes selection means, integral value calculation means, air-fuel ratio control means, and determination means. In the selection means, (one) injection mode is selected as the selection mode from at least the port injection mode and the in-cylinder injection mode based on the operating state of the internal combustion engine. In addition to the port injection mode and the in-cylinder injection mode, both the injection modes may be selected as the selection mode.

  In the integral value calculation means, when the port injection mode is selected, the port deviation integral value is calculated / updated as the deviation integral value, and when the in-cylinder injection mode is selected, the in-cylinder deviation is calculated as the deviation integral value. The integral value is calculated / updated separately from the port deviation integral value. That is, each time the selection mode is switched, the deviation integral value corresponding to the new selection mode after switching is updated from the value at the time of switching. The deviation integrated value that does not correspond to the new selection mode after switching is kept constant at the value at the time of switching.

  In the air-fuel ratio control means, the selected injection is based on at least the selected deviation integral value (deviation integral value corresponding to the selection mode among the port deviation integral value and the in-cylinder deviation integral value) and the output value of the upstream air-fuel ratio sensor. By adjusting the amount of fuel injected from the valve (the injection valve corresponding to the selection mode among the port injection valve and the in-cylinder injection valve), the air-fuel ratio (of the mixed exhaust gas) matches the stoichiometric air-fuel ratio. Feedback controlled.

  Then, the determination unit determines that “inter-cylinder air-fuel ratio variation” has occurred based on the selected deviation integral value. Specifically, for example, when the selected deviation integral value reaches a predetermined guard value in the increasing direction, it can be determined that “inter-cylinder air-fuel ratio variation” has occurred.

  According to the above configuration, the deviation integral value is calculated and updated individually for each injection mode. Therefore, when the above-mentioned “when abnormality (including occurrence of misfire) occurs only in (a part of) a plurality of port injection valves” is taken as an example again, each time the port injection mode is selected, “ The port deviation integral value can increase in the increasing direction due to the “inter-cylinder air-fuel ratio variation” (“rich deviation”) (the in-cylinder deviation integral value is kept constant). On the other hand, when the in-cylinder injection mode is selected, “in-cylinder air-fuel ratio variation” (“rich deviation”) does not occur, so that the in-cylinder deviation integrated value converges to zero, while the port deviation integrated value becomes Maintained constant.

  In this way, unlike the above-described case where “one type of deviation integral value (or learning value) is calculated / updated regardless of which injection mode is selected”, even when the in-cylinder injection mode is selected. The port deviation integral value does not decrease. Therefore, when the injection mode is frequently switched between the port injection mode and the in-cylinder injection mode, the timing at which the port deviation integral value reaches the guard value in the increasing direction, and accordingly, “inter-cylinder air-fuel ratio variation” occurs. The time when it is determined to be earlier is earlier.

  In addition to this, in the determination device according to the present invention, the timing at which it is determined that “inter-cylinder air-fuel ratio variation” has occurred is further advanced by the following configuration. That is, when the port injection mode is selected, the “value based on the port deviation integrated value” is the value in the increasing direction, and the “value based on the port deviation integrated value” in the increasing direction is “the in-cylinder deviation”. When the amount larger than the “value based on the integral value” exceeds the predetermined value, the selection mode is fixed to the port injection mode.

  Here, the “value based on the port deviation integral value” means the port deviation integral value itself, the learning value (port learning value) described above for the port deviation integral value, or the sum of the port deviation integral value and the port learning value, It is. Similarly, the “value based on the in-cylinder deviation integrated value” means the in-cylinder deviation integrated value itself, the above-described learning value (in-cylinder learning value) for the in-cylinder deviation integrated value, or the in-cylinder deviation integrated value and the in-cylinder deviation integrated value. This is the sum of the learning values.

  The above-mentioned case “when an abnormality occurs only in (a part of) a plurality of port injection valves” is taken as an example again. In this case, as described above, the in-cylinder deviation integral value is maintained at substantially zero in the in-cylinder injection mode, and is constant (at substantially zero) in the port injection mode. On the other hand, the port deviation integral value can be maintained constant in the in-cylinder injection mode while increasing in the increasing direction in the port injection mode. Based on this point of view, in the port injection mode, the port deviation integrated value is a value in the increasing direction, and the difference between the port deviation integrated value and the in-cylinder deviation integrated value exceeds the predetermined value in the increasing direction. It can be estimated that “an abnormality has occurred only in (a part of) the plurality of port injection valves”. In this case, if the injection mode is fixed to the port injection mode, the port deviation integrated value can be continuously increased, and as a result, the timing at which the port deviation integrated value reaches the guard value in the increasing direction can be made earlier. The above configuration is based on such knowledge.

  As described above, according to the determination apparatus according to the present invention, the deviation integral value is calculated and updated individually for each injection mode, and “abnormality occurs only in (a part of) a plurality of port injection valves”. By fixing the injection mode to the port injection mode when it can be estimated that the “in-cylinder” is being performed, it is possible to quickly and reliably determine the occurrence of “inter-cylinder air-fuel ratio variation” due to the abnormality of the port injection valve.

  Further, in the determination device according to the present invention, the cylinder injection mode is selected based on the same knowledge as described above in the case of “an abnormality has occurred only in (a part of) a plurality of cylinder injection valves”. In this case, the “value based on the in-cylinder deviation integral value” is the value in the increasing direction, and the “value based on the in-cylinder deviation integrated value” is the “value based on the port deviation integrated value” in the increasing direction. When the large amount exceeds the predetermined value, the selection mode is fixed to the in-cylinder injection mode. As a result, it is possible to quickly and reliably determine whether or not “in-cylinder air-fuel ratio variation” due to abnormality of the cylinder injection valve occurs.

  As described above, in the determination apparatus according to the present invention, when the “inter-cylinder air-fuel ratio variation” is caused by the abnormality of the fuel injection valve, it is determined whether the fuel injection valve in which the abnormality has occurred is a port injection valve or a cylinder injection valve. It can also be specified.

1 is a diagram schematically illustrating an inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine according to a first embodiment of the present invention. FIG. 2 is a view showing a state in which a catalyst, an upstream air-fuel ratio sensor, and a downstream air-fuel ratio sensor shown in FIG. 1 are arranged in a collective exhaust passage. 2 is a graph showing the relationship between the output of the upstream air-fuel ratio sensor shown in FIG. 1 and the air-fuel ratio. 3 is a graph showing the relationship between the output of the downstream air-fuel ratio sensor shown in FIG. 1 and the air-fuel ratio. It is the figure which showed an example of the air fuel ratio of each cylinder in case the AF dispersion | variation between cylinders has generate | occur | produced and the air fuel ratio of mixed exhaust gas corresponds with a theoretical air fuel ratio. It is the graph which showed the relationship between an air fuel ratio and the hydrogen concentration of waste gas. It is the graph which showed the relationship between the degree of AF variation between cylinders, and the hydrogen concentration of mixed exhaust gas. It is the time chart which showed an example of the change of a sum total value when the AF dispersion | variation between cylinders generate | occur | produces. It is the graph which showed the table which prescribes | regulates the relationship between an engine speed and a load factor, and the injection mode selected. It is the time chart which showed an example of change of a sum total value when AF variation between cylinders resulting from abnormality of a port injection valve occurs when injection mode changes frequently. 2 is a flowchart showing a fuel injection control routine executed by a CPU shown in FIG. 3 is a flowchart showing a routine executed by the CPU shown in FIG. 1 for calculating an air-fuel ratio feedback correction amount. 3 is a flowchart showing a routine executed by the CPU shown in FIG. 1 to calculate a sub feedback correction amount. It is the flowchart which showed the routine which CPU shown in FIG. 1 performs in order to update the learning value of a deviation integral value. It is the flowchart which showed the routine which CPU shown in FIG. 1 performs in order to select injection mode. 3 is a flowchart showing a routine executed by the CPU shown in FIG. 1 to detect inter-cylinder AF variation. It is the flowchart which showed the routine which CPU of the air-fuel ratio variation determination apparatus between cylinders concerning 2nd Embodiment of this invention performs in order to select an injection mode. It is the flowchart which showed the routine which CPU of the air-fuel ratio variation judging device between cylinders concerning 2nd Embodiment of this invention performs in order to detect AF variation between cylinders.

  Embodiments of an inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine according to the present invention will be described below with reference to the drawings.

(First embodiment)
FIG. 1 shows a schematic configuration of a system in which an inter-cylinder air-fuel ratio variation determining apparatus according to a first embodiment of the present invention is applied to a (gasoline) four-cycle spark ignition multi-cylinder internal combustion engine 10 equipped with a dual injection system. Yes. FIG. 1 shows only a cross section of a specific cylinder, but the other cylinders have the same configuration.

  The internal combustion engine 10 includes a cylinder block portion 20 including a cylinder block, a cylinder block lower case, an oil pan, and the like, a cylinder head portion 30 fixed on the cylinder block portion 20, and a gasoline mixture to the cylinder block portion 20. An intake system 40 for supplying and an exhaust system 50 for releasing exhaust gas from the cylinder block 20 to the outside are included.

  The cylinder block unit 20 includes a cylinder 21, a piston 22, a connecting rod 23, and a crankshaft 24. The piston 22 reciprocates in the cylinder 21, and the reciprocating motion of the piston 22 is transmitted to the crankshaft 24 through the connecting rod 23, whereby the crankshaft 24 rotates. The heads of the cylinder 21 and the piston 22 form a combustion chamber 25 together with the cylinder head portion 30.

  The cylinder head portion 30 includes an intake port 31 communicating with the combustion chamber 25, an intake valve 32 that opens and closes the intake port 31, an intake camshaft that drives the intake valve 32, and a phase angle of the intake camshaft and the intake valve 32. Variable valve mechanism 33 for continuously changing the maximum lift amount, actuator 33a of variable valve mechanism device 33, exhaust port 34 communicating with combustion chamber 25, exhaust valve 35 for opening and closing exhaust port 34, and driving exhaust valve 35 An exhaust camshaft 36, an ignition plug 37, an igniter 38 including an ignition coil for generating a high voltage to be applied to the ignition plug 37, a port injection valve 39P for injecting fuel at an intake port 31 upstream of the intake valve 32, a combustion chamber An in-cylinder injection valve 39 </ b> D that directly injects fuel within the cylinder 25 is provided.

  The intake system 40 includes an intake pipe 41 including an intake manifold connected to the intake port 31 of each cylinder, an air filter 42 provided at an end of the intake pipe 41, and an intake opening area within the intake pipe 41. A variable throttle valve 43 and an actuator 43a for the throttle valve 43 are provided. The intake port 31 and the intake pipe 41 constitute an intake passage.

  The exhaust system 50 includes an exhaust manifold 51 connected to the exhaust port 34 of each cylinder, an exhaust pipe 52 connected to a collection portion of the exhaust manifold 51, a catalyst 53 (three-way catalyst) disposed in the exhaust pipe 52, A catalyst 54 (three-way catalyst) disposed in the exhaust pipe 52 downstream of the catalyst 53 is provided. The exhaust port 34, the exhaust manifold 51, and the exhaust pipe 52 constitute an exhaust passage.

  On the other hand, this system includes a hot-wire air flow meter 61, a throttle position sensor 62, a cam position sensor 63, a crank position sensor 64, a water temperature sensor 65, an upstream air-fuel ratio sensor 66, a downstream air-fuel ratio sensor 67, and an accelerator opening sensor. 68.

  The air flow meter 61 outputs a signal corresponding to the mass flow rate (intake air flow rate) Ga of intake air flowing through the intake pipe 41. The throttle position sensor 62 detects the opening degree of the throttle valve 43 and outputs a signal representing the throttle valve opening degree TA.

  The cam position sensor 63 generates a signal (G2 signal) having one pulse every time the intake camshaft rotates 90 ° (that is, every time the crankshaft 24 rotates 180 °). The crank position sensor 64 outputs a signal having a narrow pulse every time the crankshaft 24 rotates 10 °, and a wide pulse every time the crankshaft 24 rotates 360 °. This signal represents the engine speed NE. The water temperature sensor 65 detects the temperature of the cooling water of the internal combustion engine 10 and outputs a signal representing the cooling water temperature THW.

  As shown in FIG. 2, the upstream air-fuel ratio sensor 66 is disposed upstream of the catalyst 53 in a collective exhaust passage formed by collecting exhaust passages extending from the cylinders. Hereinafter, the exhaust gas passing through the collective exhaust passage is referred to as “mixed exhaust gas”. The mixed exhaust gas is an exhaust gas obtained by mixing exhaust gases discharged from each cylinder.

  The upstream air-fuel ratio sensor 66 is a so-called “limit current type oxygen concentration sensor”, and detects the air-fuel ratio of the mixed exhaust gas flowing into the catalyst 53, and as shown in FIG. The signal Vabyf (V) corresponding to the (fuel ratio) is output.

  As shown in FIG. 2, the downstream air-fuel ratio sensor 67 is disposed downstream of the catalyst 53 in the collective exhaust passage. The downstream air-fuel ratio sensor 67 is a so-called “concentration cell type oxygen sensor”. As shown in FIG. 4, the downstream side air-fuel ratio sensor 67 is configured to output a maximum output value max (V) and a minimum output value min () when the air-fuel ratio of the mixed exhaust gas flowing out from the catalyst 53 is richer and leaner than the stoichiometric air-fuel ratio. V) and when the air-fuel ratio of the mixed exhaust gas flowing out from the catalyst 53 is the stoichiometric air-fuel ratio, a value approximately halfway between the maximum output value max and the minimum output value min (target air-fuel ratio equivalent target value Voxsref (V)) Is output. The accelerator opening sensor 68 outputs a signal representing the operation amount Accp of the accelerator pedal 81 operated by the driver.

  The electric control device 70 includes a CPU 71 connected to each other by a bus, a ROM 72 in which programs executed by the CPU 71, tables (maps, functions), constants, and the like are stored in advance, a RAM 73 in which the CPU 71 temporarily stores data as necessary, The microcomputer includes a backup RAM 74 that stores data while the power is turned on and retains the stored data while the power is shut off, and an interface 75 that includes an AD converter.

  The interface 75 is connected to the sensors 61 to 68, supplies signals from the sensors 61 to 68 to the CPU 71, and in accordance with instructions from the CPU 71, the actuator 33a, the igniter 38, the port injection valve 39P, Drive signals are sent to the in-cylinder injection valve 39D and the throttle valve actuator 43a.

(Outline of air-fuel ratio feedback control)
Next, an overview of air-fuel ratio feedback control by the apparatus according to the first embodiment configured as described above (hereinafter also referred to as “this apparatus”) will be described. This apparatus is configured so that the air-fuel ratio of the mixed gas flowing out from the catalyst 53 becomes the target air-fuel ratio (= theoretical air-fuel ratio). The fuel ratio, hereinafter simply referred to as “air-fuel ratio”) is controlled.

  Specifically, in this apparatus, the output value Voxs (V) of the downstream air-fuel ratio sensor 67 disposed downstream of the catalyst 53 and the target air-fuel ratio equivalent target value Voxsref (V) (constant) corresponding to the theoretical air-fuel ratio. ) To obtain a feedback correction value (sub feedback correction amount Vafsfb (%)). The air-fuel ratio is feedback controlled based on a value obtained by correcting the output value Vabyf (V) of the upstream air-fuel ratio sensor 66 by the sub feedback correction amount Vafsfb. Note that the air-fuel ratio feedback control based on the output value Voxs of the downstream air-fuel ratio sensor 67 is sometimes called “sub-feedback control”.

  Further, in the present apparatus, the “learning process” described in the background section for the integral value (deviation integral value) of the deviation included in the integral term (I term) in the PID process comes every time a predetermined timing arrives. To be executed. As a result, each time a predetermined timing arrives, the stationary component of the deviation integral value is transferred to the learning value without changing the sum of the deviation integral value and the learning value (= total value SUM). The integral term (I term) in the PID process is a value obtained by multiplying the total value SUM by a feedback gain. The above is the outline of the air-fuel ratio feedback control by this apparatus. Details of the air-fuel ratio feedback control will be described with reference to a routine described later.

(Detection of occurrence of AF variation between cylinders)
Next, an outline of detection of occurrence of “inter-cylinder AF variation” by the present apparatus will be described. “Inter-cylinder AF variation” refers to variation in the air-fuel ratio of the air-fuel mixture used for combustion between cylinders. In general, in the internal combustion engine 10 having multiple cylinders (four cylinders), variation in the injection amount from the fuel injection valve (port injection valve 39P, in-cylinder injection valve 39D) between the cylinders, variation in the maximum lift amount of the intake valve 32, Variations in the distribution of the amount of EGR gas recirculated to the intake system by an EGR mechanism (not shown) for each cylinder may occur. Due to such characteristic variation between cylinders, “inter-cylinder AF variation” may occur.

  As shown in FIG. 5, when “inter-cylinder AF variation” occurs, even if the air-fuel ratio of the mixed exhaust gas matches the stoichiometric air-fuel ratio, the air-fuel ratio of the air-fuel mixture used for combustion is the stoichiometric air-fuel ratio. There are always cylinders (rich cylinders) that become richer and cylinders (lean cylinders) in which the air-fuel ratio of the air-fuel mixture used for combustion is leaner than the stoichiometric air-fuel ratio. In FIG. 5, as an example, the case where the # 1 cylinder corresponds to the “rich cylinder” and the # 2, # 3, and # 4 cylinders correspond to the “lean cylinder” is shown.

  In general, hydrogen gas is generated in exhaust gas discharged from an internal combustion engine by a reaction between fuel and air (oxygen). As shown in FIG. 6, the hydrogen concentration of the exhaust gas is maintained substantially zero when the air-fuel ratio (the air-fuel ratio of the air-fuel mixture subjected to combustion) is leaner than the stoichiometric air-fuel ratio. In the case of rich, it has a characteristic that it increases rapidly as the degree of richness increases. Therefore, when “inter-cylinder AF variation” occurs and the air-fuel ratio of the mixed exhaust gas matches the stoichiometric air-fuel ratio, almost no hydrogen is discharged from the “lean cylinder”, while the “rich cylinder” A large amount of hydrogen can be discharged.

  On the other hand, when the air-fuel ratio of the mixed exhaust gas matches the stoichiometric air-fuel ratio, the greater the degree of “inter-cylinder AF variation”, the richer the air-fuel ratio of the “rich cylinder”. Therefore, as shown in FIG. 7, the greater the degree of “inter-cylinder AF variation”, the higher the hydrogen concentration of the exhaust gas discharged from the “rich cylinder” (and hence the hydrogen concentration of the mixed exhaust gas).

  The upstream air-fuel ratio sensor 66 is a limiting current type oxygen concentration sensor. Therefore, the higher the hydrogen concentration of the mixed exhaust gas, the higher the output value Vabyf of the upstream air-fuel ratio sensor 66 indicates an air-fuel ratio that is shifted to a richer side than the true air-fuel ratio of the mixed exhaust gas. This phenomenon is called “rich shift”. This “rich shift” is caused by a reaction in which oxygen molecules are ionized in the detection unit when hydrogen molecules smaller than oxygen molecules enter the detection unit of the upstream air-fuel ratio sensor 66 before the oxygen molecules. This is thought to be caused by difficulty.

  From the above, when the air-fuel ratio of the mixed exhaust gas matches the stoichiometric air-fuel ratio, the greater the degree of “inter-cylinder AF variation”, the greater the output value Vabyf of the upstream air-fuel ratio sensor 66 is due to “rich deviation”. Thus, the air-fuel ratio shifted to the rich side with respect to the true air-fuel ratio (= theoretical air-fuel ratio) of the mixed exhaust gas is shown.

  During the air-fuel ratio feedback control described above, based on the output value Vabyf of the upstream air-fuel ratio sensor 66, the mixed exhaust gas is set so that the air-fuel ratio indicated by the output value Vabyf of the upstream air-fuel ratio sensor 66 matches the stoichiometric air-fuel ratio. The air-fuel ratio is adjusted. Accordingly, when the “rich deviation” occurs in the output value Vabyf of the upstream air-fuel ratio sensor 66, the air-fuel ratio of the mixed exhaust gas is adjusted to a leaner air-fuel ratio than the theoretical air-fuel ratio by the “rich deviation”. Therefore, mixed exhaust gas having an air-fuel ratio leaner than the stoichiometric air-fuel ratio flows into the catalyst 53. As a result, the inside of the catalyst 53 is in a lean atmosphere, and the state where the air-fuel ratio of the mixed exhaust gas that reaches the downstream air-fuel ratio sensor 67 is leaner than the stoichiometric air-fuel ratio continues.

  As described above, when the state in which the air-fuel ratio of the mixed exhaust gas that reaches the downstream air-fuel ratio sensor 67 is leaner than the stoichiometric air-fuel ratio continues, the above-described total value SUM (= deviation integral value + learned value) becomes the fuel injection. It gradually increases in the direction of increasing the amount (increasing direction). That is, when the “rich deviation” occurs due to the “inter-cylinder AF variation” during the air-fuel ratio feedback control described above, the total value SUM described above gradually increases in the increasing direction. In this way, as the sum SUM gradually increases in the increasing direction, the air-fuel ratio of the mixed exhaust gas that has shifted to the lean side approaches the stoichiometric air-fuel ratio.

  By utilizing this fact, it is possible to detect the occurrence of “inter-cylinder AF variation” when the sum value SUM reaches a predetermined guard value Guard in the increasing direction. Hereinafter, in this example, regarding the total value SUM (or deviation integrated value, learning value), a positive value corresponds to the direction in which the fuel injection amount is increased (increase direction), and a negative value decreases the fuel injection amount. It shall correspond to the direction (weight loss direction).

  FIG. 8 shows a case where, during the air-fuel ratio feedback control described above, the sum SUM is maintained at “0” before time ts, and “inter-cylinder AF variation” starts to occur for some reason at time ts. Therefore, an example of a change in the total value SUM in the case where “rich deviation” starts to occur is shown. In the case shown in FIG. 8, after time ts, the total value SUM gradually increases in the increasing direction (positive direction), and “inter-cylinder AF variation” occurs at time te when the total value SUM reaches the guard value Guard. Can be detected.

  By the way, in the internal combustion engine 10 equipped with the dual injection system, as shown in FIG. 9, the port injection of each cylinder is performed based on the engine speed NE and the load factor KL (value corresponding to the in-cylinder intake air amount Mc). Port injection mode (P mode) in which fuel is injected only from the valve 39P (total of four injection valves), and in-cylinder injection in which fuel is injected only from the in-cylinder injection valve 39D (total of four injection valves) of each cylinder Of three injection modes: a mode (D mode) and a dual injection mode (PD mode) in which fuel is injected from the port injection valve 39P and in-cylinder injection valve 39D (total eight injection valves) of each cylinder One injection mode is selected and used. In the idling state, the P mode is selected.

  Hereinafter, a case will be considered in which “inter-cylinder AF variation” is caused by an abnormality in the fuel injection valve (port injection valve 39P, in-cylinder injection valve 39D) (for example, the occurrence of misfire of the valve). In general, it is rare that an abnormality occurs in both a part of the (four) port injection valves 39P and a part of the (four) in-cylinder injection valves 39D at the same time.

  For example, a case is assumed in which an abnormality has occurred only in a part of the plurality (four) of port injection valves 39P. It is assumed that the constituent members other than the fuel injection valve are normal, and “inter-cylinder AF variation” due to the constituent members other than the fuel injection valve does not occur. In addition, regardless of which injection mode is selected, one kind of sum value SUM (= deviation integral value SDVoxs + learning value Learn) is calculated and updated as in the case shown in FIG. 8 described above. To do.

  In this case, in the P mode, an abnormality has occurred in some of the plurality of (four) injection valves through which fuel is injected, so that “inter-cylinder AF variation” occurs. As a result, “rich deviation” occurs in the upstream air-fuel ratio sensor 66. Therefore, as described above, based on the upstream air-fuel ratio sensor output value Vabyf, the air-fuel ratio of the mixed exhaust gas tends to be adjusted to an air-fuel ratio that is leaner than the stoichiometric air-fuel ratio. As a result, when the P mode continues, the total value SUM gradually increases in the increasing direction as in the case shown in FIG.

  On the other hand, in the D mode, since all of the (four) injection valves into which fuel is injected are normal, “inter-cylinder AF variation” does not occur. As a result, no “rich deviation” occurs in the upstream air-fuel ratio sensor 66. Therefore, the air-fuel ratio of the mixed exhaust gas tends to be adjusted to the stoichiometric air-fuel ratio itself based on the upstream air-fuel ratio sensor output value Vabyf. As a result, when the D mode continues, the total value SUM converges toward zero.

  Further, in the PD mode, an abnormality has occurred in some of the plurality of (eight) injection valves into which fuel is injected, and thus “inter-cylinder AF variation” occurs. Therefore, as in the P mode, when the PD mode continues, the total value SUM gradually increases in the increasing direction. However, since the ratio of the “number of fuel injection valves in which an abnormality has occurred” to the number of all fuel injection valves is small as compared with the P mode, the degree of “inter-cylinder AF variation” (accordingly, the upstream air-fuel ratio sensor) 66 (the degree of “rich deviation”) is also small. As a result, the rate of increase (increase in gradient) of the deviation integral value SDVoxs (and hence the sum value SUM) is smaller than in the P mode.

  From the above, when the operation state in which the P mode (or PD mode) is selected continues for a relatively long time, the sum value SUM can reach the guard value Guard as in the case shown in FIG. The occurrence of “inter-cylinder AF variation” can be detected at an early stage.

  On the other hand, when the operating state of the engine frequently fluctuates, the selected injection mode can be frequently switched. FIG. 10A shows an example of a change in the total value SUM when an abnormality starts to occur in some of the (four) port injection valves 39P at time ts when the injection mode is frequently switched. FIG. 9 is a time chart corresponding to FIG. 8.

  As shown in FIG. 10A, when the injection mode is frequently switched, the total value SUM can increase in the increasing direction during the period when the P mode (or PD mode) is selected (time ts to t1). , T2-t3, t3-t4). As described above, in the PD mode, the increase gradient of the sum value SUM is smaller than that in the P mode. On the other hand, during the period when the D mode is selected, the total value SUM decreases toward zero (see time t1 to t2, t4 and after).

  Thus, since the total value SUM decreases toward zero during the period in which the D mode is selected, the total value SUM cannot reach the guard value Guard in the example illustrated in FIG. As a result, the occurrence of “inter-cylinder AF variation” cannot be detected.

  Therefore, in this apparatus, the total value SUM (= deviation integral value + learned value) is calculated and updated individually for each injection mode. Specifically, in the P mode, only the sum value SUMp (= deviation integral value SDVoxs + learning value Learnp) for the P mode is calculated and updated, and in the D mode, the sum value SUMd (= deviation integral value SDVoxs + for the D mode). Only the learning value (Learnd) is calculated and updated. In the PD mode, only the sum total value SUMpd (= deviation integral value SDVoxs + learning value Learnpd) for the PD mode is calculated and updated. The total value (= deviation integral value + learning value) corresponding to the injection mode different from the selected injection mode is constant at the last updated value in the past (within the period during which the injection mode was selected). Maintained.

  In this apparatus, when any one of the sum values SUMp, SUMd, and SUMpd (actually, the sum value corresponding to the currently selected injection mode) reaches the guard value Guard, “inter-cylinder air-fuel variation” An occurrence is detected.

  FIG. 10B is a time chart corresponding to FIG. 10A when such a configuration is employed. As can be understood from FIG. 10 (b), in this case, unlike the case shown in FIG. 10 (a), the total sum SUMp is calculated during the period in which the D mode is selected (see time t1 to t2, after t4). It remains constant (does not decrease). Therefore, as compared with the case shown in FIG. 10A, the time when the total value (specifically, the total value SUMp) reaches the guard value Guard (accordingly, occurrence of “inter-cylinder air-fuel ratio variation” occurrence is detected). The timing is likely to be earlier.

  However, in the case shown in FIG. 10B, the sum value SUMp cannot increase during the period in which the D mode or the PD mode is selected. Even during this period, if the sum value SUMp can be continuously increased, the time when the sum value SUMp reaches the guard value Guard can be made earlier.

  Here, as shown in FIG. 10B, when an abnormality has occurred in only a part of the (four) port injectors 39P, the sum value SUMd for D mode (= deviation integral value SDVoxs + The learning value (Learnd) continues to be maintained at substantially zero. On the other hand, the total sum SUMp (= deviation integral value SDVoxs + learned value Learnp) for the P mode is maintained constant in the D mode (and PD mode), but can increase in the increasing direction in the P mode.

  From such a viewpoint, when the sum value SUMp is a value in the increasing direction (positive value) and the value obtained by subtracting the sum value SUMd from the sum value SUMp exceeds the predetermined value A in the P mode, It can be estimated that only a part of the (four) port injection valves 39P has an abnormality. If this estimation is correct in the case of the estimation as described above, the sum value SUMp can be continuously increased if the injection mode is fixed to the P mode thereafter. As a result, the time when the sum value SUMp reaches the guard value Guard. Can be made faster.

  Therefore, in the present apparatus, in the P mode, after the sum value SUMp is a value in the increasing direction (positive value) and the value obtained by subtracting the sum value SUMd from the sum value SUMp exceeds a predetermined value A, The injection mode is fixed to the P mode regardless of the subsequent operation state. FIG. 10C is a time chart corresponding to FIG. 10B when such a configuration is employed. As can be understood from FIG. 10C, in this case, unlike the case shown in FIG. 10B, after the time tf when the injection mode starts to be fixed to the P mode, the injection mode determined from FIG. Even in modes other than the mode, the injection mode continues to be fixed to the P mode. Therefore, after the time tf (particularly, after the time t1 when the increase in the total sum SUMp ends in FIG. 10B), the total sum SUMp continues to increase. As a result, compared to the case shown in FIG. 10B, the time when the sum value SUMp reaches the guard value Guard (accordingly, the time when occurrence of “inter-cylinder air-fuel ratio variation” occurrence is detected, FIG. 10C). , Time te) becomes earlier.

  As described above, in this device, the total value is calculated and updated individually for each injection mode, and when it can be estimated that “an abnormality has occurred only in a part of the plurality of port injection valves 39P”, the injection mode Is fixed in the port P mode, the occurrence of “inter-cylinder AF variation” due to the abnormality of some of the plurality of port injection valves 39P can be detected early and reliably. In addition, in this case, it can also be specified that the cause of the “inter-cylinder AF variation” is the port injection valve 39P. Depending on the operating state of the engine, if the injection mode is fixed to the P mode, a good operating state (combustion state) may not be maintained. Therefore, the injection mode may be fixed to the P mode only as long as the engine operating state can be maintained satisfactorily.

  Further, in this apparatus, based on the same knowledge as described above, in the D mode, the sum value SUMd is a value in the increasing direction (positive value), and a value obtained by subtracting the sum value SUMp from the sum value SUMd is predetermined. After the time when the value A is exceeded, the injection mode is fixed to the D mode regardless of the subsequent operation state. Thereby, it is possible to detect early and surely the occurrence of “inter-cylinder AF variation” due to the abnormality of some of the plurality of in-cylinder injection valves 39D. In addition, in this case, it can be specified that the cause of occurrence of “inter-cylinder AF variation” is the in-cylinder injection valve 39D. It should be noted that the injection mode may be fixed to the D mode only as long as the engine operating state can be maintained satisfactorily. The above is the outline of detection of occurrence of “inter-cylinder AF variation” by the present apparatus. Details of detection of the occurrence of “inter-cylinder AF variation” will be described with reference to a routine described later.

(Actual operation)
Next, referring to FIGS. 11 to 16, which are flowcharts showing routines (programs) stored in the ROM 72, which are executed by the CPU 71 of the electric control device 70, with respect to the actual operation of the present apparatus configured as described above. While explaining.

  The CPU 71 performs the routine for calculating the fuel injection amount Fi and instructing the fuel injection shown in FIG. 11 every time the crank angle of an arbitrary cylinder becomes a predetermined crank angle before the intake top dead center (for example, BTDC 90 ° CA). It is designed to be executed repeatedly. Accordingly, when the crank angle of a certain cylinder (hereinafter referred to as “fuel injection cylinder”) reaches the predetermined crank angle, the CPU 71 starts processing from step 1100 and proceeds to step 1105, and is measured by the air flow meter 61. Based on the intake air flow rate Ga and the engine rotational speed NE, the amount of air (in-cylinder intake air amount Mc) to be taken into the combustion chamber 25 in the current intake stroke is obtained from the map f.

  Subsequently, the CPU 71 proceeds to step 1110, and the basic fuel for setting the air-fuel ratio to the target air-fuel ratio by dividing the obtained cylinder intake air amount Mc by the target air-fuel ratio abyfr (theoretical air-fuel ratio in this example). Obtain the injection amount Fbase. Next, the CPU 71 proceeds to step 1115 to set the total fuel injection amount Fi to a value obtained by adding the calculated basic fuel injection amount Fbase to an air-fuel ratio feedback correction amount DFi calculated in a routine described later.

  Next, the CPU 71 proceeds to step 1120 to determine whether or not the current injection mode selected by the routine to be described later is the P mode. When determining “Yes”, in step 1125, the total fuel is determined. The ratio (injection ratio R) of the total fuel injection quantity on the port injection valve 39P side to the injection quantity Fi is set to “1”. When determining “No” at step 1120, the CPU 71 proceeds to step 1130 to determine whether or not the current injection mode is the D mode. When determining “Yes”, the CPU 71 performs injection at step 1135. The ratio R is set to “0”. When it is determined as “No” in step 1130 (that is, when the current injection mode is the PD mode), in step 1140, the CPU 71 sets the engine speed NE, the in-cylinder intake air amount Mc, and the cooling water temperature THW. Based on this, the value of the injection ratio R (0 <R <1) is set from the map R.

  In step 1145, the CPU 71 sets the total fuel injection amount Fip on the port injection valve 39P side to a value obtained by multiplying the total fuel injection amount Fi by the injection ratio R, and in step 1150, the in-cylinder injection valve 39D side The total fuel injection amount Fid is set to a value obtained by multiplying the total fuel injection amount Fi by the value (1-R).

  Then, the CPU 71 proceeds to step 1155 to give instructions for injecting fuel of the total fuel injection amount Fip, Fid to the plurality (four) of port injection valves 39P and the plurality (four) of fuel injection cylinders. This is performed for the in-cylinder injection valve 39D, and the routine proceeds to step 1195 to end the present routine tentatively. As a result, with the current injection mode, the fuel of the total fuel injection amount Fi that has been air-fuel ratio feedback corrected is injected into the cylinder that reaches the intake stroke.

  Next, calculation of the air-fuel ratio feedback correction amount DFi will be described. The CPU 71 repeatedly executes the routine shown in FIG. 12 following the routine of FIG. Accordingly, when the predetermined timing is reached, the CPU 71 starts processing from step 1200 and proceeds to step 1205 to determine whether or not the air-fuel ratio feedback control condition is satisfied. The air-fuel ratio feedback control condition is, for example, that the engine coolant temperature THW detected by the water temperature sensor 65 is equal to or higher than a predetermined temperature, the intake air amount (load) per one rotation of the engine is equal to or lower than a predetermined value, and upstream This is established when the side / downstream air-fuel ratio sensors 66 and 67 are in the active state.

  First, the case where the air-fuel ratio feedback control condition is satisfied will be described. In this case, the CPU 71 makes a “Yes” determination at step 1205 to proceed to step 1210, where the current output value Vabyf (V) of the upstream air-fuel ratio sensor 66 and a sub-feedback correction amount Vafsfb (%), which will be described later, A control air-fuel ratio equivalent output value Vabyfs (V) is obtained based on the formula described in step 1210.

  Subsequently, the CPU 71 proceeds to step 1215 so as to obtain the current control air-fuel ratio abyfs based on the obtained control air-fuel ratio equivalent output value Vabyfs and the map shown in FIG. This air-fuel ratio is the “apparent air-fuel ratio” of the gas upstream of the catalyst 53.

  Next, the CPU 71 proceeds to step 1220 and divides the latest in-cylinder intake air amount Mc (for the current intake stroke) obtained in the previous step 1105 by the obtained control air-fuel ratio abyfs. The in-cylinder fuel supply amount Fc with respect to the intake stroke is obtained. Next, the CPU 71 proceeds to step 1225, and obtains the target in-cylinder fuel supply amount Fcr for the current intake stroke by dividing the in-cylinder intake air amount Mc by the target air-fuel ratio abyfr.

  Subsequently, the CPU 71 proceeds to step 1230 to set the in-cylinder fuel supply amount deviation DFc to a value obtained by subtracting the in-cylinder fuel supply amount Fc from the target in-cylinder fuel supply amount Fcr. That is, the in-cylinder fuel supply amount deviation DFc is an amount representing the excess or deficiency of fuel for the current intake stroke.

  Next, the CPU 71 proceeds to step 1235, and adds the in-cylinder fuel supply amount deviation DFc obtained in the above step 1230 to the integral value SDFc of the in-cylinder fuel supply amount deviation DFc at that time point. Update integrated value SDFc.

  Then, the CPU 71 proceeds to step 1240, where the in-cylinder fuel supply amount deviation DFc obtained in step 1230, the integrated value SDFc of the in-cylinder fuel supply amount deviation updated in step 1235, and the step 1240 are described. The air-fuel ratio feedback correction amount DFi is obtained based on the above equation. Here, Gp is a preset proportional gain, and Gi is a preset integral gain. The coefficient KFB is preferably variable depending on the engine speed NE, the in-cylinder intake air amount Mc, and the like, but is set to “1” here. Then, the CPU 71 proceeds to step 1295 to end the present routine tentatively.

  As described above, the air-fuel ratio feedback correction amount DFi is obtained by the proportional integration process (PI process), and this air-fuel ratio feedback correction amount DFi is reflected in the total fuel injection amount Fi by step 1115 in FIG.

  As a result, the excess or deficiency of the fuel supply amount for the current intake stroke is compensated, so the average value of the air-fuel ratio (and hence the air-fuel ratio of the gas flowing into the catalyst 53) is the target air-fuel ratio abyfr (= theoretical air-fuel ratio). It is made to almost agree with.

  On the other hand, if the air-fuel ratio feedback control condition is not satisfied, the CPU 71 makes a “No” determination at step 1205 to proceed to step 1245 to set the air-fuel ratio feedback correction amount DFi to “0”. Thereby, the fuel injection amount correction based on the air-fuel ratio feedback control (see step 1115 in FIG. 11) is not performed.

  Next, air-fuel ratio feedback control (ie, sub-feedback control) based on the output Voxs of the downstream air-fuel ratio sensor 67 will be described. By this sub feedback control, the sub feedback correction amount Vafsfb (%) described above is calculated.

  In order to obtain the sub feedback correction amount Vafsfb, the CPU 71 repeatedly executes the routine shown in FIG. 13 following the routine of FIG. Therefore, when the predetermined timing is reached, the CPU 71 starts the process from step 1300, proceeds to step 1305, determines whether or not the same air-fuel ratio feedback control condition as in the previous step 1205 is satisfied, and determines “No”. If so, the process immediately proceeds to step 1395 to end the present routine tentatively.

  Hereinafter, a case where the air-fuel ratio feedback control condition is satisfied will be described. In this case, the CPU 71 makes a “Yes” determination at step 1305 to proceed to step 1310, where the deviation DVoxs is a value obtained by subtracting the output value Voxs of the downstream air-fuel ratio sensor 67 from the target air-fuel ratio equivalent target value Voxsref (constant). Set to.

  Next, the CPU 71 proceeds to step 1315 to determine whether or not the injection mode that has been sequentially selected / updated by a routine to be described later has been switched, and in the case of determining “No”, in step 1320, The deviation integrated value SDVoxs is updated by adding the deviation DVoxs obtained in step 1310 to the deviation integrated value SDVoxs of the deviation DVoxs.

  On the other hand, if “Yes” is determined in step 1315, the CPU 71 proceeds to step 1325 and clears the deviation integral value SDVoxs to “0”. As a result, the value (= the steady component of the deviation integrated value SDVoxs, the learning value update amount DLearn) that is transferred to the learning value by the “learning process” described later is selected at the time of execution of the “learning process”. It can be assured that the value is based on the deviation integrated value SDVoxs obtained by integrating the deviation DVoxs in a state where the selected injection mode continues to be selected.

  Subsequently, the CPU 71 proceeds to step 1330 to determine whether or not the current injection mode is the P mode, and when determining “Yes” (that is, when the current injection mode is the P mode), in step 1335. Then, the sum value SUM and the sum value SUMp for the P mode are updated to a value obtained by adding the learning value Learnp for the P mode to the current deviation integral value SDVoxs. The sum value SUM is a sum value corresponding to the current injection mode. The learning value Learnp is calculated and updated in a routine described later in the P mode.

  When the CPU 71 determines “No” in step 1330, the CPU 71 proceeds to step 1340, determines whether or not the current injection mode is the D mode, and determines “Yes” (that is, the current injection mode is In the case of the D mode), in step 1345, the sum value SUM and the sum value SUMd for the D mode are updated to a value obtained by adding the learning value Learn for the D mode to the current deviation integral value SDVoxs. The learned value Learn is calculated and updated in a routine described later in the D mode.

  When the CPU 71 determines “No” in step 1340 (that is, when the current injection mode is the PD mode), the CPU 71 proceeds to step 1350 and determines the sum value SUM and the sum value SUMpd for the PD mode as the current deviation. The integrated value SDVoxs is updated to a value obtained by adding the learning value Learnpd for the PD mode. The learning value Learnpd is calculated and updated in a routine described later in the PD mode.

  When the CPU 71 proceeds to step 1355, the CPU 71 obtains a time differential value DDVoxs of the deviation DVoxs based on the deviation DVoxs obtained in step 1310, the previous deviation DVoxsb, and the formula described in step 1355. As the previous deviation DVoxsb, the value updated in step 1365 described later at the time of the previous execution of this routine is used. Δt is the execution interval time of this routine.

  Next, the CPU 71 proceeds to step 1360, and the deviation DVoxs obtained in step 1310 and the sum value SUM updated in any of steps 1335, 1345, 1350 (that is, the sum value corresponding to the current injection mode). ), The time differential value DDVoxs of the deviation obtained in step 1355 and the formula described in step 1360, the sub feedback correction amount Vafsfb (%) is obtained. Here, Kp is a preset proportional gain, Ki is a preset integral gain, and Kd is a preset differential gain.

  In the expression described in step 1360, the first term on the right side “Kp · DVoxs” is the proportional term, the second term on the right side “Ki · SUM” is the integral term, and the third term on the right side “Kd · DDVoxs” is the derivative term. Each corresponds. That is, the sub feedback correction amount Vafsfb is calculated by performing PID processing on the deviation DVoxs based on the proportional term, the integral term, and the derivative term.

  The CPU 71 proceeds to step 1365, sets the previous deviation DVoxsb to the deviation DVoxs obtained in step 1310, and then proceeds to step 1395 to end the present routine tentatively.

  In this way, the sub feedback correction amount Vafsfb (%) is obtained, and this value is used for the calculation of the control air-fuel ratio equivalent output value Vabyfs (V) in step 1210 of FIG. Then, the control air-fuel ratio equivalent output value Vabyfs (V) is converted into the control air-fuel ratio abyfs based on the map shown in FIG. In other words, the control air-fuel ratio abyfs is equal to the sub-feedback correction amount Vafsfb (based on the output value Voxs of the downstream air-fuel ratio sensor 67 with respect to the air-fuel ratio actually detected by the upstream air-fuel ratio sensor 66. %) Is obtained as a different air-fuel ratio.

  As a result, the in-cylinder fuel supply amount Fc calculated in step 1220 of FIG. 12 described above changes in accordance with the output value Voxs of the downstream air-fuel ratio sensor 67. Therefore, in steps 1230 to 1240, the air-fuel ratio feedback correction amount DFi Is changed in accordance with the output value Voxs of the downstream side air-fuel ratio sensor 67. As a result, the air-fuel ratio is controlled so that the air-fuel ratio downstream of the catalyst 53 matches the target air-fuel ratio abyfr (= theoretical air-fuel ratio).

  For example, since the average air-fuel ratio of the engine is lean, the output value Voxs of the downstream air-fuel ratio sensor 67 is smaller than the target air-fuel ratio equivalent target value Voxsref (constant) (that is, a value shifted to the lean side) ), The deviation DVoxs obtained in step 1310 has a positive value, and the sub-feedback correction amount Vafsfb obtained in step 1360 has a positive value. Therefore, the control air-fuel ratio abyfs obtained in step 1215 is obtained as a leaner value (a larger value) than the air-fuel ratio actually detected by the upstream air-fuel ratio sensor 66.

  Therefore, the in-cylinder fuel supply amount Fc obtained in step 1220 is a small value, and the in-cylinder fuel supply amount deviation DFc obtained in step 1230 is a large value. Accordingly, the air-fuel ratio feedback correction amount DFi becomes a large positive value. As a result, the total fuel injection amount Fi obtained in step 1115 in FIG. 11 is controlled to be larger than the basic fuel injection amount Fbase and the air-fuel ratio becomes a rich value.

  On the other hand, since the average air-fuel ratio of the engine is rich, the output value Voxs of the downstream air-fuel ratio sensor 67 is larger than the target air-fuel ratio equivalent target value Voxsref (constant) (that is, shifted to the rich side). Value), the deviation DVoxs becomes a negative value, so the sub-feedback correction amount Vafsfb becomes a negative value. Therefore, the control air-fuel ratio abyfs obtained in step 1215 is obtained as a richer value (smaller value) than the air-fuel ratio actually detected by the upstream air-fuel ratio sensor 66.

  Accordingly, since the in-cylinder fuel supply amount Fc has a large value, the in-cylinder fuel supply amount deviation DFc has a negative value. As a result, the air-fuel ratio feedback correction amount DFi becomes a negative value. Thus, the total fuel injection amount Fi is controlled to be smaller than the basic fuel injection amount Fbase so that the air-fuel ratio becomes a lean value.

  As described above, the total value (= deviation integral value + learning value) is introduced for each injection mode, and only the total value (= deviation integral value + learning value) corresponding to the current injection mode is calculated and updated. Go. The learning value corresponding to the current injection mode is updated every time the learning timing comes in a routine described later. That is, the sum total value corresponding to the current injection mode is updated every time the execution timing of the routine shown in FIG. 13 arrives, and the deviation integrated value SDVoxs that is updated every time the learning timing comes. It is the sum with the learning value corresponding to the injection mode.

  The total value (= deviation integral value + learning value) corresponding to the injection mode different from the current injection mode is maintained at the value at the time of the last update in the past (within the period during which the injection mode was selected). Is done. Then, the sub feedback correction amount Vafsfb is calculated using the total value (= deviation integral value + learning value) corresponding to the current injection mode, and the sub feedback correction amount Vafsfb and the output value Vabyf of the upstream air-fuel ratio sensor 66 are calculated. Based on this, air-fuel ratio feedback control is executed.

  Next, update (learning) of learning values (Learnp, Learnd, Learnpd) will be described. The CPU 71 repeatedly executes the routine shown in FIG. 14 following the routine of FIG. Accordingly, when the predetermined timing is reached, the CPU 71 proceeds to step 1405 to determine whether or not the learning value learning timing has come. If “No” is determined, the CPU 71 immediately proceeds to step 1495 to end this routine once. To do. For example, in this example, the learning timing comes every time the number of fuel injections reaches a predetermined number (for example, 1000 times).

  When the learning timing has arrived, the CPU 71 determines “Yes” in step 1405 and proceeds to step 1410 to set the learning value update amount DLearn to the latest value of the deviation integral value SDVoxs updated in step 1320 or 1325 described above. Is set to the value divided by the value Nref. The value Nref is a value equal to or greater than “1”, and is set to “1”, “2”, “4”, “8”, for example. Here, the learning value update amount DLearn corresponds to the “steady component of the value based on the deviation integral value”. The learning value update amount DLearn may be a value obtained by low-pass filtering the deviation integral value SDVoxs.

  Subsequently, the CPU 71 proceeds to step 1415 to determine whether or not the current injection mode is the P mode. When determining “Yes” (that is, in the P mode), in step 1420, the current P The learning value Learnp for P mode is updated by adding the learning value update amount DLearn obtained in step 1410 to the learning value Learnp for mode. If the CPU 71 determines “No” in step 1415, it determines whether or not the current injection mode is the D mode in step 1425, and determines “Yes” (that is, in the D mode). At 1430, the learned value update amount DLearn obtained in step 1410 is added to the learned value Learn for D mode at that time to update the learned value Learn for D mode. When the CPU 71 determines “No” in step 1425 (that is, in the PD mode), in step 1435, the learning value update amount obtained in step 1410 above the learning value Learnpd for PD mode at that time. The learning value Learnpd for the PD mode is updated by adding DLearn.

  As described above, the learning values Learnp, Learn, and Learnpd are individually set for each injection mode, and each time the learning timing arrives, the steady component (= DLearn) of the deviation integral value SDVoxs becomes the current injection mode. It is transferred to the corresponding learning value. Thereby, the learning value is updated individually for each injection mode.

  Then, the CPU 71 proceeds to step 1440, resets the deviation integral value SDVoxs by subtracting the learning value update amount DLearn from the deviation integral value SDVoxs at that time, and then proceeds to step 1495 to end the present routine tentatively.

  Thus, every time the learning timing arrives, the “learned value corresponding to the current injection mode” is updated without changing the sum of the deviation integral value SDVoxs and the “learned value corresponding to the current injection mode”. To go. Thus, every time the learning timing arrives, the deviation integrated value SDVoxs gradually approaches “0”, while the “learned value corresponding to the current injection mode” becomes the “rich deviation of the upstream air-fuel ratio sensor 66. Each of them approaches a value corresponding to the size of "." The speed at which the “learned value corresponding to the current injection mode” approaches the value corresponding to the “rich deviation” of the upstream air-fuel ratio sensor 66 decreases as the value Nref increases.

  Next, selection of the injection mode will be described. The CPU 71 repeatedly executes the routine shown in FIG. 15 following the routine of FIG. Therefore, when the predetermined timing is reached, the CPU 71 proceeds to step 1505 to determine whether or not the current injection mode is the P mode. If “No” is determined, in step 1510, the current injection mode is the D mode. Determine whether or not. If “No” is also determined here (ie, in the PD mode), the CPU 71 proceeds to step 1515, where the engine speed NE, the load factor KL (or in-cylinder intake air amount Mc), and FIG. Based on the map shown, an injection mode (P mode, D mode, or PD mode) corresponding to the current operating state is selected, and the routine proceeds to step 1595 to end the present routine tentatively.

  Next, the case where the current injection mode is the P mode will be described. In this case, when the CPU 71 proceeds to step 1505, it determines “Yes”, proceeds to step 1520, and the latest total value SUMp (that is, the value sequentially updated in step 1335 in FIG. 13) is a positive value. Yes, and a value obtained by subtracting the latest total value SUMp from the latest total value SUMp (that is, the value last updated in step 1345 in FIG. 13 during the period in which the D mode was selected in the past). Is greater than a predetermined value A (> 0).

  When the CPU 71 determines “No” in step 1520, the CPU 71 proceeds to step 1515 described above and selects an injection mode (P mode, D mode, or PD mode) according to the current operating state. On the other hand, when the CPU 71 determines “Yes” in step 1520 (that is, abnormality has occurred only in a part of the plurality of port injection valves 39P (and abnormality has occurred in the plurality of in-cylinder injection valves 39D). In the case where it can be estimated that it has not occurred), in step 1525, the injection mode is fixed to the P mode (regardless of the selection result by the map shown in FIG. 9). As a result, as described above, when abnormality is actually occurring only in a part of the plurality of port injection valves 39P, the sum value SUMp can be continuously increased, and as a result, the sum value SUMp becomes the guard value. The time to reach the Guard can be made earlier.

  Next, the case where the current injection mode is the D mode will be described. In this case, when the CPU 71 proceeds to step 1510, the CPU 71 determines “Yes”, proceeds to step 1530, and the latest total value SUMd (that is, the value sequentially updated in step 1345 in FIG. 13) is a positive value. Yes, and a value obtained by subtracting the latest total value SUMd from the latest total value SUMd (that is, the value last updated in step 1335 in FIG. 13 during the period in which the P mode was selected in the past). Is greater than a predetermined value A (> 0).

  When the CPU 71 determines “No” in step 1530, the CPU 71 proceeds to step 1515 described above and selects an injection mode (P mode, D mode, or PD mode) according to the current operation state. On the other hand, when the CPU 71 determines “Yes” in step 1530 (that is, an abnormality has occurred in only a part of the plurality of in-cylinder injection valves 39D (and an abnormality has occurred in the plurality of port injection valves 39P). If it is possible to estimate that it has not occurred), in step 1535, the injection mode is fixed to the D mode (regardless of the selection result by the map shown in FIG. 9). As a result, as described above, the sum value SUMd can continue to increase when only a part of the plurality of in-cylinder injection valves 39D is actually abnormal. As a result, the sum value SUMd is guarded. The time to reach the value Guard can be made earlier.

  Next, detection of occurrence of “inter-cylinder AF variation” will be described. The CPU 71 is configured to repeatedly execute the routine shown in FIG. 16 following the routine of FIG. Accordingly, when the predetermined timing is reached, the CPU 71 proceeds to step 1605 to determine whether any of the sum values SUMp, SUMd, or SUMpd has reached the guard value Guard. Proceeding to 1695, the present routine is temporarily ended. The sum values SUMp, SUMd, and SUMpd are values updated in steps 1335, 1345, and 1350 in FIG. Of the total values SUMp, SUMd, and SUMpd, the value that is currently updated is the total value corresponding to the current injection mode. Therefore, in step 1605, it is actually determined whether or not the sum value corresponding to the current injection mode has reached the guard value Guard.

  When determining “Yes” in step 1605, the CPU 71 determines in step 1610 that “inter-cylinder AF variation” has occurred. Accordingly, an abnormal code corresponding to the occurrence of “inter-cylinder AF variation” is stored in a predetermined storage area of the backup RAM 74.

  In addition, it is determined that “inter-cylinder AF variation” has occurred due to the sum value SUMp reaching the guard value Guard when the injection mode is fixed to the P mode by executing Step 1525 of FIG. In this case, it is specified that the cause of the “inter-cylinder AF variation” is the port injection valve 39P. Similarly, it is determined that “inter-cylinder AF variation” has occurred due to the sum value SUMd reaching the guard value Guard in the state where the injection mode is fixed to the D mode by executing Step 1535 of FIG. In this case, it is specified that the cause of the “inter-cylinder AF variation” is the in-cylinder injection valve 39D. Abnormal codes corresponding to these abnormal contents are also stored in a predetermined storage area of the backup RAM 74.

  On the other hand, in other states (that is, a state in which the injection mode is selected using the map shown in FIG. 9 by executing step 1515 in FIG. 13), the sum value (SUMp, SUMd or SUMpd) reaches the guard value Guard, and it is determined that “inter-cylinder AF variation” has occurred, the cause of “inter-cylinder AF variation” (for example, an abnormality in the fuel injection valves 39P and 39D) The abnormality of the maximum lift amount of the intake valve 32, the abnormality of the EGR mechanism) cannot be specified. Therefore, in this case, the storage process of the abnormality code corresponding to the abnormality content is not performed.

  As described above, in the inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine including the dual injection system according to the first embodiment of the present invention, each injection mode (P mode, D mode, PD mode) is determined. The sum value (deviation integral value + learning value) that is updated by integrating the deviation between the output value Voxs of the downstream air-fuel ratio sensor 67 and the theoretical air-fuel ratio equivalent target value Voxsref is calculated individually. Air-fuel ratio feedback control is executed based on the total value corresponding to the current injection mode and the output value Vabyf of the upstream air-fuel ratio sensor 66. While “air-fuel ratio AF variation” is occurring, a phenomenon occurs in which the “rich deviation” occurs in the upstream air-fuel ratio sensor output value Vabyf and the total value corresponding to the current injection mode increases in the increasing direction. The occurrence of “air-fuel ratio AF variation” is detected when the total value corresponding to the injection mode reaches the guard value Guard in the increasing direction. In the P mode (D mode), the sum value SUMp (SUMd) for the P mode (for D mode) is a (positive) value in the increasing direction and the sum value SUMp (SUMd for P mode (for D mode)) ) To the D mode (P mode) sum value SUMd (SUMp) exceeds a predetermined value A, the cause of occurrence of “inter-cylinder AF variation” is the port injection valve 39P (in-cylinder injection valve 39D). ) And the injection mode is fixed to the P mode (D mode).

  As a result, the total value SUMp (SUMd) can be continuously increased when the cause of the “inter-cylinder AF variation” actually occurs in the port injection valve 39P (in-cylinder injection valve 39D). As a result, in a situation where the operating state of the engine frequently fluctuates and the injection mode is essentially frequently switched, the occurrence of “inter-cylinder AF variation” due to the port injection valve 39P (in-cylinder injection valve 39D) occurs early and reliably. Can be detected.

  The present invention is not limited to the first embodiment, and various modifications can be employed within the scope of the present invention. For example, in the first embodiment, the conditions for fixing the injection mode to the P mode (D mode) are “the current injection mode is the P mode (D mode)”, “P mode (D mode)”. The sum value SUMp (SUMd) is a value in the direction of increase (positive value) ”and“ from the sum value SUMp (SUMd) for P mode (for D mode) to D mode (for P mode) ” “The value obtained by subtracting the sum SUMd (SUMp) exceeds a predetermined value A” is employed. On the other hand, as a condition for fixing the injection mode to the P mode (D mode), “P mode (D mode) sum value SUMp (SUMd)> PD mode sum value SUMpd> D mode (P mode) The fact that the sum value SUMd (SUMp) of (for) is established may be further added. This additional condition is that, when the cause of occurrence of “inter-cylinder AF variation” is in the port injection valve 39P (in-cylinder injection valve 39D), as described above, the increasing gradient of the sum value SUMpd is equal to the sum value SUMp (SUMd). This is based on the fact that the sum value SUMd (SUMp) is maintained at substantially zero, and becomes smaller than the increasing gradient.

(Second Embodiment)
Next, an inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine according to a second embodiment of the present invention will be described. This second embodiment is the sum value between injection modes only in that the injection mode is fixed to the P mode or D mode using the magnitude relationship of the learning values (Learnp, Learn, Learnpd) between the injection modes. This is different from the first embodiment in which the injection mode is fixed to the P mode or the D mode using the magnitude relationship of (SUMp, SUMd, SUMpd). Only such differences will be described below.

  The CPU 71 of the second embodiment executes the routines of the routines of FIGS. 11 to 116 executed by the CPU 71 of the first embodiment except the routines of FIGS. 15 and 16 as they are, and FIGS. In place of this routine, the routines shown in the flowcharts of FIGS. 17 and 18 corresponding to the routines of FIGS. 15 and 16 are executed. In the new routine, the same steps as those in the previous routine are denoted by the same step numbers as those in the previous routine, and the description thereof is omitted.

  The routine shown in FIG. 17 differs from the routine shown in FIG. 15 only in that steps 1520 and 1530 in FIG. 15 are replaced with steps 1705 and 1710, respectively. Accordingly, in the second embodiment, as conditions for fixing the injection mode to the P mode, “the current injection mode is the P mode”, “the learning value Learnp for the P mode is a value in the increasing direction (positive value). And “the value obtained by subtracting the learning value Learn for the D mode from the learning value Learn for the P mode exceeds the predetermined value A” is employed. Similarly, in the second embodiment, as conditions for fixing the injection mode to the D mode, “the current injection mode is the D mode”, “the learning value Learn for the D mode is a value in the increasing direction (positive value). And “the value obtained by subtracting the learning value Learnp for the P mode from the learning value Learn for the D mode exceeds the predetermined value A” is employed.

  The routine shown in FIG. 18 differs from the routine shown in FIG. 16 only in that step 1605 in FIG. 16 is replaced with step 1805. Thus, in the second embodiment, it is determined that “inter-cylinder AF variation” has occurred when any one of the learned values Learnp, Learnd, Learnpd has reached the guard value Guard. The learning values Learnp, Learnd, and Learnpd are values updated in steps 1420, 1430, and 1435 of FIG. Of the learning values Learnp, Learn, and Learnpd, the learning value corresponding to the current injection mode is continuously updated every time the learning timing arrives. Therefore, in step 1805, it is actually determined whether or not the learning value corresponding to the current injection mode has reached the guard value Guard.

  As described above, also according to the second embodiment, it is possible to obtain the same actions and effects as those of the first embodiment. In addition, since the learning value is a steady component of the value based on the deviation integral value, even if the deviation integral value is temporarily rough due to some disturbance or the like, it is not easily affected. On the other hand, the total value (= deviation integral value + learned value) is greatly influenced by the deviation integral value temporarily rough due to some disturbance or the like. From this point of view, the second embodiment in which the injection mode is fixed by using the magnitude relationship between the learning values between the injection modes is more likely to cause the occurrence of “inter-cylinder AF variation” in the port injection valve 39P or in the cylinder. “Injection mode fixed” is executed to detect the occurrence of “inter-cylinder AF variation” due to the port injection valve 39P or the in-cylinder injection valve 39D at an early stage. Can be executed at a more appropriate timing.

  The present invention is not limited to the second embodiment, and various modifications can be adopted within the scope of the present invention. For example, in the first embodiment, the conditions for fixing the injection mode to the P mode (D mode) are “the current injection mode is the P mode (D mode)” and “for the P mode (for the D mode). The learning value Learnp (Learnd) of the current value is a value in the increasing direction (positive value) ", and the learning value Learnp (Learnd) for P mode (for D mode) to D mode (for P mode) “The value obtained by subtracting the learning value Learn (Learnp) exceeds the predetermined value A” is employed. On the other hand, as a condition for fixing the injection mode to the P mode (D mode), “P mode (D mode) learning value Learnp (Learnd)> PD mode learning value Learnpd> D mode (P mode) The learning value “Learnd (Learnp)” is established. This additional condition is that, when the cause of occurrence of “inter-cylinder AF variation” is in the port injection valve 39P (in-cylinder injection valve 39D), as described above, the increasing gradient of the sum value SUMpd is equal to the sum value SUMp (SUMd). This is based on the fact that the sum value SUMd (SUMp) is maintained at substantially zero, and becomes smaller than the increasing gradient.

  25 ... Combustion chamber, 39P ... Port injection valve, 39D ... In-cylinder injection valve, 53 ... Catalyst, 61 ... Air flow meter, 66 ... Upstream air-fuel ratio sensor, 67 ... Downstream air-fuel ratio sensor, 70 ... Electric control device, 71 ... CPU

Claims (5)

A catalyst disposed in a collective exhaust passage formed by collecting respective exhaust passages extending from each cylinder in an internal combustion engine having two or more cylinders;
An upstream air-fuel ratio sensor disposed in the collective exhaust passage upstream of the catalyst;
A downstream air-fuel ratio sensor disposed in the collective exhaust passage downstream of the catalyst;
A port injection valve that injects fuel in the intake passage upstream of the intake valve;
An in-cylinder injection valve for injecting fuel in the combustion chamber;
A multi-cylinder internal combustion engine inter-cylinder air-fuel ratio variation determination device applied to an internal combustion engine comprising:
Based on the operating state of the internal combustion engine, an injection mode is selected from at least a port injection mode in which fuel is injected only from the port injection valve and a cylinder injection mode in which fuel is injected only from the cylinder injection valve A selection means for selecting as a mode;
When the selection mode is the port injection mode, the value corresponding to the deviation between the output value of the downstream air-fuel ratio sensor and the target value of the output value corresponding to the theoretical air-fuel ratio is integrated and updated. In addition to calculating a port deviation integral value, when the selection mode is the in-cylinder injection mode, a value corresponding to the deviation between the output value of the downstream air-fuel ratio sensor and the target value is integrated and updated. Integral value calculating means for calculating the in-cylinder deviation integral value separately from the port deviation integral value;
Based on at least a selected deviation integrated value that is a deviation integrated value corresponding to the selection mode among the port deviation integrated value and the in-cylinder deviation integrated value, and an output value of the upstream air-fuel ratio sensor, the port injection valve And an air-fuel ratio that feedback-controls the air-fuel ratio to match the theoretical air-fuel ratio by adjusting the amount of fuel injected from the selected injection valve that is an injection valve corresponding to the selection mode among the in-cylinder injection valves Control means;
Determination means for determining that variation in the air-fuel ratio between the two or more cylinders has occurred based on the selected deviation integral value;
A cylinder-to-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine comprising:
The selection means includes
When the selection mode is the port injection mode, the value based on the port deviation integral value is a value in an increasing direction that is a direction in which the amount of fuel injected from the port injection valve is increased, and When the amount based on the port deviation integral value in the increasing direction exceeds a predetermined value with respect to the value based on the in-cylinder deviation integral value, the selection mode is fixed to the port injection mode. An inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine. An inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine according to claim 1,
When the selection mode is the port injection mode, a port learning value representing a steady component of the port deviation integrated value is calculated and updated using the port deviation integrated value, and a change in the port learning value due to the update is calculated. An amount corresponding to the amount is subtracted from the port deviation integral value, and when the selection mode is the in-cylinder injection mode, the in-cylinder deviation integral value is used to represent a steady component of the in-cylinder deviation integral value. Learning means for calculating and updating the in-cylinder learning value and subtracting the amount corresponding to the amount of change in the in-cylinder learning value due to the update from the in-cylinder deviation integrated value;
The selection means includes
When the selection mode is the port injection mode, the port learning value is a value in the increase direction, and the amount of the port learning value in the increase direction is greater than the in-cylinder learning value. An inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine configured to fix the selection mode to the port injection mode when a predetermined value is exceeded. A catalyst disposed in a collective exhaust passage formed by collecting respective exhaust passages extending from each cylinder in an internal combustion engine having two or more cylinders;
An upstream air-fuel ratio sensor disposed in the collective exhaust passage upstream of the catalyst;
A downstream air-fuel ratio sensor disposed in the collective exhaust passage downstream of the catalyst;
A port injection valve that injects fuel in the intake passage upstream of the intake valve;
An in-cylinder injection valve for injecting fuel in the combustion chamber;
A multi-cylinder internal combustion engine inter-cylinder air-fuel ratio variation determination device applied to an internal combustion engine comprising:
Based on the operating state of the internal combustion engine, an injection mode is selected from at least a port injection mode in which fuel is injected only from the port injection valve and a cylinder injection mode in which fuel is injected only from the cylinder injection valve A selection means for selecting as a mode;
When the selection mode is the port injection mode, the value corresponding to the deviation between the output value of the downstream air-fuel ratio sensor and the target value of the output value corresponding to the theoretical air-fuel ratio is integrated and updated. In addition to calculating a port deviation integral value, when the selection mode is the in-cylinder injection mode, a value corresponding to the deviation between the output value of the downstream air-fuel ratio sensor and the target value is integrated and updated. Integral value calculating means for calculating the in-cylinder deviation integral value separately from the port deviation integral value;
Based on at least a selected deviation integrated value that is a deviation integrated value corresponding to the selection mode among the port deviation integrated value and the in-cylinder deviation integrated value, and an output value of the upstream air-fuel ratio sensor, the port injection valve And an air-fuel ratio that feedback-controls the air-fuel ratio to match the theoretical air-fuel ratio by adjusting the amount of fuel injected from the selected injection valve that is an injection valve corresponding to the selection mode among the in-cylinder injection valves Control means;
Determination means for determining that variation in the air-fuel ratio between the two or more cylinders has occurred based on the selected deviation integral value;
A cylinder-to-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine comprising:
The selection means includes
When the selection mode is the in-cylinder injection mode, the value based on the in-cylinder deviation integral value is a value in an increasing direction that is a direction in which the amount of fuel injected from the in-cylinder injection valve is increased, When the value based on the in-cylinder deviation integrated value in the increasing direction exceeds a predetermined value with respect to the value based on the port deviation integrated value, the selection mode is fixed to the in-cylinder injection mode. An inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine configured as described above. An inter-cylinder air-fuel ratio variation determining apparatus for a multi-cylinder internal combustion engine according to claim 3,
When the selection mode is the port injection mode, a port learning value representing a steady component of the port deviation integrated value is calculated and updated using the port deviation integrated value, and a change in the port learning value due to the update is calculated. An amount corresponding to the amount is subtracted from the port deviation integral value, and when the selection mode is the in-cylinder injection mode, the in-cylinder deviation integral value is used to represent a steady component of the in-cylinder deviation integral value. Learning means for calculating and updating the in-cylinder learning value and subtracting the amount corresponding to the amount of change in the in-cylinder learning value due to the update from the in-cylinder deviation integrated value;
The selection means includes
When the selection mode is the in-cylinder injection mode, the in-cylinder learning value is a value in the increasing direction, and the in-cylinder learning value is larger than the port learning value in the increasing direction. An air-fuel ratio variation determination device for a cylinder of a multi-cylinder internal combustion engine configured to fix the selection mode to the in-cylinder injection mode when the engine pressure exceeds the predetermined value. In the multi-cylinder internal combustion engine inter-cylinder air-fuel ratio variation determination apparatus according to any one of claims 1 to 4,
The determination means includes
A cylinder of a multi-cylinder internal combustion engine configured to determine that a variation in air-fuel ratio occurs between the two or more cylinders when the selected deviation integral value reaches a predetermined guard value in the increasing direction. Inter-air-fuel ratio variation determination device.

Images