JP5461049B2 - Engine control device - Google Patents

Description

  The present invention relates to an engine control device having a loss torque learning function for learning engine loss torque characteristics.

  As this type of engine control device, for example, the technique of Patent Document 1 is known. Patent document 1 relates to an engine rotation stop control device that stops combustion at a desired crank angle position by stopping combustion when an engine stop request is generated. In the control device, the engine rotation is a target stop crank. The target trajectory, which is the rotational behavior until stopping at the corner, is calculated using the learned value of the loss torque characteristic, and the load on the engine auxiliary machine is adjusted so that the actual engine rotational behavior matches the target trajectory when the engine rotation is stopped. Had to control. Further, the loss torque characteristic of the engine is learned based on at least the actual engine rotational behavior.

  And according to the said structure, the dispersion | variation in the engine rotation behavior in an engine stop process can fully be compensated, and it was supposed that the rotation stop crank angle of an engine could be controlled accurately.

JP 2008-215182 A

  In the above Patent Document 1, in consideration of the fact that loss torque may fluctuate momentarily due to momentary fluctuations in friction loss, etc. in the learning process of loss torque characteristics, the influence of the momentary loss torque fluctuations is eliminated. A configuration is adopted in which the learning value of the loss torque characteristic is smoothed as much as possible. However, with such a configuration, it is considered that when the engine friction changes greatly due to the exchange of engine oil or the like, it takes time to converge the learning value with respect to the actual loss torque to be originally learned.

  The main object of the present invention is to provide an engine control apparatus that can realize suitable learning about the loss torque characteristics of the engine.

  Hereinafter, means for solving the above-described problems and the effects thereof will be described.

  The engine control apparatus of the present invention includes storage means for storing the engine loss torque characteristic as a learning value, and loss torque learning means for updating the learning value of the storage means based on the loss torque characteristic calculated based on the actual engine rotational behavior. In particular, the loss torque learning means is stored in the storage means and first learning means for updating the learning value based on a learning value stored in the storage means and a newly calculated loss torque characteristic. Whether the learning value is updated by the second learning means that updates the learning value based on the newly calculated loss torque characteristic without using the learning value, and the first learning means or the second learning means. Switching means for switching.

  In the above configuration, according to the first learning means, the learning value is updated while continuously using the past learning value. Therefore, it is possible to eliminate the influence when the loss torque characteristic temporarily changes, and to improve reliability. High learning value can be acquired. On the other hand, according to the second learning means, a learning value corresponding to the actual loss torque characteristic at that time can be acquired quickly. Then, by appropriately switching between the first learning means and the second learning means, for example, when it is required to gradually update the learning value from the viewpoint of improving reliability, and sometimes from the viewpoint of improving convergence. In any case where it is required to update the learning value only with the actual loss torque characteristic, appropriate measures can be taken. As a result, it is possible to realize suitable learning about the loss torque characteristics.

  Note that the learning value is updated by the first learning unit, while the amount of change is limited for each update by the past learning value (the learning value stored in the storage unit), whereas the learning value is updated by the second learning unit. Is not limited by past learning values. Therefore, it can be said that the learning by the second learning means is fast learning because the learning value converges (arrives) quickly with respect to the actual loss torque characteristic to be originally learned.

More specifically, as the first and second learning means, the first learning means reflects the newly calculated loss torque characteristic in the smoothing process with respect to the learning value stored in the storage means. The learning value may be updated, and the second learning means may update the learning value based on the newly calculated loss torque characteristic without performing the annealing process ( sixth invention ). Note that the annealing process is also referred to as a first-order lag process, a weighted averaging process, and a filter process.

In the second invention, it is determined that the difference between the learning value stored in the storage means and the current calculated value of the loss torque characteristic is a predetermined value or more, and it is determined that the difference is a predetermined value or more. The learning value is updated by the second learning means.

  In short, when the learning value is gradually updated while reflecting the stored learning value (corresponding to the first learning means), the difference between the stored learning value and the current calculated value of the loss torque characteristic is not less than a predetermined value. Therefore, it takes time to converge the learning value with respect to the actual loss torque characteristic to be learned. In this regard, in the above configuration, when the difference between the stored learned value and the current calculated value of the loss torque characteristic is greater than or equal to a predetermined value, a desired learned value can be acquired quickly by performing learning using the second learning means.

  Incidentally, after the learning by the second learning means, if the difference between the stored learned value and the current calculated value of the loss torque characteristic becomes less than a predetermined value, the learning by the first learning means is performed thereafter (normal learning) Return to processing). That is, considering that the loss torque characteristic rarely changes suddenly, learning by the second learning means is temporarily performed.

Here, as in the third aspect of the invention, whether or not the learning execution condition for the loss torque characteristic is satisfied is determined, and the fact that the difference is equal to or greater than a predetermined number in a learning execution period associated with the establishment of the learning execution condition is a predetermined number of times When the determination is made, the learning value may be updated by the second learning means. Thereby, learning of the loss torque characteristic becomes more suitable.

  The second learning means may update the learning value with an average value of a plurality of calculated values of loss torque characteristics calculated based on the actual engine rotational behavior.

The following conditions are assumed as conditions for updating the learning value by the second learning means. That is, (1) it is determined that a factor change affecting the friction loss of the engine has occurred, and when the factor change of the friction loss has occurred, the learning value is updated by the second learning means ( first Invention of 4 ).
(2) When it is determined that the learning value stored in the storage means has disappeared or the learning value is an unlearned initial value, and the learning value is lost or unlearned is determined, The learning value is updated by the second learning means ( fifth invention ).

  Specifically, regarding (1) above, when engine oil is changed or parts (piston rings, etc.) related to engine friction are changed, it is better to determine the factor of friction loss based on the history information. .

  More specifically, in the above configuration (2), when the storage means stores and holds stored data by supplying battery power, the history when the battery is replaced or the battery power terminal is disconnected. It may be determined that the learning value is lost or not learned based on information or the like.

The learned value of the loss torque characteristic may be used for engine rotation stop control for stopping combustion by stopping combustion when an engine stop request is generated. That is, in the seventh invention, the target trajectory calculating means for calculating the target trajectory that is the rotational behavior until the engine rotation stops at the target stop crank angle using the learned value of the loss torque characteristic stored in the storage means; And stop control means for executing engine rotation stop control for controlling a load of an auxiliary machine of the engine so that the actual engine rotation behavior is matched with the target trajectory when stopping the engine rotation.

  In the engine rotation stop control described above, the engine rotation stop crank angle can be appropriately controlled using the learning value of the loss torque characteristic that is suitably calculated as described above. In such a case, even if the actual loss torque characteristic changes due to individual differences or factors over time, the target trajectory can be calculated with high accuracy while compensating for the deviation of the loss torque characteristic by learning. And, by controlling the auxiliary load of the engine so that the actual engine rotation behavior matches the target trajectory when stopping the engine rotation, it is possible to sufficiently compensate for variations in the actual engine rotation behavior during the engine stop process, The engine rotation stop crank angle can be accurately controlled within the target crank angle range.

1 is a schematic configuration diagram of an engine control system in an embodiment of the invention. (A) is a figure explaining an alternator load characteristic, (b) is a figure explaining an apparent alternator load characteristic at the time of engine rotation stop control. (A) is a time chart for explaining a comparative example in which the engine rotation stop control is performed by setting the quasi-load torque Tref = 0, and (b) is an engine rotation stop control by setting the reference load torque Tref to half of the maximum load. The time chart explaining the Example which performed. The flowchart which shows the flow of a process of the main routine of engine rotation stop learning control. The flowchart which shows the flow of a process of a loss torque characteristic learning routine. The flowchart which shows the flow of a process of target trajectory calculation routine. The flowchart which shows the flow of a process of an engine rotation stop control routine. The figure explaining the calculation method of target engine speed.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, embodiments of the invention will be described with reference to the drawings. In the present embodiment, an engine control system is constructed for an in-vehicle multi-cylinder gasoline engine. In the control system, fuel injection amount control, ignition timing control, idle stop control, and the like are performed with an electronic control unit (hereinafter referred to as ECU) as a center. A schematic configuration diagram of the engine control system is shown in FIG.

  In the engine 10 shown in FIG. 1, a throttle valve 14 whose opening degree is adjusted by a throttle actuator 15 such as a DC motor is provided in an intake pipe 11 (intake passage). The opening degree of the throttle valve 14 (throttle opening degree) is detected by a throttle opening degree sensor built in the throttle actuator 15. A surge tank 16 is provided downstream of the throttle valve 14, and an intake pipe pressure sensor 17 for detecting the intake pipe pressure is provided in the surge tank 16. The surge tank 16 is connected to an intake manifold 18 that introduces air into each cylinder of the engine 10. In the intake manifold 18, an electromagnetically driven fuel injection that injects fuel near the intake port of each cylinder. A valve 19 is attached.

  An intake valve 21 and an exhaust valve 22 are provided at an intake port and an exhaust port of the engine 10, respectively. The air-fuel mixture is introduced into the combustion chamber 23 by the opening operation of the intake valve 21, and the exhaust gas after combustion is discharged to the exhaust pipe 24 (exhaust passage) by the opening operation of the exhaust valve 22.

  A spark plug 27 is attached to the cylinder head of the engine 10 for each cylinder. A high voltage is applied to the spark plug 27 at a desired ignition timing through an ignition device (not shown) including an ignition coil. By applying this high voltage, a spark discharge is generated between the opposing electrodes of each spark plug 27, and the air-fuel mixture introduced into the combustion chamber 23 is ignited and used for combustion.

  The exhaust pipe 24 is provided with a catalyst 31 such as a three-way catalyst for purifying CO, HC, NOx and the like in the exhaust. An A / F sensor 32 is provided on the upstream side of the catalyst 31 for detecting the air-fuel ratio (oxygen concentration) of the air-fuel mixture using exhaust gas as a detection target. Further, the engine 10 includes a coolant temperature sensor 33 that detects the coolant temperature, and a crank angle sensor 34 that outputs a rectangular crank angle signal at predetermined crank angles (for example, at a cycle of 30 ° CA) as the crankshaft rotates. In addition, a cam angle sensor 35 that outputs a cam angle signal for each predetermined cam angle as the cam shaft rotates is attached. In addition, an accelerator sensor 36 for detecting the accelerator opening, a brake sensor 37 for detecting the depression amount of the brake pedal, a vehicle speed sensor 38 for detecting the vehicle speed, and the like are attached to the present system.

  The rotation of the crankshaft is transmitted to an alternator (generator) 39, which is a typical auxiliary machine of the engine 10, via a belt. As a result, the alternator 39 is rotationally driven by the power of the engine 10 to generate power. The load on the alternator 39 can be controlled by duty-controlling the power generation control current (field current) of the alternator 39.

  As is well known, the ECU 40 is mainly composed of a microcomputer composed of a CPU, ROM, RAM, and the like, and executes various control programs stored in the ROM, so that various controls of the engine 10 are performed according to the engine operating state each time. To implement. That is, the ECU 40 inputs detection signals from the various sensors described above, calculates the fuel injection amount, ignition timing, etc. based on the various detection signals, and controls the drive of the fuel injection valve 19 and the ignition device. Or, idle stop control is performed.

  As the idle stop control, the ECU 40 automatically stops the engine 10 by stopping fuel injection and ignition when a predetermined engine stop condition is satisfied during idle operation. When a predetermined engine restart condition is satisfied while the engine is stopped, the engine 10 is initially rotated by cranking a starter (not shown), and then fuel injection and ignition are resumed to restart the engine 10 automatically. Perform a restart. Here, as the predetermined engine stop condition, for example, the accelerator is off, the brake is on, the vehicle speed is zero, and the like. The predetermined engine restart condition is, for example, brake off.

  Further, the ECU 40 functions as a target trajectory calculating means for calculating a target trajectory that is a rotational behavior until the engine rotation stops at the target stop crank angle, and matches the engine rotational behavior to the target trajectory when stopping the engine rotation. Thus, it functions as a stop control means for controlling the load of the alternator 39, and further functions as a loss torque learning means for learning a loss torque characteristic based at least on the actual engine rotational behavior.

  Here, the target trajectory uses a relational expression of an energy conservation law that takes into account the engine's loss torque, and the target stop crank angle of the engine and the target at a predetermined crank angle in the engine stop process (during the rotation decrease accompanying the engine stop request). The crank angle is calculated in the direction of increasing the crank angle based on the engine speed. The relational expression of the energy conservation law is expressed by the following expression.

Ne (i + 1) ² = Ne (i) ² −2 / J × {Tloss−Tref (Ne (i))}
Ne (i + 1) is the engine rotational speed at a time point (i + 1) before a predetermined crank angle (180 ° CA in this embodiment) before the current time point (i), and Ne (i) is the engine speed at the current time point (i). Rotational speed, J is the moment of inertia of the engine 10, and Tloss is a loss torque obtained by adding the pumping loss and the friction loss at a predetermined crank angle (for example, TDC). The ECU 40 includes a backup RAM 41 that is a rewritable storage means that retains stored data even when the power is turned off. The backup RAM 41 stores a loss torque Tloss as a learned value of the loss torque characteristic. Tref (Ne (i)) is the reference load torque of the alternator 39 at the current engine speed Ne (i).

  The target trajectory may be obtained by calculating the relationship between the crank angle up to the target stop crank angle and the target engine speed at a predetermined crank angle interval (for example, 180 ° CA interval) and assigning it to a table or the like. Good.

  In the present embodiment, the reference load torque Tref (Ne (i)) of the alternator 39 is set to half (1/2) of the maximum load that can be controlled by the alternator 39, as shown in FIG. . In this way, unlike the motor generator, the alternator 39 can virtually control the load torque of the alternator 39 in both the positive and negative directions even if there is a situation where the assist torque cannot be output (below the reference load Tref). The load torque of the alternator 39 can be controlled by virtually setting the load torque of the alternator 39 as a negative load torque and a load torque equal to or higher than the reference load Tref as a positive load torque). Can be improved.

  The reference load torque Tref (Ne (i)) of the alternator 39 is not limited to half (1/2) of the maximum load, for example, 1/3, 1/4, 2/3, 3 / of the maximum load. 4 or the like. In short, an appropriate load smaller than the maximum controllable load of the alternator 39 and larger than 0 may be set as the reference load torque Tref (Ne (i)) (0 < Tref (Ne (i)) <alternator maximum load).

  FIG. 3A shows a comparative example in which the engine rotation stop control is performed by setting the reference load torque Tref (Ne (i)) = 0. In this comparative example, since the load torque of the alternator 39 can be controlled only in the positive direction, if the actual engine rotation behavior overshoots, the actual engine rotation behavior cannot be matched with the target trajectory. On the other hand, if the reference load torque Tref (Ne (i)) of the alternator 39 is set to an appropriate load smaller than the maximum load as in the present embodiment, as shown in FIG. Since the load torque of the alternator 39 can be controlled in both positive and negative directions, even if the actual rotational behavior overshoots, as shown in FIG. Can do.

  Furthermore, in this embodiment, when calculating the target trajectory, the target trajectory is calculated based on the loss torque Tloss. During engine rotation stop control, the reference load torque Tref (Ne (i)) corresponding to the engine rotation speed Ne (i) is calculated, and the target engine rotation speed and the actual engine at a predetermined crank angle θ (for example, TDC) are calculated. The base load torque is calculated so as to reduce the deviation from the rotational speed, and the required load torque Talt is calculated by adding the reference load torque Tref (Ne (i)) to this base load torque (in practice, the required load The torque Talt is multiplied by the pulley ratio Ratio to convert it to the required shaft torque Tfinal). Then, a power generation command is calculated in accordance with the current required load torque Talt (required shaft torque Tfinal) and the engine speed Ne (i), and the power generation control current (field current) of the alternator 39 is controlled by this power generation command. Thus, the load torque of the alternator 39 is controlled.

  Such control of the load torque of the alternator 39 is periodically executed at a predetermined crank interval (for example, 180 ° CA interval) until the actual engine rotational speed decreases below the power generation limit rotational speed Nelow (see FIG. 2) of the alternator 39. By doing so, the load torque of the alternator 39 is feedback-controlled so that the actual engine rotational behavior matches the target trajectory.

  By the way, the loss torque characteristic (that is, the loss torque Tloss) used when calculating the target trajectory changes due to manufacturing variation of the engine 10 or a change with time. Therefore, if the target trajectory is calculated using a preset standard loss torque characteristic. When there is a deviation between the actual loss torque characteristics and the standard loss torque characteristics due to variations in manufacturing of the engine 10, changes over time, changes in oil temperature, and the like, the accuracy of calculating the target trajectory decreases.

  As a countermeasure, in the present embodiment, focusing on the effect of the change in the loss torque characteristic appearing in the actual engine rotation behavior, the loss torque characteristic is learned based on the actual engine rotation behavior, updated and stored in the backup RAM 41, and this loss torque. The target trajectory is calculated using the characteristics.

  In the present embodiment, a loss torque between a predetermined crank angle (for example, TDC) from the actual engine rotation behavior in a period from when combustion is stopped in response to an engine stop request (idle stop signal) to when engine rotation stop control is started. The consumed energy is calculated, and the loss torque characteristic is learned based on the consumed energy. That is, when the combustion is stopped in response to the engine stop request, the actual engine rotation speed decreases due to the energy consumed by the loss torque (pumping loss or friction loss) while the engine 10 rotates by inertia. In this case, according to the actual engine rotation behavior, the energy consumed by the loss torque between predetermined crank angles (for example, between TDCs) can be calculated with high accuracy, and the loss torque characteristics can be learned with high accuracy based on the energy consumption by this loss torque.

  In other words, the drive of the alternator 39 may be stopped during a period when the actual engine rotation speed decreases. In this case, it is considered that the energy consumed by the loss torque can be calculated with higher accuracy by inertial rotation of the engine 10 while the alternator 39 is not driven, and the learning accuracy of the loss torque characteristic can be improved.

  Note that the loss torque Tloss, which is the loss torque characteristic, differs depending on the crank angle (that is, the piston position) of the engine 10 even in the same stroke. Therefore, a configuration may be adopted in which a table or the like is set at a predetermined crank angle interval (for example, 30 ° CA interval) and the loss torque Tloss is stored for each crank angle at the predetermined crank angle interval. It is also possible to use an EEPROM instead of the backup RAM 41 as the storage means.

  The learning of the loss torque characteristic and the engine rotation stop control according to the present embodiment described above are executed by the ECU 40 according to the routines shown in FIGS. The processing contents of these routines will be described below.

[Main routine of engine rotation stop learning control]
The main routine of the engine rotation stop learning control in FIG. 4 is executed at a predetermined cycle during engine operation. When this routine is started, first, in step S100, a loss torque characteristic learning routine of FIG. 5 described later is executed, and the loss torque characteristic is learned when the learning execution condition is satisfied. Thereafter, in step S200, a target trajectory calculation routine of FIG. 6 described later is executed to calculate a target trajectory based on the loss torque characteristic learned by the loss torque characteristic learning routine of FIG. Further, in step S300, an engine rotation stop control routine of FIG. 7 described later is executed to control the load torque of the alternator 39 so that the actual engine rotation behavior is matched with the target trajectory when the engine rotation is stopped.

[Loss torque characteristics learning routine]
The loss torque characteristic learning routine of FIG. 5 is a subroutine executed in step S100 of the main routine of FIG. 4, and plays a role as loss torque learning means.

When this routine is started, first, in step S101, it is determined whether or not a predetermined learning execution condition is satisfied. Here, the learning execution condition includes the time after completion of warm-up of the engine 10 and at the time of idling stop (when the engine is automatically stopped when the engine stop condition is satisfied). As a warm-up completion determination method, for example, any of the following methods may be used.
When the coolant temperature detected by the coolant temperature sensor 33 rises to a predetermined warm-up determination value, it is determined that the warm-up has been completed.
When the integrated value of the intake air amount after starting the engine (after turning on the ignition switch) reaches a predetermined warm-up determination value, it is determined that the warm-up has been completed.
When the integrated value of the fuel injection amount after starting the engine (after turning on the ignition switch) reaches a predetermined warm-up determination value, it is determined that the warm-up has been completed.

  The learning execution condition may include the first idle stop time after the coolant temperature reaches the warm-up completion temperature. However, it is possible to learn not only at the first idle stop but also at the second and subsequent idle stops. Or you may make it learn at the time of an appropriate idle stop at any time interval for every predetermined idle stop frequency | count, every predetermined | prescribed driving frequency | count, every predetermined integrated driving | running distance, and every predetermined period progress. If it is determined in step S101 that the learning execution condition is not satisfied, the present routine is terminated as it is.

  On the other hand, if it is determined in step S101 that the learning execution condition is satisfied, the process proceeds to step S102, where the current crank angle θ and engine speed are determined based on the output pulses of the crank angle sensor 34 and the cam angle sensor 35. And calculate. At this time, the crank angle θ is calculated by the intake ATDC.

  Thereafter, in step S103, it is determined whether or not the current crank angle θ is TDC (that is, ATDC 0 ° CA) which is a learning value calculation timing. If not, the present routine is terminated. On the other hand, if the current crank angle θ is TDC, the process proceeds to step S104, and the calculated engine speed is stored in the RAM of the ECU 40 as the engine speed Ne (i) at the current TDC.

  Thereafter, in step S105, the energy consumption amount ΔE consumed by the loss torque between TDCs (180 ° CA) is calculated by the following equation.

ΔE = 1/2 · J · Ne (i-1) ² −1 / 2 · J · Ne (i) ²
J is the moment of inertia of the engine 10, and Ne (i-1) is the engine speed at the previous TDC.

  Thereafter, in step S106, the temporary learning correction torque Tgg and the actual loss torque Tlossg are calculated by the following equations using the calculated energy consumption ΔE due to the loss torque between the TDCs.

Tgg = ΔE-ΔEtg
Tlossg = ΔE
ΔEtg is the energy consumption due to the loss torque between TDCs used for the previous calculation of the target trajectory. ΔEtg may be stored in the backup RAM 41. The temporary learning correction torque Tgg and the actual loss torque Tlossg are
Tgg = (ΔE−ΔEtg) / Δθ
Tlossg = ΔE / Δθ
May be calculated as In this case, Δθ is a crank angle between TDCs (180 ° CA).

  Here, the provisional learning correction torque Tgg is calculated based on an error (ΔE−ΔEtg) between the previous value and the current value of the energy consumption. The actual loss torque Tlossg is calculated based on the energy consumption amount ΔE actually generated this time.

  Thereafter, in step S107, it is determined whether the ratio of the provisional learning correction torque Tgg to the loss torque Tloss used for the previous calculation of the target trajectory (that is, the error ratio of the current calculated value) is equal to or less than a predetermined determination value. Specifically, it is determined whether the (Tgg / Tloss) value is within ± 0.15 (within ± 15%). Step S107 corresponds to means for determining whether or not the difference between the learned value of the loss torque characteristic stored in the backup RAM 41 and the calculated value in the current loss torque characteristic learning is greater than or equal to a predetermined value.

  If | Tgg / Tloss | ≦ 0.15, the process proceeds to step S108, and the learning correction torque Tg is calculated by subjecting the provisional learning correction torque Tgg to, for example, the following equation. In this case, the learning correction torque Tg is gradually corrected while the change of the learning correction torque Tg from the previous value is limited.

Tg (current value) = {a × Tg (previous value) + b × Tgg} / (a + b)
Here, a and b are coefficients. Tg (previous value) may be stored in the backup RAM 41.

  Thereafter, in step S109, a new loss torque Tloss is calculated by adding the learning correction torque Tg to the loss torque Tloss (Tloss stored in the backup RAM 41) used for the previous calculation of the target trajectory. Update and store.

  On the other hand, if | Tgg / Tloss |> 0.15, the process proceeds to step S110 and the n counter is incremented by one. This n counter is initialized to 0 when the learning execution condition is first established (when learning of the loss torque characteristic is started) and when affirmative in step S107, and when step S107 is continuously denied after starting learning of the loss torque characteristic, It is incremented by one. In step S111, it is determined whether or not the n counter is greater than or equal to a predetermined value (3 or greater in the present embodiment).

  If n <3, the process proceeds to step S108 to calculate the learning correction torque Tg by smoothing the temporary learning correction torque Tgg as described above, and add the learning correction torque Tg to the previous value of the loss torque Tloss. The learning value in the backup RAM 41 is updated with the new loss torque Tloss obtained in this way.

  If n ≧ 3, that is, if the negative determination in step S107 is made three times in succession, the process proceeds to step S112. In step S112, the actual loss torque Tlossg calculated in step S106 is set as a new loss torque Tloss, and this is updated and stored in the backup RAM 41. That is, in step S112, a new loss torque Tloss is calculated without using the previous value of the loss torque Tloss or the learning correction torque Tg used for the previous calculation of the target trajectory.

  In this embodiment, step S109 corresponds to the first learning means, step S112 corresponds to the second learning means, and step S107 corresponds to the switching means.

[Target trajectory calculation routine]
The target trajectory calculation routine of FIG. 6 is a subroutine executed in step S200 of the main routine of FIG. 4 and plays a role as target trajectory calculation means.

  First, the outline of the target trajectory calculation process will be described. In the present embodiment, based on the target stop crank angle of the engine, the target rotational speed at the predetermined crank angle (TDC) in the engine rotation stop process (or target rotational energy), and the loss torque Tloss (learning value of the loss torque characteristic). The target trajectory is set. In particular, in the case where the actual engine rotational speed is matched with the target engine rotational speed on the target trajectory, two target trajectories that are shifted back and forth by 180 ° CA are set, and among these target trajectories, the actual engine The target engine rotational speed is calculated by selecting the target trajectory with the smaller amount of energy required for correcting the rotational speed.

  More specifically, as shown in FIG. 8, when the target stop crank angle of the engine is set in two ways as “φ + 180 × k” and “φ + 180 × (k + 1)”, the target stop crank angle with respect to each target stop crank angle is set. Two rotational trajectories are defined (trajectory 1, trajectory 2). In this case, the energy difference at the crank angle shifted by 180 ° CA in each of the tracks 1 and 2 corresponds to the loss torque characteristic of the engine. The difference between the current rotational energy based on the current engine rotational speed Ne and the two target rotational energy based on the respective target rotational speeds Nt (i + 1) and Nt (i) on the two target trajectories ( Energy difference), a rotation trajectory having a smaller absolute value of the amount of energy required for the current rotational energy is selected, and a target engine rotational speed is calculated from the rotation trajectory and each crank angle position. .

  In the routine of FIG. 6, first, in step S201, it is determined whether or not the current crank angle θ is TDC which is the target trajectory calculation timing. If it is not TDC, this routine is ended as it is. On the other hand, if the current crank angle θ is TDC, the process proceeds to step S202, and the square value of the target rotational speed Nt (i + 1) is calculated by the following equation.

Nt (i + 1) ² = Nt (i) ² + 2 / J ・ Tloss
Here, Nt (i) is a target rotational speed at a predetermined TDC while the engine rotation is stopped (Ne> 0), and Nt (i + 1) goes back by 180 ° CA with respect to Nt (i). This is the target rotational speed at the time (before 180 ° CA). For example, 200 rpm is set as the initial value of Nt (i). According to step S202, target rotational speeds at two positions separated by 180 ° CA (large and small target rotational speeds) are calculated.

  Thereafter, in step S203, it is determined whether or not the current engine speed Ne is smaller than the target speed Nt (i + 1) (square root of the square value of Nt (i + 1)) calculated in step S202. To do. In this case, if Ne ≧ Nt (i + 1), the current engine rotational speed Ne is excessive with respect to the target rotational speed Nt (i + 1) calculated this time, and a larger target rotational speed Nt (i +). Since it is necessary to calculate 1), the process proceeds to step S204, and after incrementing the i counter by 1, the target rotational speed Nt (i + 1) is calculated again in step S202. As a result, the target rotational speed Nt (i + 1) is calculated further backward by 180 ° CA. The calculation of the target rotational speed Nt (i + 1) in step S202 is repeated until step S202 becomes YES.

Then, if Ne <Nt (i + 1), the process proceeds to step S205, where the current engine speed Ne is adjusted to one of the two target speeds Nt (i + 1) and Nt (i). Determine if the amount of energy required is small. In particular,
Nt (i + 1) -Ne> Ne-Nt (i)
Whether or not is satisfied is determined. In this case, if Nt (i + 1) −Ne> Ne−Nt (i), it can be determined that the amount of energy required to match the target rotational speed Nt (i) is smaller, and Nt (i + 1) − If Ne ≦ Ne−Nt (i), it can be determined that the amount of energy required to match the target rotational speed Nt (i + 1) is smaller.

  When step S205 is YES, the process proceeds to step S206, and the target rotational speed Nt (i) is set as the target engine rotational speed. In the subsequent step S207, the i counter is decremented by 1 from “i” (i = i−1). If step S205 is NO, the process proceeds to step S208, where the target rotational speed Nt (i + 1) is set as the target engine rotational speed. In the subsequent step S209, the i counter is decremented by 1 from “i + 1” (i = (i + 1) −1).

[Engine rotation stop control routine]
The engine rotation stop control routine of FIG. 7 is a subroutine executed in step S300 of the main routine of FIG. 4 and plays a role as stop control means.

  When this routine is started, first, in step S301, it is determined whether or not an engine stop request (idle stop signal) has been generated. If no engine stop request has been generated, this routine is terminated and the engine operation is terminated. Continue (fuel injection control and ignition control). Then, when it is determined that an engine stop request has been generated, the process proceeds to step S302, where the current crank angle θ and the engine speed Ne are calculated.

  Thereafter, in step S303, it is determined whether or not the current crank angle θ is the load torque control timing (TDC) of the alternator 39. In step S304, it is determined whether or not the current engine speed Ne is lower than the maximum engine speed Nemax at which engine rotation stop control can be executed. If either of steps S303 and S304 is NO, this routine is terminated as it is, and if both of steps S303 and S304 are YES, the process proceeds to subsequent step S305.

  In step S305, it is determined whether or not the engine is still in the combustion state after the engine stop request. If it is in the combustion state, the process proceeds to step S306, and the reference load torque Tref (Ne) of the alternator 39 calculated based on the current engine speed Ne is set as the required load torque Talt of the alternator 39 (Talt = Tref). (Ne)).

  If not in the combustion state, the process proceeds to step S307, and the required load torque Talt for making the engine rotational speed Ne coincide with the target rotational speed Netg using the relation of the energy conservation law and the reference load torque Tref (Ne) of the alternator 39. Is calculated. Specifically, the required load torque Talt is calculated using the following equation.

Here, J is the moment of inertia of the engine 10, K is the feedback gain, and Δθ is the crank angle change amount (180 ° CA). The target engine speed Netg may be calculated by map calculation or mathematical expression calculation. It is also possible to provide a dead zone for the deviation in engine rotation speed.

  Thereafter, in step S308, the required load torque Talt is converted into the required shaft torque Tfinal of the alternator 39 by multiplying the required load torque Talt by the pulley ratio Ratio between the engine crankshaft side and the alternator shaft side.

Tfinal = Talt × Ratio
In the subsequent step S309, the alternator rotational speed Nalt is calculated from the engine rotational speed Ne and the pool ratio Ratio.

Nalt = Ne × Ratio
Thereafter, in step S310, the battery voltage is detected. In step S311, the excitation current IF is calculated based on the required shaft torque Tfinal of the alternator 39, the alternator rotation speed Nalt, and the battery voltage. Specifically, a required load torque characteristic map corresponding to the current battery voltage is selected from a plurality of required load torque characteristic maps created for each battery voltage at a predetermined interval, and the current required axis is selected using the characteristic map. An exciting current IF is calculated according to the torque Tfinal and the alternator rotation speed Nalt. The excitation current IF is converted into a power generation command (duty duty), and the load torque of the alternator 39 is controlled by the duty duty.

  According to the embodiment described in detail above, the following excellent effects can be obtained.

  Considering that the loss torque characteristic (Tloss) used when calculating the target trajectory changes due to manufacturing variations of the engine 10 and changes over time, the loss torque characteristic is learned based on the actual engine rotational behavior and updated and stored in the backup RAM 41. Therefore, even if actual loss torque characteristics vary due to variations in manufacturing of the engine 10, changes over time, changes in oil temperature, etc., the variations in the loss torque characteristics can be compensated by learning, and the target trajectory is calculated. Accuracy can be improved. Thus, as in this embodiment, when the load torque of the alternator 39 is controlled so that the actual engine rotation behavior matches the target trajectory when the engine rotation is stopped, the variation in the actual engine rotation behavior during the engine stop process can be sufficiently increased. Therefore, the engine rotation stop position can be accurately controlled within the target crank angle range.

  In the loss torque characteristic learning process, first learning means for updating the learning value based on the learning value (Tloss) stored in the backup RAM 41 and the newly calculated loss torque characteristic, and the learning value stored in the backup RAM 41 ( (Tloss) is used, and the second learning means for updating the learning value based on the newly calculated loss torque characteristic is appropriately switched, so that it is required to gradually update the learning value from the viewpoint of improving reliability. Therefore, it is possible to appropriately cope with both the case where it is required to update the learning value only with the actual actual loss torque characteristic from the viewpoint of improving convergence. As a result, it is possible to realize suitable learning about the loss torque characteristics.

  In particular, when supplementing the learning value update by the second learning means, the learning means obtains the ratio of the provisional learning correction torque Tgg to the loss torque Tloss used for the previous calculation of the target trajectory (that is, the error ratio of the current calculated value). When the value is larger than the predetermined determination value, the learning value is updated by the actual loss torque Tlossg calculated this time without using the loss torque Tloss. As a result, it takes less time for the learning value to converge with respect to the actual loss torque characteristic that should be learned, and a desired learning value can be acquired quickly.

(Other embodiments)
The present invention is not limited to the description of the above embodiment, and may be implemented as follows, for example.

  In the loss torque characteristic learning process, the operation of the alternator 39 is stopped in the process of stopping the engine to put the engine in the inertial rotation state, and various calculations (loss of energy consumption) on the loss torque Tloss based on the actual engine rotation behavior in the inertial rotation state It is also possible to employ a configuration in which ΔE calculation, learning correction torque Tg calculation, etc.

In the loss torque characteristic learning process, the second learning means may update the learning value with an average value of a plurality of calculated values of the loss torque characteristic calculated based on the actual engine rotational behavior. Specifically, in step S112 of the loss torque characteristic learning routine (FIG. 5), the calculated values (Tlossg _n , Tlossg _n−1 , Tlossg _n−1 , The average value of Tlossg _n-2 ) is calculated, and the average value is updated and stored in the backup RAM 41 as a new loss torque Tloss. The calculation formula is as follows.

Tloss = (Tlossg _n + Tlossg _n-1 + Tlossg _n-2 ) / 3
According to the above configuration, the learning accuracy of the loss torque characteristic can be ensured even when the difference between the stored learned value and the current calculated value of the loss torque characteristic is greater than or equal to a predetermined value.

  -It is determined that a factor change affecting the engine friction loss has occurred, and when the factor change of friction loss occurs, the learning value is updated by the second learning means instead of the first learning means. It is good also as composition to do. For example, when a change in engine oil or replacement of parts related to engine friction (piston ring or the like) occurs, it is preferable to determine a factor change in friction loss based on the history information. Specifically, in step S107 of the loss torque characteristic learning routine (FIG. 5), history information such as engine oil replacement and piston ring replacement is confirmed, and if the history exists, learning by the second learning means is performed. The value is updated (step S112).

  First learning means when it is determined that the learning value stored in the backup RAM 41 has disappeared or the learning value is an unlearned initial value, and the learning value has been lost or unlearned. Instead of this, the learning value may be updated by the second learning means. For example, in the backup RAM 41 that stores and holds stored data by supplying battery power, when the battery is replaced or the battery power terminal is disconnected, the learning value is lost or not learned based on the history information. It is good to judge. As a specific configuration, in step S107 of the loss torque characteristic learning routine (FIG. 5), the history information on the battery replacement or the battery power terminal disconnection is confirmed, and if the history exists, the second learning means The learning value is updated (step S112).

  A first-order lag process, a weighted averaging process, a filter process, and the like can be arbitrarily applied as the smoothing process regarding the learned value (Tloss) of the loss torque characteristic. In any case, any newly calculated loss torque characteristic may be reflected on the learning value stored in the backup RAM 41 by the smoothing process.

  -This invention is not limited to the structure which controls the load of alternator 39 (generator) during engine rotation stop control, It is made to control loads, such as auxiliary machines other than alternator 39, for example, the compressor of an air conditioner. Good.

  DESCRIPTION OF SYMBOLS 10 ... Engine, 39 ... Alternator, 40 ... ECU (Loss torque learning means, 1st learning means, 2nd learning means, switching means, target trajectory calculation means, stop control means), 41 ... Backup RAM (storage means).

Claims (5)

In an engine control device comprising: storage means for storing a loss torque characteristic of an engine as a learning value; and loss torque learning means for updating a learning value of the storage means with a loss torque characteristic calculated based on an actual engine rotational behavior.
The loss torque learning means includes
First learning means for updating the learning value based on the learning value stored in the storage means and the newly calculated loss torque characteristic;
Second learning means for updating the learning value based on the newly calculated loss torque characteristic without using the learning value stored in the storage means;
Switching means for switching whether the learning value is updated by either the first learning means or the second learning means;
Means for determining that a change in a factor affecting the friction loss of the engine has occurred,
The switching means causes the second learning means to update the learning value when a factor change of the friction loss occurs,
Means for determining that the difference between the learning value stored in the storage means and the current calculated value of the loss torque characteristic is greater than or equal to a predetermined value;
Means for determining success or failure of a learning execution condition for the loss torque characteristic,
In the learning execution period associated with the establishment of the learning execution condition, the switching means determines that the difference is equal to or greater than a predetermined number of times, and the second learning means performs a plurality of loss torque characteristics. using the average value of the calculated values to update the learning value, after learning by the second learning means has been performed, the learning execution period due to establishment of the learning execution condition is that the said difference is less than the predetermined engine control device, characterized in that when it is determine a constant, but to learn the learning value by after the first learning means. Means for determining that the learning value stored in the storage means has disappeared or that the learning value is an unlearned initial value;
The engine control apparatus according to claim 1, wherein the switching unit causes the second learning unit to update the learning value when it is determined that the learning value has disappeared or has not yet been learned. In an engine control device comprising: storage means for storing a loss torque characteristic of an engine as a learning value; and loss torque learning means for updating a learning value of the storage means with a loss torque characteristic calculated based on an actual engine rotational behavior.
The loss torque learning means includes
First learning means for updating the learning value based on the learning value stored in the storage means and the newly calculated loss torque characteristic;
Second learning means for updating the learning value based on the newly calculated loss torque characteristic without using the learning value stored in the storage means;
Switching means for switching whether the learning value is updated by either the first learning means or the second learning means;
Means for determining that the difference between the learning value stored in the storage means and the current calculated value of the loss torque characteristic is greater than or equal to a predetermined value;
Means for determining success or failure of a learning execution condition for the loss torque characteristic,
In the learning execution period associated with the establishment of the learning execution condition, the switching means determines that the difference is equal to or greater than a predetermined number of times, and the second learning means performs a plurality of loss torque characteristics. using the average value of the calculated values to update the learning value, after learning by the second learning means has been performed, the learning execution period due to establishment of the learning execution condition is that the said difference is less than the predetermined If it is determine constant, the engine control unit, characterized in that to learn the learning value by after the first learning means. The first learning means updates the learning value by reflecting the newly calculated loss torque characteristic in the smoothing process with respect to the learning value stored in the storage means,
The engine control apparatus according to any one of claims 1 to 3, wherein the second learning unit updates the learning value based on a newly calculated loss torque characteristic without performing the annealing process. In an engine control device that stops combustion by stopping combustion when an engine stop request is generated,
Target trajectory calculation means for calculating a target trajectory that is rotational behavior until engine rotation stops at a target stop crank angle, using a learning value of loss torque characteristics stored in the storage means;
Stop control means for executing engine rotation stop control for controlling a load of an auxiliary machine of the engine so that the actual engine rotation behavior is matched with the target trajectory when stopping the engine rotation;
The engine control apparatus according to any one of claims 1 to 4, further comprising: