JP2012082735A - Engine revolution stop control device - Google Patents

Abstract

In a system that learns target trajectory information (the amount of deviation between a reference rotational speed and loss torque used for calculating a target trajectory) based on actual engine rotational behavior when engine rotation stop control is executed, an error in target trajectory information is obtained. Help prevent learning.
If the learning value calculation data of the current target trajectory information deviates more than a predetermined value from the stored data of the learning value of the previous target trajectory information (reference rotational speed or loss torque deviation amount), the state Is not consecutive for a predetermined number of times, it is determined that the learning value (calculated data) of the target trajectory information may have temporarily changed due to some external load, etc. Hold to value. On the other hand, when the state continues for a predetermined number of times, it is determined that the friction or the like of the engine 11 has changed and the learning value (calculated data) of the target trajectory information has steadily changed. Update.
[Selection] Figure 1

Description

  The present invention relates to an engine rotation stop control device having a function of controlling an engine rotation stop position (stop crank angle).

  In recent years, for example, as described in Patent Document 1 (Japanese Patent Laid-Open No. 2008-215230), in a vehicle equipped with an engine automatic stop / start system (idle stop system), in order to improve restartability, Targeting the rotational behavior until the engine rotation stops at the target stop crank angle for the purpose of controlling the engine rotation stop position (stop crank angle) to a crank angle range suitable for starting when the engine is stopped (idle stop) There is an engine rotation stop control that is calculated as a track and controls the load torque of a generator (alternator) so that the actual engine rotation behavior matches the target track when the engine rotation is stopped. Specifically, the required load torque of the generator is calculated so that the actual engine rotation behavior matches the target trajectory, and the generator load torque characteristics (the relationship between the power generation command value, the engine rotation speed, and the load torque) are used. The power generation command value corresponding to the current engine speed and the required load torque is calculated, and the power generation control current (field current) of the generator is controlled by this power generation command value to control the load torque of the generator. Yes.

JP 2008-215230 A

  By the way, when the engine rotation speed is reduced below the power generation limit rotation speed of the generator during the engine rotation stop control, the load torque of the generator hardly occurs (see FIG. 3). In such a rotation speed region, the engine rotation speed decreases and the engine rotation stops almost without being affected by the load torque of the generator. Therefore, the engine rotation speed when passing a predetermined reference timing (for example, TDC). The engine rotation stops at the stop crank angle corresponding to.

  By paying attention to such characteristics, the present applicant uses the relationship between the engine speed at the reference timing (the engine speed when passing the reference timing) and the stop crank angle to determine that the stop crank angle is the target stop crank. The engine speed at the reference timing that is the corner is obtained as the reference speed, and the reference speed is set as the initial value using the energy conservation law relational expression that takes into account engine loss torque (torque including pumping loss and friction loss). We are researching a system that sets the target trajectory in the direction of turning the crank angle, but the following new issues were found during the research process.

  When engine friction or the like changes due to engine oil change or changes over time, the reference rotational speed and loss torque used to calculate the target trajectory also change. Therefore, in order to ensure the calculation accuracy of the target trajectory, the learning value of the reference rotational speed and the loss torque based on the actual engine rotational behavior when the engine rotational speed decreases below the power generation limit rotational speed of the generator by the engine rotational stop control. We are studying a system that learns the amount of deviation between the reference rotational speed and loss torque by calculating and storing the learned value of the amount of deviation. However, if the actual engine rotation behavior changes temporarily when the engine rotation stop control is executed due to some external load, etc., the reference rotation speed calculated based on the actual engine rotation behavior is affected. Since the learning value and the learning value of the loss amount of the loss torque change temporarily, there is a possibility of mislearning the deviation amount of the reference rotation speed and the loss torque, and the learning accuracy of the deviation amount of the reference rotation speed and the loss torque decreases. there's a possibility that.

  Therefore, the problem to be solved by the present invention is to prevent engine false stop of target trajectory information (information used for calculation of target trajectory) and to improve learning accuracy of target trajectory information. To provide an apparatus.

  In order to solve the above-mentioned problem, the invention according to claim 1 is directed to a target trajectory calculating means for calculating a target trajectory of engine rotation behavior so that engine rotation stops at a target stop crank angle, and an engine in response to an engine stop request. An engine rotation stop control device comprising stop control means for executing engine rotation stop control for controlling a load on a generator so that an actual engine rotation behavior is matched with a target trajectory when rotation is stopped. A learning value of information used for calculating a target trajectory (hereinafter referred to as “target trajectory information”) is calculated based on the actual engine rotational behavior when executed, and the target trajectory is calculated based on the learning value calculation data of the target trajectory information. The learning means for learning the target trajectory information by updating the stored data of the learning value of the information, the learning means, the learning value of the target trajectory information When the learning data of the target trajectory information calculated this time differs from the stored data by a predetermined value or more, the update permission condition is satisfied based on the learning data of the learning value of the target trajectory information for multiple times thereafter If the update permission condition is satisfied, the stored data of the learning value of the target trajectory information is updated.

  When the calculation data of the learning value of the target trajectory information calculated this time changes relatively with respect to the stored data of the learning value of the target trajectory information, the engine friction changes due to the oil change of the engine, etc. The learning value (calculation data) of information may change constantly, but the learning value (calculation data) of target trajectory information may temporarily change due to some external load or the like.

  In view of this, the present invention provides learning of target trajectory information multiple times after that when the calculated data of the target trajectory information calculated this time deviates from the stored data of the target trajectory information by a predetermined value or more. By monitoring the calculated value data, whether the learning value (calculated data) of the target trajectory information has changed constantly due to changes in engine friction or the like, or the learned value of the target trajectory information due to some external load, etc. It can be determined whether (calculation data) has changed temporarily, and it can be determined whether an update permission condition is satisfied.

  If it is determined that there is a possibility that the learning value (calculation data) of the target trajectory information has temporarily changed due to some external load or the like, and it is determined that the update permission condition is not satisfied, the learning of the target trajectory information is performed. The stored value data can be held at the previous value without being updated. On the other hand, when it is determined that the learning value (calculated data) of the target trajectory information has changed steadily due to changes in engine friction or the like and it is determined that the update permission condition is satisfied, the learning value of the target trajectory information is stored. Data can be updated. As a result, erroneous learning of the target trajectory information can be prevented, and the learning accuracy of the target trajectory information can be improved.

  Specifically, as in claim 2, when the calculated data of the target trajectory information calculated this time differs from the stored data of the target trajectory information by a predetermined value or more, the state is determined a predetermined number of times. It may be determined that the update permission condition is satisfied when consecutive. In this way, when the state in which the learning data of the target trajectory information calculated this time deviates more than a predetermined value from the stored data of the learning value of the target trajectory information continues for a predetermined number of times, engine friction or the like may occur. It can be determined that the update permission condition is satisfied by determining that the learning value (calculated data) of the target trajectory information has changed steadily.

  Further, as in claim 3, a gain for learning the target trajectory information when the update permission condition is satisfied (hereinafter referred to as “learning gain”) is made larger than a value before the update permission condition is satisfied. Also good. In this way, the learning value (stored data) of the target trajectory information can be quickly brought close to the true value of the target trajectory information.

  Further, as in claim 4, when the learning value of the target trajectory information is in a converged state after the learning gain is made larger than the value before the update permission condition is established, the learning gain is set before the update permission condition is established. You may make it return to a value. In this way, when the learning value of the target trajectory information is in a converged state (almost constant), it is determined that the learning value of the target trajectory information has converged near the true value of the target trajectory information, and the learning gain is set. It is possible to restore the value (normal value) before the update permission condition is satisfied.

  Further, as in claim 5, the target trajectory is calculated based on the engine rotational speed at the reference timing (hereinafter referred to as “reference rotational speed”) at which the engine stop crank angle becomes the target stop crank angle, and the engine loss torque. In the case of a system that does this, it is preferable to learn the deviation amount between the reference rotational speed and the loss torque as the target trajectory information. In this way, in the system that calculates the target trajectory based on the reference rotational speed and the loss torque, the learning accuracy of the deviation amount between the reference rotational speed and the loss torque can be improved, and the calculation accuracy of the target trajectory can be improved.

FIG. 1 is a diagram showing a schematic configuration of an engine control system in one embodiment of the present invention. FIG. 2 is a diagram for explaining a method for calculating a target trajectory. FIG. 3 is a diagram for explaining alternator load characteristics. FIG. 4 is a diagram for explaining an apparent alternator load characteristic during engine rotation stop control. FIG. 5A is a time chart for explaining a comparative example in which the engine rotation stop control is performed by setting the reference load torque Tref (Ne (i)) = 0, and FIG. 5B is a reference load torque Tref ( 6 is a time chart illustrating an embodiment in which engine rotation stop control is performed with Ne (i)) set to half of the maximum load. FIG. 6 is a block diagram illustrating an engine rotation stop control function of the engine ECU. FIG. 7 is a diagram schematically showing an example of a map of load torque characteristics. FIG. 8 is a diagram for explaining the outline of the entire control. FIG. 9 is a diagram for explaining a loss torque learning method. FIG. 10 is a diagram for explaining a learning method of the reference rotation speed. FIG. 11 is a diagram conceptually illustrating an example of a map of the energy shift amount Δe. FIG. 12 is a time chart for explaining a learning value (stored data) update method. FIG. 13 is a time chart showing the behavior of the learning value (stored data) when the update permission condition is satisfied. FIG. 14 is a flowchart for explaining the processing flow of the target trajectory calculation routine. FIG. 15 is a flowchart for explaining the processing flow of the alternator F / B stop control routine. FIG. 16 is a flowchart for explaining the flow of processing of the learning control routine. FIG. 17 is a flowchart for explaining the processing flow of the loss torque learning routine. FIG. 18 is a flowchart for explaining the processing flow of the reference rotational speed learning routine. FIG. 19 is a flowchart for explaining the flow of the learning value update determination routine.

Hereinafter, an embodiment embodying a mode for carrying out the present invention will be described.
First, the overall configuration of the engine control system will be schematically described with reference to FIG.
A throttle valve 14 is provided in the middle of the intake pipe 13 connected to the intake port 12 of the engine 11, and the opening (throttle opening) of the throttle valve 14 is detected by a throttle opening sensor 15. Further, an intake pipe pressure sensor 18 for detecting an intake pipe pressure is provided on the downstream side of the throttle valve 14, and fuel injection for injecting fuel toward the intake port 12 in the vicinity of the intake port 12 of each cylinder. A valve 19 is attached.

  On the other hand, an exhaust gas purifying catalyst 22 is installed in the middle of the exhaust pipe 21 connected to the exhaust port 20 of the engine 11. The cylinder block of the engine 11 is provided with a cooling water temperature sensor 23 that detects the cooling water temperature. A crank angle sensor 26 is installed facing the outer periphery of the signal rotor 25 attached to the crankshaft 24 of the engine 11. The crank angle sensor 26 synchronizes with the rotation of the signal rotor 25 from the crank angle sensor 26 at every predetermined crank angle (for example, 30 ° C.). A crank pulse signal is output every A). A cam angle sensor 29 is installed opposite to the outer periphery of the signal rotor 28 attached to the cam shaft 27 of the engine 11, and the cam pulse is generated at a predetermined cam angle in synchronization with the rotation of the signal rotor 28 from the cam angle sensor 29. A signal is output.

  Further, rotation of a crank pulley 34 connected to the crankshaft 24 is transmitted to the alternator 33 (generator) via a belt 35. Thereby, the alternator 33 is rotationally driven by the power of the engine 11 to generate power. The load on the alternator 33 can be controlled by duty-controlling the power generation control current (field current) of the alternator 33.

  Outputs of the various sensors described above are input to an engine control circuit (hereinafter referred to as “engine ECU”) 30. The engine ECU 30 is mainly composed of a microcomputer, and controls the fuel injection amount and fuel injection timing of the fuel injection valve 19 and the ignition timing of the spark plug 31 according to the engine operating state detected by various sensors, and the engine ECU 30. Stop combustion (fuel injection and / or ignition) when a predetermined automatic stop condition (for example, fully closed accelerator, brake operation, idle operation, etc.) is established during operation and an engine stop request is generated. An idle stop that stops the engine rotation is executed, and during the engine stop due to this idle stop (during idle stop), the driver prepares for starting the vehicle (brake release, operation of the shift lever to the drive range, etc.) When a start operation (accelerator depression, etc.) is performed, or from the control system of the in-vehicle device When the dynamic request occurs, by energizing a predetermined automatic start condition is satisfied in the starter (not shown) to restart the engine 11 by cranking.

  Further, the engine ECU 30 functions as a target trajectory calculating means for calculating a target trajectory of the engine rotation behavior so that the engine rotation stops at the target stop crank angle by executing routines shown in FIGS. 14 and 15 described later. At the same time, it functions as stop control means for executing engine rotation stop control for controlling the actual engine rotation behavior to match the target trajectory when stopping engine rotation in response to an engine stop request. In the present embodiment, as engine rotation stop control, alternator F / B stop control is performed in which the load of the alternator 33 is feedback-controlled so that the actual engine rotation behavior matches the target trajectory. Furthermore, before the combustion of the engine 11 is stopped (during combustion), ignition F / B stop control for feedback control of the ignition timing may be executed so that the actual engine rotational behavior matches the target trajectory.

  During engine rotation stop control (alternator F / B stop control), if the engine rotation speed falls below the power generation limit rotation speed of the alternator 33, almost no load torque is generated in the alternator 33 (see FIG. 3). In such a rotation speed region, the engine rotation speed decreases and the engine rotation stops almost without being affected by the load torque of the alternator 33. Therefore, the stop according to the engine rotation speed when the predetermined reference timing is passed. The engine stops at the crank angle. Here, the reference timing is a timing at which the crank angle becomes a predetermined position (for example, TDC) in a rotational speed region where the engine rotational speed is equal to or lower than the power generation limit rotational speed of the alternator 33.

  Focusing on such characteristics, in this embodiment, the stop crank angle is set to the target stop crank by using the relationship between the engine speed at the reference timing (the engine speed when passing the reference timing) and the stop crank angle. The engine rotation speed at the reference timing that is an angle is obtained as the reference rotation speed, and the target trajectory is calculated by calculating the relationship between the crank angle and the target engine rotation speed up to the reference rotation speed at predetermined crank angle intervals (see FIG. (Not shown). This target trajectory is calculated for each predetermined crank angle Δθ (for example, every TDC) in the direction of turning the crank angle using the reference rotational speed as an initial value, for example, using an energy conservation law relational expression that considers loss torque (see FIG. 2).

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))} × Δθ
Here, Ne (i + 1) is the engine speed at the time (i + 1) before the predetermined crank angle Δθ before the current time (i), and Ne (i) is the engine speed at the current time (i). is there. J is the moment of inertia of the engine 11. Tloss is the total torque of the pumping loss and the friction loss, and Tref (Ne (i)) is the reference load torque of the alternator 33 at the current engine speed Ne (i).

  In the relational expression of the energy conservation law, “Tloss−Tref (Ne (i))” corresponds to a loss torque obtained by summing the pumping loss, the friction loss, and the reference load torque Tref (Ne (i)) of the alternator 33.

  In this embodiment, the reference load torque Tref (Ne (i)) of the alternator 33 is set to half (1/2) of the maximum load that can be controlled by the alternator 33 as shown in FIG. In this way, unlike the motor generator, the alternator 33 can virtually control the load torque of the alternator 33 in both the positive and negative directions even if there is a situation where the assist torque cannot be output (reference load Tref). It is possible to control the load torque of the alternator 33 using the following load torque as a virtually negative load torque and a load torque greater than the reference load Tref as a positive load torque), and following the engine rotation behavior to the target track Can be improved.

Note that the reference load torque Tref (Ne (i)) of the alternator 33 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 may be used. In short, an appropriate load smaller than the maximum controllable load of the alternator 33 and larger than 0 may be set as the reference load torque Tref (Ne (i)).
0 <Tref (Ne (i)) <maximum load

  FIG. 5A shows a comparative example in which the engine rotation stop control (alter F / B stop control) is performed with the reference load torque Tref (Ne (i)) = 0 set. In this comparative example, since the load torque of the alternator 33 can be controlled only in the positive direction, when 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 33 is set to an appropriate load smaller than the maximum load as in this embodiment, the alternator is virtually set as shown in FIG. Since the load torque 33 can be controlled in both positive and negative directions, as shown in FIG. 5B, even when the actual rotational behavior overshoots, the actual rotational behavior can be matched with the target trajectory.

  Furthermore, in this embodiment, as shown in FIG. 6, when calculating the target trajectory, the target trajectory corresponding to the reference load torque Tref (Ne (i)) of the alternator 33 is calculated, and the engine rotation stop control is executed. While calculating the reference load torque Tref (Ne (i)) according to the engine speed Ne (i), the target engine speed and the actual engine speed at the crank angle θ (i) at the present time (i) are calculated. And calculating the required load torque Talt by adding the reference load torque Tref (Ne (i)) to the feedback load torque (actually, the required load torque Talt). Is multiplied by the pulley ratio Ratio to convert it to the required shaft torque Talt.final).

  Thereafter, using the load torque characteristics of the alternator 33 shown in FIG. 7, the required load torque Talt (required shaft torque Talt.final) of the alternator 33 and the engine rotational speed Ne (or the engine rotational speed Ne) are multiplied by the pulley ratio Ratio. The power generation command value (duty duty) corresponding to the rotation speed Nalt) of the alternator 33 determined in this way is calculated. At this time, the power generation command value (duty duty) may be directly calculated from the required load torque Talt and the engine rotational speed Ne (or the rotational speed Nalt of the alternator 33), but the required load torque Talt and the engine rotational speed Ne ( Alternatively, the required field current (required excitation current) may be calculated from the rotation speed Nalt of the alternator 33, and the power generation command value (duty duty) may be calculated from the required field current.

  The load torque characteristics shown in FIG. 7 are characteristics when the output voltage of the alternator 33 is constant at a predetermined value (for example, 13.5 V), and similar characteristics are set for each output voltage. Based on this power generation command value (duty duty), the power generation control current (field current) of the alternator 33 is controlled to control the load torque of the alternator 33.

  Such control of the load torque of the alternator 33 is periodically executed at predetermined crank angle intervals until the engine rotational speed decreases below the power generation limit rotational speed Nelow (see FIG. 3) of the alternator 33. Engine rotation stop control (alternator F / B stop control) is performed to feedback control the load torque of the alternator 33 so that the behavior matches the target trajectory.

  During engine rotation stop control, the engine ECU 30 calculates a power generation command value at a predetermined crank angle cycle, and the power generation system ECU 36 (see FIG. 1) calculates this power generation command value at a predetermined time cycle by CAN (Controller Area Network) communication or the like. Send to. Further, the power supply system ECU 36 transmits the received power generation command value to the alternator 33 at a predetermined time period by LIN (Local Interconnect Network) communication or the like.

By the way, when the friction of the engine 11 changes due to the oil change of the engine 11 or a change with time, the reference rotational speed and the loss torque used for calculating the target trajectory also change. Therefore, in this embodiment, in order to ensure the calculation accuracy of the target trajectory, as shown in FIG. 8, engine rotation stop control that feedback-controls the load torque of the alternator 33 so that the actual engine rotational behavior matches the target trajectory. And the learning value of the reference rotational speed and the learning amount of the loss torque are calculated based on the actual engine rotational behavior when the engine rotational speed decreases below the power generation limit rotational speed of the alternator 33 by the engine rotational stop control. Thus, the deviation amount between the reference rotation speed and the loss torque is learned, and the target trajectory is calculated using the learning value of the reference rotation speed and the learned value of the loss torque obtained from the learned value of the loss torque deviation amount.
Hereinafter, the learning method of the loss torque and the reference rotation speed will be described.

[Loss torque learning method]
As shown in FIG. 9, when the engine rotation speed is reduced below the power generation limit rotation speed of the alternator 33 by the engine rotation stop control, after the actual engine rotation speed Ne1 at the reference timing (for example, the first TDC) is detected, a predetermined value is obtained. The actual engine rotational speed Ne2 at the detection timing (for example, TDC next to the reference timing) is detected. The relationship between the actual engine rotational speed Ne1 at the reference timing, the actual engine rotational speed Ne2 at the detection timing, and the loss torque Tloss can be expressed by the following equation (1) using an energy conservation law.

The loss torque Tloss is calculated by the above equation (1).
Further, the stored data of the loss torque learning value Tloss stored in the rewritable nonvolatile memory (the rewritable memory holding the stored data even when the engine ECU 30 is powered off) such as the backup RAM 37 of the engine ECU 30 is used as the previous loss torque. Is read as a learned value Tloss (i-1).

Thereafter, using the loss torque Tloss calculated this time and the learned value Tloss (i-1) of the previous loss torque, the learned value ΔTloss (i) of the current loss torque deviation amount is calculated by the following equation.
ΔTloss (i) = Tloss-Tloss (i-1)

  Thereafter, the stored data of the loss torque deviation amount learning value ΔTloss stored in the nonvolatile memory such as the backup RAM 37 is normally updated with the calculated loss torque deviation amount learning value ΔTloss (i) calculated this time. To do.

Thereafter, when the loss torque learning value Tloss (i) is calculated, the previous loss torque learning value Tloss (i-1), the learning gain Gain, and the loss torque stored in the non-volatile memory such as the backup RAM 37 or the like. The learning value Tloss (i) of the current loss torque is calculated by the following equation using the learning value ΔTloss of the deviation amount.
Tloss (i) = Tloss (i-1) + Gain × ΔTloss

  Thereafter, the stored data of the learned value Tloss of the loss torque stored in the nonvolatile memory such as the backup RAM 37 is updated with the calculated data of the learned value Tloss (i) of the lost torque calculated this time.

[How to learn the reference rotation speed]
As shown in FIG. 10, when the engine rotation speed is reduced below the power generation limit rotation speed of the alternator 33 by the engine rotation stop control, after the actual engine rotation speed Ne1 at the reference timing (for example, the first TDC) is detected, the engine The stop crank angle X at which the rotation has stopped is detected, and the stop position deviation amount ΔX is calculated by the following equation using the stop crank angle X and the target stop crank angle Xtgt.
ΔX = X−Xtgt

  Thereafter, the energy deviation amount Δe corresponding to the stop position deviation amount ΔX is calculated with reference to the map of the energy deviation amount Δe shown in FIG. Here, the energy deviation amount Δe is an energy deviation amount with respect to a reference point (a point at which the reference rotational speed is reached at the reference timing), and a map of the energy deviation amount Δe is created in advance based on test data, design data, and the like. It is stored in the ROM of the engine ECU 30.

Further, the storage data of the reference rotation speed learning value Netgt stored in the nonvolatile memory such as the backup RAM 37 is read as the previous reference rotation speed learning value Netgt (i−1).
Thereafter, using the learning value Netgt (i-1) of the previous reference rotational speed, the energy deviation amount Δe, and the correction coefficient α for the loss torque learning, the reference timing at which the engine rotation stops at the stop crank angle X is estimated. The engine speed Neest is calculated by the following equation (2).

Thereafter, using the learning value Netgt (i-1) of the previous reference rotational speed, the learning gain Gain, the actual engine rotational speed Ne1 and the estimated engine rotational speed Neest at the reference timing, The learning value Netgt (i) is calculated.
Netgt (i) = Netgt (i-1) + Gain x (Ne1-Neest)

  Thereafter, the stored data of the reference rotational speed learning value Netgt stored in the non-volatile memory such as the backup RAM 37 is normally updated with the calculated data of the reference rotational speed learned value Netgt (i) calculated this time.

  However, if the actual engine rotation behavior changes temporarily when the engine rotation stop control is executed due to some external load, etc., the reference rotation speed calculated based on the actual engine rotation behavior is affected. Since the learning value (calculated data) and the learning value (calculated data) of the loss torque deviation amount change temporarily, the reference rotational speed and the loss torque deviation amount (hereinafter collectively referred to as “target trajectory information”) are erroneously learned. The learning accuracy of the target trajectory information may be reduced.

  As a countermeasure, in this embodiment, when the learning data of the target trajectory information calculated this time deviates more than a predetermined value from the stored data of the target trajectory information (reference rotational speed or loss torque deviation amount). In addition, by monitoring the calculation data of the learning value of the target trajectory information for a plurality of times thereafter, whether the learning value (calculation data) of the target trajectory information has changed steadily due to changes in the friction of the engine 11, Alternatively, it is determined whether the learning value (calculated data) of the target trajectory information has temporarily changed due to some external load or the like, and it is determined whether the update permission condition is satisfied.

  Specifically, as shown in FIG. 12, the calculated data of the learning value of the current target trajectory information is a predetermined value with respect to the stored data of the learning value of the previous target trajectory information (reference rotational speed or loss torque deviation amount). If there is a divergence as described above (see A and B in FIG. 12), it is determined whether or not the update permission condition is satisfied depending on whether or not the state continues for a predetermined number of times (for example, four times).

  When the learning value calculation data of the current target trajectory information has deviated more than a predetermined value from the stored data of the previous learning value of the target trajectory information for a predetermined number of times, the target trajectory is caused by some external load or the like. It is determined that the learning value (calculated data) of the information may have changed temporarily, it is determined that the update permission condition is not satisfied, and the last time without updating the learning data of the target trajectory information Hold to value.

  On the other hand, when the state in which the calculation data of the learning value of the current target trajectory information deviates by a predetermined value or more from the stored data of the learning value of the previous target trajectory information continues for a predetermined number of times, the friction of the engine 11 occurs at that time t1. And the learning value (calculation data) of the target trajectory information is steadily changed, it is determined that the update permission condition is satisfied, and the storage data of the learning value of the target trajectory information is updated.

  At that time, as shown in FIG. 13, at a time t1 when it is determined that the update permission condition is satisfied, the learning gain is made larger than the normal value (value before the update permission condition is satisfied). As a result, the learning trajectory value (stored data) of the target trajectory information is increased (the update amount per time is increased), and the learning value (stored data) of the target trajectory information is quickly updated to the target trajectory information. Approach the true value. Further, it is determined whether or not the learning value (stored data) of the target trajectory information is in a converged state (substantially constant) after the learning gain is made larger than the normal value, and the learning value (stored data) of the target trajectory information is converged. At the time t2 when it is determined that the state has been reached, it is determined that the learning value (stored data) of the target trajectory information has converged near the true value of the target trajectory information, and the learning gain is returned to the normal value.

  The engine rotation stop control and the related control of the present embodiment described above are executed by the engine ECU 30 according to the routines shown in FIGS. The processing contents of these routines will be described below.

[Target trajectory calculation routine]
The target trajectory calculation routine shown in FIG. 14 is repeatedly executed at a predetermined cycle (predetermined crank angle cycle) while the engine ECU 30 is powered on. When this routine is started, first, in step 101, it is determined whether or not the target trajectory calculation completion flag is set to “0”, which means that the target trajectory is not calculated. If it is set to “1” indicating completion of trajectory calculation, this routine is terminated without performing the subsequent processing.

  On the other hand, if it is determined in step 101 that the target trajectory calculation completion flag = 0 (before the target trajectory is calculated), the process proceeds to step 102 where the loss torque Tloss and the reference load torque Tref (Ne (i)) of the alternator 33 are set. Then, the square of the target engine speed Ne (i + 1) at the next time point (i + 1) is calculated using the relational expression of the energy conservation law expressed by the following formula.

Ne (i + 1) ² = Ne (i) ² + 2 / J × {Tloss−Tref (Ne (i))} × Δθ
Here, J is the moment of inertia of the engine 11, and Tloss is the loss torque obtained by adding the pumping loss and the friction loss. In the relational expression of the energy conservation law, “Tloss−Tref (Ne (i))” corresponds to a loss torque obtained by summing the pumping loss, the friction loss, and the reference load torque Tref (Ne (i)) of the alternator 33.

  Initial values are i = 0, θ (0) = crank angle at reference timing, and Ne (0) = reference rotation speed. This reference rotational speed Ne (0) is the engine rotational speed at the reference timing at which the stop crank angle becomes the target stop crank angle. The target trajectory is calculated every predetermined crank angle Δθ (for example, every TDC) in the direction of turning the crank angle with the reference rotational speed Ne (0) as an initial value.

  After this, the routine proceeds to step 103, where it is determined whether or not the square of the target engine speed Ne (i + 1) exceeds the square of the maximum engine speed Nemax at which the engine rotation stop control can be executed. If it does not exceed the square of Nemax, the process proceeds to step 104, and the target trajectory calculation completion flag is maintained at “0” (reset).

  Thereafter, the routine proceeds to step 106, where the square root of the square of the target engine speed Ne (i + 1) is calculated to obtain the target engine speed Ne (i + 1), which is obtained as a target trajectory table (not shown). To complete this routine. In order to reduce the calculation load on the engine ECU 30, the square of the engine speed may be assigned to the table as it is. The target trajectory table is stored in the memory of the engine ECU 30.

  The above processing is repeated to calculate the square of the target engine speed Ne (i + 1) for each predetermined crank angle (for example, every TDC) in the direction of turning the crank angle with the reference rotational speed Ne (0) as an initial value. Then, the process of assigning the target engine speed Ne (i + 1) to the target trajectory table is repeated. When it is determined in step 103 that the square of the target engine rotational speed Ne (i + 1) exceeds the square of the maximum engine rotational speed Nemax that can execute the engine rotational stop control, the process proceeds to step 105, and the target The trajectory calculation completion flag is set to “1” which means the completion of the target trajectory calculation, and the routine proceeds to step 106 where the square root of the last target engine speed Ne (i + 1) is calculated to calculate the target engine speed Ne. (i + 1) is obtained, assigned to the target trajectory table, and this routine is terminated.

[Alter F / B stop control routine]
The alternator F / B stop control routine shown in FIG. 15 is repeatedly executed at a predetermined cycle (predetermined crank angle cycle) while the engine ECU 30 is powered on. When this routine is started, first, at step 201, it is determined whether or not an engine stop request (idle stop request) has occurred. If no engine stop request has occurred, the subsequent processing is performed. This routine ends.

  Thereafter, when it is determined in step 201 that an engine stop request has been generated, the routine proceeds to step 202, where the current crank angle θ and engine rotational speed Ne are calculated. Thereafter, the process proceeds to step 203, where it is determined whether or not the current crank angle θ is the load torque control timing of the alternator 33 (for example, TDC). This routine is terminated without performing.

  If it is determined in step 203 that the current crank angle θ is the timing for controlling the load torque of the alternator 33, the routine proceeds to step 204, where the current engine speed Ne is the maximum engine speed at which engine rotation stop control can be executed. It is determined whether the speed is lower than the speed Nemax. If the current engine speed Ne is equal to or higher than the maximum engine speed Nemax, this routine is terminated without performing the subsequent processing.

Thereafter, if it is determined in step 204 that the current engine speed Ne is lower than the maximum engine speed Nemax, the routine proceeds to step 205, where it is determined whether or not the engine 11 is in combustion. If it is determined in step 205 that the engine 11 is still burning immediately after the engine stop request is generated, the process proceeds to step 206, where the required load torque Talt of the alternator 33 when starting the engine rotation stop control is initialized. Set to a value (for example, reference load torque Tref (Ne)).
Talt = Tref (Ne)

  Thereafter, if it is determined in step 205 that the combustion of the engine 11 has stopped, the process proceeds to step 207, where the target engine speed Netg corresponding to the current control timing is obtained with reference to the target trajectory table. .

Thereafter, the routine proceeds to step 208, where the energy difference ΔE between the current actual engine speed Ne and the target engine speed Netg is calculated by the following equation.
ΔE = J / 2 × (Ne ² −Netg ² )
Here, J is the moment of inertia of the engine 11.

Thereafter, the routine proceeds to step 209, where the required load torque Talt is calculated by the following equation using the energy difference ΔE and the reference load torque Tref (Ne) of the alternator 33.
Talt = K × ΔE / Δθ + Tref (Ne)

Here, “K × ΔE / Δθ” is a feedback load torque, K is a feedback gain, and Δθ is a crank angle change amount.
Thereafter, the process proceeds to step 210, where the required load torque Talt is multiplied by the pulley ratio Ratio to convert it to the required shaft torque Talt.final of the alternator 33.

  Thereafter, the process proceeds to step 211, and after detecting the battery voltage, the process proceeds to step 212, and the load corresponding to the current battery voltage is selected from a plurality of load torque characteristic maps (see FIG. 7) created for each battery voltage. A torque characteristic map is selected, and a power generation command value (duty duty) corresponding to the current required load torque Talt (required shaft torque Talt.final) and the engine rotational speed Ne (or the rotational speed Nalt of the alternator 33) is calculated.

  By controlling the power generation control current (field current) of the alternator 33 on the basis of this power generation command value (duty duty) and controlling the load torque of the alternator 33, the alternator 33 is controlled so as to match the actual engine rotational behavior with the target trajectory. Alter F / B stop control for feedback control of load torque is performed.

[Learning control routine]
The learning control routine shown in FIG. 16 is repeatedly executed at a predetermined cycle (predetermined crank angle cycle) while the power of the engine ECU 30 is turned on, and serves as learning means in the claims. When this routine is started, first, at step 301, it is determined whether or not an engine stop request (idle stop request) has occurred. If no engine stop request has occurred, the subsequent processing is not performed. This routine ends.

  Thereafter, when it is determined in step 301 that an engine stop request has occurred, the process proceeds to step 302 to determine whether or not the learning execution condition is satisfied, for example, whether the engine 11 is in a warm-up state (cooling water temperature is predetermined). If the learning execution condition is not satisfied, this routine is terminated without performing the subsequent processing.

  On the other hand, if it is determined in the above step 302 that the learning execution condition is satisfied, the process proceeds to step 303, where a loss torque learning routine of FIG. Is learned based on the actual engine rotation behavior when the engine speed decreases below the power generation limit rotation speed of the alternator 33. Further, in the next step 304, a reference engine speed learning routine of FIG. 18 described later is executed, so that the actual engine speed behavior when the engine speed is reduced below the power generation limit speed of the alternator 33 by the engine speed stop control. The reference rotational speed is learned based on

[Loss torque learning routine]
The loss torque learning routine shown in FIG. 17 is a subroutine executed in step 303 of the learning control routine shown in FIG. When this routine is started, first, in step 401, when the engine rotation speed is reduced below the power generation limit rotation speed of the alternator 33 by the engine rotation stop control, the current crank angle is changed to the reference timing (for example, the first TDC). When it is determined that it is the reference timing, the routine proceeds to step 402, where the actual engine speed Ne1 at the reference timing is detected.

  Thereafter, the process proceeds to step 403, where it is determined whether or not the current crank angle is at the detection timing (for example, the TDC next to the reference timing). The actual engine rotational speed Ne2 at the timing is detected.

  Thereafter, the process proceeds to step 405, and the loss torque Tloss is calculated by the above equation (1) using the actual engine speed Ne1 at the reference timing and the actual engine speed Ne2 at the detection timing. Then, the process proceeds to step 406 and the engine ECU 30 The stored data of the loss torque learning value Tloss stored in the non-volatile memory such as the backup RAM 37 is read as the previous loss torque learning value Tloss (i-1).

Thereafter, the process proceeds to step 407, and the learning value ΔTloss (i) of the current loss torque deviation amount is calculated by the following equation using the currently calculated loss torque Tloss and the previous loss torque learning value Tloss (i-1). .
ΔTloss (i) = Tloss-Tloss (i-1)

  Thereafter, the process proceeds to step 408, and a learning value update determination routine of FIG. 19 described later is executed, so that the stored data of the learning value ΔTloss of the loss amount of loss torque stored in the nonvolatile memory such as the backup RAM 37 is stored this time. The calculated loss torque deviation amount is updated with the calculation data of the learned value ΔTloss (i) or held at the previous value.

Thereafter, when the loss torque learning value Tloss (i) is calculated, the previous loss torque learning value Tloss (i-1), the learning gain Gain, and the loss torque stored in the non-volatile memory such as the backup RAM 37 or the like. The learning value Tloss (i) of the current loss torque is calculated by the following equation using the learning value ΔTloss of the deviation amount.
Tloss (i) = Tloss (i-1) + Gain × ΔTloss

  Thereafter, the stored data of the learned value Tloss of the loss torque stored in the nonvolatile memory such as the backup RAM 37 is updated with the calculated data of the learned value Tloss (i) of the lost torque calculated this time.

[Reference rotational speed learning routine]
The reference rotational speed learning routine shown in FIG. 18 is a subroutine executed in step 304 of the learning control routine of FIG. When this routine is started, first, at step 501, when the engine rotation speed is reduced below the power generation limit rotation speed of the alternator 33 by the engine rotation stop control, the current crank angle is set to the reference timing (for example, the first TDC). When it is determined that it is the reference timing, the process proceeds to step 502, and the actual engine speed Ne1 at the reference timing is detected.

  Thereafter, the process proceeds to step 503 to determine whether or not the engine rotation has stopped. When it is determined that the engine rotation has stopped, the process proceeds to step 504 to detect the stop crank angle X at which the engine rotation has stopped.

Thereafter, the process proceeds to step 505, and the stop position deviation amount ΔX is calculated by the following equation using the stop crank angle X and the target stop crank angle Xtgt.
ΔX = X−Xtgt

  Thereafter, the process proceeds to step 506, the energy shift amount Δe corresponding to the stop position shift amount ΔX is calculated with reference to the map of the energy shift amount Δe shown in FIG. 11, and then the process proceeds to step 507, where the backup RAM 37 of the engine ECU 30 is stored. The reference rotation speed learning value Netgt stored in the non-volatile memory is read as the previous reference rotation speed learning value Netgt (i-1).

  Thereafter, the process proceeds to step 508, where the engine rotation is stopped at the stop crank angle X using the learning value Netgt (i-1) of the previous reference rotational speed, the energy deviation amount Δe, and the correction coefficient α for the loss torque learning. The estimated engine speed Neest of the reference timing to be calculated is calculated by the equation (2).

Thereafter, the process proceeds to step 509, where the previous reference rotational speed learning value Netgt (i-1), the learning gain Gain, the actual engine rotational speed Ne1 at the reference timing, and the estimated engine rotational speed Neest are used as follows: The learning value Netgt (i) of the current reference rotational speed is calculated.
Netgt (i) = Netgt (i-1) + Gain x (Ne1-Neest)

  Thereafter, the process proceeds to step 510, and a learning value update determination routine shown in FIG. 19 described later is executed, whereby the stored data of the learning value Netgt of the reference rotation speed stored in the nonvolatile memory such as the backup RAM 37 is calculated this time. It is updated with the calculated data of the learned value Netgt (i) of the reference rotational speed or held (held) at the previous value.

[Learning value update determination routine]
The learning value update determination routine shown in FIG. 19 is a subroutine executed in step 408 of the loss torque learning routine of FIG. 17 and step 510 of the reference rotational speed learning routine of FIG. In practice, the learning value update determination routine for the loss amount of the loss torque is executed in step 408 in FIG. 17, and the learning value update determination routine for the reference rotational speed is executed in step 510 in FIG. Now, the amount of deviation between the reference rotational speed and the loss torque will be collectively referred to as “target trajectory information”.

  When this routine is started, first, in step 601, it is determined whether or not the gain change flag is set to “1”. If it is determined that the gain change flag is set to “0”, Proceeding to step 602, the difference (absolute value) between the calculated data of the current target trajectory information (reference rotational speed or loss torque deviation amount) and the stored data of the previous target trajectory information learned value is a predetermined value. It is determined whether or not the calculation data of the learning value of the current target trajectory information has deviated by a predetermined value or more from the stored data of the learning value of the previous target trajectory information depending on whether or not it is K1 or more.

  If it is determined in step 602 that the difference between the calculated data of the learned value of the target trajectory information this time and the stored data of the learned value of the previous target trajectory information is smaller than the predetermined value K1, the process proceeds to step 603. Then, the storage data of the learning value of the target trajectory information stored in the nonvolatile memory such as the backup RAM 37 is updated with the calculation data of the learning value of the target trajectory information calculated this time.

  On the other hand, in the step 602, the difference between the learning value calculation data of the current target trajectory information and the stored data of the learning value of the previous target trajectory information is equal to or greater than a predetermined value K1 (the learning value of the previous target trajectory information If it is determined that the learning data of the current target trajectory information has deviated by a predetermined value or more from the stored data, the process proceeds to step 604, where the learning data of the current target trajectory information is calculated and the previous data The difference between the learning data of the target trajectory information and the stored data of the target trajectory information is greater than or equal to the predetermined value K1. Whether or not the update permission condition is satisfied is determined based on whether or not the state has been continued a predetermined number of times (for example, four times).

  If it is determined in step 604 that a state in which the difference between the learning data for the current target trajectory information and the stored data for the previous target trajectory information is equal to or greater than the predetermined value K1 has not continued a predetermined number of times. Is determined that there is a possibility that the learning value (calculation data) of the target trajectory information has temporarily changed due to some external load or the like, it is determined that the update permission condition is not satisfied, and the process proceeds to Step 605, The stored data of the learning value of the target trajectory information stored in the nonvolatile memory such as the backup RAM 37 is held (held) at the previous value without being updated.

  On the other hand, when it is determined in the above step 604 that a state in which the difference between the learning value calculation data of the current target trajectory information and the stored data of the previous learning value of the target trajectory information is equal to or greater than the predetermined value K1 has continued a predetermined number of times. Is determined that the learning permission value (calculation data) of the target trajectory information has been constantly changed due to a change in the friction of the engine 11 and the like, and the update permission condition is satisfied.

  In this case, first, the process proceeds to step 606, the learning gain is set larger than the normal value (the value before the update permission condition is satisfied), the gain change flag is set to “1”, the process proceeds to step 607, and the backup RAM 37 The learning data of the target trajectory information stored in the non-volatile memory is updated with the calculated learning data of the target trajectory information calculated this time. As a result, the learning trajectory value (stored data) of the target trajectory information is increased (the update amount per time is increased), and the learning value (stored data) of the target trajectory information is quickly updated to the target trajectory information. Approach the true value.

  Thereafter, the process proceeds to step 608, and whether or not the difference (absolute value) between the learning data of the current learning value of the target trajectory information and the storage data of the learning value of the previous target trajectory information is equal to or smaller than a predetermined value K2. Thus, it is determined whether or not the learning value (stored data) of the target trajectory information has converged (substantially constant).

  If it is determined in step 608 that the learning value (stored data) of the target trajectory information is not in a converged state, this routine is terminated while the learning gain is kept larger than the normal value. In this case, it is determined in step 601 that the gain change flag is set to “1”, and the process proceeds to step 607 where the stored data of the learning value of the target trajectory information is calculated this time. After updating with data, the process of determining whether or not the learning value (stored data) of the target trajectory information is in a converged state is repeated (steps 607 and 608).

  Thereafter, when it is determined in step 608 that the learning value (memory data) of the target trajectory information has converged, the learning value (memory data) of the target trajectory information has converged near the true value of the target trajectory information. Then, the process proceeds to step 609, where the learning gain is returned to the normal value, and the gain change flag is reset to “0”.

  In the present embodiment described above, when the calculation data of the learning value of the current target trajectory information deviates by a predetermined value or more from the stored data of the learning value of the previous target trajectory information (reference rotational speed or loss torque deviation amount). If the condition has not continued for a predetermined number of times, it is determined that the learning value (calculation data) of the target trajectory information may have temporarily changed due to some external load or the like, and the update permission condition is satisfied. The stored data of the learning value of the target trajectory information is held (held) at the previous value without being updated. On the other hand, when the state in which the calculation data of the learning value of the current target trajectory information deviates by a predetermined value or more from the storage data of the learning value of the previous target trajectory information continues for a predetermined number of times, the friction of the engine 11 changes. It is determined that the learning value (calculated data) of the target trajectory information has been constantly changed, it is determined that the update permission condition has been satisfied, and the stored data of the learning value of the target trajectory information is updated. Thereby, erroneous learning of the target trajectory information can be prevented, the learning accuracy of the target trajectory information can be improved, and the calculation accuracy of the target trajectory can be improved.

  In the above-described embodiment, the update permission condition is satisfied when a state in which the calculation data of the learning value of the current target trajectory information deviates more than a predetermined value from the stored data of the learning value of the previous target trajectory information continues a predetermined number of times. However, the present invention is not limited to this. For example, the number of times that the calculation data of the learning value of the current target trajectory information deviates more than a predetermined value from the stored data of the learning value of the previous target trajectory information. It may be determined that the update permission condition is satisfied when the value exceeds a predetermined value within a predetermined period. In short, the current target trajectory information is stored with respect to the stored data of the learning value of the previous target trajectory information. When the learning value calculation data deviates by a predetermined value or more, it may be determined whether the update permission condition is satisfied based on the learning value calculation data of the target trajectory information for a plurality of times thereafter.

  Further, in the above embodiment, the present invention is applied to a system that learns both the reference rotational speed and the loss amount of the loss torque as target trajectory information (information used for calculating the target trajectory). The present invention is applied to a system that learns only one of the reference rotational speed and loss torque deviation amount as the target trajectory information, or other information other than the reference rotation speed and loss torque deviation amount as the target trajectory information ( For example, the present invention may be applied to a system that learns pumping loss, friction loss, and the like.

  Further, the present invention is not limited to the intake port injection type engine as shown in FIG. 1, and includes an in-cylinder injection type engine, and a fuel injection valve for intake port injection and a fuel injection valve for in-cylinder injection. It can also be applied to dual-injection engines.

  Further, the scope of application of the present invention is not limited to a general vehicle having only an engine as a power source of the vehicle, and the present invention may be applied to a hybrid vehicle having an engine and a motor as the power source of the vehicle. .

  DESCRIPTION OF SYMBOLS 11 ... Engine (internal combustion engine), 13 ... Intake pipe, 14 ... Throttle valve, 18 ... Intake pipe pressure sensor, 19 ... Fuel injection valve, 21 ... Exhaust pipe, 30 ... Engine ECU (target track calculation means, stop control means, Learning means), 33 ... alternator (generator), 37 ... backup RAM

Claims (5)

Target trajectory calculating means for calculating the target trajectory of the engine rotation behavior so that the engine rotation stops at the target stop crank angle, and adjusting the actual engine rotation behavior to the target trajectory when stopping the engine rotation in response to the engine stop request An engine rotation stop control device comprising stop control means for executing engine rotation stop control for controlling the load of the generator as described above,
A learning value of information used for calculating the target trajectory (hereinafter referred to as “target trajectory information”) is calculated based on actual engine rotational behavior when the engine rotation stop control is executed, and the learning value of the target trajectory information is calculated. Learning means for learning the target trajectory information by updating storage data of learning values of the target trajectory information based on the calculated data;
When the learning value calculation data of the target trajectory information calculated this time deviates by a predetermined value or more from the stored data of the learning value of the target trajectory information, the learning means learns the target trajectory information multiple times thereafter. It is determined whether or not an update permission condition is satisfied based on value calculation data, and when it is determined that the update permission condition is satisfied, the stored data of the learning value of the target trajectory information is updated. Engine rotation stop control device.   The learning means updates the update data when the learning value calculation data of the target trajectory information calculated this time differs from the storage data of the learning value of the target trajectory information by a predetermined value or more when the state continues for a predetermined number of times. The engine rotation stop control device according to claim 1, wherein it is determined that a permission condition is satisfied.   The learning means is characterized in that when the update permission condition is satisfied, a gain for learning the target trajectory information (hereinafter referred to as “learning gain”) is larger than a value before the update permission condition is satisfied. The engine rotation stop control device according to claim 1 or 2.   The learning means sets the learning gain before the update permission condition is satisfied when the learning value of the target trajectory information is in a converged state after the learning gain is made larger than a value before the update permission condition is satisfied. The engine rotation stop control device according to claim 3, wherein the engine rotation stop control device returns to the value of The target trajectory calculating means calculates the target trajectory based on the engine rotational speed at a reference timing (hereinafter referred to as “reference rotational speed”) at which the stop crank angle of the engine rotation becomes the target stop crank angle, and the loss torque of the engine. ,
5. The engine rotation stop control device according to claim 1, wherein the learning unit learns a deviation amount between the reference rotation speed and the loss torque as the target trajectory information.

Images