JP6187057B2 - Control device for hybrid vehicle - Google Patents

Description

  The present invention relates to a control apparatus for a hybrid vehicle that uses an engine starter as an engine starter and cranks the engine via a first clutch when there is an engine start request in EV mode.

  Conventionally, only when the four learning permission determination conditions (MG rotational speed stability condition, MG torque stability condition, low gear position condition, CL2 transmission torque stability condition) are satisfied, An apparatus that performs learning correction of a torque-stroke map of one clutch CL1 is known (see, for example, Patent Document 1).

JP 2010-30428 A

  However, in the conventional apparatus, learning correction of the torque-stroke map of the first clutch CL1 is not permitted unless any one of the four stability conditions is satisfied. Therefore, when the friction coefficient of the first clutch changes, it takes a long time to obtain a learning correction value that reflects the friction coefficient change. For this reason, when the transmission torque of the first clutch that is slip-engaged deviates from a desired torque when the engine is started, a plurality of engine crankings due to torque deviation are obtained until a learning correction value that reflects a change in the friction coefficient is obtained. There was a problem that I had to experience it once.

  The present invention has been made paying attention to the above problem, and provides a control device for a hybrid vehicle capable of performing a learning correction reflecting the decrease in the clutch friction coefficient with good response when the friction coefficient of the first clutch is decreased. The purpose is to do.

To achieve the above object, the present invention comprises an engine, a first clutch, a motor, and drive wheels in a drive system, and when there is an engine start request in an EV mode using the motor as a drive source. The engine is used as an engine starter, and the first clutch is engaged with a predetermined slip to perform cranking control of the engine.
The hybrid vehicle control device includes engine start control means and learning correction control means.
The engine start control means starts the cranking control when the engine start request is made, and measures an elapsed time from the start of the cranking control. When the elapsed time exceeds the start request time, at least the engine starts. Until the explosion is completed, the engine is started by performing backup control for raising the command torque to the first clutch.
The learning correction control means, when experiencing engine start intervening with the backup control, performs learning correction for increasing the command torque to the first clutch from the next engine start. In the learning correction, the learning correction value is set to a larger value as the instruction torque increase time in the backup control for increasing the instruction torque to the first clutch is longer.

Therefore, on the engine start control side, the elapsed time from the start of cranking control is measured, and when the elapsed time exceeds the start request time, backup control is performed to raise the command torque to the first clutch. On the learning correction control side, when the engine start in which the backup control intervenes is experienced, learning correction for increasing the command torque to the first clutch is performed from the next engine start.
That is, since engine start has a time request (start request time) that it is desired to start by a certain time at the latest, backup control is performed when the elapsed time from the start of cranking control exceeds the start request time. The cranking control exceeds the start request time because the actual transmission torque of the first clutch is lower than the command torque. Therefore, it can be estimated that the friction coefficient of the first clutch is reduced by the backup control intervening during the cranking control.
On the other hand, when the engine start in which the backup control intervenes is experienced, the estimation that the friction coefficient of the first clutch is reduced is immediately reflected, and learning correction for increasing the command torque to the first clutch is performed. For this reason, from the time of the next engine start, it is possible to start the engine whose elapsed time from the start of cranking control is equal to or less than the start request time.
As a result, when the friction coefficient of the first clutch decreases, learning correction reflecting the decrease of the clutch friction coefficient can be performed with good response.
In addition, in the learning correction, the longer the instruction torque increase time in the backup control for starting up the instruction torque to the first clutch is set to a learning correction value with a larger value. Thus, it is possible to perform appropriate and accurate learning correction according to the degree of decrease in the friction coefficient μ of the first clutch.

1 is an overall system diagram illustrating an FF hybrid vehicle to which a control device according to a first embodiment is applied. 3 is a flowchart illustrating a flow of an engine start control process executed by the hybrid control module according to the first embodiment. 6 is a flowchart illustrating a flow of CL1μ learning correction control processing executed by the hybrid control module according to the first embodiment. FIG. 5 is a schematic diagram illustrating a hybrid drive system for explaining calculation of a CL1 estimated torque in a CL1μ learning correction control process in the first embodiment. 6 is a time chart showing characteristics of CL1 command torque, CL1 estimated torque, motor rotation, and engine rotation when the FF start torque is increased in backup hydraulic pressure. Power generation request at P range shift when learning CL1μ with FF torque up operation detection ・ CL1 status ・ First-explosion completion flag ・ CL1 command torque stability judgment ・ CL1 actual torque stability judgment ・ CL1 estimated temperature ・ CL1 estimated temperature judgment ・ CL1μ learning permission Flag • CL1μ learning value acquisition permission • Normal torque magnitude determination NE threshold detection • Learning value calculation according to normal torque magnitude determination • Forced torque small determination • Learning value calculation according to forced torque small determination • μ learning value (offset) • μ learning consideration μ This is a time chart showing characteristics of the number of start times, CL1 command torque, CL1 estimated torque, target motor rotation, actual motor rotation, and actual engine rotation.

  Hereinafter, the best mode for realizing a control device for a hybrid vehicle of the present invention will be described based on a first embodiment shown in the drawings.

First, the configuration will be described.
The configuration of the FF hybrid vehicle (an example of a hybrid vehicle) to which the control device according to the first embodiment is applied is divided into “entire system configuration”, “detailed configuration of engine start control”, and “detailed configuration of CL1μ learning correction control”. explain.

[Overall system configuration]
FIG. 1 shows an overall system of an FF hybrid vehicle. Hereinafter, the overall system configuration of the FF hybrid vehicle will be described with reference to FIG.

  As shown in FIG. 1, the drive system of the FF hybrid vehicle includes a starter motor 1, a horizontally mounted engine 2, a first clutch 3 (abbreviated as “CL1”), a motor / generator 4, and a second clutch 5 ( An abbreviation “CL2”) and a belt-type continuously variable transmission 6 (abbreviation “CVT”). The output shaft of the belt type continuously variable transmission 6 is drivingly connected to the left and right front wheels 10R and 10L via a final reduction gear train 7, a differential gear 8, and left and right drive shafts 9R and 9L. The left and right rear wheels 11R and 11L are driven wheels.

  The starter motor 1 is a cranking motor that has a gear that meshes with an engine starting gear provided on a crankshaft of the horizontally mounted engine 2 and that rotates the crankshaft when the engine is started.

  The horizontal engine 2 is an engine disposed in the front room with the crankshaft direction as the vehicle width direction, and includes an electric water pump 12 and a crankshaft rotation sensor 13 that detects reverse rotation of the horizontal engine 2.

  The first clutch 3 is a normally open dry multi-plate friction clutch that is hydraulically interposed between the horizontal engine 2 and the motor / generator 4, and is fully engaged / slip engaged / released by the first clutch oil pressure. Is controlled. A stroke sensor is not provided.

  The motor / generator 4 is a three-phase AC permanent magnet type synchronous motor connected to the transverse engine 2 through a first clutch 3. The motor / generator 4 uses a high-power battery 21 described later as a power source, and an inverter 26 that converts direct current into three-phase alternating current during power running and converts three-phase alternating current into direct current during regeneration is connected to the stator coil. Connected through.

  The second clutch 5 is a wet-type multi-plate friction clutch by hydraulic operation that is interposed between the motor / generator 4 and the left and right front wheels 10R and 10L that are driving wheels. Slip fastening / release is controlled. The second clutch 5 of the first embodiment uses the forward clutch 5a and the reverse brake 5b provided in the forward / reverse switching mechanism of the belt-type continuously variable transmission 6 using planetary gears. That is, the forward clutch 5 a is the second clutch 5 during forward travel, and the reverse brake 5 b is the second clutch 5 during reverse travel.

  The belt-type continuously variable transmission 6 is a transmission that obtains a continuously variable transmission ratio by changing the belt winding diameter by the transmission hydraulic pressure to the primary oil chamber and the secondary oil chamber. The belt type continuously variable transmission 6 includes a main oil pump 14 (mechanical drive), a sub oil pump 15 (motor drive), and a line pressure PL generated by adjusting pump discharge pressure from the main oil pump 14. And a control valve unit (not shown) that generates the first and second clutch hydraulic pressures and the transmission hydraulic pressure with the pressure as the original pressure. The main oil pump 14 is rotationally driven by a motor shaft (= transmission input shaft) of the motor / generator 4. The sub oil pump 15 is mainly used as an auxiliary pump for producing lubricating cooling oil.

  The first clutch 3, the motor / generator 4 and the second clutch 5 constitute a one-motor / two-clutch drive system, and there are “EV mode” and “HEV mode” as main drive modes by this drive system. The “EV mode” is an electric vehicle mode in which the first clutch 3 is disengaged and the second clutch 5 is engaged and only the motor / generator 4 is used as a drive source. Driving in the “EV mode” is referred to as “EV driving”. . The “HEV mode” is a hybrid vehicle mode in which both the clutches 3 and 5 are engaged and the horizontal engine 2 and the motor / generator 4 are used as driving sources, and traveling in the “HEV mode” is referred to as “HEV traveling”.

  The regenerative cooperative brake unit 16 shown in FIG. 1 is a device that controls the total braking torque in accordance with the regenerative operation in principle when the brake is operated. The regenerative cooperative brake unit 16 includes a brake pedal, a negative pressure booster that uses the intake negative pressure of the horizontally placed engine 2, and a master cylinder. Then, during the brake operation, cooperative control for the regenerative / hydraulic pressure is performed such that the amount of subtraction of the regenerative braking force from the required braking force based on the pedal operation amount is shared by the hydraulic braking force.

  As shown in FIG. 1, the power system of the FF hybrid vehicle includes a high-power battery 21 as a motor / generator power source and a 12V battery 22 as a 12V system load power source.

  The high-power battery 21 is a secondary battery mounted as a power source for the motor / generator 4. For example, a lithium ion battery in which a cell module constituted by a large number of cells is set in a battery pack case is used. The high-power battery 21 has a built-in junction box in which relay circuits for supplying / cutting off / distributing high-power are integrated, and further includes a cooling fan unit 24 having a battery cooling function, a battery charging capacity (battery SOC) and a battery. And a lithium battery controller 86 for monitoring the temperature.

  The high-power battery 21 and the motor / generator 4 are connected via a DC harness 25, an inverter 26, and an AC harness 27. The inverter 26 is provided with a motor controller 83 that performs power running / regenerative control. That is, the inverter 26 converts a direct current from the DC harness 25 into a three-phase alternating current to the AC harness 27 during power running for driving the motor / generator 4 by discharging the high-power battery 21. Further, the three-phase alternating current from the AC harness 27 is converted into a direct current to the DC harness 25 during regeneration in which the high-power battery 21 is charged by power generation by the motor / generator 4.

  The 12V battery 22 is a secondary battery mounted as a power source for a 12V system load that is an auxiliary machine, and for example, a lead battery mounted in an engine car or the like is used. The high voltage battery 21 and the 12V battery 22 are connected via a DC branch harness 25a, a DC / DC converter 37, and a battery harness 38. The DC / DC converter 37 converts a voltage of several hundred volts from the high-power battery 21 into 12V, and the charge amount of the 12V battery 22 is controlled by controlling the DC / DC converter 37 by the hybrid control module 81. The configuration is to be managed.

  As shown in FIG. 1, the control system of the FF hybrid vehicle includes a hybrid control module 81 (abbreviation: “HCM”) as an integrated control unit that has a function of appropriately managing energy consumption of the entire vehicle. Control means connected to the hybrid control module 81 include an engine control module 82 (abbreviation: “ECM”), a motor controller 83 (abbreviation: “MC”), and a CVT control unit 84 (abbreviation: “CVTCU”). And a lithium battery controller 86 (abbreviation: “LBC”). These control means including the hybrid control module 81 are connected via a CAN communication line 90 (CAN is an abbreviation for “Controller Area Network”) so that bidirectional information can be exchanged.

  The hybrid control module 81 performs various controls based on input information from each control means, an ignition switch 91, an accelerator opening sensor 92, a vehicle speed sensor 93, and the like. The engine control module 82 performs fuel injection control, ignition control, fuel cut control, and the like of the horizontally placed engine 2. The motor controller 83 performs power running control, regeneration control, and the like of the motor generator 4 by the inverter 26. The CVT control unit 84 performs engagement hydraulic pressure control of the first clutch 3, engagement hydraulic pressure control of the second clutch 5, shift hydraulic pressure control of the belt type continuously variable transmission 6, and the like. The lithium battery controller 86 manages the battery SOC, battery temperature, and the like of the high-power battery 21.

[Detailed configuration of engine start control]
FIG. 2 shows a flow of engine start control processing executed by the hybrid control module 81 of the first embodiment (engine start control means). Hereinafter, each step of FIG. 2 showing the engine start control processing configuration will be described. This engine start control process is repeatedly executed until the completion of the first explosion is determined.

  In step S1, when there is an engine start request, precharge control is performed to fill the piston chamber of the first clutch 3 with hydraulic oil by the precharge hydraulic pressure, and the process proceeds to step S2.

  In step S2, following the precharge control in step S1, cranking control for cranking the front engine 2 by starting the motor / generator 4 as an engine starter and slip-engaging the first clutch 3 is started. When cranking control is started, computations in the FF torque computation unit and the FB torque computation unit (only the torque reduction side FB is described) are processed in parallel. The FF torque refers to the transmission torque of the first clutch 3 acquired by feedforward control, and the FB torque refers to the transmission torque of the first clutch 3 acquired by feedback control (PI control or the like).

  In step S3, following the start of cranking control in step S2, the time from the start of cranking control is counted up, and the process proceeds to step S4.

In step S4, following the time measurement from the start of cranking control in step S3, FF torque corresponding to the elapsed time from the start of cranking control is calculated, and the process proceeds to step S8.
Here, when the cranking control is started, the FF torque corresponding to the elapsed time from the start of the cranking control increases the CL1 command torque diagonally until the target cranking torque is reached. When the target cranking torque is reached, the CL1 command torque for obtaining the target cranking torque is maintained. Further, when the elapsed time from the start of cranking control exceeds the required start time TD, backup control is performed in which the instruction torque to the first clutch 3 is suddenly increased until the horizontal engine 2 completes the first explosion.

  In step S5, following the start of cranking control in step S2, the target motor speed is compared with the actual motor speed, and the process proceeds to step S6.

  In step S6, following the comparison of the motor rotational speed in step S5, it is determined that the actual motor rotational speed has become lower than the target motor rotational speed (actual torque extra large), and the process proceeds to step S7.

In step S7, following the determination that the actual motor rotation speed is lower than the target motor rotation speed in step S6, the CL1 command torque is calculated by PI control according to the rotation speed deviation (target motor rotation speed−actual motor rotation speed). And go to step S8.
In the FB torque calculation unit, when the rotational speed deviation amount is small, the CL1 command torque is slightly increased or decreased depending on the deviation (rotational speed deviation amount) so that the actual motor rotational speed matches the target motor rotational speed. Is calculated.

  In step S8, following the FF torque calculation in step S4 and the FB torque calculation in step S7, the target transmission torque of the first clutch 3 is calculated by FF torque + FB torque, and the process proceeds to step S9.

In step S9, following the calculation of the target transmission torque to the first clutch 3 in step S8, the target transmission torque obtained in step S8 is obtained using the latest μ learning value updated and stored by the CL1 μ learning correction process described later. The CL1 command torque is calculated from the above, and the CL1 command torque is converted into the CL1 command hydraulic pressure, and the process proceeds to step S10.
Here, as the calculation formula of CL1 command torque,
Target transmission torque = CL1 command torque x μ learning value x N x D
Is used. That is, when the target transmission torque is the same, the CL1 instruction torque increases when the μ learning value decreases, and conversely, the CL1 instruction torque decreases when the μ learning value increases. However, N (number of clutch plates) and D (torque transmission area) are given by known specification values. Conversion from the CL1 command torque to the CL1 command oil pressure is performed using a torque-hydraulic characteristic representing the relationship between the two.

  In step S10, following the conversion of the CL1 command hydraulic pressure in step S9, a hydraulic pressure control command for obtaining the CL1 command hydraulic pressure converted in step S9 is output, and the process proceeds to return.

[Detailed configuration of CL1μ learning correction control]
FIG. 3 shows the flow of CL1μ learning correction control processing executed by the hybrid control module 81 of the first embodiment (learning correction control means). Hereinafter, each step of FIG. 3 representing the CL1μ learning correction control processing configuration will be described.

In step S21, it is determined that the engine start is permitted for μ learning, and the process proceeds to step S22.
Here, μ-learning-permitted engine start is confirmed during the cranking control when learning permission conditions (CL1μ learning permission condition, CL1 indicated torque stability condition, actual torque stability condition, CL1 temperature condition) are satisfied, and CL1μ learning value is acquired. Judgment is made by setting a permission flag. The establishment of the CL1μ learning permission condition is determined by a CL1μ learning permission flag that is set at the start of cranking control.

  In step S22, following the engine start determination in which μ learning is permitted in step S21, it is determined in the FB torque calculation section whether or not the rotational speed feedback due to the actual torque extra large has been activated. If YES (revolution speed FB operation), the process proceeds to step S23, and if NO (revolution speed FB non-operation), the process proceeds to step S24.

  In step S23, following the determination that the rotational speed FB operation is in step S22, the learning amount for when the CL1 actual torque is very large is set, and the process proceeds to step S29.

  In step S24, following the determination that the rotational speed FB is not activated in step S22, it is determined whether or not there is a torque increase due to the actual torque characteristic in the calculation control of the FF torque in the FF calculation unit by the timer. If YES (torque up), the process proceeds to step S25. If NO (torque up), the process proceeds to step S24.

In step S25, following the determination that there is torque up in step S24, in accordance with the elapsed time from the time when the torque-up operation is started in the FF torque calculation to the time when the engine start control is ended (the initial explosion completion flag is ON). The learning amount determined is set, and the process proceeds to step S29.
Here, the learning amount determined according to the elapsed time is a value that decreases the friction coefficient μ of the first clutch 3 as the elapsed time is longer.

  In step S26, following the determination that there is no torque up in step S24, it is determined that the engine speed has increased, and the process proceeds to step S27.

In step S27, following the engine speed increase determination in step S26, the difference between the target torque (CL1 command torque) and the CL1 estimated torque is calculated, and the process proceeds to step S28.
Here, as shown in FIG. 4, the CL1 estimated torque is in a system state in which the first clutch 3 is slip-engaged and the second clutch 5 is released or the torque is constant as shown in FIG. For this reason,
Motor torque = CL1 actual torque + auxiliary machine load (1)
When CL1 actual torque is replaced with CL1 estimated torque in equation (1),
CL1 Estimated Torque = Motor Torque-Auxiliary Load (2)
It becomes. Here, since the motor torque is a value that can be known from, for example, the motor current, and the auxiliary machine load is also a value that can be obtained from auxiliary machine operation load detection, the CL1 estimated torque can be calculated by the above equation (2). it can.

In step S28, following the torque divergence calculation in step S27, the learning amount is calculated based on whether the torque divergence is positive or negative, and the process proceeds to step S29.
Here, when the torque divergence is calculated by the formula of (CL1 command torque−CL1 estimated torque), when the value of the torque divergence is positive, the learning amount for decreasing the friction coefficient μ of the first clutch 3 is used as the value of the torque divergence. Is negative, the learning amount is set to increase the friction coefficient μ of the first clutch 3. In this normal learning correction, the learning amount for one time is given by a constant value smaller than the learning amount set in step S23 or step S25.

  In step S29, following the set (or calculation) of the learning amount in any of step S23, step S25, or step S28, the update amount (= set or calculated learning amount) to be learned this time is held, and the process proceeds to step S30. move on.

In step S30, following the current learning update amount holding, when the state of the first clutch 3 becomes the lock-up engagement state, the current learning update amount is reflected in the current μ and the process proceeds to the end.
Reflecting this time μ is the previous μ + current update amount, and this value is held as the friction coefficient μ of the first clutch 3 used at the next engine start.

Next, the operation will be described.
The operation in the control device of the FF hybrid vehicle of the first embodiment is divided into [engine start control operation], [CL1μ estimation operation by backup hydraulic pressure increase operation of FF start torque], and [CL1μ learning correction control operation by forced small determination]. explain.

[Engine start control action]
Hereinafter, the engine start control operation will be described with reference to FIG.
When there is an engine start request, the process proceeds from step S1 to step S2 in the flowchart of FIG. In step S1, precharge control is performed, and in step S2, cranking control for cranking the front engine 2 by starting the motor / generator 4 as an engine starter and slip-engaging the first clutch 3 is started.

When cranking control is started, the calculations in the FF torque calculation unit (steps S3 and S4) and the FB torque calculation unit (steps S5 to S7) are processed in parallel.
The FF torque calculation process in the FF torque calculation unit is performed by proceeding from step S2 to step S3 → step S4. In step S3, the time from the start of cranking control is counted up, and in step S4, the FF torque is calculated according to the elapsed time from the start of cranking control. When the cranking control is started, the FF torque rises the CL1 command torque diagonally up to the target cranking torque, and when the CL1 command torque reaches the target cranking torque, the CL1 command torque by the target cranking torque is maintained. Further, when the elapsed time from the start of the cranking control exceeds the start request time TD (see FIG. 6), the backup control for raising the CL1 command torque to the first clutch 3 is performed.

  On the other hand, the calculation process of the lower side FB torque in the FB torque calculation unit is performed by proceeding from step S2 to step S5 → step S6 → step S7. In step S5, the target motor speed and the actual motor speed are compared. In step S6, it is determined that the actual motor rotational speed has become lower than the target motor rotational speed due to the actual torque oversize. In step S7, a negative FB torque that decreases the CL1 command torque is calculated by PI control according to the rotational speed deviation (target motor rotational speed−actual motor rotational speed).

  When the FF torque calculation in step S4 and the FB torque calculation in step S7 are performed, the process proceeds from step S8 to step S9 to step S10 to return in the flowchart of FIG. In step S8, the target transmission torque of the first clutch 3 is calculated from FF torque + FB torque. In step S9, the CL1 instruction torque is calculated from the target transmission torque acquired in step S8 using the latest μ learning value updated and stored by the CL1 μ learning correction process, and the CL1 instruction torque is converted into the CL1 instruction hydraulic pressure. Is done. In step S10, a hydraulic pressure control command for obtaining the CL1 command hydraulic pressure converted in step S9 is output. The engine start control is terminated when the first clutch 3 is locked up and the transverse engine 2 is in a self-sustaining operation state due to a complete explosion.

[Effect of CL1μ estimated by FF starting torque backup hydraulic pressure increase operation]
The FF hybrid vehicle of the first embodiment uses a normally open dry multi-plate friction clutch that is hydraulically operated as the first clutch 3. In the case of the first clutch 3, the friction coefficient may greatly deviate from the initial value due to variations due to having a plurality of friction engagement surfaces, and the friction coefficient varies depending on the degree of progress of facing wear of each plate. The width also increases.

  On the other hand, for example, a comparison example in which normal learning correction control is performed in which the CL1 estimated torque and the CL1 command torque are compared to provide a single learning amount with a small constant value is used. In this comparative example, although the friction coefficient can follow a gradual change, it cannot follow a sudden change in the friction coefficient, and it takes a long time to obtain a learning correction value that reflects the sudden change in the friction coefficient. It will take. For this reason, when the transmission torque of the first clutch that is slip-engaged deviates from a desired torque when the engine is started, the engine cranking due to the torque deviation is obtained until a learning correction value that reflects the change in the friction coefficient is obtained. You will have to experience multiple times.

  On the other hand, in the engine start control, there is a time requirement that the engine start is started by a certain time at the latest because of the lag requirement for starting the engine. Therefore, when the actual transmission torque of the first clutch 3 is low, the engine start time becomes long. Therefore, as a response to this, an instruction to raise the CL1 instruction torque when a certain time (start request time TD) is reached in the FF calculation unit. Profile (FF backup control) is taken.

  However, since this backup state is a transient state in which the command hydraulic pressure of the first clutch 3 is increased by itself, the relationship of the above equation (1) becomes a transient state, and there is a problem that the estimated CL1 torque shifts, and the CL1 command torque The greater the slope of raising That is, as shown in the characteristics at the time of backup hydraulic pressure increase operation in the FF starting torque of FIG. 4, in the backup hydraulic pressure increase operation region, as indicated by the arrow A in FIG. 4, both the CL1 command torque and the CL1 estimated torque are in a transient state. , Accurate correction is not possible. On the other hand, in the CL1 command torque stable and actual torque stable region, as shown by the arrow B in FIG. 4, since both the CL1 command torque and the CL1 estimated torque are stable, accurate correction can be performed.

  Therefore, when the engine start time is within a certain time (start request time TD), normal learning correction is performed by comparing the CL1 estimated torque with the CL1 command torque. On the other hand, when the engine start time exceeds a certain time (start request time TD) and the CL1 command torque is increased to keep the engine start time, the friction of the first clutch 3 is increased by increasing the CL1 command torque. It is determined that the coefficient μ has decreased, and the capacity of the first clutch 3 is learned and corrected. At this time, as the learning correction amount, a learning correction amount different from that at the time of normal learning correction performed by comparing the CL1 command torque and the CL1 estimated torque can be selected. In other words, if there is a backup hydraulic pressure raising operation, the longer the CL1 command torque raising time is, the larger it is.

[CL1μ learning correction control action by forced small judgment]
If neither engine speed FB operation nor backup hydraulic pressure increase operation is performed in the engine start control, the flow proceeds to step S21 → step S24 → step S26 → step S27 → step S28 → step S29 → step S30 → end in the flowchart of FIG. It becomes. That is, in step S26, it is determined that the engine speed has increased. In step S27, the difference between the target torque (CL1 command torque) and the CL1 estimated torque is calculated. In step S28, the learning amount is calculated based on whether the torque deviation is positive or negative. In step S29, following the calculation of the learning amount in step S28, the update amount (= the calculated learning amount) to be learned this time is held. In step S30, when the state of the first clutch 3 becomes the lock-up engagement state, the update amount of the current learning is reflected in the current μ and is held as the friction coefficient μ of the first clutch 3 used at the next engine start ( Normal CL1μ learning correction control).

  On the other hand, when there is a backup hydraulic pressure raising operation in the engine start control, the flow proceeds to step S21 → step S22 → step S24 → step S25 → step S29 → step S30 → end in the flowchart of FIG. That is, in step S25, following the determination that there is torque up in step S24, the elapsed time from the time when the torque increase operation is started in the FF torque calculation to the time when the engine start control is completed (initial explosion completion flag ON). The learning amount determined according to is set. In step S29, following the set of learning amounts in step S25, an update amount (= set learning amount) to be learned this time is held. In step S30, when the state of the first clutch 3 becomes the lock-up engagement state, the update amount of the current learning is reflected in the current μ and is held as the friction coefficient μ of the first clutch 3 used at the next engine start ( CL1μ learning correction control by forced small judgment).

  The time chart shown in FIG. 6 shows the CL1μ learning correction control based on the forced small determination on the time axis.

  Time t1 is the output timing of the P range shift power generation request (engine start request). At the time t1 when the engine start request is output, the first clutch 3 is set to the standby mode by the precharge control, the CL1μ learning permission flag is set, and a command for setting the motor speed to the target motor speed is output.

  Time t2 is the engine cranking start timing. At time t2, CL1 estimated temperature determination is set, and the CL1 command torque starts to rise toward the target cranking torque during μ learning.

  Time t3 is the stability determination timing of the CL1 command torque at which the CL1 command torque reaches the target cranking torque during μ learning. The time t4 is a timing when a predetermined time has elapsed from the time t3. When the time t4 is reached, a CL1 learning value acquisition permission flag is set. Time t5 is the rising timing of the engine speed.

  Time t6 is the timing when the FF start torque backup hydraulic pressure raising operation is started as indicated by an arrow C in FIG. 6 because the elapsed time from time t2 has become the start request time TD. At this time t6, the forced torque small determination is turned ON, and the learning value is calculated according to the forced torque small determination. That is, when the forced torque small determination is made somewhere while the CL1 learning permission flag is set, the learning value at that time is set to a large learning value regardless of the normal torque small determination.

  Time t7 is the timing when the engine cranking is finished and preparation for the lockup engagement of the first clutch 3 is started. At this time t7, the CL1μ learning permission flag and the CL1 learning value acquisition permission flag are lowered, and the forced torque small determination is turned OFF.

  Time t8 is the lockup engagement start timing of the first clutch 3. In response to the timing t8 when the pre-lock-up to the lock-up is switched as indicated by the arrow D in FIG. 6, the learning value is calculated according to the forced torque small determination, the μ learning value (offset) is determined, and the μ learning consideration μ is determined. It is done.

In the first embodiment, the CL1μ learning correction control employs a configuration in which learning correction is performed to increase the CL1 command torque to the first clutch 3 after the next engine start when experiencing engine start in which backup control intervenes.
That is, since there is a time request for starting the engine by a certain time at the latest, backup control is performed when the elapsed time from the start of cranking control exceeds the start request time TD. The reason why the cranking control exceeds the start request time TD is that the actual transmission torque of the first clutch 3 is lower than the CL1 command torque. Therefore, it can be estimated that the friction coefficient μ of the first clutch 3 is reduced by the backup control intervening during the cranking control.
On the other hand, when the engine start in which the backup control intervenes is experienced, the learning correction that immediately reflects the estimation that the friction coefficient μ of the first clutch 3 is reduced and raises the CL1 command torque to the first clutch 3. Is done. For this reason, from the time of the next engine start, it becomes possible to start the engine whose elapsed time from the start of cranking control is equal to or shorter than the start request time TD.
As a result, when the friction coefficient μ of the first clutch 3 decreases, learning correction reflecting the decrease in the clutch friction coefficient μ can be performed with good response. In addition, learning correction can be performed even in an inexpensive system that does not use a stroke sensor for the first clutch 3.

In the first embodiment, a configuration is adopted in which the learning correction value is set to a larger value as the CL1 instruction torque increase time in the backup control for starting up the CL1 instruction torque to the first clutch 3 is longer.
That is, a long CL1 instruction torque increase time in the backup control indicates that the amount of transmission torque shortage of the first clutch 3 is large and the decrease in the friction coefficient μ of the first clutch 3 is correspondingly large.
Therefore, when experiencing engine start in which backup control intervenes, appropriate and highly accurate learning correction can be performed according to the degree of decrease in the friction coefficient μ of the first clutch 3.

In the first embodiment, a configuration is adopted in which learning correction is performed by positive / negative of the torque divergence between the CL1 command torque to the first clutch 3 and the CL1 estimated torque when no engine start involving backup control is experienced.
That is, when the engine start without the backup control is repeated, if the learning correction is performed based on the sign of the torque deviation, it is possible to cope with a gradual decrease or increase in the friction coefficient μ of the first clutch 3.
Therefore, learning control corresponding to both a gradual change in the friction coefficient μ of the first clutch 3 and a sudden decrease in the friction coefficient μ of the first clutch 3 can be performed.

In the first embodiment, a CL1μ learning correction is adopted in which the clutch friction coefficient μ used when calculating the CL1 command torque is used as a learning correction value based on the target transmission torque of the first clutch 3 during the cranking control. did.
For example, if an inclination of the hydraulic-torque characteristic is used as a learning correction value and an attempt is made to perform highly accurate learning correction, it is necessary to change a value representing a fine relationship between the two, and the memory capacity increases. On the other hand, if the memory capacity is reduced, highly accurate learning correction cannot be expected.
On the other hand, when the forced small determination that the friction coefficient μ of the first clutch 3 is reduced is made, the memory correction is reduced and the accuracy is high by the learning correction that directly decreases the value of the friction coefficient μ. Learning correction can be performed.

Next, the effect will be described.
In the control apparatus for the FF hybrid vehicle according to the first embodiment, the effects listed below can be obtained.

(1) The drive system includes an engine (horizontal engine 2), a first clutch 3, a motor (motor / generator 4), and drive wheels (left and right front wheels 10L, 10R).
When there is an engine start request in an EV mode using the motor (motor / generator 4) as a drive source, the engine (motor / generator 4) is used as an engine starter, the first clutch 3 is engaged with a predetermined slip, and the engine is started. In the control device for the hybrid vehicle that performs cranking control of the (horizontal engine 2),
When the engine start request is issued, the cranking control is started and the elapsed time from the start of cranking control is measured. When the elapsed time exceeds the start request time TD, at least the engine (horizontal engine 2) is first started. Engine start control means (hybrid control module 81) for starting up the engine (horizontal engine 2) by performing backup control for raising the instruction torque to the first clutch 3 until the explosion is completed;
Learning experience control means (hybrid control module 81) for performing learning correction for raising the command torque to the first clutch 3 from the next engine start when experiencing the engine start in which the backup control intervenes,
(Fig. 3).
For this reason, when the friction coefficient of the first clutch 3 decreases, learning correction reflecting the decrease in the clutch friction coefficient μ can be performed with good response.

(2) The learning correction control means (hybrid control module 81) sets the learning correction value to a larger value as the instruction torque increase time in the backup control for increasing the instruction torque to the first clutch 3 becomes longer (FIG. 3).
For this reason, in addition to the effect of (1), when experiencing engine start in which backup control intervenes, appropriate and highly accurate learning correction according to the degree of decrease in the friction coefficient μ of the first clutch 3 can be performed.

(3) When the learning correction control means (hybrid control module 81) does not experience the engine start in which the backup control intervenes, the learning correction control means (hybrid control module 81) corrects the learning correction based on the sign of the torque difference between the command torque to the first clutch 3 and the estimated torque. Perform (FIG. 3).
Therefore, in addition to the effect of (1) or (2), learning control corresponding to both a gradual change in the friction coefficient μ of the first clutch 3 and a sudden decrease in the friction coefficient μ of the first clutch 3 It can be performed.

(4) The learning correction control means (hybrid control module 81) uses a clutch friction coefficient used when calculating a command torque of the first clutch 3 based on a target transmission torque of the first clutch 3 during the cranking control. CL1μ learning correction is performed using μ as a learning correction value (FIG. 3).
For this reason, in addition to the effects (1) to (3), the memory capacity can be reduced and highly accurate learning correction can be performed.

  As mentioned above, although the control apparatus of the hybrid vehicle of this invention was demonstrated based on Example 1, it is not restricted to this Example 1 about a concrete structure, The invention which concerns on each claim of a claim Design changes and additions are permitted without departing from the gist of the present invention.

  In the first embodiment, as an engine start control unit, an example is shown in which FF torque calculation for performing backup hydraulic pressure increase control and FB torque calculation by the rotational speed FB control on the torque reduction side is performed. However, the engine start control means may be an example of performing FF torque calculation for performing at least backup hydraulic pressure increase control.

  In the first embodiment, the learning correction control means is set to a learning correction value (variable value) with a larger value as the instruction torque increase time in the backup control for starting up the instruction torque to the first clutch 3 is longer. . However, the learning correction control means may be an example in which a learning correction value (fixed value) determined in advance is updated and stored when experiencing engine start in which backup control intervenes.

  In the first embodiment, as an example of learning correction control means, CL1μ learning correction is performed in which the clutch friction coefficient μ of the first clutch 3 is used as a learning correction value. However, the learning correction control means may be an example in which the inclination of the torque-hydraulic characteristic is used as the learning correction value. In short, any parameter that increases the CL1 command torque by determining the decrease of the clutch friction coefficient μ and decreases the CL1 command torque by determining the increase of the clutch friction coefficient μ can be a learning correction target.

  In Example 1, the example which applies the control apparatus of this invention to FF hybrid vehicle was shown. However, the control device of the present invention can be applied not only to FF hybrid vehicles but also to FR hybrid vehicles and 4WD hybrid vehicles. In short, any hybrid vehicle having an engine, a first clutch, a motor, and drive wheels in the drive system can be applied.

1 Starter motor 2 Horizontal engine (engine)
3 First clutch 4 Motor / generator (motor)
5 Second clutch 6 Belt type continuously variable transmission 10R, 10L Left and right front wheels (drive wheels)
11R, 11L Left and right rear wheels 21 High power battery 22 12V battery 81 Hybrid control module (engine start control means, learning correction control means)

Claims (3)

The drive system includes an engine, a first clutch, a motor, and drive wheels,
When there is an engine start request in an EV mode using the motor as a driving source, the motor is an engine starter, the first clutch is engaged with a predetermined slip, and the engine is cranked and controlled.
When there is an engine start request, the cranking control is started and the elapsed time from the start of cranking control is measured, and when the elapsed time exceeds the start request time, at least until the engine completes the first explosion Engine start control means for starting the engine by performing backup control for starting up an instruction torque to the first clutch;
Learning correction control means for performing learning correction to increase the command torque to the first clutch from the next engine start when experiencing the engine start intervening in the backup control ,
The hybrid vehicle control apparatus , wherein the learning correction control means sets a learning correction value with a larger value as the instruction torque increase time in the backup control for increasing the instruction torque to the first clutch is longer . In the hybrid vehicle control device according to claim 1 ,
The learning correction control means, when not experiencing engine start the backup control intervention have line learning correction by the sign of the torque deviation indicated torque and the estimated torque to the first clutch,
The control apparatus for a hybrid vehicle, wherein the estimated torque is a torque obtained by subtracting an auxiliary machine load from a motor torque of the motor . In the hybrid vehicle control device according to claim 1 or 2 ,
The learning correction control means is a CL1μ learning correction in which a clutch friction coefficient μ used when calculating an instruction torque of the first clutch is a learning correction value based on a target transmission torque of the first clutch during the cranking control. A control device for a hybrid vehicle, characterized in that: