Control device and control system

The control device addresses performance degradation in multi-unit plants by using model-based feedforward and feedback control to adjust unit performance, ensuring consistent plant operation through targeted compensation.

JP7581899B2Active Publication Date: 2024-11-13MAZDA MOTOR CORP
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
JP2021006569
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-01-19
Publication Date
2024-11-13
Estimated Expiration
2041-01-19

AI Technical Summary

Technical Problem

Existing control systems for plants with multiple units struggle to maintain constant performance over time due to aging and other factors, as they lack effective methods to compensate for changes in unit performance.

Method used

A control device that includes a model control section to generate target values, a unit specifying section to identify performance changes, and feedback sections to correct output signals, using model-based feedforward and feedback control to adjust unit performance and maintain plant performance.

Benefits of technology

The system effectively compensates for performance changes in individual units, maintaining overall plant performance by adjusting target values and feedback signals, thereby reducing the impact on other units and ensuring consistent operation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007581899000001
    Figure 0007581899000001
  • Figure 0007581899000002
    Figure 0007581899000002
  • Figure 0007581899000003
    Figure 0007581899000003
Patent Text Reader

Abstract

To provide a control device for controlling a plant equipped with multiple units so as to keep the performance of the plant constant.SOLUTION: A control device (PCM 100) controls a vehicle C equipped with multiple units 20. The control device comprises: a model control unit 13 for creating target values of performance that the respective units are to realize based on models to be set corresponding respectively to the multiple units; a unit-identification unit (performance change determination unit 10a) for identifying one or more units having had a change in unit specific performance among the multiple units; and a target value correction unit (FF update unit 10b) for correcting the target values with regard to the units identified by the unit-identification unit.SELECTED DRAWING: Figure 3
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a control device for controlling a plant having a plurality of units, and a control system including the plurality of plants. [Background technology]

[0002] For example, Non-Patent Document 1 discloses a development methodology that combines so-called Model Based Development (MBD) with a Database-Driven (DD) control system design methodology. Specifically, Non-Patent Document 1 discloses a methodology for adjusting adjustable parameters of a modeled system by a database mechanism as an example of a design methodology for a system to achieve desired performance. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Toru Yamamoto: "Smart MBD (S-MBD) Method Based on Control Engineering Perspective - Proposal of a New Development Platform -", IEEJ Transactions on Control Engineering, Electronics, Information and Systems Division, June 29, 2019, CT19099, pp.25-26 Summary of the Invention [Problem to be solved by the invention]

[0004] The inventors of the present application considered applying a model-based design method as described in Non-Patent Document 1 to system control in general, and operating a system equipped with multiple units (e.g., an engine and a motor), such as an automobile hybrid system or steering system, as a controlled object (plant).

[0005] It is believed that the performance of such plants will change due to aging and other factors as they are operated for a long period of time. In order to keep the performance of the plant constant, the inventors of the present application have sought a mechanism for automatically compensating for the changed performance by appropriately modifying the configuration of a model corresponding to the plant or the output value of the model. In doing so, the inventors of the present application have focused on the fact that each plant has multiple units, and have arrived at the present disclosure.

[0006] The present disclosure has been made in consideration of the above points, and an object of the present disclosure is to maintain constant plant performance in a control device for controlling a plant having a plurality of units. [Means for solving the problem]

[0007] Specifically, a first aspect of the present disclosure relates to a control device for controlling a plant including a plurality of units, the control device including a model control unit that generates target values ​​of characteristics to be realized by each of the plurality of units based on a model set corresponding to each of the plurality of units, a unit identification unit that identifies a unit among the plurality of units in which a change has occurred in unit-specific performance, and a target value correction unit that corrects the target value for the unit identified by the unit identification unit.

[0008] Also, according to the first aspect, the target value correction unit corrects the target value by modifying the model corresponding to the unit identified by the unit identification unit.

[0009] Also, according to the first aspect, the control device includes a plurality of feedback sections corresponding to each of the plurality of units, and the plurality of feedback sections generate, based on an output signal from a unit corresponding to each feedback section, a feedback signal for correcting an output signal of the model control section so as to compensate for a difference between a characteristic actually realized by the unit and the target value, and the feedback section corrects the output signal by adjusting an FB feature quantity indicating either the feedback signal or an FB parameter indicating a coefficient contributing to an increase or decrease in the feedback signal, based on the output signal from the unit.

[0010] According to the first aspect, the target value correction unit corrects the target value of a unit identified by the unit identification unit among the multiple units, that is, a unit determined to have a change in unit-specific performance. This allows the plant performance to be spontaneously compensated for and kept constant. In addition, by performing compensation on a unit-by-unit basis rather than on a plant-by-plant basis, it is possible to realize performance compensation while suppressing changes in the control aspects of other units as much as possible.

[0011] Furthermore, according to the first aspect, by modifying a model corresponding to a unit whose performance has changed, it is possible to compensate for the performance change of the unit, which is advantageous in keeping the performance of the plant constant.

[0012] Furthermore, according to the first aspect, the feedforward control based on the model and the feedback control are combined, which is advantageous in keeping the plant performance constant.

[0013] Furthermore, according to a second aspect of the present disclosure, the control device may include a plurality of characteristic estimation units provided corresponding to each of the plurality of units, each estimating a characteristic realized by each unit, and a plant behavior estimation unit estimating a plant output of the plant based on the characteristic estimated by each of the plurality of characteristic estimation units, and the target value correction unit may correct the target value so that the plant output realizes a desired plant output.

[0014] Here, the term "plant output" refers to the general output of a plant, such as the vehicle speed when an automobile is used as the plant.

[0015] According to the second aspect, the target value correction unit corrects the target value of the characteristic to be realized by the unit, thereby realizing a desired plant output via this target value. In this manner, by adjusting the plant output via compensation on a unit-by-unit basis, it is possible to realize the adjustment of the plant output while suppressing changes in the control mode related to other units.

[0016] Furthermore, according to a third aspect of the present disclosure, the multiple units may include a first unit that contributes to increasing or decreasing a predetermined plant output, and a second unit that contributes to increasing or decreasing a plant output common to the first unit, and when the unit identification unit determines that performance of one of the first and second units has changed, the model control unit may increase or decrease the target value for the other of the first and second units so as to compensate for the change.

[0017] According to the third aspect, for example, when the performance of the first unit changes irreversibly and significantly, the target value for the second unit is also increased or decreased to achieve a desired plant output. In this way, the other units spontaneously recover from the change in performance, which is advantageous in keeping the plant performance constant.

[0018] In addition, the first aspect of the present disclosure 4 According to this aspect, the feedback unit may input the FB feature amount to the unit identification unit, the unit identification unit may determine a change in performance specific to the unit based on the change in the FB feature amount in each unit, and the target value correction unit may modify the model based on the change in the FB feature amount in each unit.

[0019] The above 4 According to the above aspect, the model is corrected based on the change in the FB feature amount, which is advantageous in keeping the plant performance constant.

[0020] In addition, the first aspect of the present disclosure 5 According to this aspect, the target value correction unit may correct the target value by adjusting the FB feature when a moving average of the FB feature is less than a predetermined threshold, and may correct the target value by modifying the model when the moving average of the FB feature is equal to or greater than the predetermined value.

[0021] The above 5 According to the embodiment, for a temporary or irreversible performance change, the target value is corrected by feedback control when the change is relatively small, while for an irreversible performance change, the target value is corrected by correcting the model when the change is relatively large. In this way, by using two correction methods, more flexible control can be realized, which is advantageous in maintaining constant plant performance.

[0022] In addition, the first aspect of the present disclosure 6According to this aspect, the control device may include a measurement unit that detects a measurement signal indicative of an operating environment of the plant, and the unit identification unit may determine a change in performance specific to the unit based on the measurement signal of the measurement unit and the FB feature.

[0023] The above 6 According to the above aspect, it is possible to realize control that associates the performance change of the unit with the operating environment of the plant, thereby making it possible to more appropriately compensate for the performance of each unit.

[0024] In addition, the first aspect of the present disclosure 7 According to this aspect, the control device may include a map generation unit that associates and stores an operation environment of the plant, the FB feature amount, and an FF parameter that is a parameter characterizing the model, and the target value correction unit may match the FF parameter corresponding to the FB feature amount based on a detection signal of the measurement unit and the FB feature amount.

[0025] The above 7 According to the above aspect, it is possible to realize control that associates the performance change of the unit, the operating environment of the plant, and information related to the correction of the model, thereby making it possible to more appropriately compensate the performance of each unit.

[0026] In addition, the first aspect of the present disclosure 8 According to this aspect, the map generation unit may update the relationship between the operational environment of the plant, the FB feature amount, and the FF parameter in real time while the plant is in operation.

[0027] The above 8 According to the above aspect, knowledge that associates the performance change of the unit, the operating environment of the plant, and information related to the correction of the model can be obtained. This contributes to spontaneous updating of the plant.

[0028] In addition, the first aspect of the present disclosure 9According to this aspect, when a plurality of units are identified by the unit identification section, the timing at which the correction of the target value is reflected may be adjusted so that it is approximately the same for each unit.

[0029] The above 9 According to this aspect, by correcting the target values ​​as simultaneously as possible, it is possible to suppress the occurrence of unintended plant outputs, which is advantageous in maintaining constant plant performance.

[0030] In addition, the first aspect of the present disclosure 10 According to this aspect, the plant may be an automobile, and the multiple units may include one or more of an engine and a motor that output torque for driving the automobile, a brake unit for braking the automobile, and a steering system for steering the automobile.

[0031] In addition, the first aspect of the present disclosure 11 The above aspect relates to a control system including a plurality of plants controlled by the control device. In this control system, each of the plurality of plants may include the plurality of units, and a result of identification by the unit identification unit in any one of the plurality of plants may be shared among the other plants.

[0032] The above 11 According to the above aspect, it is possible to share information related to changes in the performance of the units between plants. In this way, by sharing knowledge obtained by the unit identifying unit between plants, it is possible to realize more proactive control.

[0033] In addition, the first aspect of the present disclosure 12 According to this aspect, at least a portion of the control device may be implemented in an external server, and the plurality of plants may communicate with each other via the external server.

[0034] The above 12According to the aspect of the present invention, information related to changes in the performance of the units can be transmitted between plants via an external server. In this way, by sharing knowledge obtained by the unit identification unit between plants, more proactive control can be realized. Effect of the Invention

[0035] As described above, according to the present disclosure, in a control device for controlling a plant having a plurality of units, the performance of the plant can be kept constant. [Brief description of the drawings]

[0036] [Figure 1] FIG. 1 is a diagram illustrating an example of the overall configuration of the control system. [Diagram 2] FIG. 2 is a diagram illustrating an example of the configuration of each plant in the control system. [Diagram 3] FIG. 3 is a diagram illustrating an example of the configuration of a control device for each plant. [Figure 4] FIG. 4 is a diagram for explaining the control of the plant by the control device. [Diagram 5] FIG. 5 is a diagram for explaining estimation of unit characteristics by the characteristic estimating unit. [Figure 6] FIG. 6 is a diagram for explaining feedback control of unit characteristics by the feedback section. [Figure 7] FIG. 7 is a flowchart illustrating a main part of the control performed by the control device. [Figure 8] FIG. 8 is a flowchart illustrating a main part of a procedure for correcting a target value. [Figure 9] FIG. 9 is a diagram for explaining the distribution of functions performed when controlling a plant. [Figure 10] FIG. 10 is a diagram illustrating the change over time in unit performance. [Figure 11] FIG. 11 is a diagram for explaining cooperation with an external server. [Figure 12]FIG. 12 is a diagram for explaining cooperation between plants in a control system. [Figure 13] FIG. 13 is a diagram for explaining cooperation between a plant and a factory. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0037] Hereinafter, embodiments of the present disclosure will be described. Note that the following description is merely an example.

[0038] That is, in this specification, an automobile equipped with an engine, a motor, etc. will be described as an example of a plant equipped with a plurality of units, but the application of the technology disclosed herein is not limited to automobiles. The present disclosure can be applied to mechanical systems in general equipped with a plurality of units.

[0039] In particular, in this specification, an engine, a motor, a brake, and a steering unit for driving an automobile are described as examples of units included in a plant, but the units according to the present disclosure include units related to the operation of a plant in general. Elements other than mechanical units, such as an automobile driver, can also be considered as units.

[0040] In addition, in this specification, a PCM mounted on an automobile will be described as an example of a plant control device, but the control device according to the present disclosure is not limited to a module mounted on a plant. The control device according to the present disclosure includes general devices that can be connected to a plant by wire or wirelessly, such as an external server. As in a modified example described below, a part of the control device can be realized by a PCM, and the other part of the control device can be realized by an external server.

[0041] <Overall composition> FIG. 1 is a diagram illustrating an example of the overall configuration of a control system S. FIG. 2 is a diagram illustrating an example of the configuration of each plant (automobile C) in the control system S. As shown in FIG. 1, the control system S according to this embodiment controls a plurality of automobiles C1 to C2 (n automobiles in the illustrated example). nIt consists of several automobiles C1 to C n Each of the above exemplifies a "plant" in this embodiment. Hereinafter, when there is no limitation to a specific vehicle, it will be simply referred to as "vehicle C."

[0042] Multiple cars C1~C n are connected to each other via an external server Cs. One vehicle C transmits and receives electrical communications with other vehicles C via the external server Cs. Note that the external server Cs is not essential. The vehicles C may be configured to communicate with each other without going through the external server Cs.

[0043] As shown in FIG. 2, each vehicle C includes a plurality of units 201 to 20 (N units in the illustrated example). N Specifically, each automobile C includes a first unit 201 including an engine, a second unit 202 including a motor, a third unit 203 including a brake unit, and an Nth unit 204 including a steering system. N and N-4 units 204-20 N-1 The first unit 201, which is an engine, and the second unit 202, which is a motor, each output a torque for driving the automobile C. The third unit 203, which is a brake unit, brakes the automobile C. The Nth unit 20, which is a steering system, N The first unit 201 to the Nth unit 20 N Each of the units 20 is an example of a "unit" in this embodiment. Hereinafter, when there is no limitation to a specific unit, it will be simply described as a "unit 20." These units 20 configure the hardware system of the automobile C.

[0044] Each unit 20 is selected from among elements capable of controlling the behavior of the automobile C. Note that the behavior of the automobile C here includes any indicator related to the dynamic behavior of the automobile C.

[0045] In general, the "behavior of the plant" refers to the output value controlled by the PCM 100, i.e., the plant output in general. When model predictive control is performed by the PCM 100 as described below, the behavior of the plant corresponds to the plant output controlled to follow a predetermined set value trajectory. Hereinafter, the "behavior of the plant" and the "behavior of the automobile C" are also referred to as the "plant output."

[0046] For example, the behavior of the automobile C in this embodiment includes physical quantities related to the unit according to the present disclosure among any physical quantities that characterize the movement of the automobile C, such as the forward / backward speed, left / right speed, forward / backward acceleration, left / right acceleration, and yaw rate of the automobile C. Note that "forward / backward" in this specification is equivalent to the forward direction and reverse driving direction of the automobile C. Similarly, "left / right" in this specification is equivalent to the left / right turning direction of the automobile C.

[0047] More specifically, as described above, if the engine, motor, and brake unit of the automobile C are regarded as a unit, the behavior of the automobile C can include physical quantities related to the propulsion of the automobile C, such as longitudinal speed and longitudinal acceleration. On the other hand, if the steering system is included as a unit, the behavior of the automobile C can include physical quantities related to the steering of the automobile C, such as yaw rate.

[0048] Each unit 20 may have a plurality of sub-units. The sub-units here are selected from elements capable of controlling the behavior of the corresponding unit 20. In the example shown in FIG. 2, the first unit 201 has devices related to the operation of the engine as sub-units. Specifically, the first unit 201 has a throttle valve 201a and an EGR valve 201b as sub-units. Although not shown, the second unit 202 may have an inverter, a DC / DC converter, or the like as sub-units related to the drive of the motor. According to the present disclosure, the sub-units of the unit 20 may be indirectly controlled through the control of the unit 20.

[0049] Each automobile C is equipped with a PCM 100 as a control device, as an element for controlling the automobile C as a plant. The PCM 100 includes a CPU 100a for executing various calculations, a memory 100b for at least temporarily storing data necessary for the calculations performed by the CPU 100a, and an input / output bus 100c that forms a path for transmitting and receiving data. The PCM 100 is connected to the external server Cs described above so as to be able to transmit and receive electrical signals. In other words, the automobile C is connected to the external server Cs via the PCM 100 for controlling the automobile C.

[0050] Further, sensors for detecting measurement signals related to the driving environment (operating environment) of the automobile C are wirelessly connected to the PCM 100 as a control device. More specifically, sensors that are not physically connected to each unit 20 are connected to the PCM 100 according to this embodiment. That is, as the sensors according to this embodiment, in addition to sensors that are physically directly attached to the unit 20 such as the engine, such as a crank angle sensor and an in-cylinder pressure sensor, sensors that can be attached to the outside of the unit 20 (so-called external sensors), such as a GPS sensor, can be selected.

[0051] Such external sensors are adapted to detect measurement signals related to information characterizing the environment in which vehicle C is located (driving conditions), such as air pressure, temperature, altitude, etc., rather than information related to the dynamic behavior of vehicle C, such as the speed and acceleration of vehicle C.

[0052] Specifically, an outside air temperature sensor SW1 as an external sensor (measurement unit) that detects a measurement signal indicative of the operating environment of the plant, and a vehicle speed sensor SW2 as a non-external sensor are connected to the PCM 100 according to this embodiment. These sensors SW1 to SW2, together with the PCM 100, constitute the control system of the automobile C.

[0053] A display device 30 for displaying various information is also connected by wire or wirelessly to the PCM 100. The display device 30 can be configured using a liquid crystal display, an organic EL display, or the like.

[0054] Fig. 3 is a diagram illustrating the configuration of the control device (PCM100) of each plant (automobile C). Fig. 4 is a diagram for explaining the control of the plant (automobile C) by the control device (PCM100). Fig. 4 shows the connection relationship between each functional block illustrated in Fig. 3 and the relationship with the dynamic behavior (dynamics) of the plant and units. Below, each function implemented in the PCM100 will be roughly described with reference to Figs. 3 and 4.

[0055] As shown in FIG. 3, the PCM 100 according to this embodiment includes a first characteristic estimation unit 101 to an Nth characteristic estimation unit 102. N a plant behavior estimation unit 11, a plant behavior target generation unit 12, a model control unit 13, and a first feedback unit 141 to an N-th feedback unit 14 N These functional blocks are realized by a predetermined program (for example, a program coded in advance or a program inputted to function as an interpreter), and are read into the memory 100b as necessary. Among these, the first characteristic estimation unit 101 to the Nth characteristic estimation unit 10 N The first feedback unit 141 to the Nth feedback unit 14 N are both implemented so that the number of the functional blocks is equal to the number of the units 20. The programs corresponding to the functional blocks may be automatically or manually updated after the plant (car C) starts operating.

[0056] The basic concept of the processing realized by each functional block will be described below.

[0057] <Basic concept> Among the above-mentioned functional blocks, the first characteristic estimation unit 101 to the Nth characteristic estimation unit 10 NThe unit 20 determines the inherent performance of each unit 20 (hereinafter, referred to as "unit performance"), and based on the determined unit performance, the first unit 201 to the Nth unit 20 N A model is set to characterize each characteristic (hereinafter, referred to as a "unit characteristic"). N " may be collectively referred to as "characteristic estimation unit 10."

[0058] The first characteristic estimation unit 101 to the Nth characteristic estimation unit 10 N Based on the set model, each characteristic estimation unit estimates a unit characteristic that should be realized by the unit 20 corresponding to the characteristic estimation unit. N inputs the respective estimation results to the plant behavior estimation unit 11. Hereinafter, the estimated unit characteristics are also referred to as "estimated characteristics".

[0059] Specifically, the model set by the characteristic estimation unit 10 is configured to output unit characteristics according to unit performance. The "unit characteristics" here refer to physical quantities that contribute to an increase or decrease in the plant output. In other words, the unit characteristics according to the present disclosure include general characteristics that contribute to an increase or decrease in the plant output.

[0060] For example, if an engine, a motor and a brake unit are used as unit 20 and the longitudinal acceleration or longitudinal speed of automobile C is used as the plant output, the unit characteristics corresponding to these settings can be, for example, engine torque, motor torque and brake torque.

[0061] Similarly, when a steering unit and a brake unit are used as the unit 20 and the yaw rate of the automobile C is used as the plant output, the unit characteristics corresponding to these settings can be, for example, the rack position of the steering rack in the steering unit and the brake torque.

[0062] The plant according to this embodiment has a plurality of units 20. The plurality of units 20 can include a first unit 201 that contributes to increasing or decreasing a predetermined plant output, and a second unit 202 that contributes to increasing or decreasing a common plant output with the first unit 201. In the above example, when the plant output is set to the longitudinal acceleration or longitudinal speed of the automobile C, the first and second units 201 and 202 correspond to any combination of an engine, a motor, and a brake unit, respectively. Similarly, when the plant output is set to the yaw rate of the automobile C, the first and second units 201 and 202 correspond to a steering unit and a brake unit, respectively.

[0063] The unit performance determined by the characteristic estimating unit 10 includes performance that contributes to the increase or decrease of unit characteristics such as engine torque, and in turn includes general coefficients obtained by parameterizing such performance.

[0064] For example, when engine torque is used as the unit characteristic, the unit performance corresponding to the unit characteristic can be, for example, tire radius, gear ratio, vehicle weight, air resistance, gradient resistance, brake pad resistance (friction resistance of brake pad), hub resistance (friction resistance of hub), and moment of inertia of tire. The characteristic estimation unit 10 sets a model in which these unit performances are parameters, engine acceleration is input, and engine torque is output.

[0065] The characteristic estimation unit 10 judges the change in the unit performance as described above. Based on the judgment result of the change in the unit performance, the characteristic estimation unit 10 can correct the model (hereinafter, also referred to as the "unit model") so as to compensate for the change.

[0066] The characteristic estimation unit 10 has a performance change determination unit (unit identification unit) 10a and an FF update unit (target value correction unit) 10b as functional blocks for respectively realizing the determination of unit performance and the correction of the model. The FF update unit 10b identifies a unit 20 in which a change has occurred in unit performance among the multiple units 20, for example, based on an FB feature amount (described later) input from a corresponding feedback unit 14. The FF update unit 10b corrects a target value (target characteristic) (described later) output by the model control unit 13, for example, by correcting a unit model corresponding to the unit 20 identified by the performance change determination unit 10a. Details of these functional blocks will be described later.

[0067] The plant behavior estimation unit 11 includes a first characteristic estimation unit 101 to an N-th characteristic estimation unit 10 N The plant behavior estimation unit 11 estimates the behavior of the plant (vehicle C) based on the estimated characteristics output from the model control unit 13.

[0068] As described above, the plant output is closely related to the unit characteristics of each unit 20. The plant output can be modeled as a function using the unit characteristics of each of the multiple units 20 as input. The function obtained by modeling the plant output (hereinafter, also referred to as a "plant model") is read from the memory 100b or the like and used by the plant behavior estimation unit 11 and the model control unit 13 as appropriate.

[0069] For example, when engine torque and motor torque are used as unit characteristics, the corresponding plant output can be the longitudinal acceleration or longitudinal speed of the automobile C, as described above. In this case, the plant output can be described by a model in which the combined value of the engine torque and motor torque is input after formulating the connection structure from the power train consisting of the engine and motor to the tires (particularly the drive wheels).

[0070] In other words, the characteristic estimator 10 and the plant behavior estimator 11 can be regarded as a model of the relationship between the unit characteristics and the plant output. By tracing this relationship in reverse, it is possible to autonomously correct the unit characteristics (target values ​​of the characteristics that each unit should achieve) corresponding to the desired plant output, and further the model settings for achieving the unit characteristics.

[0071] The plant behavior target generating unit 12 generates a target value for the behavior that the automobile C as a plant should achieve. As described later, when the model control unit 13 performs model predictive control, this target value corresponds to a so-called set value trajectory. When the longitudinal acceleration of the automobile C is used as the plant output, the plant behavior target generating unit 12 generates a target value for the longitudinal acceleration (specifically, the longitudinal acceleration that should be achieved in the future). Such a target value can be generated based on a measurement signal related to an interface for driving the automobile C, such as the accelerator opening of the automobile C. For example, when the yaw rate of the automobile C is used as the plant output, the related target value can be generated based on the rotation angle of the steering wheel, etc.

[0072] The model control unit 13 generates target values ​​for the characteristics (unit characteristics) that each unit 20 should achieve, based on a model (unit model) set corresponding to each of the multiple units 20. Specifically, the model control unit 13 calculates a command value for each unit 20, based on the plant output (estimated behavior) estimated by the plant behavior estimation unit 11, the unit characteristics (estimated characteristics) estimated by the characteristic estimation unit 10, and a model similarly corrected by the characteristic estimation unit 10. Note that, when the multiple units 20 include a unit 20 having a subunit, the model control unit 13 can also be configured to calculate a command value corresponding to the subunit. The model control unit 13 calculates at least the same number of command values ​​as the number of units 20. The command values ​​calculated by the model control unit 13 are fed to the first feedback unit 141 to the Nth feedback unit 142. N is input to one of the units 20.

[0073] The command value calculated by the model control unit 13 corresponds to the target value of the unit characteristic of each unit 20. In other words, when the longitudinal acceleration of the automobile C is used as the estimated behavior, the model control unit 13 calculates the target value of the engine torque, the target value of the motor torque, etc. Such calculations are realized by back-calculating the unit characteristic corresponding to the plant output to be realized based on the plant model and the unit model.

[0074] In particular, the model control unit 13 according to this embodiment can execute model predictive control. When configured in this way, the input to the model control unit 13 is discretized to be a multi-level system. Specifically, the model control unit 13 receives the unit characteristics, plant behavior, and model estimated or corrected at a timing one step ahead of the current time, and the unit characteristics, plant behavior, and model estimated or corrected at a timing two steps ahead of the current time, and corrects the target values ​​of the unit characteristics of each unit 20 using the target values ​​generated by the plant behavior target generation unit 12 as the set value trajectory.

[0075] Here, as described above, when the multiple units 20 include a first unit 201 that contributes to increasing or decreasing a predetermined plant output and a second unit 202 that contributes to increasing or decreasing a common plant output with the first unit 201, the plant output as the plant behavior increases or decreases according to the sum of the unit characteristics. For example, when engine torque and motor torque are used as unit characteristics, the corresponding plant output can be described by a model in which the sum of the engine torque and the motor torque is input, as described above. In this case, the model control unit 13 can output a target value of the sum of the engine torque and the motor torque.

[0076] Therefore, in order to calculate the target value of the engine torque and the target value of the motor torque, respectively, it is necessary to appropriately distribute the target value of the added value to each unit 20. Therefore, the model control unit 13 according to this embodiment changes the distribution of the plant output according to the unit performance determined by each characteristic estimating unit 10. Specifically, when the characteristic estimating unit 10 determines that the performance of one of the first and second units 201, 202 has changed, the model control unit 13 increases or decreases the target value for the other of the first and second units 201, 202 so as to compensate for the change. For example, when the performance of the engine as the first unit 201 changes over time and the target value of the unit characteristic is corrected, the target value of the motor torque in the motor as the second unit 202 is set to a higher value, assuming that an engine torque lower than a desired engine torque is output.

[0077] FIG. 9 is a diagram for explaining the distribution of functions performed when controlling a plant, particularly an automobile C. As shown in FIG. 9, the PCM 100 reads a map that specifies the distribution of the plant output. In FIG. 9, the sum of the engine torque (ENG torque) and the motor torque (MG torque) is constant on the straight line Lt. The straight line Lt shifts according to the driving state of the automobile C. In the example shown in FIG. 9, as shown by the dashed line, when the automobile C accelerates or runs at high speed, the straight line Lt shifts upward. Here, it is assumed that the motor torque and the engine torque are distributed as shown in the plot P2 at the initial point in time. Here, as described above, when aging deterioration of the engine is determined, the model control unit 13 changes the distribution of torque from the plot P2 to the plot P1. As a result, the engine torque is set low, while the motor torque is set high. In addition, a map that specifies the current distribution may be stored in association with, for example, the driving environment (driving conditions) of the automobile C. In this case, each time vehicle C is started, PCM 100 can load the map appropriate for the current driving environment and use it to allocate plant power.

[0078] The first feedback unit 141 to the Nth feedback unit 14 N are provided corresponding to each of the multiple units 20, and correct the target value of the unit characteristic calculated for each unit 20. N " may be collectively referred to as "feedback section 14."

[0079] In detail, the first feedback unit 141 to the N-th feedback unit 14 N Based on the output signal from the unit 20 corresponding to each feedback section 14, the model control section 13 corrects the output signal so as to reduce the difference between the unit characteristic (hereinafter also referred to as "actual characteristic") actually realized by the unit 20 and the target value (hereinafter also referred to as "target characteristic") input from the model control section 13. In order to perform this correction, the first feedback section 141 to the Nth feedback section 14 N generates a feedback signal (corresponding to the FB feature in FIG. 5) corresponding to each unit 20.

[0080] The feedback signal may be a signal based on the difference or ratio between the actual characteristic and the target characteristic. For example, when PID control is performed based on the difference between the actual characteristic and the target characteristic, the feedback signal is a signal obtained by adding together the product (proportional term) of the difference and the proportional gain, the product (integral term) of the integral value of the difference and the integral gain, and the product (differential term) of the differential value of the difference and the differential gain. When the difference between the actual characteristic and the target characteristic is used, the feedback unit 14 corrects the target characteristic by adding the feedback signal thus calculated to an electrical signal corresponding to the target characteristic. The feedback unit 14 is provided for each unit 20. Therefore, this correction is performed for each unit 20.

[0081] Furthermore, the feedback section 14 changes an FB feature quantity indicating either a feedback signal or an FB parameter indicating a coefficient contributing to an increase or decrease in the feedback signal, based on an output signal from the corresponding unit 20.

[0082] For example, in the case of PID type feedback control, the FB parameters correspond to a proportional gain, an integral gain, and a differential gain. The feedback unit 14 adjusts the FB feature amount to correct the target characteristic and compensate for the change in unit performance. For example, when the actual characteristic of the corresponding unit 20 changes relatively greatly, the FB feature amount is adjusted to a larger amount compared to when the actual characteristic changes relatively small. This makes it possible to realize feedback control that follows the change in unit performance.

[0083] More specifically, the feedback unit 14 according to the present embodiment adjusts the FB feature quantity by data-driven control. In this case, the unit characteristics realized by the corresponding unit 20 are directly input to each feedback unit 14, as in adaptive feedback control, or the unit characteristics are indirectly input via a database, as in so-called database-type data-driven control. The feedback unit 14 can implement both.

[0084] Moreover, the feedback unit 14 inputs the FB feature quantity to the corresponding performance change determination unit 10a. The performance change determination unit 10a determines the change in unit performance based on the change in the FB feature quantity in each unit 20. For example, when the FB feature quantity changes significantly, it can be determined that the unit performance is more likely to have changed than when the FB feature quantity changes slightly. Also, when the change in the FB feature quantity is not temporary but occurs continuously, it can be determined that the unit performance is more likely to have changed irreversibly. Therefore, the performance change determination unit 10a determines whether or not the moving average of the FB feature quantity exceeds a predetermined threshold value, and counts the number of times the moving average exceeds the predetermined threshold value. The performance change determination unit 10a determines whether or not the counted number of times exceeds a predetermined number of times within a predetermined period, and when this determination is NO (or when the moving average is less than the predetermined threshold value), it adjusts the FB feature quantity by, for example, data-driven control, and executes correction of the target characteristic to compensate for the change in the unit performance.

[0085] On the other hand, if the determination is YES (or if the moving average is equal to or greater than the predetermined threshold), the performance change determination unit 10a corrects the target characteristics and thus compensates for the change in unit performance by modifying the unit model via the FF update unit 10b as described above. Here, the FF update unit 10b updates the model by increasing or decreasing parameters that characterize the unit model. The FF update unit 10b inputs a signal indicating that the model has been updated and the updated parameters to the model control unit 13. The model control unit 13 can change the distribution of the plant output based on the signal thus input.

[0086] The target characteristic corrected by the feedback section 14 is input to the corresponding unit 20 as a command value corresponding to the corresponding unit 20 .

[0087] In this embodiment, the command value corrected by the first feedback section (first FB section) 141 is input to the first unit (engine) 201. The first unit 201 outputs engine torque as the actual characteristic of the engine based on dynamics D1 corresponding to the unit performance at the time when the command value is input. A signal indicating the engine torque is input to the first feedback section (first FB section) 141, which executes feedback control based on the FB feature amount, correction of the engine unit model, and the like.

[0088] Similarly, the command value corrected by the second feedback section (first FB section) 142 is input to the second unit (motor) 202. The second unit 202 outputs the motor torque that is actually realized as the actual characteristics of the motor based on the dynamics D2 corresponding to the unit performance at the time when the command value is input. A signal indicating the engine torque is input to the second feedback section (second FB section) 142, which executes feedback control based on the FB feature amount, correction of the motor unit model, etc.

[0089] In addition, the automobile C as a plant has dynamics D corresponding to the engine torque output by the engine as the first unit 201, the motor torque output by the motor as the second unit, and the performance of the first and second units 201 and 202 at the time when the engine torque and the motor torque are input. p Based on this, the longitudinal speed, longitudinal acceleration, etc. are output as the plant output.

[0090] In other words, the unit characteristics estimated by the first characteristic estimation unit 101, the unit characteristics estimated by the second characteristic estimation unit 102, and the plant output (estimated behavior) estimated by the plant behavior estimation unit 11 correspond to the unit characteristics realized by the first unit 201, the unit characteristics realized by the second unit 202, and the plant output (actual behavior) realized by the automobile, respectively.

[0091] When the inherent performance of each unit 20 is kept constant and model predictive control by the model control unit 13 is performed with high accuracy, the unit characteristics and the plant output will be approximately equal to each other. The above-mentioned unit models correspond to modeled dynamics D1, D2 of the corresponding units 20, and the plant model corresponds to modeled dynamics D1, D2 of the entire plant. p Therefore, when a performance change occurs in any of the units 20 due to aging or the like, it is possible to compensate for the performance change by simply modifying the model corresponding to the unit 20 in which the performance change has occurred, without changing the plant model or the unit models of all the units 20.

[0092] Further details of each functional block will now be described in turn.

[0093] <Detailed configuration> -Characteristics Estimation Unit 10- Fig. 5 is a diagram for explaining estimation of unit characteristics by the characteristic estimation unit 10. As shown in Fig. 5, the characteristic estimation unit 10 has a unit characteristic estimation unit 10c, a map generation unit 10d, and a model generation unit 10e in addition to the performance change determination unit 10a and the FF update unit 10b described above. In the example shown in the figure, only the first characteristic estimation unit 101 is illustrated, but the second characteristic estimation unit 102 has a similar configuration.

[0094] Of these, the performance change determination unit 10a, as described above, determines the change in the corresponding unit performance based on the multi-level command value and the FB feature amount. Here, the measurement signal of the outside air temperature sensor SW1 as an external sensor is also input to the performance change determination unit 10a. That is, for example, air resistance can change depending on the air density and therefore the outside air temperature. As described above, air resistance can affect the optimization of the unit performance and therefore the unit model, but such an effect is considered to be a temporary effect. It is not appropriate to consider the effects caused by the driving environment (external environment) of the automobile C, including air resistance, as aging changes in the unit performance.

[0095] Therefore, the performance change determination unit 10a determines the change in the unit performance based on the driving environment (operation environment of the plant) of the automobile C detected by an external sensor such as the outside air temperature sensor SW1 and the FB feature amount. For example, the PCM 100 or a secondary storage device connected to the PCM 100 stores a database, a map, a model, etc. that associates the FB feature amount and the driving environment with an index for determining the change in the unit performance. The performance change determination unit 10a inputs the FB feature amount input from the feedback unit 14 and the current driving environment into the database, etc., to determine whether the input FB feature amount indicates a temporary change in the unit performance or indicates a continuous change (irreversible change) in the unit performance.

[0096] The information relating to the driving environment may include the altitude and air pressure at the point where the automobile C is traveling. Such information may be measured, for example, by a GPS sensor serving as an external sensor. The PCM 100 may use the altitude, air pressure, etc., instead of or in addition to the outside air temperature to determine changes in unit performance. Also, instead of or in addition to a measurement signal from an external sensor, the mileage of the automobile C (or, in the case of a more general plant, the operation period of the plant) may be recorded, and changes in unit performance may be determined by referring to the recorded contents.

[0097] As described in the explanation of the schematic configuration of the characteristic estimation unit 10, the unit characteristic estimation unit 10c estimates how much each unit performance has changed among a plurality of unit performances set for each unit 20. This estimation can be performed based on the amount of change in the FB feature amount (particularly, the difference between the moving average of the FB feature amount and the predetermined threshold value) and the driving environment or driving state of the automobile C.

[0098] For example, in the case of an engine as the first unit 201, the unit characteristic estimation unit 10c can estimate a performance change of the engine by associating an opening change of the throttle valve 201a as a subunit constituting the engine with a change amount of the FB feature amount that directly or indirectly contributes to the engine torque. More specifically, the unit characteristic estimation unit 10c can estimate a performance change in detail by associating a control target of the subunit constituting each unit 20 with a time change of the FB feature amount. Note that the "FB feature amount" in this example refers to an FB feature amount that contributes to the opening degree of the throttle valve 201a (more specifically, an FB feature amount of the torque acting on the throttle valve 201a). Since the amount of intake air of the engine and therefore the engine torque are determined according to the opening degree of the throttle valve 201a, this FB feature amount can be regarded as an FB feature amount that indirectly contributes to the engine torque.

[0099] For example, if the FB feature value increases uniformly regardless of the opening degree of the throttle valve 201a, it can be determined that a change has occurred in the reaction force acting on the valve body of the throttle valve 201a. In this case, it can be determined that a change has occurred in the performance of a return spring constituting the throttle valve 201a (for example, the return spring has become worn). Similarly, for example, if the FB feature value is constant when the opening speed of the throttle valve 201a is constant, while the FB feature value increases and decreases when the opening speed increases and decreases, it can be determined that a change has occurred in the moment of inertia of the throttle valve 201a. In this case, it can be determined that a foreign object has adhered to the valve body or that the shape of the valve body has changed. By determining the change in the reaction force and the change in the moment of inertia in combination, it can also be determined how much each change has progressed.

[0100] The determination result by the unit characteristic estimation unit 10c can be displayed on the display device 30 of the automobile C. This allows the driver to understand the current performance of the automobile C. As described later, the determination result by the unit characteristic estimation unit 10c can also be shared with other automobiles C via an external server Cs as knowledge obtained by the automobile C, or can be transmitted to a factory or the like.

[0101] The map generating unit 10d stores a map that associates the driving environment of the automobile C (the operating environment of the plant), the FB feature amount, and the FF parameters that characterize the model. The FF parameters refer to parameters used in formulating the unit model, such as tire radius and gear ratio. Instead of the map, a simpler database may be used. The map generating unit 10d can update the relationship between the driving environment of the automobile C, the FB feature amount, and the FF parameters in real time while the automobile C is being driven.

[0102] For example, the FF parameters are updated in real time as the unit performance changes, as described above. This corrects the unit model of the engine. When the unit model of the engine is corrected, the PCM 100 can also change the distribution of the plant output based on the unit characteristics corresponding to the unit model and the plant output, as described above. For example, assume a case where the maximum value of the engine torque as a unit characteristic is reduced as a result of clogging of the injector of the engine. In such a case, depending on the setting of the target torque, the desired plant output may not be realized. This possibility is particularly noticeable when maximization of the plant output or maximization of the energy efficiency of the plant is required. In this case, as described with reference to FIG. 9, the PCM 100 realizes the desired plant output by compensating for the change in the maximum value of the engine torque with the motor torque. On the other hand, for example, when the target torque is not so large and maximization of the energy efficiency is not required, compensation by the motor torque is not required, and therefore the above-mentioned change in the distribution is not required. In this way, the PCM 100 is configured to change the distribution of the plant output according to the target setting of the plant output.

[0103] By using the map generator 10d, it becomes possible to reflect various factors, such as changes in unit performance that occur in units 20 other than the corresponding unit 20, in the model control. By reflecting the performance change of the corresponding unit 20 "alone" in the model control, it becomes possible to more reliably achieve the plant performance that is in line with the developer's intention. By accumulating knowledge of the characteristics of each unit 20 related to the deterioration and performance variation of each unit 20, it becomes possible to make the model-based control function more reliably.

[0104] The model generating unit 10e generates a model corresponding to the map by performing regression processing on the map generated by the map generating unit 10d. The model generated by the model generating unit 10e is a model in which the driving environment (operation environment of the plant) of the automobile C, the FB feature amount, and the FF parameters corresponding to each unit performance are associated with each other. Note that both the map generating unit 10d and the model generating unit 10e are not essential.

[0105] The FF update unit 10b reflects the FF parameters newly obtained by checking against a map, a model, etc., in the unit model, and executes updating of the unit model. When the FF parameters characterizing the unit model also affect the state of the plant model, the FF update unit 10b outputs the updated FF parameters to the plant behavior estimation unit 11 and the model control unit 13. By this output, the unit model and the plant model are each updated.

[0106] -Feedback section 14- 6 is a diagram for explaining feedback control of unit characteristics by the feedback section 14. As shown in FIG. 6, the first feedback section 141 as the feedback section 14 has an FB update section 14a, an FB command value generation section 14b, a target characteristic correction section 14c, and a timing adjustment section 14d. As shown in the figure, the second feedback section 142 is configured similarly to the first feedback section 141. Hereinafter, only the configuration of the first feedback section 141 will be described, except for the configuration related to cooperation between the feedback sections 14.

[0107] The FB update unit 14a updates the FB parameters for increasing or decreasing the FB feature amount based on the difference or ratio between the target characteristic output from the model control unit 13 and the actual characteristic in the first unit 201. As described above, for example, when performing PID type feedback control, the FB parameters correspond to the proportional gain, the integral gain, and the differential gain. The FB parameter can be updated using a general method included in data-driven control. For example, the FB parameters may be updated by general adaptive control, or may be updated by referring to a database, a map, a model, or the like.

[0108] Although not shown, the FB update unit 14a according to this embodiment can update in real time the map and model associated with one or more of the driving environment, target characteristics, and actual characteristics of the automobile C as inputs and output FB parameters, similar to the map generation unit 10d and model generation unit 10e in the characteristic estimation unit 10. The FB update unit 14a can also derive a relational expression by machine learning that inputs one or more of the driving environment, target characteristics, and actual characteristics of the automobile C and output FB parameters, regardless of the form of a model set in advance. In this case, the FB update unit 14a also functions as a regression learner.

[0109] The FB update unit 14a inputs the updated FB parameters to the FB command value generation unit 14b. When the FB parameters are used as FB features, the FB parameters updated by the FB update unit 14a are branched between the FB update unit 14a and the FB command value generation unit 14b, and input to a characteristic estimation unit (specifically, a first characteristic estimation unit 101) 10 related to a unit 20 common to this feedback unit 14. Instead of directly inputting the FB parameters, a change amount of the FB parameters may be input to the characteristic estimation unit 10.

[0110] The FB command value generating unit 14b generates a feedback signal indicating a correction value (FB value) of the target characteristic based on the FB parameter updated by the FB updating unit 14a and the difference between the target characteristic and the actual characteristic.

[0111] As already explained, when performing PID type feedback control, the FB value is the sum of a proportional term, an integral term, and a differential term with the difference as an argument, but the feedback unit 14 according to the present embodiment is not limited to such a value. For example, a signal indicating the FB value can be generated based on the ratio between the target characteristic and the actual characteristic.

[0112] When the FB value is used as the FB feature, the FB value calculated by the FB command value generating unit 14b branches off between the FB command value generating unit 14b and the target characteristic correcting unit 14c, and is input to a characteristic estimating unit (specifically, a first characteristic estimating unit 101) 10 associated with a unit 20 common to this feedback unit 14.

[0113] The target characteristic correction unit 14c corrects the target characteristic based on the target characteristic (FF value) output from the model control unit 13 and the correction value (FB value) generated by the FB command value generation unit 14b.

[0114] Here, in the feedback section 14, when PID control based on the difference between the target characteristic and the actual characteristic is performed, the target characteristic correction section 14c functions as an adder that outputs an electrical signal corresponding to the sum of the FF value and the FB value. On the other hand, when feedback control based on the ratio between the target characteristic and the actual characteristic is performed, the target characteristic correction section 14c functions as a multiplier that outputs an electrical signal corresponding to the product of the FF value and the FB value. In the illustrated example, only the former configuration is disclosed, but the present disclosure also includes the latter configuration.

[0115] When there is one or more units 20 identified by the unit identifying unit, that is, one or more units 20 determined by the performance change determining unit 10a to have changed in performance among the multiple units 20, the timing adjusting unit 14d sets the timing at which the correction of the target characteristic is reflected to approximately the same timing for each unit 20. More specifically, the timing adjusting unit 14d executes the above-mentioned timing setting for the units 20 among the multiple units 20 that contribute to a common increase or decrease in the plant output. Such combinations of units 20 include a combination of a steering unit and a brake unit in addition to the combination of a motor, an engine, and a brake unit as shown in the example.

[0116] When the unit specifying unit specifies a plurality of units, the target value correcting unit sets the timing at which the correction of the target value is reflected to be approximately the same timing for each unit.

[0117] <Control example> The main parts of the control realized by the control device configured as described above will be described below. In the example shown below, the control device corresponds to the PCM 100 mounted on the automobile C as a plant, and the unit 20 controlled by the PCM 100 corresponds to the first unit 201 configured by the engine and the second unit 202 configured by the motor, as described above.

[0118] Fig. 7 is a flow chart illustrating the main part of the control by the control device (PCM 100). Fig. 8 is a flow chart illustrating the main part of the procedure for correcting the target characteristic. The flows shown in Fig. 7 and Fig. 8 are repeatedly executed while the automobile C is being driven.

[0119] First, as shown in step S1 of Fig. 7, the characteristic estimation unit 10 estimates the unit characteristic. The unit characteristic is estimated for each unit 20. Next, in step S2, the plant behavior estimation unit 11 estimates the plant behavior (plant output) based on the estimated unit characteristic. Note that when the model control unit 13 performs model predictive control, the estimation of the unit characteristic and the estimation of the plant behavior are performed for each multi-level time.

[0120] Then, in step S3, the model control unit 13 generates a target value (target characteristic) by, for example, model predictive control based on the estimated unit characteristic and plant behavior. The generation of the target value is performed for each unit 20.

[0121] In step S4, before or after the generation of the target value in step S3, or in parallel with the generation of the target value, the performance change determination unit 10a in the characteristic estimation unit 10 identifies the units 20 whose performance has changed among the units 20 constituting the automobile C as a plant, based on the actual unit characteristics (real characteristics). Note that the order of the steps in Fig. 7 is merely an example. For example, it may be configured so that step S4 is executed prior to step S1.

[0122] In step S5, the FF update unit 10b in the characteristic estimation unit 10 modifies the model related to the unit identified in step S4. Through the model modification by the FF update unit 10b, the target characteristic is corrected.

[0123] In step S6, as described with reference to Fig. 9, the allocation of functions among the units 20 is changed. Specifically, the model control unit 13 adjusts the ratio between the target characteristics of the engine as the first unit 201 and the target characteristics of the motor as the second unit 202 so that the behavior (plant output) that the automobile C should achieve is constant. This process can be performed based on parameters that characterize the driving state of the automobile C, such as the measurement signal of the vehicle speed sensor SW2.

[0124] In step S7, the corrected target characteristic is input to the unit 20. Each unit 20 realizes a characteristic that reflects the correction result of the target characteristic for the unit characteristic estimated in step S1, and in response to the thus-corrected characteristic, the automobile C realizes a plant behavior (actual behavior) similar to the plant behavior estimated in step S2.

[0125] Here, a specific adjustment procedure for the target characteristic is as shown in Fig. 8. First, as shown in step S11, the characteristic estimation unit 10 sets a normative characteristic of the unit 20 based on the current driving environment, etc. Next, as shown in step S12, the characteristic estimation unit 10 reads the FB feature amount calculated by the feedback unit 14.

[0126] Then, in the next step S13, the characteristic estimation unit 10 judges whether or not the characteristic estimated by the characteristic estimation unit 10 is judged to be a standard characteristic based on the characteristic of the unit 20 estimated based on the current FB feature amount. If the judgment is YES, the characteristic estimation unit 10 judges that the current system behavior is appropriate, and the model control unit 13 generates an FF value and returns. In this case, the PCM 100 executes model predictive control in a state in which the FB parameters are fixed and the model corresponding to each unit 20 is not changed. This state corresponds to a state in which the unit 20 does not change over time and the unit performance is kept constant, or a state in which the change over time is compensated for by updating the FF parameters or FB parameters after the change over time occurs (a state in which the standard characteristic reflecting the update of the FF parameters or FB parameters is achieved).

[0127] On the other hand, if the determination in step S13 is NO, the characteristic estimation unit 10 determines in step S14 whether the response characteristic of the unit 20 can be improved (whether it can be dealt with by adjusting the FB feature amount). If this determination is YES, as shown in step S15, the feedback unit 14 corresponding to the unit 20 updates the FB parameter, and the model control unit 13 generates an FF value and returns. In this case, the PCM 100 executes model predictive control with the FB parameter fixed and the model corresponding to each unit 20 unchanged. This state corresponds to a state in which the unit performance is deteriorated due to the change over time of the unit 20, but the influence of the deterioration is compensated by updating the FB parameter, as shown by the solid line in (2) of FIG. 10. Note that the dashed line in FIG. 10 corresponds to a performance curve in the case where neither the FB parameter nor the FF parameter is updated.

[0128] Also, if the determination in step S14 is NO, the characteristic estimation unit 10 determines that the unit performance of the unit 20 has changed significantly and irreversibly. In this case, the characteristic estimation unit 10 updates the FF parameters, and the model control unit 13 generates an FF value and returns. In this case, as shown in step S16, the PCM 100 executes model predictive control in a state in which the FB parameters have been changed and in which the FF parameters characterizing the model corresponding to each unit 20 have been changed. This state corresponds to a state in which the unit performance has been significantly deteriorated due to the change over time of the unit 20, but the effect of the deterioration has been compensated for by updating the FB parameters and changing the model (updating the FF parameters), as shown in (3) of FIG. 10.

[0129] As described above, according to this embodiment, the FF update unit 10b executes correction of the characteristic target for the unit 20 identified by the performance change determination unit 10a among the multiple units 20, that is, for the unit 20 determined to have undergone a change in the unit-specific performance. This allows the performance of the automobile C to be spontaneously compensated for and kept constant. In addition, by performing compensation on a unit 20 basis rather than on a vehicle C basis, it is possible to realize compensation of performance while suppressing changes in the control mode of the other units 20 as much as possible.

[0130] Furthermore, the FF update unit 10b realizes a desired plant output by correcting the target characteristics through modification of the model corresponding to the identified unit 20. In this manner, by adjusting the plant output through compensation on a unit 20 basis, it is possible to realize the adjustment of the plant output while suppressing changes in the control aspects related to the other units 20.

[0131] 9, for example, when the performance of the first unit 201 changes irreversibly and significantly, the desired plant output can be achieved by increasing or decreasing the target value for the other second unit 202. In this way, the other units 20 spontaneously recover from the change in performance, which is advantageous in keeping the performance of the automobile C constant.

[0132] In addition, for temporary or irreversible performance changes, the characteristic targets are corrected by feedback control when the change is relatively small, while for irreversible performance changes, the characteristic targets are corrected by correcting the model when the change is relatively large. In this way, by using two different correction methods, more flexible control can be achieved, which is advantageous in maintaining constant performance of vehicle C.

[0133] In addition, by using an external sensor such as the outside air temperature sensor SW1, it is possible to realize control that correlates the performance change of the unit 20 with the operating environment of the automobile C. This makes it possible to more appropriately compensate for the performance of each unit 20.

[0134] <Other embodiments> (About other units 20) In the above embodiment, the control related to the combination of the engine as the first unit 201 and the motor as the second unit 202 has been described, but the present disclosure is not limited to such control. The present disclosure is also applicable to the control of a combination of a brake unit in addition to the engine and the motor. Moreover, the present disclosure is also applicable to the control related to the combination of a steering unit and a brake unit instead of the engine and the motor.

[0135] (Linkage with external server Cs) In the above embodiment, the PCM 100 of each automobile C constituting the control system S functions as a control device according to the present disclosure, but the present disclosure is not limited to such a configuration. At least a part of the functional blocks functioning as control devices may be implemented on an external server Cs, and the automobiles C may communicate with each other via the external server Cs. In this case, a calculation system combining the PCM 100 and the external server Cs functions as a control device according to the present disclosure.

[0136] In this case, as in the PCM 100′ shown in FIG. 11, the first characteristic estimation unit 101 to the Nth characteristic estimation unit 10 N can be implemented on the external server Cs. N A part of the characteristics estimation unit 101 to the Nth characteristics estimation unit 10 may be implemented in the PCM 100′ and the other parts may be implemented in the external server Cs. N Other functional blocks, for example the plant behavior estimation unit 11, may be implemented in the external server Cs.

[0137] (Cooperation between plants) Also, instead of the configuration shown in Fig. 11, it is also possible to configure so that the identification result of the unit identification section (characteristics estimation section 10) in any one of the multiple plants is shared between the other plants. In this case, as shown in Fig. 12, the characteristic estimation section 10 is implemented only in a predetermined PCM 1001'', and the other PCM 1002'' receives the identification result by the PCM 1001'' and performs correction of the target characteristics, etc. In this case, cooperation via an external server Cs is not essential.

[0138] Not limited to the illustrated example, the characteristic estimation unit 10 can also be implemented in other PCMs 1002" and configured to mutually transmit and receive information on units whose performance has changed between plants. For example, databases, maps, models, etc. generated based on the outside air temperature can be shared between automobiles C.

[0139] (Collaboration with Factories F1 and F2) Furthermore, for example, as shown in FIG. 13, instead of or in addition to linking automobiles C with each other via an external server Cs, automobile C can be linked with a parts factory F1 for automobile C and a system repair factory F2 for automobile C.

[0140] In this case, as shown in process P1, when information indicating the knowledge gained by the automobile C (more specifically, information indicating the change over time in the unit performance of each unit 20) is transmitted to the external server Cs, the external server Cs places an order for the manufacture of the unit 20 to the parts factory F1 based on the change over time in the unit performance, or places an order for the manufacture of sub-units that make up each unit 20, such as the throttle valve 201a and the EGR valve 201b (see process P2).

[0141] Then, as shown in processes P3 and P4, parts such as subunits are delivered from the parts factory F1 to the system repair factory F2, and the parts factory F1 notifies the external server Cs of the delivery of the parts. Then, the external server Cs reserves the maintenance work at the system repair factory F2 (see process P5). Finally, the system repair factory F2 or the external server Cs notifies the car C of the reservation details (see process P6).

[0142] In this way, the control device according to the present disclosure performs control based on changes in unit performance, and is extremely useful when information indicating those changes is output outside the plant or shared between plants. [Explanation of symbols]

[0143] S Control System C Automobile (Plant) Cs External Server 201 Engine (1st Unit) 202 motor (second unit) 203 Brake unit (third unit) 20 N Steering unit (Nth unit) 100 PCM (control device) 10. Characteristics Estimation Section 10a Performance change determination unit (unit identification unit) 10b FF update section (target value correction section) 10c Unit characteristic estimation section 10d Map Generation Section 12 Plant behavior estimation section 13 Model control section 14 Feedback section SW1 Outside air temperature sensor (measurement part)

Claims

1. A control device for controlling a plant having a plurality of units, comprising: a model control unit that generates target values ​​of characteristics to be realized by each of the plurality of units based on a model that is set corresponding to each of the plurality of units; a unit identification unit that identifies a unit in which a change has occurred in unit-specific performance among the plurality of units; a target value correction unit that corrects the target value for the unit identified by the unit identification unit, the target value correction unit corrects the target value by modifying the model corresponding to the unit identified by the unit identification unit; The feedback control unit further includes a plurality of feedback sections corresponding to the plurality of units, the plurality of feedback sections generate, based on an output signal from a unit corresponding to each feedback section, a feedback signal for correcting the output signal of the model control section so as to compensate for a difference between a characteristic actually realized by the unit and the target value; The feedback section corrects the output signal by adjusting an FB feature quantity indicating either the feedback signal or an FB parameter indicating a coefficient contributing to an increase or decrease of the feedback signal, based on the output signal from the unit. A control device comprising:

2. 2. The control device according to claim 1, a plurality of characteristic estimating units provided corresponding to the plurality of units, each estimating a characteristic realized by each unit; a plant behavior estimating unit that estimates a plant output of the plant based on the characteristics estimated by each of the plurality of characteristic estimating units, The target value correction unit corrects the target value so that the plant output realizes a desired plant output. A control device comprising:

3. 3. The control device according to claim 1, the plurality of units include a first unit that contributes to increasing or decreasing a predetermined plant output, and a second unit that contributes to increasing or decreasing a plant output common to the first unit, When the unit specifying unit determines that the performance of one of the first and second units has changed, the model control unit increases or decreases the target value for the other of the first and second units so as to compensate for the change. A control device comprising:

4. 2. The control device according to claim 1, The feedback unit inputs the FB feature quantity to the unit identification unit, The unit identifying unit determines a change in performance specific to the unit based on a change in the FB feature amount in each unit; The target value correction unit corrects the model based on a change in the FB feature amount in each unit. A control device comprising:

5. 2. The control device according to claim 1, The target value correction unit When the moving average of the FB feature quantity is less than a predetermined threshold, the target value is corrected through adjustment of the FB feature quantity; When the moving average of the FB feature amount is equal to or greater than the predetermined value, the target value is corrected by correcting the model. A control device comprising:

6. 6. The control device according to claim 4 or 5, a measurement unit that detects a measurement signal indicative of an operating environment of the plant; The unit identifying unit determines a change in performance specific to the unit based on the measurement signal of the measuring unit and the FB feature amount. A control device comprising:

7. 7. The control device according to claim 6, a map generating unit that stores an operation environment of the plant, the FB feature amount, and an FF parameter that characterizes the model in association with each other, The target value correction unit collates the FF parameters corresponding to the FB feature amount based on the detection signal of the measurement unit and the FB feature amount. A control device comprising:

8. 8. The control device according to claim 7, The map generation unit updates the relationship between the operating environment of the plant, the FB feature amount, and the FF parameter in real time during operation of the plant. A control device comprising:

9. The control device according to any one of claims 1 to 8, When the unit specifying unit specifies a plurality of units, the timing at which the correction of the target value is reflected is adjusted so that the timing is substantially the same for each unit. A control device comprising:

10. The control device according to any one of claims 1 to 9, The plant is an automobile, The plurality of units includes one or more of an engine and a motor that output torque for driving the vehicle, a brake unit that brakes the vehicle, and a steering system that steers the vehicle. A control device comprising:

11. A control system including a plurality of plants controlled by the control device according to any one of claims 1 to 10, Each of the plurality of plants includes the plurality of units, The identification result of the unit identification section in any one of the plurality of plants is shared among the other plants. A control system comprising:

12. 12. The control system according to claim 11, At least a part of the control device is implemented in an external server; The plurality of plants communicate with each other via the external server. A control system comprising:

Citation Information

Patent Citations

  • Plant monitoring device

    JP1992313195A

  • Data providing system of power generating facility

    JP2002259505A

  • Control device for plant and control device for thermal power plant

    JP2011210215A

  • Denitrification controller

    JP2015048975A