JP2003150205A - Plant control equipment - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To use a control target model defined using a reference value of a plant input or output, to set a more versatile reference value and improve control performance. Provide a plant control device that can be used. SOLUTION: A model parameter of a control target model includes:
Model parameter c1 'not related to plant input / output
Is adopted, and the model parameter c1 ′ calculated by the model parameter identifier 22 is supplied to the reference value setting unit 25. The reference value setting unit 25 sets the model parameter c1 ′
Is performed, and a default opening deviation thdefadp indicating the deviation of the reference value is calculated. The default opening THDEF, which is a reference value, is corrected using the default opening deviation thdefadp, and a corrected reference value THDEFR is calculated.

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] [0001] The present invention relates to a control device for a plant.
Control model, especially for plant
Model parameters were identified in real time and identified
For controlling plant using model parameters
I do. [0002] 2. Description of the Related Art A control target model in which a plant is modeled
Model parameters were identified in real time and identified
Control device for controlling plant using model parameters
Is disclosed in, for example, JP-A-2000-179384.
ing. The control device disclosed in this publication is a model
Identifier for identifying parameters and estimating plant output
Estimator and model parameters identified by the identifier
And the plant output estimated by the estimator
The control input to the plant is controlled by the riding mode control.
And a sliding mode controller for calculating.
The identifier and estimator include the output of the plant and its target.
Deviation from the plant value and the deviation between the plant input and the reference value.
The reference value is used to manipulate the input to the plant.
Is variably set in accordance with the operation amount. The reference value
By setting variably according to the manipulated variable, the plant
Increase convergence speed in control to converge output to target value
Effect can be obtained. [0004] SUMMARY OF THE INVENTION
Coming devices are more specifically sliding mode control
Set the reference value according to the adaptive control input calculated by the heater
This reference value setting method uses
Only for controllers using adaptive sliding mode control
It was applicable. Therefore, more versatile
A new reference value setting technique was desired. The present invention has been made by paying attention to this point.
Yes, defined using plant input or output reference values
More versatility when using the controlled model
High reference values can be set to improve control performance.
It is an object of the present invention to provide a control device for a plant that can be cut. [0006] [MEANS FOR SOLVING THE PROBLEMS] To achieve the above object
The present invention uses a predetermined reference value (THDEF) for the plant.
Multiple model parameters of the controlled model
Identification means for identifying data, and
Control the plant using the model parameters
A control device for a plant, comprising:
Number of model parameters are related to the input and output of the plant.
Model parameter (c1)
Based on model parameters (c1) not related to input / output
Correction means for correcting the predetermined reference value (THDEF)
It is characterized by having. According to this configuration, the model of the control target model is
Parameters are models that are not related to plant input / output.
Parameters that are not related to the input / output of the plant.
Predetermined reference values are corrected based on Dell parameters
You. That is, the model parameters of the control target model itself
Since the reference value is corrected by
It can be applied to any control method.
is there. In addition, modeling errors (characteristics of actual plant
And the characteristics of the controlled model)
Change the reference value of the controlled model to the reference value of the actual plant
Can be matched, reducing modeling errors,
Control performance can be improved. [0008] The correction means is used for input and output of the plant.
Statistical processing of unrelated model parameters (c1)
To calculate a correction value (thdefadp).
Correcting the predetermined reference value (THDEF) according to the value
It is desirable. The controlled object model is the plant
First model parameters (a1, a2) related to the output of
And a second model parameter related to a control input to the plant.
Parameter (b1), the control input and the output of the plant
The third model parameter (c
It is desirable to be defined by 1). The plant is a throttle valve for an internal combustion engine.
And a driving means for driving the throttle valve.
A throttle valve driving device, wherein the control device includes:
So that the throttle valve opening matches the target opening.
The parameters that determine the control input to the
It is desirable to calculate. [0010] The throttle valve driving device may include:
First biasing means for biasing the throttle valve in a valve closing direction;
A second urging means for urging the throttle valve in the valve opening direction.
The drive means does not drive the throttle valve.
The first slot and the second biasing means.
The throttle valve is maintained at the default opening.
The predetermined reference value may be set to the default opening.
desirable. [0011] BRIEF DESCRIPTION OF THE DRAWINGS FIG.
It will be described with reference to FIG. (First Embodiment) FIG. 1 shows a first embodiment of the present invention.
It is a figure showing composition of such a throttle valve control device. Internal combustion
In an intake passage 2 of an engine (hereinafter referred to as “engine”) 1,
A throttle valve 3 is provided. For throttle valve 3
Is a first urging mechanism for urging the throttle valve 3 in the valve closing direction.
Return spring 4 as a step and throttle valve 3
Member 5 as a second urging means for urging the valve in the valve opening direction
And are attached. The throttle valve 3 is driven
Via a gear (not shown) by a motor 6 as a means
It is configured to be able to be driven. Drive by motor 6
When no force is applied to the throttle valve 3, the slot
The opening degree TH of the valve 3 is determined by the urging force of the return spring 4.
And the default opening T at which the biasing force of the elastic member 5 is balanced.
It is held at HDEF (for example, 5 degrees). The motor 6 includes an electronic control unit (hereinafter referred to as “E”).
CU ”) and its operation is controlled by the ECU.
7. Throttle valve 3 has a throttle
A throttle valve opening sensor 8 for detecting the valve opening TH is provided.
The detection signal is supplied to the ECU 7. The engine 1 is mounted on the ECU 7.
Of the accelerator pedal that detects the required output of the driver of the vehicle
An accelerator sensor 9 for detecting the depression amount ACC is connected.
The detection signal is supplied to the ECU 7. EC
U7 is a throttle valve opening sensor 8 and an accelerator sensor
The input circuit to which the detection signal is supplied, and the input signal is digitized.
A / D conversion circuit that converts to digital signals and performs various arithmetic processing
Central processing unit (CPU)
Maps and tables referenced in programs and programs
ROM (Read Only Memory) for storing
A memory consisting of a RAM (Random Access Memory)
Output circuit for supplying drive current to the motor circuit and the motor 6
It has. The ECU 7 depresses the accelerator pedal
The target opening THR of the throttle valve 3 is determined according to the amount ACC.
And the detected throttle valve opening TH is equal to the target opening THR.
The control amount DUT of the motor 6 is determined so that
An electric signal corresponding to the control amount DUT is supplied to the motor 6. In this embodiment, the throttle valve 3, the retarder
Slot consisting of spring 4, elastic member 5, and motor 6
The throttle valve driving device 10 is set as a control target, and the control target is
Duty ratio of the electric signal applied to the motor 6
A DUT and the output of the controlled object is a throttle valve opening sensor
8 as the throttle valve opening TH detected. The response frequency characteristic of the throttle valve driving device 10
The gain characteristic and phase shown by the solid line in FIG.
Characteristics are obtained. Therefore, the module defined by the following equation (1)
Dell was set as the control target model. The response of this model
The frequency response is as shown by the broken line in FIG.
It is confirmed that the characteristics are similar to those of the
Have been.   DTH (k + 1) = a1 × DTH (k) + a2 × DTH (k−1)                     + B1 × DUT (k-d) + c1 (1) Here, k is a parameter representing the discretized time.
DTH (k) is a slot defined by the following equation (2).
It is a tor valve opening deviation amount. DTH (k) = TH (k) -THDEF (2) Here, TH is the detected throttle valve opening, THDEF
Is the default opening. Also, a1, a of the formula (1)
2, b1 and c1 are models that determine the characteristics of the control target model.
And d is the dead time. The model defined by the above equation (1) is
Discrete-time system adopted to facilitate control application
DARX model (delayed autoregressive model with
exogeneous input (autoregressive model with external input)
is there. In equation (1), the output deviation DTH
Model parameters a1 and a2, input duty ratio D
In addition to the model parameter b1 related to the UT,
The unknown model parameter c1 is employed. this
The model parameter c1 is the default opening THDEF
A parameter that indicates displacement and disturbance applied to the throttle valve drive.
Data. That is, the model parameter identifier
Model parameters a1, a2, b1
By identifying the parameter c1, the default opening
Deviations and disturbances can be identified. FIG. 3 shows a slot realized by the ECU 7.
FIG. 3 is a functional block diagram of a tor valve control device,
Is connected to the adaptive sliding mode controller 21 and the mode.
Dell parameter identifier 22 and after dead time d
Predicted throttle valve opening deviation (hereinafter “predicted deviation”)
To calculate PREDTH (k) (= DTH (k + d))
State predictor 23 and the accelerator pedal depression amount ACC.
To set the target opening THR of the throttle valve 3
Setting unit 24 and a reference for setting the reference value THDEFR
It comprises a value setting section 25 and subtracters 26 and 27. Adaptive sliding mode controller 2
1 indicates that the detected throttle valve opening TH is equal to the target opening THR.
Adaptive sliding mode control so that
And calculate the duty ratio DUT.
Output the ratio DUT. Adaptive sliding mode control
By using the roller 21, the throttle valve opening TH
The response characteristic following the target opening THR
Data can be changed as needed using the data (VPOLE).
As a result, the throttle valve 3 is moved from the open position to the fully closed position.
Impact when moving (impact on throttle fully closed stopper
Collision) and engine response to accelerator operation
It is possible to make the resistance variable. In addition, the model parameters
It is possible to ensure stability against errors. The model parameter identifier 22 is a modified model
Parameter vector θL (θL^T= [A1, a2, b
1, c1]), and calculates the adaptive sliding mode control.
It is supplied to the roller 21. More specifically, model parameters
The meter identifier 22 determines the throttle valve opening TH and the duty.
Based on the tee ratio DUT, the model parameter vector θ
(Θ^T= [A1 ', a2', b1 ', c1'])
I do. Furthermore, for the model parameter vector θ
Model parameters by performing limit processing
Calculate the vector θL and calculate the corrected model parameter vector
ΘL to the adaptive sliding mode controller 21
Supply. In this way, the throttle valve opening TH is targeted.
Optimal model parameters to follow the opening THR
a1, a2, b1 are obtained, and disturbance and default
The model parameter c1 indicating the deviation of the opening THDEF is obtained.
Can be Identify model parameters in real time
By using the model parameter identifier 22
Adaptation to changes in engine operating conditions, hardware characteristics
Fluctuation compensation, power supply voltage fluctuation compensation, and hardware
It is possible to adapt to the aging of characteristics. The state predictor 23 calculates the throttle valve opening TH
And the duty ratio DUT,
Rottle valve opening TH (predicted value), more specifically, predicted bias
Calculate the difference amount PREDTH and use the adaptive sliding mode
It is supplied to the controller 21. Predicted deviation PREDTH
To control the dead time of the controlled object
The robustness of the system is ensured, and
It is possible to improve controllability near the tilt opening THDEF.
it can. The reference value setting unit 25 uses the identifier 23 to
Model parameter c1 '(model parameter c
(The value before limit processing of 1 is displayed as "c1 '".)
Is used to indicate the deviation of the default opening THDEF.
The default opening deviation thdefadp is calculated and the default opening
Default opening THD using the deviation thdefadp
By correcting the EF, the correction reference value THDEFR
(= THDEF-thdefadp) is calculated. However
And a throttle valve according to the above equation (2) and the following equation (3).
The definition of the opening deviation amount DTH and the target value DTHR is defined as THD
EFR = THDEF (that is, default open)
Degree deviation thdefadp is 0)
Things. The subtracters 26 and 27 output the correction reference value THD
EFR, throttle valve opening TH and target opening THR
Of the throttle valve opening deviation DTH and the target value
Calculated as DTHR. Next, adaptive sliding mode control
The operation principle of the laser 21 will be described. First, according to the following equation (3)
The target value DTHR (k) is changed to the target opening THR (k) by default.
G is defined as a deviation from the opening degree THDEF. DTHR (k) = THR (k) -THDEF (3) Here, the throttle valve opening deviation amount DTH and the target value DT
If the deviation e (k) from HR is defined by the following equation (4),
Switching function value σ (k) of sliding mode controller
Is set as in the following equation (5). e (k) = DTH (k) -DTHR (k) (4) σ (k) = e (k) + VPOLE × e (k-1) (5) = (DTH (k) -DTHR (k)) + VPOLE × (DTH (k)
(k-1) -DTHR (k-1)) Here, VPOLE is a value larger than −1 and smaller than 1.
Is a switching function setting parameter that is set to. The vertical axis is the deviation e (k), and the horizontal axis is the previous deviation e (k).
On the topological plane defined as (k-1), σ (k) = 0
The combination of the deviation e (k) to be added and the previous deviation e (k-1) is
This straight line is generally called the switching straight line.
You. Sliding mode control is based on this
This control focuses on the behavior of the difference e (k).
σ (k) becomes 0, that is, the deviation e (k) and the previous deviation
The combination of e (k-1) is on the switching line on the phase plane
Control in such a way that disturbances and modeling errors (actual plan
Between the characteristics of the model and the characteristics of the model
(Difference). So
As a result, the throttle valve opening deviation amount DTH becomes the target value DT
Controlled with good robustness to follow HR
It is. Further, the switching function setting parameter V in equation (5)
By changing the value of POLE, as shown in FIG.
The damping characteristic of the deviation e (k), that is, the throttle valve opening
Change the characteristic of the deviation amount DTH following the target value DTHR
be able to. Specifically, VPOLE = −1
Characteristics that do not follow at all.
The smaller the absolute value of VPOLE, the faster the following speed
Can be In the throttle valve control device, the following requirements
The requirements A1 and A2 are required to be satisfied. A1) When moving the throttle valve 3 to the fully closed position,
Avoid collision with the fully closed stopper A2) Non-linear characteristics near the default opening THDEF
(The urging force of the return spring 4 and the urging of the elastic member 5)
Changes in elastic characteristics due to balance with force, motor
Backlash of the gear interposed between the throttle valve 6 and the throttle valve 3
Throttle valve opening even if the duty ratio DUT changes
To improve the controllability of the dead zone where Therefore, in the vicinity of the fully closed position of the throttle valve, the deviation e
(k), and the default opening THD
In the vicinity of the EF, the convergence speed needs to be increased. According to the sliding mode control, the switching
By changing the function setting parameter VPOLE,
In the present embodiment, the convergence speed can be easily changed.
Rotational valve opening TH and change amount DDT of target value DTHR
Switching according to HR (= DTHR (k) -DTHR (k-1))
The function setting parameter VPOLE is set.
As a result, the above requirements A1 and A2 can be satisfied.
You. As described above, the sliding mode control
Then, a combination of the deviation e (k) and the previous deviation e (k-1) (hereinafter referred to as a combination)
(Referred to as "deviation state quantity") on the switching straight line.
Therefore, the deviation e (k) can be obtained at the specified convergence speed,
Robustly converges to 0 against the delling error. did
Therefore, in sliding mode control,
It is important to put the difference state quantity on the switching straight line and restrain it there.
Become. From such a viewpoint, input to the control target
(Output of controller) DUT (k) (also described as Usl (k)
Is equivalent to the equivalent control input U, as shown in the following equation (6).
eq (k), reaching law input Urch (k) and adaptive law input Uad
It is configured as the sum of p (k).   DUT (k) = Usl (k)            = Ueq (k) + Urch (k) + Uadp (k) (6) The equivalent control input Ueq (k) represents the deviation state quantity.
This is an input for constraining on the switching straight line.
rch (k) is for placing the deviation state quantity on the switching straight line.
The adaptive law input Uadp (k) is the modeling error
And suppress the influence of disturbances and put the deviation state quantity on the switching straight line
Input. Hereinafter, each input Ueq (k), Urch (k)
And a method of calculating Uadp (k) will be described. The equivalent control input Ueq (k) represents the deviation state quantity.
Since this is an input to constrain on the switching straight line,
Condition is given by the following equation (7). σ (k) = σ (k + 1) (7) Expression (7) is calculated using Expression (1) and Expressions (4) and (5).
Calculating the duty ratio DUT (k) that satisfies
(9) is obtained, which is the equivalent control input Ueq (k).
You. Further, the reaching law input Urch (k) and the adaptive law input U
adp (k) is given by the following equations (10) and (11), respectively.
More defined. (Equation 1) Here, F and G are the reaching law control gain and
Is the control gain, which is set as described below.
You. ΔT is a control cycle. In the calculation of the above equation (9), the elapsed time d
Subsequent throttle valve opening deviation amount DTH (k + d) and the corresponding
The target value DTHR (k + d + 1) is required. So, when you waste
The amount of deviation DTH (k + d) of the throttle valve opening after d has elapsed
And the predicted deviation PRE calculated by the state predictor 23
Using DTH (k) as the target value DTHR (k + d + 1),
Of the target value DTHR is used. Next, the reaching law input Urch and the adaptive law input U
The deviation state quantity is stably placed on the switching straight line by adp.
So that the reaching law control gain F and the adaptive law control gain
Make a G decision. Specifically, assuming a disturbance V (k), the disturbance
Conditions for the switching function value σ (k) to be stable with respect to V (k)
, The conditions for setting the gains F and G are obtained.
You. As a result, the combination of gains F and G is
Satisfying (12) to (14), in other words, FIG.
Within the area indicated by the hatching is the stability condition.
Obtained. F> 0 (12) G> 0 (13) F <2- (ΔT / 2) G (14) As described above, the equations (9) to (11) give the equivalent control input.
Force Ueq (k), reaching law input Urch (k) and adaptive law input U
adp (k) and calculate the sum of those inputs as
The duty ratio DUT (k) can be calculated. The model parameter identifier 22 is described above.
(DUT (k)) and output (TH
(k)), the model parameters of the controlled object model
Calculate the vector. Specifically, model parameters
The detector 22 is a sequential type identification algorithm according to the following equation (15).
(Generalized sequential least squares algorithm)
Calculate the model parameter vector θ (k).   θ (k) = θ (k-1) + KP (k) ide (k) (15)   θ (k)^T= [A1 ', a2', b1 ', c1'] (16) Here, a1 ', a2', b1 'and C
1 'is a model pattern before executing a limit process described later.
Parameters. Also, ide (k) is given by the following equation (17):
The identification error defined by (18) and (19)
You. DTHHAT (k) is the latest model parameter vector
Throttle opening deviation calculated using the torque
The estimated value of the difference amount DTH (k) (hereinafter referred to as “estimated throttle valve opening
Deviation amount). KP (k) is calculated by the following equation (20).
Is a gain coefficient vector defined by Also, the formula
P (k) in (20) is calculated by the following equation (21).
This is a fourth-order square matrix. (Equation 2) [Equation 3] According to the setting of the coefficients λ1 and λ2 in the equation (21),
Thus, the identification algorithm according to equations (15) to (21) is
One of the following four identification algorithms:
You. λ1 = 1, λ2 = 0 Fixed gain algorithm λ1 = 1, λ2 = 1 Least squares algorithm λ1 = 1, λ2 = λ Decreasing gain algorithm (λ is
(Predetermined value other than 0, 1) λ1 = λ, λ2 = 1 Weighted least squares algorithm
(Λ is a predetermined value other than 0 and 1) On the other hand, in the present embodiment, the following B1), B
It is required to satisfy the requirements of 2) and B3). B1) Quasi-static dynamic characteristic change and hardware characteristic variation
Adaptation “Quasi-static dynamic characteristic change” refers to, for example, fluctuations in power supply voltage or
Slow characteristic change such as hardware aging
Means B2) Adaptation to fast changes in dynamic characteristics Specifically, the dynamics corresponding to the change in the throttle valve opening TH
It means adaptation to characteristic changes. B3) Drift prevention of model parameters Non-linearities of controlled objects that should not be reflected in model parameters
Caused by the effects of identification errors due to characteristics, etc.
Drift, the absolute value of model parameters
Prevent big things. First, the requirements of B1) and B2) are satisfied.
Therefore, the coefficients λ1 and λ2 are respectively set to predetermined values λ and
By setting to “0”, the weighted least squares algorithm
Adopt a gorhythm. Next, the drift of model parameters
Will be described. As shown in FIG.
After the data has converged to some extent, the friction characteristics of the throttle valve
There were residual identification errors caused by which nonlinear characteristics
Where disturbances with non-zero mean values are constantly applied.
In this case, residual identification errors accumulate and the model parameter
Cause a lift. Such a residual identification error is caused by a model parameter.
Since it should not be reflected in the data values,
A dead zone process is performed using a dead zone function Fnl as shown in FIG.
U. Specifically, the corrected identification error is calculated by the following equation (23).
idenl (k) is calculated, and the corrected identification error idenl (k) is calculated.
(k) is used to calculate the model parameter vector θ (k).
U. That is, the following equation (15) is used instead of the above equation (15).
Use a). As a result, the above requirement B3) can be satisfied.
Can be.   idenl (k) = Fnl (ide (k)) (23)   θ (k) = θ (k-1) + KP (k) idenl (k) (15a) The dead zone function Fnl is shown in FIG.
It is not limited to what is shown, for example, as shown in FIG.
Such a discontinuous dead band function, or as shown in FIG.
A simple incomplete dead band function may be used. However, incomplete
Completely prevent drift when using the zonal function
It is not possible. The residual identification error is determined by the throttle valve opening.
The amplitude changes according to the variation of TH. So the book
In the embodiment, a dead zone defining the width of the dead zone shown in FIG.
The width parameter EIDNRLMT is calculated by the following equation (24).
Of the change amount of the target throttle valve opening THR
Set according to the root mean square value DDTHRSQA (specifically
The more the mean square value DDTHRSQA increases,
The setting is made so that the sensing zone width parameter EIDNRLMT increases.
). This allows the model parameters
The identification error that should be reflected in the data
Can be prevented from being ignored. Equation (2
4) DDTHR is a change in the target throttle valve opening THR.
And is calculated by the following equation (25). (Equation 4)Here, the throttle valve opening deviation amount DTH is
Adaptive sliding mode control to target value DTHR
Is controlled by the equation (21).
Change target value DTHR to throttle valve opening deviation amount DTH
And the change amount DDTH of the throttle valve opening deviation amount DTH is
Calculated and obtained by substituting DDTH for DDTH in equation (24).
Dead band width according to the mean square value DDTHRSQA
The parameter EIDNRLMT can also be set. Further, the robustness of the control system is further improved.
For this purpose, the adaptive sliding mode controller 21
It is effective to make it more stable. Therefore, the present embodiment
Now, the model parameter calculated by the above equation (15)
Elements a1 ', a2', b1 'and
c1 'is subjected to the limit processing, and the modified model parameter
Data vector θL (k) (θL (k)^T= [A1, a2, b
1, c1]). And adaptive sliding
The mode controller 21 has a modified model parameter vector.
Sliding mode control is performed using the torque θL (k).
Run. For details on limit processing, refer to
This will be described later with reference to a chart. Next, the deviation PR predicted by the state predictor 23
A method of calculating EDTH will be described. First, the following equation (26)
According to (29), the matrices A and B and the vector X
(k) and U (k) are defined. (Equation 5) These matrices A and B and vectors X (k) and U (k)
Write the above equation (1) that defines the control target model using
In other words, the following equation (30) is obtained. X (k + 1) = AX (k) + BU (k-d) (30) When X (k + d) is obtained from Expression (30), the following is obtained.
Equation (31) is obtained. (Equation 6) Here, the model parameters a1 ', a
2 ', b1' and c1 ', the matrix A' and
When B 'is defined by the following equations (32) and (33),
The prediction vector XHAT (k + d) is given by the following equation (34).
Can be (Equation 7) The first row of the predicted vector XHAT (k + d)
The prime DTHHAT (k + d) is the predicted deviation PREDT
H (k), which is given by the following equation (35).   PREDTH (k) = DTHHAT (k + d)         = Α1 × DTH (k) + α2 × DTH (k−1)           + Β1 × DUT (k-1) + β2 × DUT (k-2) + ... + βd × DUT (k-d)           + Γ1 + γ2 +... + Γd (35) Here, α1 is the matrix A ′^d1 row and 1 column element of α2
Is the matrix A '^d1 row, 2 column elements, βi is a matrix
A '^diB ', one row and one column element, .gamma.i is the matrix A'^di
B 'is a 1-row, 2-column element. The predicted deviation amount P calculated by the equation (35)
Applying REDTH (k) to the above equation (9),
The values DTHR (k + d + 1), DTHR (k + d), and DTHR (k + d
-1) to DTHR (k), DTHR (k-1), and DT
By replacing with HR (k-2), the following equation (9a) is obtained.
can get. From equation (9a), the equivalent control input Ueq (k)
Is calculated. (Equation 8) Further, the prediction bias calculated by the equation (35)
Using the difference PREDTH (k), the following equation (36) is used.
Define the predictive switching function value σpre (k) and enter the reaching law input Ur
ch (k) and adaptive law input Uadp (k) are expressed by the following equations, respectively.
It is calculated by (10a) and (11a).   σpre (k) = (PREDTH (k) −DTHR (k−1))                 + VPOLE (PREDTH (k-1) -DTHR (k-2)) (36) (Equation 9) Next, the model parameter c 1 ′ is
Shows the deviation of the default opening THDEF and disturbance.
Parameters. Therefore, as shown in FIG.
Although it fluctuates due to disturbance, the default opening deviation is relatively
It can be regarded as almost constant within a short period. So, the real
In the embodiment, the model parameter c1 'is statistically processed, and
The central value of the fluctuation of the default opening deviation thdefadp
And the throttle valve opening deviation amount DTH and the target
The value DTHR was used for calculation. Generally, the least square method is used as a statistical processing method.
As is known, the statistical processing by the least squares method is generally performed.
Usually, data within a certain period of time, that is, the model
All parameters c1 'are stored in the memory, and
At some point, it is performed by performing a batch operation. Toko
However, this batch operation stores all data.
Requires a huge amount of memory, and the inverse matrix
The calculation is required, resulting in an increase in the amount of calculation. Therefore, in this embodiment, the above equations (15) to (15)
Iterative least squares algorithm of adaptive control shown in (21)
The rhythm is applied to statistical processing, and the model parameters c1
The least square center value is set to the default opening deviation thdefad
It is calculated as p. Specifically, the above formulas (15) to (21)
θ (k) and θ (k)^TTo thdefadp, and ζ (k) and
Bye (k)^TIs replaced with “1”, and ide (k) is placed in ec1 (k).
KP (k) is replaced with KPTH (k), and P (k) is replaced with PTH
(k), and λ1 and λ2 are respectively λ1 ′ and λ
2 ′, the following formulas (37) to (40)
Get. (Equation 10) By setting the coefficients λ1 ′ and λ2 ′,
You can select any of the four algorithms
In the equation (39), the coefficient λ1 ′ is set to a value other than 0 or 1.
By setting the coefficient to a predetermined value and setting the coefficient λ2 ′ to 1,
Weighted least squares method. In the calculations of the above equations (37) to (40),
Is the value to be stored is thdefadp (k + 1) and PTH
Only (k + 1), and no inverse matrix operation is required. did
Therefore, adopt the recursive least squares algorithm
To overcome the drawbacks of the general least-squares method,
Statistical processing of model parameter c1 by the square method
Can be. No default opening obtained as a result of statistical processing
The thdefadp is applied to the above equations (2) and (3).
Then, instead of Expressions (2) and (3), the following Expressions (41) and
According to (42), the throttle valve opening deviation amount DTH (k) and
And a target value DTHR (k) are calculated. Subtraction shown in FIG.
The units 26 and 27 calculate the following equations (41) and (42)
Is performed, the throttle valve opening deviation amount DTH (k)
And a target value DTHR (k).   DTH (k) = TH (k) -THDEF + thdefadp            = TH (k) -THDEFR (41)   DTHR (k) = THR (k) -THDEF + thdefadp              = THR (k) -THDEFR (42) Using equations (41) and (42)
Therefore, the default opening THDEF is
Deviation from the design value due to variability in characteristics or aging
Even if it does, it is necessary to compensate for the deviation and perform accurate control.
Can be. That is, the modeling error (the actual plant
Difference between the characteristics and the characteristics of the controlled model)
The reference value of the controlled object model to be
Uses R to match actual plant reference values
Can be done. As a result, modeling errors are reduced
Thus, control performance can be improved. Next, the adaptive sliding mode command described above is used.
Controller 21, model parameter identifier 22 and state
CPU of ECU 7 for realizing the function of predictor 23
The calculation processing in will be described. FIG. 9 shows the overall flow of the throttle valve opening control.
This processing is performed for a predetermined time (for example, 2 ms).
This is executed by the CPU of the ECU 7 every ec). Step S
At 11, the state variable setting process shown in FIG. 10 is executed.
That is, the operations of Expressions (41) and (42) are executed, and
Rotor valve opening deviation amount DTH (k) and target value DTHR (k)
Is calculated (FIG. 10, steps S21 and S22). What
(K) indicating the current value is omitted.
There is. In step S12, the model shown in FIG.
The operation of the parameter identifier, that is, the equation (15a)
Calculation of model parameter vector θ (k)
And execute limit processing to modify the model parameters.
The data vector θL (k) is calculated. In the following step S13, the state shown in FIG.
The calculation of the state predictor is executed, and the predicted deviation amount PREDTH (k) is calculated.
Is calculated. Next, the correction model calculated in step S12
FIG. 22 using the parameter vector θL (k).
The arithmetic processing of the control input Usl (k) is executed (step S
14). That is, the equivalent control input Ueq and the reaching law input U
rch (k) and adaptive law input Uadp (k)
The control input Usl (k) (= Dute
DUT (k)) is calculated. In the following step S16, the process shown in FIG.
Executes stability judgment processing of the riding mode controller
I do. In other words, the stability based on the differential value of the Lyapunov function
Make a determination and set the stability determination flag FSMCSTAB.
Do. This stability determination flag FSMCSTAB is "1".
Adaptive sliding mode controller when set to
21 indicates that it is unstable. Stability judgment flag
FSMCSTAB is set to “1” and the adaptive slide
When the switching mode controller 21 becomes unstable,
The switching function setting parameter VPOLE is stabilized to a predetermined value XP.
OLESTB (FIG. 24, step S231, step S231).
At the same time, the equivalent control input Ueq is set to “0”.
And only the reaching law input Urch and the adaptive law input Uadp
Control by switching to control by
(See FIG. 22, steps S206 and S208). Adaptation
Sliding mode controller 21 becomes unstable
The reaching law input Urch and the adaptive law input U
Change the formula for calculating adp. That is, the reaching law control gay
Value of the adaptive law control gain G
To a value that stabilizes
The reaching law input Urch and the adaptive law without using the
The input Uadp is calculated (see FIGS. 27 and 28). More than
Adaptive stabilizing mode control
The unstable state of the controller 21 is terminated early,
You can return to the state. In step S17, FIG.
Execute the thdefadp calculation process shown in
The opening deviation thdefadp is calculated. FIG. 11 shows the model parameter identifier 22.
It is a flowchart of a calculation process. In step S31
Calculates the gain coefficient vector KP (k) by equation (20).
And then, using equation (18), the estimated throttle valve opening deviation
The difference DTHHAT (k) is calculated (step S32).
In step S33, the performance of idenl (k) shown in FIG.
Calculation processing and calculate the estimated slot calculated in step S32.
Tolerance DTHHAT (k) is applied to equation (17).
The calculation of the identification error ide (k) using FIG.
Perform dead zone processing using the function shown in FIG.
Calculate the error idenl. In the following step S34, equation (15a)
To calculate the model parameter vector θ (k),
Executes stabilization processing of model parameter vector θ (k)
(Step S35). That is, each model parameter
Modified model parameter vector
Calculate θL (k). FIG. 12 is executed in step S33 of FIG.
Is a flowchart of idenl (k) calculation processing performed.
You. In step S51, the identification error i is calculated according to equation (17).
Calculate de (k). Next, in step S53, the ink
The value of the counter CNTIDST being incremented is
Predetermined value XCNTID set according to elephant dead time d
ST (for example, set to “3” corresponding to dead time d = 2)
Is determined (step S5).
2). The initial value of the counter CNTIDST is “0”
Therefore, the process first proceeds to step S53, and the counter CNTI
DST is incremented by “1” and the identification error ide
(k) is set to "0" (step S54), and
Proceed to S55. Identification of model parameter vector θ (k)
Immediately after the start of the calculation, the correct identification
Since no error is obtained, steps S52 to S54
In other words, the identification error id is calculated without using the calculation result obtained by the equation (17).
e (k) is set to “0”. The answer to step S52 is affirmative (YES).
Then, the process immediately proceeds to step S55. In step S55
Performs low-pass filtering of the identification error ide (k).
U. Specifically, the model of the control target having low-pass characteristics
Parameter identification, the least squares identification algorithm
The identification weight of the system with respect to the identification error ide (k) is shown in FIG.
(A) has a frequency characteristic as shown by a solid line L1,
This is indicated by a broken line L2 by the low-pass filter processing.
The characteristic is such that high frequency components are attenuated. This is
For reasons. An actual control object having a low-pass characteristic;
The frequency characteristics of the control target model that models this
As shown by solid lines L3 and L4 in FIG.
You. In other words, low-pass characteristics
Model parameters for controlled objects with
When the model parameters are identified by the
Model parameters greatly affect high-frequency rejection
Control model in the low frequency range.
In is lower than the actual characteristics. As a result,
Of the control input by the
It becomes correction. Therefore, identification is performed by low-pass filter processing.
The frequency characteristics of the algorithm weights are broken as shown in FIG.
By setting the characteristic as indicated by the line L2,
The frequency characteristic of Dell is shown by a broken line L5 in FIG.
Characteristics, match the actual frequency characteristics, or
The low-frequency gain of the target model is slightly higher than the actual gain.
It was decided to fix it. This allows the control
To prevent overcorrection by the controller 21 and improve the robustness of the control system.
The control system can be further stabilized by increasing the value. It should be noted that the low-pass filter processing is performed for the identification error.
Past value ide (k-i) (for example, 1 corresponding to i = 1 to 10)
0 past values) are stored in the ring buffer.
It is executed by multiplying the last value by a weighting factor and adding it.
You. Further, the identification error ide (k) is calculated by the equation (17),
Since it is calculated using (18) and (19),
The throttle valve opening deviation amount DTH (k) and the estimated throttle valve opening
Low-pass filter similar to the degree deviation DTHHAT (k)
Processing, or the throttle valve opening deviation amount D
TH (k-1) and DTH (k-2), and the duty ratio DUT (k
-d-1) to perform the same low-pass filter processing.
The same effect can be obtained. Returning to FIG. 12, in the following step S56,
The dead zone process shown in FIG. 14 is executed. Steps in FIG.
In S61, for example, n = 5 in the equation (24) is set.
Mean square value of the amount of change in the target throttle valve opening THR
Calculate DDTHRSQA and then calculate the root mean square value DDTH
EIDNRLMT table shown in FIG. 15 according to RSQA
Search for the dead band width parameter EIDNRLMT
It is issued (step S62). In step S63, the identification error ide (k)
Is greater than the dead zone parameter EIDNLLMT?
And if ide (k)> EIDNRLMT
Is the corrected identification error idenl by the following equation (43).
(k) It is calculated (step S67).   idlen (k) = ide (k) -EIDNRLMT (43) The answer to step S63 is negative (NO)
In this case, the identification error ide (k) is further reduced by the dead band width parameter.
Whether EIDNRLMT is smaller than the value with a minus sign
Is determined (step S64), and ide (k) <− EIDN
If it is RLMT, the modified equation is obtained by the following equation (44).
The constant error idenl (k) is calculated (step S65).   idlen (k) = ide (k) + EIDNRLMT (44) Further, the identification error ide (k) is within the range of ± EIDNRLMT.
, The modified identification error idenl (k) is set to “0”.
(Step S66). FIG. 16 is executed in step S35 of FIG.
9 is a flowchart of a process of stabilizing θ (k). S
In step S71, a flag FA1 used in this processing is set.
STAB, FA2STAB, FB1LMT and FC1L
Initialization is set by setting MT to “0”, respectively.
Do. Then, in step S72, a shown in FIG.
The limit processing of 1 'and a2' is executed, and step S7
In 3, the limit processing of b1 'shown in FIG.
In step S74, the limit processing of c1 'shown in FIG.
Execute the process. FIG. 17 is executed in step S72 of FIG.
A1 'and a2' limit processing flow chart
It is. FIG. 18 is a diagram for explaining the processing of FIG.
This is referred to with FIG. In FIG. 18, a limit process is required.
The combination of model parameters a1 'and a2' is "x"
And stable model parameters a1 'and a1'
The range of the combination of 2 ′ is a hatched area (hereinafter referred to as a hatched area)
The "stable region" below). Processing of FIG.
Are the model parameters a1 'and a
The combination of 2 ′ is within the stable region (the position indicated by “○”).
This is the process of moving to. At step S81, model parameters a
2 ′ is equal to or larger than a predetermined a2 lower limit value XIDA2L
I do. The predetermined a2 lower limit value XIDA2L is larger than “−1”.
Set to a negative value. Predetermined a2 lower limit value XIDA2L
Is a stable modified model parameter even if it is set to -1.
Are obtained, which are defined by the above equation (26).
Matrix n becomes unstable (this is due to a1 'and a1'
2 'does not diverge but vibrates)
Therefore, the value is set to a value larger than “−1”. In step S81, if a2 '<XIDA2L
In some cases, the modified model parameter a2 is
XIDA2L and a2 stabilization flag F
Set A2STAB to “1”. a2 stabilization flag F
When A2STAB is set to “1”, the modified model parameter
Indicates that the meter a2 has been set to the lower limit value XIDA2L.
You. In FIG. 18, steps S81 and S82
Modification of model parameters by the mitt process P1 is "P
This is indicated by an arrow line with an “1” (a line with an arrow). If the answer to step S81 is affirmative (YES),
That is, when a2 '≧ XIDA2L, the modified model
Parameter a2 is set to model parameter a2 '
(Step S83). Step S84 and step S8
5, the model parameter a1 'is set to a predetermined a1 lower limit value X
Range where IDA1L and predetermined a1 upper limit value XIDA1H can be set
It is determined whether it is within. Predetermined a1 lower limit value XIDA1
L is set to a value greater than or equal to −2 and smaller than 0,
One upper limit value XIDA1H is set to 2, for example. If the answers in steps S84 and S85 are both
When affirmative (YES), that is, XIDA1L ≦ a
When 1 ′ ≦ XIDA1H, the modified model parameter
The parameter a1 is set to the model parameter a1 '(step
S88). On the other hand, when a1 '<XIDA1L,
Set the modified model parameter a1 to the lower limit value XIDA1L
And a1 stabilization flag FA1STAB is set.
It is set to "1" (steps S84, S86). And a
1 ′> XIDA1H, the modified model parameter
A1 is set to the upper limit value XIDA1H, and a1
Set the stabilization flag FA1STAB to “1” (step
Steps S85 and S87). a1 stabilization flag FA1STA
When B is set to “1”, the modified model parameter a
1 as the lower limit value XIDA1L or the upper limit value XIDA1H.
Indicates that it has been set. In FIG. 18, step S84
Of model parameters by the limit processing P2 of S87
The correction is indicated by the arrow with “P2”. In step S90, the modified model parameters
The sum of the absolute value of the parameter a1 and the corrected model parameter a2 is
It is determined whether or not the value is equal to or less than a fixed stability determination value XA2STAB.
You. The predetermined stability determination value XA2STAB is close to “1”.
The value is set to a value smaller than “1” (for example, 0.99). The straight lines L1 and L2 shown in FIG.
This is a straight line that satisfies (45). a2 + | a1 | = XA2STAB (45) Therefore, step S90 is performed by using the modified model parameters
The combination of a1 and a2 corresponds to the straight line L1 and the line L1 shown in FIG.
And whether it is on or below the line of L2
You. When the answer to step S90 is affirmative (YES)
Is the combination of the modified model parameters a1 and a2
Is within the stable region of FIG.
Complete. On the other hand, if the answer to step S90 is negative (NO),
In some cases, the corrected model parameter a1 is
From the fixed value XA2STAB, the predetermined a2 lower limit value XIDA2L
The subtracted value (XIDA2L <0, so XA2ST
AB-XIDA2L> XA2STAB holds)
It is determined whether or not this is the case (step S91). And the correction model
Parameter a1 is (XA2STAB-XIDA2L)
If the following, the modified model parameter a2 is set to (XA
2STAB- | a1 |) and a2 stable
The flag FA2STAB to “1” (step
S92). At step S91, the modified model parameter a
When 1 is greater than (XA2STAB-XIDA2L)
Sets the modified model parameter a1 to (XA2STAB-X
IDA2L) and set the modified model parameter a2
A2 lower limit XIDA2L
The stabilization flag FA2 and the stabilization flag FA2
STAB are both set to "1" (step S9)
3). In FIG. 18, steps S91 and S91
Correction of model parameters by 92 limit processing P3
Is indicated by an arrow with “P3”, and
Model by the limit processing P4 in steps S91 and S93.
Parameter modifications are indicated by arrows with “P4”
You. As described above, the model shown in FIG.
Parameters a1 'and a2' are within the stable region shown in FIG.
Limit processing is performed so that the
Meters a1 and a2 are calculated. FIG. 19 is executed in step S73 of FIG.
It is a flowchart of the limit processing of b1 'performed.
In steps S101 and S102, the model parameters
b1 'is a predetermined b1 lower limit value XIDB1L and a predetermined b1 upper limit value
It is determined whether or not the value XIDB1H is within a range that can be determined.
You. The predetermined b1 lower limit value XIDB1L is a positive predetermined value (for example,
0.1) and the predetermined b1 upper limit value XIDB1H
Is set to, for example, “1”. No answer in steps S101 and S102
When these are also affirmative (YES), that is, XIDB1L
≦ b1 ′ ≦ XIDB1H, the modified model parameter
The meter b1 is set to the model parameter b1 '(the
Step S105). On the other hand, if b1 '<XIDB1L
The modified model parameter b1 is set to the lower limit value XIDB1L.
And the b1 limit flag FB1LMT
Is set to "1" (steps S101, S104).
When b1 '> XIDB1H, the modified model
Set the parameter b1 to the upper limit value XIDB1H
, The b1 limit flag FB1LMT is set to “1”.
(Steps S102 and S103). b1 limit hula
When the FB1LMT is set to “1”, the modified model
Parameter b1 is set to lower limit value XIDB1L or upper limit value XI
DB1H is set. FIG. 20 is executed in step S74 of FIG.
Of the limit processing of the model parameter c1 'to be performed
It is a chart. In steps S111 and S112,
When the model parameter c1 'is a predetermined c1 lower limit value XIDC1
L and the predetermined c1 upper limit value XIDC1H
It is determined whether or not. The predetermined c1 lower limit value XIDC1L is an example.
For example, it is set to -60, and the predetermined c1 upper limit value XIDC1H
Is set to 60, for example. No answer in steps S111 and S112
When these are also affirmative (YES), that is, XIDC1L
≦ c1 ′ ≦ XIDC1H, the modified model parameter
The meter c1 is set to the model parameter c1 '(the
Step S115). On the other hand, if c1 '<XIDC1L,
When the modified model parameter c1 is set to the lower limit value XIDC1L
And the c1 limit flag FC1LMT
Is set to "1" (steps S111, S114).
When c1 '> XIDC1H, the modified model
Set parameter c1 to the upper limit value XIDC1H
, Set the c1 limit flag FC1LMT to “1”.
(Steps S112 and S113). c1 limit hula
FC1LMT is a modified model when set to "1".
The parameter c1 is set to the lower limit XIDC1L or the upper limit XI.
Indicates that DC1H has been set. FIG. 21 is a flowchart showing the processing executed in step S13 of FIG.
6 is a flowchart of a calculation process performed by a state predictor. S
In step S121, a matrix operation is performed to
(35) matrix elements α1, α2, β1 and β2, and γ1
Γγd is calculated. In step S122, equation (35)
To calculate the predicted deviation amount PREDTH (k). FIG. 22 is a flowchart showing the processing executed in step S14 of FIG.
Control input Usl to the throttle valve drive 10
It is a flowchart of the process which calculates (= DUT).
In step S201, the prediction switching function value σ shown in FIG.
pre is performed, and in step S202,
Calculation of integrated value of predicted switching function value σpre shown in 26
Execute In step S203, the equation (9) is used.
Then, the equivalent control input Ueq is calculated. Step S204
Now, the calculation process of the reaching law input Urch shown in FIG.
In step S205, the adaptive law input shown in FIG.
Executes Uadp arithmetic processing. In step S206, the processing shown in FIG.
The stability determination flag FSMCSTAB set in the process is
It is determined whether it is "1". Stability determination flag FSM
When CSTAB is set to "1", the adaptive slide
That the switching mode controller 21 is unstable
Is shown. In step S206, FSMCSTAB = 0
And the adaptive sliding mode controller 21
If it is stable, calculate in steps S203 to S205
Control inputs Ueq, Urch and Uadp
To calculate the control input Usl (step S
207). On the other hand, when FSMCSTAB = 1 and the adaptive
Riding mode controller 21 becomes unstable
When there is, reaching law input Urch and adaptive law input Uad
The sum of p is calculated as the control input Usl. That is,
The equivalent control input Ueq is used to calculate the control input Usl.
Not to be. This can cause the control system to become unstable.
And can be prevented. In the following steps S209 and S210,
The calculated control input Usl is equal to a predetermined upper / lower limit XUSLH and
It is determined whether it is within the range of XUSLL, and the control input U
If sl is within the range of the predetermined upper and lower limits, immediately
End the process. On the other hand, when the control input Usl is equal to the predetermined lower limit XU
When the value is equal to or less than SLL, the control input Usl is set to a predetermined lower limit value.
XUSLL is set (steps S209 and S212),
When the control input Usl is equal to or more than the predetermined upper limit value XUSLH
Sets the control input Usl to a predetermined upper limit value XUSLH
(Steps S210, S211). FIG. 23 is a flowchart showing the operation performed in step S201 of FIG.
Float of calculation processing of predicted switching function value σpre to be executed
It is a chart. In step S221, the disconnection shown in FIG.
Execute the calculation process of the conversion function setting parameter VPOLE,
Next, the predictive switching function value σpre is calculated by the equation (36).
The calculation of (k) is executed (step S222). In the following steps S223 and S224,
The calculated prediction switching function value σpre (k) is equal to the predetermined upper / lower limit X
Determines if it is within the range of SGMH and XSGML
And the predicted switching function value σpre (k) is within a predetermined upper / lower range.
If the value is within the range, the present process is immediately terminated. Meanwhile, prediction
When the switching function value σpre (k) is equal to or less than a predetermined lower limit value XSGML,
In some cases, the predicted switching function value σpre (k) is
XSGML (steps S223 and S225),
The predicted switching function value σpre (k) is equal to or less than a predetermined upper limit value XSGMH.
When it is above, the predicted switching function value σpre (k) is
Limit value XSGMH (steps S224 and S22).
6). FIG. 24 is a flowchart showing the operation performed in step S221 of FIG.
Calculation processing of the executed switching function setting parameter VPOLE
It is a flowchart of FIG. In step S231, the stable
It is determined whether the determination flag FSMCSTAB is "1".
Separately, FSMCSTAB = 1 and adaptive sliding
When the mode controller 21 is unstable
Sets the switching function setting parameter VPOLE to a stabilized predetermined value.
It is set to XPOLESTB (step S232). Cheap
The fixed predetermined value XPOLESTB is larger than “−1”.
Set to a value very close to "-1" (eg -0.999)
Is done. When FSMCSTAB = 0 and the adaptive slide
When the sliding mode controller 21 is stable,
The change amount DD of the target value DTHR (k) is calculated by the following equation (46).
THR (k) is calculated (step S233).   DDTHR (k) = DTHR (k) -DTHR (k-1) (46) In step S234, the throttle valve opening
Deviation amount DTH and target value calculated in step S233
VPOLE map according to the amount of change DDTHR
Then, a switching function setting parameter VPOLE is calculated. V
The POLE map, as shown in FIG.
When the torque valve opening deviation amount DTH takes a value near 0 (slope
The throttle valve opening TH is a value near the default opening THDEF
Increases), and the throttle valve increases at values other than near 0.
It becomes a substantially constant value with respect to the change in the opening deviation amount DTH.
It is set as follows. The VPOLE map is the same
As shown by the solid line in (b), the target value change amount DDTHR
Is set so that the VPOLE value increases as
However, the throttle valve opening deviation amount DTH is close to zero.
When taking the value, change the target value as shown by the broken line in the figure.
Increase when the oxidation amount DDTHR takes a value near 0.
Is set to That is, the target value DT of the throttle valve opening degree
When the change in the direction of decrease in HR is large, the switching function is set.
Parameter VPOLE is set to a relatively small value.
As a result, the throttle valve 3 is fully closed
Collision can be prevented. Also, default
In the vicinity of the tilt opening THDEF, the switching function setting parameter
Meter VPOLE is set to a relatively large value,
The controllability near the tilt opening THDEF can be improved.
Can be. Note that, as shown in FIG.
The valve opening TH is near the fully closed position or near the fully opened position
At this time, the switching function setting parameter VPOLE is decreased.
May be set. This opens the throttle valve
When the degree TH is in the vicinity of the fully closed position or in the vicinity of the fully opened position
Means that the speed following the target opening THR is slow,
Fully-closed stopper for the throttle valve 3
Function) can be more reliably prevented
You. In the following steps S235 and S236,
The calculated switching function setting parameter VPOLE is up or down
Whether it is within the limits XPOLEH and XPOLEL
Is determined, and the switching function setting parameter VPOLE is
If the value is within the range of the upper and lower limits, the process is immediately terminated.
You. On the other hand, the switching function setting parameter VPOLE is lower than a predetermined value.
When the value is equal to or less than the limit value XPOLEL, the switching function setting parameter is set.
Set the meter VPOLE to the predetermined lower limit XPOLEL
(Steps S236, S238), switching function setting parameters
Data VPOLE is greater than or equal to a predetermined upper limit XPOLEH.
The switching function setting parameter VPOLE is set to a predetermined upper limit value.
Set to XPOLEH (steps S235, S23)
7). FIG. 26 is a flowchart showing the operation performed in step S202 of FIG.
Integrated value SUMSI of predicted switching function value σpre
It is a flowchart of a process of calculating GMA. Integrated value
SUMSIGMA enters an adaptive rule in the process of FIG.
Used to calculate the force Uadp (see the above equation (11a)).
See). In step S241, the following equation (47) is used.
Then, the integrated value SUMSIGMA is calculated. Δ in the following equation
T is the execution cycle of the operation.   SUMSIGMA (k) = SUMSIGMA (k-1) + σpre × ΔT (47) In the following steps S242 and S243, the calculated integration
The value SUMSIGMA is a predetermined upper and lower limit value XSUMSH and X
It is determined whether or not the sum is within the range of SUMSL, and the integrated value SU is determined.
When MSIGMA is within the range of predetermined upper and lower limits,
This process is terminated at a later time. On the other hand, the integrated value SUMSIGMA
Is less than or equal to a predetermined lower limit value XSUMSL, the integrated value S
UMSIGMA is set to a predetermined lower limit value XSUMSL (
Steps S242 and S244), integrated value SUMSIGMA
Is greater than or equal to the predetermined upper limit value XSUMSH, the integrated value S
Set UMSIGMA to a predetermined upper limit value XSUMSH
(Steps S243 and S245). FIG. 27 is a flowchart showing the operation performed in step S204 of FIG.
Flowchart of Arithmetic Processing of Arrival Law Input Urch to be Executed
It is. In step S261, the stability determination flag FS
It is determined whether or not MCSTAB is “1”. Stability
If the separate flag FSMCSTAB is "0" and the adaptive slide
When the sliding mode controller 21 is stable,
The control gain F is set to a predetermined normal gain XKRCH (the
Step S262), the following equation (48) (the above equation (10a)
Is calculated by the same formula as above).
(Step S263). Urch = −F × σpre / b1 (48) On the other hand, when the stability determination flag FSMCSTAB is
"1" for adaptive sliding mode controller
21 becomes unstable when the control gain F
Set to the stabilization gain XKRCHSTB (step S26)
4), the following equation (49) without using the model parameter b1
Is used to calculate the reaching law input Urch (Step S26)
5). Urch = −F × σpre (49) In the following steps S266 and S267,
The calculated reaching law input Urch is equal to a predetermined upper / lower limit value XURCH
H and XURCHL are determined to be within the range.
When the reaching rule input Urch is within the range of predetermined upper and lower limits
Ends this process immediately. On the other hand, the reaching law input Urc
When h is equal to or less than a predetermined lower limit value XURCHL, the reaching law
The input Urch is set to a predetermined lower limit value XURCHL (step
Steps S266, S268), reaching law input Urch is predetermined
If it is not less than the upper limit value XURCHH, the reaching law input Ur
is set to a predetermined upper limit value XURCHH (step S
267, S269). As described above, the adaptive sliding mode control
When the controller 21 becomes unstable, the control gain F
Set the predetermined stabilization gain XKRCHSTB and
, The reaching law input U without using the model parameter b1
By calculating rch, the adaptive sliding mode
Controller 21 can be returned to a stable state.
Identification by the model parameter identifier 22 becomes unstable
The adaptive sliding mode controller 21
Becomes unstable, so the unstable model parameters
By not using b1, adaptive sliding mode
Controller 21 can be stabilized. FIG. 28 is a flowchart showing the operation performed in step S205 of FIG.
Of the adaptive law input Uadp to be executed
It is. In step S271, the stability determination flag FS
It is determined whether or not MCSTAB is “1”. Stability
If the separate flag FSMCSTAB is "0" and the adaptive slide
When the sliding mode controller 21 is stable,
The control gain G is set to a predetermined normal gain XKADP (the
Step S272), the following equation (50) (the above equation (11a)
The adaptive law input Uadp is calculated by the equation corresponding to
(Step S273).   Uadp = −G × SUMSIGMA / b1 (50) On the other hand, the stability determination flag FSMCSTAB is
"1" for adaptive sliding mode controller
When 21 becomes unstable, the control gain G is reduced to a predetermined level.
Set to the stabilization gain XKADPSTB (step S27)
4), the following equation (51) without using the model parameter b1
Is used to calculate the adaptive law input Uadp (step S27)
5).   Uadp = −G × SUMSIGMA (51) As described above, the adaptive sliding mode control
When the controller 21 becomes unstable, the control gain G
Set the predetermined stabilization gain XKADPSTB and
The adaptive law input U without using the model parameter b1
By calculating adp, the adaptive sliding mode
Controller 21 can be returned to a stable state. FIG. 29 is a flowchart of the processing executed in step S16 of FIG.
Processing of the sliding mode controller
It is a flowchart of FIG. In this process, the Lyapunov
Performs stability judgment based on the differential term of the number, and responds to the stability judgment result.
Then, the stability determination flag FSMCSTAB is set. In step S281, the following equation (52) is used.
Then, the switching function change amount Dσpre is calculated, and then the following equation is calculated.
According to (53), the stability determination parameter SGMSTAB is obtained.
Is calculated (step S282).   Dσpre = σpre (k) −σpre (k-1) (52)   SGMSTAB = Dσpre × σpre (k) (53) In step S283, the stability determination parameter is set.
Data SGMSTAB is the stability judgment threshold value XSGMSTAB
It is determined whether or not SGMSTAB> XSGMSTA
If B, controller 21 may be unstable
The instability detection counter CNTSMCS
T is incremented by "1" (step S28)
5). Also, SGMSTAB ≦ XSGMSTAB
When it is determined that the controller 21 is stable,
The count value of the fixed detection counter CNTSMCST is
It is held without being incremented (step S284). In step S286, the instability detection count
The value of CNTSMCST is a predetermined count value XSSTAB
It is determined whether or not: CNTSMCST ≦ XSSTA
B, it is determined that the controller 21 is stable.
And set the first determination flag FSMCSTAB1 to "0".
(Step S287). On the other hand, CNTSMCST>
If XSSTAB, controller 21 is unstable
And the first determination flag FSMCSTA
B1 is set to "1" (step S288). In addition,
The instability detection counter CNTSMCST is set to the ignition
When the switch is turned on, the count value is initialized to "0".
It is. In the following step S289, a stability determination period
Decrement the counter CNTJUDST by "1"
And then the stability determination period counter CNTJUDST
Is determined whether or not the value of is “0” (step S29)
0). The stability determination period counter CNTJUDST
Predetermined discrimination count value XCJ when Nishon switch is on
Initialized to UDST. So, first step
The answer to S290 is negative (NO), and the process immediately proceeds to step S290.
Proceed to 295. Thereafter, the stability determination period counter CNTJUD
When ST becomes “0”, the process proceeds from step S290 to step S290.
Proceeding to S291, the first determination flag FSMCSTAB1 is cleared.
It is determined whether it is "1". Then, the first determination flag
If FSMCSTAB1 is "0", the second determination
The flag FSMCSTAB2 is set to "0" (step
S293), the first determination flag FSMCSTAB1 is
If it is "1", the second determination flag FSMCSTAB
2 is set to "1" (step S292). In the following step S294, a stability determination period
The value of the counter CNTJUDST is set to a predetermined determination count value X.
Set to CJUDST and set the instability detection counter
The value of CNTSMCST is set to “0”, and step S2
Go to 95. In step S295, the stability determination flag F
SMCSTAB is set to a first determination flag FSMCSTAB1
And the second judgment flag FSMCSTAB2
You. The second determination flag FSMCSTAB2 is determined in step S
The answer to 286 is affirmative (YES), and the first determination flag F
Even if SMCSTAB1 is set to "0", the stability determination period
Until the value of the interval counter CNTJUDST becomes "0"
Is maintained at “1”. Therefore, the stability determination flag
FSMCSTAB is also a stability determination period counter CNTJU
Until the value of DST becomes "0", it is maintained at "1".
You. FIG. 30 is a flowchart of the processing executed in step S17 of FIG.
Of the default opening deviation thdefadp
It is a flowchart. In step S251, the following expression
The gain coefficient KPTH (k) is calculated by (54).   KPTH (k) = PTH (k-1) / (1 + PTH (k-1)) (54) Here, PTH (k-1) is a step at the time of the previous execution of this processing.
This is the gain parameter calculated in S253. In step S252, the model shown in FIG.
Model parameters calculated by the parameter
Data c1 'and the gain coefficient calculated in step S251.
Apply KPTH (k) to the following equation (55) to open the default
The degree deviation thdefadp (k) is calculated.   thdefadp (k) = thdefadp (k-1)                       + KPTH (k) × (c1′-thdefadp (k−1))                                                               (55) In step S253, the following equation (56) is used.
Then, a gain parameter PTH (k) is calculated.   PTH (k) = (1-PTH (k-1) / (XDEFADPW + PTH (k-1)))                 × PTH (k-1) / XDEFADPW (56) Equation (56) is obtained by calculating λ1 ′ and λ in equation (39).
2 ′ to predetermined values XDEFADPW and “1”, respectively.
It is set. By the processing in FIG. 30, the model parameters c
1 'is statistically processed by the sequential weighted least squares method,
The default opening deviation thdefadp is calculated. Book
In the embodiment, the throttle valve driving device 10 and the ECU 7
(The output circuit that supplies the drive current to the motor 6)
The model parameter identifier 22 corresponds to the runt
Corresponding to the step, adaptive sliding mode controller 2
1 corresponds to the control unit, and the reference value setting unit 25 corresponds to the correction unit.
Equivalent to. More specifically, the processing of FIG.
22 corresponds to the control means, and the processing in FIG.
The processing constitutes a part of the correction means. (Second Embodiment) FIG. 31 shows a second embodiment of the present invention.
Control device of plant according to second embodiment, ie, oil
FIG. 2 is a diagram illustrating a configuration of a pressure positioning device and a control device thereof.
You. Such a hydraulic positioning device is used, for example, for an internal combustion engine.
Continuous to continuously change the valve timing of intake and exhaust valves
Used for variable valve timing mechanisms. Continuously variable valve
The timing mechanism controls the rotation of the cam that drives the intake and exhaust valves.
By changing the phase, the opening and closing timing of the intake and exhaust valves
Shift, improve filling efficiency and reduce pumping loss
Things. The hydraulic positioning device comprises a piston 64 and a piston.
The hydraulic cylinder 61 in which the stone 64 is fitted, and the electric
Valve 67, hydraulic pump 65, and hydraulic pump 65
Hydraulic supply path 66 for supplying hydraulic pressure to electric spool valve 67
And the first hydraulic pressure P1 is applied to the first hydraulic chamber 6 of the hydraulic cylinder 61.
2 and the second oil pressure P2
Second oil passage 6 to be supplied to second hydraulic chamber 63 of cylinder 61
9 and the operating oil discharged from the electric spool valve 67
A hydraulic discharge path 70 for returning to lupine (not shown).
You. Further, the position PACT of the piston 64 is detected.
A potentiometer 71 is provided at the detection position P.
A signal indicating ACT is sent to the electronic control unit (ECU) 72.
Supplied. The target position PCMD is input to the ECU 72.
The ECU 72 sets the detection position PACT to the target position PC.
The control amount DUT is calculated so as to match the MD, and the control amount is calculated.
Supply an electric signal corresponding to the quantity DUT to the electric spool valve 67
I do. The electric spool valve 67 responds to the control amount DUT.
Move the position of the valve element (not shown)
The corresponding first and second hydraulic pressures P1 and P2 are output. First
And the differential pressure DP between the second hydraulic pressures P1 and P2 (= P1-P2)
Is positive, the piston 64 moves to the right in the figure.
When the differential pressure DP is a negative value, the piston 6
4 moves leftward in the figure. Detection position PACT is the target position
The differential pressure DP is maintained at “0” when the pressure matches the PCMD.
Be held. FIG. 14 shows the hydraulic positioning device shown in FIG.
Control using adaptive sliding mode controller
FIG. 3 is a block diagram illustrating a configuration of a control system when the control is performed. control
The device 80 comprises an identifier 81 and an adaptive sliding mode.
Controller 82, predictor 83, reference value setting unit 84
And the subtractors 85 and 86.
This is realized by the arithmetic processing executed in U. The reference value setting section 84 is the same as in the first embodiment.
The model parameter c identified by the identifier 81
1 ′ and the following equations (37a) and (40a).
Thus, a correction value pbadp of the reference value PBASE is calculated.   pbadp (k + 1) = pbadp (k) + KPTH (k) ec1 (k) (37a)   ec1 (k) = c1 '(k) -pbadp (k) (40a) Further, the reference value setting section 84 sets the correction value p
The reference value PBASE is corrected by badp and the corrected reference value is corrected.
Calculate PBSR (= PBASE-pbadp). Decrease
The calculator 85 calculates the correction reference value PBSR from the detection position PACT.
Is subtracted to calculate the detected position deviation amount DPACT.
The subtractor 86 calculates the correction reference value from the target position PCMD.
The target value DPCMD is calculated by subtracting the PBSR.
Put out. It should be noted that the reference value PBASE is a hydraulic positioning device.
Is set in advance to an optimum value based on the operation characteristics of the. The identifier 81 has a control amount D as a control input.
UT and detection position deviation DPACT as control output
Accordingly, the model parameter identifier 22 of the first embodiment
Similarly, the modified model parameter vector θL (k) is calculated.
Put out. That is, the identification error ide (k) is calculated by the following equation (17)
a) and (18a). Where
The parameter vector ζ (k) is defined by the following equation (19a)
Is done. Then, the identification error ide (k) is calculated by the above equation (15).
Apply and use (20) and (21) above
Is used to calculate the model parameter vector θ (k).
It is. Further calculated model parameter vector θ
For (k), the same limit processing as in the first embodiment is performed.
By doing so, the corrected model parameter θL (k) is calculated
Is done.   ide (k) = DPACT (k) -DPACTHAT (k) (17a)   DPACTHAT (k) = θ (k-1)^Tζ (k) (18a)   ζ^T(k) =     [DPACT (k-1), DPACT (k-2), DUT (k-d-1), 1] (19a) [0131] The estimator 83 operates in the same manner as in the first embodiment.
Similarly to the state predictor 23, the prediction is performed by the following equation (35a).
The deviation amount PREDPACT (k) is calculated.   PREDPACT (k)         = Α1 × DPACT (k) + α2 × DPACT (k-1)           + Β1 × DUT (k-1) + β2 × DUT (k-2) + ... + βd × DUT (k-d)           + Γ1 + γ2 +... + Γd (35a) Adaptive Sliding Mode Controller 8
2 is the adaptive sliding mode controller of the first embodiment.
Similarly to the roller 21, the detected position deviation amount DPACT and
Applying the predicted deviation PREREDACT to the following equation (9b)
Thus, the equivalent control input Ueq (k) is calculated. Further adaptation
The riding mode controller 82 uses the following equation (36)
a) calculates a prediction switching function value σpre (k) according to
Equations (10a) and (11a) show the predicted switching function value σpre
Applying (k), reaching law input Urch (k) and adaptive law input
Uadp (k) is calculated.   Ueq (k) = (1 / b1) {(1-a1-VPOLE) PREDPACT (k)               + (VPOLE-a2) PREDPACT (k-1)               -C1 + DPACT (k) + (VPOLE-1) DPACT (k-1)               −VPOLE × DPACT (k−2)} (9b)   σpre (k) = (PREDPACT (k) -DPCMD (k-1))                 + VPOLE (PREDPACT (k-1) -DPCMD (k-2))                                                             (36a) Therefore, according to the control device 80, the first
Output TH and target opening degree THR in the embodiment of FIG.
To the control output PACT and the target position PCMD, respectively.
Since the replaced control is executed, it is the same as the first embodiment.
Control output PACT follows target position PCMD
Is controlled with good robustness. Further, the reference value PBASE is input to the plant.
Correction according to model parameter c1 'not related to output
The correction reference value PBSR obtained by performing
Therefore, even if the reference value PBASE shifts due to a change over time,
The reference value of the controlled model is
Control performance by reducing modeling errors
Can be improved. In this embodiment, the hydraulic positioning device shown in FIG.
Is equivalent to a plant, and the control device 80 controls the plant.
Corresponds to the device. Also, the identifier 81 corresponds to an identification unit,
The adaptive sliding mode controller 82 controls the
, And the reference value setting unit 84 corresponds to a correction unit. The present invention is limited to the above embodiment.
Instead, various modifications are possible. For example, as described above
The default opening THDEF in the embodiment is the first opening THDEF.
The urging force of the return spring 4 as the urging means;
A slot that balances the urging force of the elastic member 5 as the urging means
Although it was defined as a throttle valve opening, Japanese Patent Application Laid-Open No. 9-72231
As shown in the publication, the driving force is applied to the throttle valve.
If not, use the intermediate lever stopper to
Throttle valve driving device having a structure in which the valve is maintained at a predetermined opening.
When the position is to be controlled, the intermediate lever stopper
The throttle opening determined by the default opening TH
What is necessary is just to define it as DEF. Further, in the second embodiment, the hydraulic positioning is performed.
Pneumatic system using pneumatic pressure instead of hydraulic pressure
Regarding the positioning device, the control device shown in the second embodiment
The control by the device 80 may be applied. Also supplement
The positive reference value THREFR or PBSR is the model parameter.
Data in the adaptive sliding mode.
Since parameters unique to the controller are not involved,
The controller is an adaptive sliding mode controller
Sliding mode that does not use adaptive law input
Controller and other controllers, for example, as well
Control and design that can be designed for
Same as sliding mode control by adjusting parameters
Backstepping control to achieve various control results
Performing modern and / or robust control of
It may be. [0138] As described in detail above, according to the present invention,
The model parameters of the target model are the input and output of the plant.
Model parameters that are not relevant to
Based on model parameters that are not related to output,
The reference value is corrected. In other words, the model of the control target model
Since the reference value is corrected by the
The description is based on what control method the control device uses.
Even applicable. Moreover, there is a direct effect on the modeling error.
The reference value of the controlled object model
Modeling error,
And control performance can be improved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a throttle valve driving device for an internal combustion engine and a control device thereof according to a first embodiment of the present invention. FIG. 2 is a diagram showing a frequency characteristic of the throttle valve driving device shown in FIG. FIG. 3 is a functional block diagram showing functions realized by an electronic control unit (ECU) of FIG. FIG. 4 is a diagram showing a relationship between control characteristics of a sliding mode controller and values of a switching function setting parameter (VPOLE). FIG. 5 is a diagram showing a setting range of a control gain (F, G) of a sliding mode controller. FIG. 6 is a diagram for explaining drift of model parameters. FIG. 7 is a diagram showing a function for correcting an identification error. FIG. 8 is a diagram for explaining that a default opening deviation of a throttle valve is reflected on a model parameter (c1 ′). FIG. 9 is a flowchart of a throttle valve opening control process. FIG. 10 is a flowchart of a process for setting a state variable in the process of FIG. 9; FIG. 11 is a flowchart of a process of executing a calculation of a model parameter identifier in the process of FIG. 9; FIG. 12 is a flowchart of a process of executing an operation of an identification error (ide) in the process of FIG. 11; FIG. 13 is a diagram for explaining low-pass filtering of an identification error (ide). FIG. 14 is a flowchart of a dead zone process in the process of FIG. FIG. 15 is a diagram showing a table used in the processing of FIG. 14; FIG. 16 is a flowchart of a process of stabilizing a model parameter vector (θ) in the process of FIG. 11; FIG. 17 shows a model parameter (a) in the processing of FIG.
It is a flowchart of the limit processing of 1 ′, a2 ′). FIG. 18 is a diagram for explaining a change in a value of a model parameter due to the processing of FIG. 16; FIG. 19 shows a model parameter (b) in the processing of FIG.
It is a flowchart of the limit process of 1 '). FIG. 20 shows a model parameter (c) in the processing of FIG.
It is a flowchart of the limit process of 1 '). FIG. 21 is a flowchart of a process of executing a calculation of a state predictor in the process of FIG. 9; FIG. 22 is a flowchart of a process for executing a calculation of a control input (Usl) in the process of FIG. 9; FIG. 23 shows a prediction switching function value (σp
It is a flowchart of the process which performs the calculation of re). FIG. 24 is a flowchart of a process of executing a calculation of a switching function setting parameter (VPOLE) in the process of FIG. 23; FIG. 25 is a diagram showing a map used in the processing of FIG. 24; FIG. 26 shows a prediction switching function value (σp
It is a flowchart of the process which performs the calculation of the integrated value of re). FIG. 27 shows a reaching law input (Urc) in the processing of FIG. 22;
It is a flowchart of the process which performs the calculation of h). FIG. 28 shows an adaptive law input (Uad) in the processing of FIG. 22;
It is a flowchart of the process which performs the calculation of p). FIG. 29 is a flowchart of a process for executing a stability determination of the sliding mode controller in the process of FIG. 9; FIG. 30 shows a default opening deviation (t
15 is a flowchart of a process for executing the calculation of (hdefadp). FIG. 31 is a diagram showing a hydraulic positioning device and a control device thereof according to a second embodiment of the present invention. FIG. 32 is a functional block diagram of a control system shown in FIG. 31. DESCRIPTION OF SYMBOLS 1 Internal combustion engine 3 Throttle valve 7 Electronic control unit 10 Throttle valve driving device 21 Adaptive sliding mode controller (control means) 22 Model parameter identifier (identification means) 25 Reference value setting section (correction means) 82 Adaptive sliding Mode controller (control means) 81 Identifier (identification means) 84 Reference value setting unit (correction means)

Continuation of front page    (72) Inventor Jun Takahashi             1-4-1 Chuo, Wako-shi, Saitama Stock Association             Inside the Honda Research Laboratory F term (reference) 3G065 CA26 DA05 FA09 GA41 GA46                       KA33                 5H004 GA08 GA40 GB12 HA07 HB07                       KA45 KA74 KC24 KC26 KC28                       KC43 KC45

Claims (1)

Claims: 1. An identification means for identifying a plurality of model parameters of a control target model obtained by modeling a plant using a predetermined reference value, and using the model parameters identified by the identification means. A control device for controlling the plant, wherein the plurality of model parameters include a model parameter not related to the input / output of the plant, based on the model parameter not related to the input / output of the plant. A control device for a plant, comprising a correction unit for correcting the predetermined reference value.