JP2003167604A - Parameter identification device in controller of hydraulic equipment - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately estimate disturbance by accurately identifying the control parameter of a control system. <P>SOLUTION: A position control signal u is inputted to a real controlled object 10, and hydraulic equipment (a valve 12) is operated, and positioned at a position y. In a control model 31, the real controlled object 10 is described with a control parameter such as a coefficient of friction c, and an output when a position control signal u is inputted to the real controlled object 10 is calculated, and a position ym is outputted. Then, a disturbance observer 30 generates a signal w for offsetting disturbance to be added to the real controlled object 10 based on an error e between the position y at which the real controlled object 10 actually operates and the position ym outputted from the control model 31, and corrects the position control signal u by the signal w. A parameter identifier 50 identifies the control parameter of a control model Pm by regarding the real controlled object 10 and the disturbance observer 30 as an apparent controlled object 40. Thus, it is possible to accurately identify the control parameter such as the coefficient of friction c. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for controlling hydraulic equipment such as an operation valve, and more particularly to a device for estimating control parameters such as control model parameters and control gains.

[0002]

2. Prior Art and Problems to be Solved by the Invention FIG.
FIG. 3 is a hydraulic circuit diagram for explaining a control target common to the conventional technology and the present invention, showing a hydraulic cylinder 7 that operates in response to an operation of the operation lever 1.

When the operating lever 1 is operated as shown in FIG. 4, a position command signal S indicating the operation amount S is output and input to the control device 3. The control device 3 generates a command current i for operating the valve 12, which is the operation valve, and applies it to the actuator 14 of the valve 12. Valve 1
2 is an electromagnetic proportional control valve, and the actuator 14 is an electromagnetic solenoid.

When a command current i is applied to the actuator 4 of the valve 12, the valve 12 responds to the command current i.
The spool moves and the valve position changes.

The valve 12 has valve positions 12a, 12b, 12
and the valve position 12c is in the neutral position. When the valve 12 is located at the valve position 12c, the supply of pressure oil from the hydraulic pump 5 to the hydraulic cylinder 7 is cut off.

When the valve 12 is located at the valve position 12a,
The pressure oil discharged from the hydraulic pump 5 is supplied to the A chamber of the hydraulic cylinder 7 via the valve 12 and is supplied to the B chamber of the hydraulic cylinder 7.
The pressure oil in the chamber is discharged to the tank 6 via the valve 12.
When the valve 12 is in the valve position 12b, the hydraulic pump 5
The pressure oil discharged from is supplied to the B chamber of the hydraulic cylinder 7 via the valve 12, and the pressure oil of the A chamber of the hydraulic cylinder 7 is discharged to the tank 6 via the valve 12.

FIG. 5 shows a specific structure of the valve 12.

The valve 12 comprises a sensor circuit 13, an actuator 14, and a valve body 15.

A spool 19 is slidably accommodated in the block of the valve body 15. The spool 19 is supported by springs 17 and 18.

The actuator 14 is a voice coil motor consisting of a coil 8, a magnet 9 and a plunger 8a,
The plunger 8a is formed integrally with the spool 19.

The sensor circuit 13 includes a stroke sensor 16 for detecting a stroke amount of the spool 19, that is, a movement position.
And an amplifier 16a that amplifies the electric signal output from the stroke sensor 16 and outputs it as the spool position y.
It consists of and. A position control signal u corresponding to the position command signal S of the operating lever 1 is input to the current amplifier 11, and the position control signal u is amplified and output as a command current i. The command current i output from the current amplifier 11 is input to the actuator 14 and the coil 8 is energized. As a result, the plunger 8a is biased, and the spool 19 integrated with the plunger 8a moves to a position corresponding to the command current i.

The state shown in FIG. 5 shows the state in which the spool 19 is located at the neutral position 12c in FIG. When the spool 19 moves to the left side in the drawing, the discharge port of the hydraulic pump 5 communicates with the A chamber of the hydraulic cylinder 7, and the B chamber of the hydraulic cylinder 7 communicates with the tank 6. Therefore, the pressure oil discharged from the hydraulic pump 5 is supplied to the A chamber of the hydraulic cylinder 7 via the notch of the spool 19, and
The pressure oil in the B chamber of the hydraulic cylinder 7 is discharged to the tank 6 through the notch of the spool 19.

When the spool 19 moves to the right side in the figure, the discharge port of the hydraulic pump 5 and the chamber B of the hydraulic cylinder 7 communicate with each other, and the chamber A of the hydraulic cylinder 7 communicates with the tank 6. Therefore, the pressure oil discharged from the hydraulic pump 5 is supplied to the B chamber of the hydraulic cylinder 7 through the notch of the spool 19, and the pressure oil of the A chamber of the hydraulic cylinder 7 is passed through the notch of the spool 19. It is discharged to the tank 6.

FIG. 7 shows a concrete example of the configuration of the control device 3 shown in FIG.

The control device shown in FIG. 7 has a position controller 20.
And an actual controlled object 10 including a current amplifier 11, a valve 12, and a sensor circuit 13.

The position control unit 20 sets the spool position command S output from the operation lever 1 as a target value, and sets the spool position y at which the spool 19 of the valve 12 has actually moved as a feedback amount, and makes these deviations zero. A position control signal v is output. In the position control unit 20, the position of the spool 19 is proportional (P), integral (I), differential (D).
Control gains to be controlled are set, and a position control signal v corresponding to these control gains is generated.

A position control signal u corresponding to the position control signal v is input to the current amplifier 11, as will be described later, and this position control signal u is amplified and applied to the actuator 14 of the valve 12 as a command current i.

As a result, the spool 19 of the valve 12 operates and is positioned at a position corresponding to the operation amount S of the operation lever 1. The spool position y at which the spool 19 of the valve 12 has actually moved is detected by the sensor circuit 13 and is fed back to the position controller 20 as described above.

Assuming that the electric system of the actual controlled object 10 shown in FIG. 7 is a primary delay system and the mechanical system is a secondary system of spring-mass, the movement of the valve 12 is expressed by the equation (1) below. Can be represented.

[0020] (1) where u is a position control signal indicating the command voltage to the current amplifier 11, and y is the sensor circuit 13
Output, that is, the spool position where the valve 12 has actually moved, and Pm is a control model of the actual controlled object 10. That is, the control model Pm of the actual controlled object 10 shows the transfer function which is the ratio of the input u and the output y.

K is a thrust coefficient generated in the actuator 14 when a unit input signal is input to the current amplifier 11, T is a time constant of the system composed of the current amplifier 11 and the actuator 14, and m Is the mass of the moving part, k is the spring constant, and c is the coefficient of friction.

Generally, in a control device in which a hydraulic device such as the valve 12 is a control target, the control target is modeled by the equation (1). The parameters in the equation are determined by design values and identification experiments. Here, the friction coefficient c largely depends on the viscosity of the oil. And the viscosity of oil greatly depends on the temperature of oil. The oil temperature is several tens from the start of operation (cold) to the continuous operation (warm).
The oil viscosity rises by a few degrees, and the oil viscosity decreases by a factor of several. Therefore, the coefficient of friction c, which is the coefficient of viscosity of the spool 19, also greatly changes to a fraction.

When the valve 12 is controlled, even in a control system in which the control parameters are appropriately adjusted in a certain state, the characteristics of the actual controlled object represented by the control parameters of the control system change and are represented by the control parameters. When a difference occurs between the characteristic and the characteristic of the actual controlled object, the control performance deteriorates. Here, the control parameter is a parameter of the control model Pm when a disturbance observer described later is used, and P (proportional), I (integral), D in the case of a hoodback control system or a feedforward control system. These are (differential) control gains and feedforward parameters.

FIG. 6 shows the response when the control parameters of the control system are optimally adjusted according to the friction coefficient when the oil temperature is high, and the control parameters of the control system according to the friction coefficient when the oil temperature is low. The response when the oil pressure is optimally adjusted is shown in comparison between the case where the oil temperature actually becomes high and the case where the oil temperature becomes low.

FIG. 6 (a) shows the response when the oil temperature actually becomes high, and FIG. 6 (b) shows the response when the oil temperature actually becomes low. In FIGS. 6A and 6B, the alternate long and short dash line indicates the input command, the broken line indicates the response when optimized at high temperature, and the solid line indicates the response when optimized at low temperature.

As shown in FIG. 6A, when the oil temperature is actually adjusted to a high temperature when the friction coefficient c is optimally adjusted when the oil temperature is high, the response is good as shown by the broken line. Becomes However, when optimally adjusted according to the friction coefficient c when the oil temperature is low, when the oil temperature actually increases, the actual friction coefficient c decreases and the response becomes oscillating and the controllability deteriorates.

As shown in FIG. 6B, when the oil temperature is actually adjusted to a low temperature when the friction coefficient c is optimally adjusted when the oil temperature is low, the response is good as shown by the broken line. Becomes However, when optimally adjusted according to the friction coefficient c when the oil temperature is high, when the oil temperature actually becomes low, the actual friction coefficient c increases and the response becomes slow and controllability deteriorates.

Therefore, in order to always realize the optimum responsiveness regardless of the oil temperature, the friction coefficient c is constantly monitored, and the control parameter of the control system is optimally changed according to the friction coefficient c at that time. I need to do it.

On the other hand, conventionally, a control method for incorporating an "observer" into a control device is widely known. The observer is a mechanism that estimates the state variables from the input and output data of the plant during control operation and the state variable model of the plant.Under certain conditions, even if there is a deviation in the estimation at first. It is a dynamics system that approaches a true state value as time goes by. In normal state feedback control, it is assumed that all state variables of the plant are used, but some of the physical quantities used as state variables are difficult to measure, and many state variable detectors are required. Therefore, this is an effective mechanism in such a case.

The disturbance observer is a mechanism for estimating a state variable when disturbance is applied to the plant under control operation.

That is, the valve 12 is used for adjusting the flow rate of oil and switching the direction. When the oil flows through the valve 12, a fluid force called flow force is generated by the flow, and the spool 19 in the valve 12 is caused. Works as a disturbance force.

In general, the flow force acts in the direction of pushing back the spool 19, and even if the same current is applied to the actuator 14, the flow force also changes when the flow rate of the valve 12 changes, and the position of the spool 19 also depends on the flow rate. Will change.

In the control device shown in FIG. 8, a disturbance observer 30 is incorporated, the disturbance due to the flow force is estimated from the information of the input u and the output y of the controlled object 10, and the feedback is fed back to the control input so as to cancel the estimated disturbance. is doing.

As shown in FIG. 8, the disturbance observer 30
Is composed of a control model unit 31, an error calculation unit 32, an inverse control model unit 33, and a low-pass filter 34.

In the control model unit 31, the control model Pm of the actual controlled object 10 is described by various parameters as described above, and the operating position y of the spool 19 when the position control signal u is input to the actual controlled object 10 is described. The predicted value ym is calculated and output.

The error calculator 32 calculates an error e between the spool position y at which the spool 19 of the valve 12 of the actual controlled object 10 actually operates and the predicted position ym of the spool 19 calculated by the control model Pm. .

In the inverse control model unit 33, the error e is input to the inverse system 1 / Pm of the control model Pm and the filter calculation process is executed. Thereby, a signal obtained by converting the error e generated by the influence of the disturbance into a control input (position control signal u), that is, a signal that cancels the disturbance applied to the actual controlled object 10 is obtained. However, since the filter calculation process executed by the inverse control model unit 33 includes a high-order differential calculation process, a high-frequency signal component is amplified and adversely affects the system. In order to avoid this, the low-pass filter 34 is inserted in the subsequent stage of the inverse control model unit 33 to generate an error signal w from which high frequency components have been removed, and this is added to the subtraction unit 41.

The subtracting unit 41 executes a process of subtracting the error w output from the low pass filter 34 from the position control signal v output from the position control unit 20. As a result, the position control signal v is corrected so as to cancel the disturbance applied to the actual control target 10, and is newly output to the actual control target 10 as the position control signal u.

As described above, in the disturbance observer 30, the actual controlled object 1 is based on the control model Pm of the actual controlled object 10.
The disturbance applied to 0 is estimated and the disturbance is canceled.

However, the disturbance cannot be accurately estimated unless the control model Pm of the actual controlled object 10 is accurate.

An actual controlled object 10 especially for hydraulic equipment
In this case, since the parameters such as the friction coefficient c greatly change depending on the oil temperature as described with reference to FIG. 6, the disturbance observer 30 must be adjusted unless the parameters of the control model Pm are adjusted to appropriate values according to the oil temperature. Does not work accurately, and particularly affects the transient response. If the control model Pm is not accurate, the disturbance cannot be accurately estimated and the spool 19
Responsiveness deteriorates.

Here, it is conceivable to separately prepare a sensor or the like in order to detect a deviation between the actual characteristic of the actual controlled object 10 and the characteristic represented by the parameter in the control model Pm. However, the system becomes complicated and the device cost increases.

The present invention has been made in view of the above circumstances, and in a control device in which a hydraulic device is a control target, a deviation between the actual characteristic of the actual control target and the characteristic represented by the parameter of the control model is The first problem to be solved is to make it possible to accurately estimate the disturbance even if it is caused by the influence of the environment such as the oil temperature.

If the various control gains set in the position control section 20 of the control device of FIG. 8 are not adjusted optimally according to the environment such as the oil temperature, the problem described in FIG. 6 occurs and the optimum response is obtained. May not be realized.

Therefore, according to the second aspect of the present invention, in the control device for controlling the hydraulic equipment, the control parameter in the position control section is optimally adjusted according to the environment such as the oil temperature.
It is a problem to be solved.

As a document showing the conventional general state of the art,
There are Japanese Patent No. 3131241 and Japanese Patent Laid-Open No. 7-71574.

In the above-mentioned Japanese Patent No. 3131341, in a control system in which a servo valve is controlled, the fluid force acting on the servo valve is regarded as a disturbance, and the disturbance is estimated by a disturbance observer to cancel the disturbance. Is listed.

Further, the above-mentioned Japanese Patent Application Laid-Open No. 7-71574 describes an invention in which the oil temperature is estimated from the resistance value of the solenoid of the solenoid valve and the cooling water temperature without using the oil temperature sensor.

[0049]

According to a first aspect of the present invention, an actual control target including a hydraulic device that operates by inputting a control signal, and the actual control target are described by control parameters.
A control model that calculates and outputs an output when a control signal is input to the actual controlled object, based on an error between the output of the actual controlled object and the output of the control model, in addition to the actual controlled object. A disturbance observer that generates a signal that cancels the disturbance to be corrected and corrects the control signal, and the actual control target and the disturbance observer as apparent control targets, based on the input signal and output signal of this apparent control target. And a parameter identifier for identifying a control parameter of the control model.

That is, as shown in FIG. 1 (a), the position control signal u is input to the actual controlled object 10 and the hydraulic device (valve 12 in FIG. 4) is operated to be positioned at the position y.

In the control model 31, the actual controlled object 10 is described by control parameters such as the friction coefficient c, and the output when the position control signal u is input to the actual controlled object 10 is calculated to calculate the position ym. Output.

The disturbance observer 30 cancels the disturbance applied to the actual controlled object 10 based on the error e between the position y at which the actual controlled object 10 actually operated and the position ym output from the control model 31. The signal w is generated and the position control signal u is corrected by this. Specifically, the error signal w output from the disturbance observer 30 is subtracted from the position control signal v output from the position control unit 20, the position control signal v is corrected to the position control signal u, and the actual control target 10 is obtained. Output.

The parameter identifier 50 is used for the actual controlled object 10
And the disturbance observer 30 as an apparent controlled object 40
As a result, the control parameter of the control model Pm is identified based on the input signal v and the output signal y of the apparent controlled object 40. Thereby, the control parameter such as the friction coefficient c is accurately identified. Therefore, the disturbance observer 3
At 0, this accurately identified control parameter can be used to accurately estimate the disturbance.

As described above, according to the present invention, the valve 12
In a control device that controls hydraulic equipment such as, even if there is a deviation between the characteristics of the actual control object 10 and the characteristics represented by the parameters of the control model Pm due to the influence of the environment such as the oil temperature, the parameters Is accurately identified, the disturbance applied to the actual controlled object 10 can be accurately estimated.

A second aspect of the present invention inputs an actual control target including a hydraulic device that operates by inputting a control signal and a command signal, and based on this command signal, the output of the actual control target, and a control parameter. A control unit that generates a control signal and outputs the control signal to the actual control target, and a signal that cancels a disturbance applied to the actual control target based on an input signal and an output signal of the actual control target. A disturbance observer that generates and corrects the control signal, the actual control target and the disturbance observer are set as apparent control targets, and the control of the control unit is performed based on the input signal and the output signal of the apparent control target. And a parameter identifier for identifying a parameter.

That is, as shown in FIG. 2, the position controller 20
The position command signal S is inputted to the position command signal S
And the position control signal u is generated based on the output of the actual controlled object 10 and the control parameter.
It outputs to the actual controlled object 10 including hydraulic equipment such as the valve 12.

The position control signal u is input to the actual controlled object 10 and the hydraulic device (valve 12 in FIG. 4) is operated to position it at the position y.

Then, the disturbance observer 30 generates a signal for canceling the disturbance applied to the actual controlled object 10 based on the input signal u and the output signal y of the actual controlled object 10, and corrects the position control signal u.

The parameter identifier 50 is used for the actual controlled object 10
And the disturbance observer 30 as an apparent control target, the control parameters such as the control gain in the position control unit 20 are identified based on the input signal u and the output signal y of the apparent control target.

As described above, according to the present invention, various control gains set in the position control section 20 are identified and optimally adjusted according to the environment such as the oil temperature. Can always optimize the response.

A third invention is characterized in that, in the first invention, the control parameter is a friction coefficient of the control model.

In a fourth aspect based on the first or second aspect, the hydraulic device is an operation valve operated by an operation means, and the operation means is operated to detect the operation of the operation means. Each time that is detected, the parameter identifier is operated to identify the control parameter.

That is, as shown in FIG. 4, the hydraulic equipment is
The operation valve 12 is operated by the operation means 1. Figure 1
(B), as shown in FIG.
Is detected, and each time it is detected that the operating means 1 is operated, the switch unit 92 is turned on to operate the parameter identifier 50 to identify the control parameter.

According to the present invention, the control parameter can be identified by operating the parameter identifier 50 each time the operating means 1 is operated.

According to a fifth invention, in the first or second invention, a signal generator for generating a signal for identifying a control parameter is prepared, and the signal generator generates the signal before the operation of the hydraulic equipment is started. The control parameter is identified by inputting the actual control target or the position control unit.

That is, as shown in FIG. 1A, a signal S for controlling parameter identification is generated by the signal generator 90.
Before starting the operation of the hydraulic device 12, the signal S is output by the signal generator 90.
Is generated, and the corresponding position control signal u is input to the actual controlled object 10 in FIG. Alternatively, before the operation of the hydraulic equipment 12 is started, the signal S is generated by the signal generator 90, and
As shown in, the signal S is input to the position control unit 20. As a result, it becomes possible to identify the control parameter with the parameter identifier 50 before the operation of the hydraulic device 12 is started.

In a sixth aspect based on the first or second aspect, a dither signal is superposed on the position control signal input to the actual control target while the hydraulic equipment is operating, and the dither signal is applied to the position control signal. The control parameter is identified by inputting the control parameter to the actual control target.

That is, as shown in FIG. 1A, the dither signal d is superposed on the position control signal u input to the actual controlled object 10 during the operation of the hydraulic device 12. Therefore, the dither signal d is input to the actual controlled object 10 as a signal for identifying the control parameter. As a result, the parameter identifier 5
It becomes possible to identify the control parameter at zero.

[0069]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1A shows a configuration example of the control device 3 of the embodiment.

The control device shown in FIG. 1 (a) is roughly composed of an operation lever 1, a position control section 20, an actual controlled object 10,
It consists of a disturbance observer 30 and a parameter identifier 50.

The actual controlled object 10 of the present embodiment includes the valve 12 as already described with reference to FIGS. 4 and 5, and its specific configuration is shown in FIG.

FIG. 7 shows the internal configuration of the position control unit 20 and the actual control target 10 by extracting the position control unit 20 and the actual control target 10 from FIG.

In FIG. 7, the spool position command S output from the operating lever 1 is used as a target value, and the spool position y at which the spool 19 of the valve 12 is actually moved is used as a feedback amount. In addition to the feedback control as described above, a feedforward amount corresponding to the spool position command S is added to the deviation to perform feedforward control.

The position controller 20 is composed of a PID controller 21, a feedforward controller 22, a deviation calculator 23, and an adder 24. The actual controlled object 10 is composed of a current amplifier 11, a valve 12 and a sensor circuit 13.

The deviation calculation unit 23 of the position control unit 20 is inputted with the spool position command S which is a target value and the spool position y which is a feedback amount, and these deviations are calculated and outputted to the PID controller 21. . The PID controller 21 is set with control gains for proportional (P), integral (I), and derivative (D) control of the position of the spool 19 of the valve 12, and a position control signal corresponding to these control gains is generated. To be done. That is, a position control signal is generated and output to the adder 24 by performing a proportional (P) operation, an integral (I) operation, and a differential (D) operation on the input deviation.

On the other hand, the spool position command S which is the target value is
The feedforward controller 22 inputs the feedforward amount required for the spool 19 to move to a position corresponding to the spool position command S. The calculated feedforward amount is output to the addition unit 24.

In the adder 24, the position control signal input from the PID controller 21 and the feedforward controller 22 are added.
The feed-forward amount input from is added, and the added value is output as the position control signal v. The position control signal v is corrected to the position control signal u by the disturbance observer 30 shown in FIG. 1A as described later, and then added to the current amplifier 11 of the actual control target 10.

The position control signal u is amplified by the current amplifier 11 and output as the command current i to the actuator 14 of the valve 12. This allows spool 1 of valve 12
9 is actuated and positioned at a position corresponding to the operation amount of the operation lever 1. Moving position y of spool 19 of valve 12
Is detected by the sensor circuit 13 and fed back to the position controller 20 as described above.

Now, assuming that the electric system of the actual controlled object 10 shown in FIG. 7 is a primary delay system and the mechanical system is a spring-mass secondary system,
The movement of the valve 12 is based on the equation (1) already described. It is expressed by the equation shown by. As shown in the equation (1), the characteristic of the valve 12 is represented by each parameter such as the mass m, the spring constant k, the damping coefficient c in the control model Pm.

As described with reference to FIG. 6, the parameters in the control model Pm, in particular the friction coefficient c, greatly depend on environmental changes such as oil temperature (oil temperature change during operation, secular change). fluctuate. Therefore, the parameter identifier 50 of FIG. 1 (a) identifies the friction coefficient c using known information.

By modifying the above equation (1), the following equation (2) is obtained.

[0083] (2) The left side of the above equation (2) is obtained by subjecting the input u and the output y of the actual controlled object 10 to an appropriate filtering process configured by known parameters. Therefore (2)
The left side of the equation is known. On the other hand, the right side of the equation (2) has a coefficient obtained by subjecting the output y to an appropriate filtering process and a friction coefficient c which is an unknown parameter to be identified.
Is the product of Let Y be the left side of the equation (2), Z be the known part of the right side, and Θ be the unknown parameter part of the right side.
Equation (2) can be rewritten as equation (3) below.

[0084] (3) However, α is an appropriate positive number.

The unknown parameter portion Θ is estimated from the known portions Y and Z of the equation (3). The following variables are defined with Θ ^ being an estimated value of Θ.

Y ^ = ZΘ ^: Estimated value of Y e = Y ^ −Y: Error of estimated value of Y At this time, if the adaptive identification rule shown in the following equation (4) is used, the estimated value Θ ^ of Θ is It is known to asymptotically approach Θ.

[0087] (4) However, the initial value Γ (0) of Γ in the above equation (4) is Γ
Assume that (0)> 0. The weighting factor γ is γ ≧
Assume 0.

The case where only the friction coefficient c is estimated has been described above. However, the thrust coefficient K may change due to, for example, heat generation of the actuator 14.
Therefore, the thrust coefficient K may also be identified. By treating the parameters to be identified as vectorized as described below, it can be handled in the same manner as above.

[0089] (5) However, α is an appropriate positive number.

If Z and Θ are treated as vectors, the estimated value Θ ^ of Θ approaches asymptotically Θ by using the same adaptive identification rule shown in the following equation (6).

[0091] (6) However, the initial value Γ (0) of the matrix Γ in the above equation (6)
Is assumed to be Γ (0)> 0. Also, the weighting factor γ
Is assumed to be γ ≧ 0.

When other parameters change due to changes in the environment and must be identified, they can be identified by the same procedure as above.

As already described with reference to FIG. 5, when oil is caused to flow through the valve 12, a fluid force called flow force acts, which acts as a disturbance and acts on the actual controlled object 10. On the other hand, the parameter identification rule described above does not consider the influence of disturbance. Therefore, if a disturbance is applied to the actual controlled object 10, the identified parameter cannot be correctly estimated due to the influence of the disturbance. Therefore, when the above-mentioned parameter identification rule is used, the disturbance observer 30 is first applied to the actual controlled object 10 to remove the influence of the disturbance, and a system composed of the actual controlled object 10 and the disturbance observer 30 is newly added. The identification rule is used after assuming that it is the control target 40.

The internal structure of the disturbance observer 30 shown in FIG. 1A is the same as that shown in FIG.

In the disturbance observer 30, the actual controlled object 10
The disturbance due to the flow force is estimated from the information of the input u and the output y to the control input and is fed back to the control input so as to cancel the estimated disturbance.

As shown in FIG. 8, the disturbance observer 30
Is composed of a control model unit 31, an error calculation unit 32, an inverse control model unit 33, and a low-pass filter 34.

In the control model section 31, the control model Pm of the actual controlled object 10 is described by various parameters as described above, and the operating position y of the spool 19 when the position control signal u is input to the actual controlled object 10 is set. A prediction calculation is performed and the predicted value ym is output.

The error calculator 32 calculates an error e between the spool position y at which the spool 19 of the valve 12 of the actual controlled object 10 actually operates and the predicted position ym of the spool 19 output from the control model Pm. .

In the inverse control model unit 33, the error e is input to the inverse system 1 / Pm of the control model Pm and the filter calculation process is executed. Thereby, a signal obtained by converting the error e generated by the influence of the disturbance into a control input (position control signal u), that is, a signal that cancels the disturbance applied to the actual controlled object 10 is obtained. However, since the filter calculation process executed by the inverse control model unit 33 includes a high-order differential calculation process, a high-frequency signal component is amplified and adversely affects the system. In order to avoid this, the low-pass filter 34 is inserted in the subsequent stage of the inverse control model unit 33 to generate an error signal w from which high frequency components have been removed, and this is added to the subtraction unit 41.

The subtracting section 41 executes a process of subtracting the error w output from the low pass filter 34 from the position control signal v output from the position control section 20. As a result, the position control signal v is corrected so as to cancel the disturbance applied to the actual control target 10, and is newly output to the actual control target 10 as the position control signal u.

As described above, the disturbance observer 30 uses the control model Pm of the actual control target 10 to determine the actual control target 1
The disturbance applied to 0 is estimated and the disturbance is canceled.

However, the disturbance cannot be accurately estimated unless the control model Pm of the actual controlled object 10 is accurate.

The actual control target 10 especially for hydraulic equipment
In this case, since the parameters such as the friction coefficient c greatly change depending on the oil temperature as described with reference to FIG. 6, the disturbance observer 30 must be adjusted unless the parameters of the control model Pm are adjusted to appropriate values according to the oil temperature. Does not work accurately, and particularly affects the transient response.

Therefore, in the parameter identifier 50, the system composed of the actual controlled object 10 and the disturbance observer 30 is newly regarded as the apparent controlled object 40, and then the equations (2) to (6) are used. By applying the described identification rule, the parameter that varies depending on the environment such as the oil temperature in the actual controlled object 10 is identified.

Next, the operation of the control device shown in FIG. 1A will be described.

As shown in FIG. 1A, the actual controlled object 1
The position control signal u is input to 0, the valve 12 of the actual control target 10 is operated, and the spool 19 of the valve 12 is positioned at the position y.

In the control model unit 31 of the disturbance observer 30, the control model Pm of the actual controlled object 10 is described by the control parameters such as the friction coefficient c as shown in the above equation (1). A calculation for predicting the operating position y of the spool 19 when the position control signal u is input is performed and the predicted position ym is output.

The disturbance observer 30 has a position y at which the spool 19 of the valve 12 of the actual controlled object 10 actually operates,
Based on the error e from the predicted position ym of the spool 19 output from the control model Pm, a signal w that cancels the disturbance applied to the actual controlled object 10 is generated, and the position control signal u is corrected by this. That is, the subtraction unit 41 executes a process of subtracting the error signal w output from the disturbance observer 30 from the position control signal v output from the position control unit 20, and corrects so as to cancel the disturbance applied to the actual controlled object 10. The generated position control signal u is output to the actual controlled object 10.

The parameter identifier 50 is used for the actual controlled object 10
And the disturbance observer 30 as an apparent controlled object 40
Based on the apparent input signal v and output signal y of the controlled object 40, the control parameters of the control model Pm are identified by applying the identification rules described using the equations (2) to (6). To do. This allows parameters such as the coefficient of friction c to be accurately identified. Therefore, the disturbance observer 3
At 0, this accurately identified parameter can be used to accurately estimate the disturbance.

As described above, according to this embodiment, in the control device for controlling the hydraulic equipment such as the valve 12, the parameters of the control model Pm can be accurately identified according to the change of the environment such as the oil temperature. Therefore, the disturbance applied to the actual controlled object 10 can be accurately estimated. In particular, the friction coefficient c, which is the viscosity coefficient of the spool 19, is identified, and the term of the viscosity coefficient c of the control model Pm used for the estimation calculation of the disturbance observer 30 is changed according to the identification result. The accuracy can be stabilized regardless of temperature changes.

Specifically, the vibration shown by the solid line in FIG. 6A does not occur, and the same response as the response shown by the broken line is obtained, and the response is good. Further, in FIG. 6 (b), the delay of the response shown by the solid line does not occur, and the response similar to the response shown by the broken line is obtained, and the responsiveness becomes good.

In the control device shown in FIG. 1A, as shown in FIG. 5, the electric command i is directly input to the electric actuator 14, and the spool 19 is generated by the force generated by the electric actuator 14. It is assumed that the valve 12 drives the. However, in the case of a valve through which a large amount of pressure oil passes, the force generated by the electric actuator may not be sufficient to drive the spool 19.
Therefore, the electric command i may be input to the electric actuator 14, the electric actuator 14 may drive a smaller pressure valve, and the spool 19 may be driven by the hydraulic pressure generated by the small pressure valve. . By adopting the valve 12 having such a configuration, a large thrust can be obtained even if a large amount of pressure oil passes.

Even if such a valve 12 is adopted, a control model P similar to the above equation (1) is used.
It is only necessary to construct m, select a parameter that fluctuates due to the influence of the environment such as oil temperature from the control model Pm, and identify the selected parameter by applying the above-mentioned identification rule.

By the way, when the parameters are identified by the parameter identifier 50, the position control signal u is sent to the actual controlled object 10.
If is not entered, identification cannot be performed. Specifically, identification cannot be performed unless the operation lever 1 is operated by the operator.

A) Real-time estimation during operation Assuming that the construction machine is equipped with the control device shown in FIG. 1A, the operator operates the operation lever 1 in the operator's cab (cabin). The lever signal generated by the operation lever 1 is subjected to some electrical processing or is not processed at all, and is constituted by the position control unit 20 and the actual controlled object 10 as the spool position command S. It is input to the servo control system. Since the lever operation is an operation arbitrarily performed by the operator, the spool position command S is
It is a signal with various frequency components. If the lever is not operated, nothing is input to the servo control system, and as a result, the spool 19 does not move and the output y cannot be obtained. Even if parameters are identified by the parameter identifier 50 in such a state, the identification cannot be performed correctly.

Therefore, in the structure shown in FIG. 1A, a means for detecting the lever operation, that is, the input signal to the servo control system is added, and only when there is the input signal. Parameter identification calculation can be performed. When the lever is not operated and there is no input signal, the immediately preceding identification result is held and the parameter is not updated. When the lever is operated and there is an input signal, the identification calculation is executed, and the parameters are updated with the parameters obtained as a result of the identification calculation.

FIG. 1B shows a configuration in which the parameter identifier 50 is operated to execute the parameter identification calculation when the operating lever 1 is actually operated. FIG. 1B shows only a part of FIG.
Illustration of the configuration overlapping with (a) is omitted.

As shown in FIG. 1B, the operation lever 1 is provided with a conversion portion 91 for converting the operation amount of the operation lever 1 into the spool position S and outputting the spool position S. The conversion unit 91
It is composed of, for example, a potentiometer, and when the operation lever 1 is operated, it outputs an electric signal indicating the spool position S. A switch unit 92 is provided between the parameter identifier 50 and the disturbance observer 30. A signal indicating the spool position S is input from the conversion unit 91 to the switch unit 92. The switch unit 92 operates to close when a signal indicating the spool position S, that is, a signal indicating that the operation lever 1 is operated, is input. As a result, the parameter identifier 50 and the disturbance observer 30 are connected, and the parameter observer 3
A calculation process for identifying the parameter in the control model Pm at 0 is executed. The control model Pm of the disturbance observer 30 is calculated by the parameters obtained as a result of the identification calculation process.
The parameters inside are updated with the new ones. However, when the signal indicating the spool position S is not input from the conversion unit 91 to the switch unit 92, that is, when the operation lever 1 is not operated, the switch unit 92 moves between the parameter identifier 50 and the disturbance observer 30. Works to open.
Therefore, when the operation lever 1 is not operated, the parameter identifying unit 50 does not execute the arithmetic processing for identifying the parameter.

As described above, according to the configuration shown in FIG. 1B, the parameter identifier 5 is operated every time the operating lever 1 is operated.
0 can be activated, which allows real-time identification of parameters during operation.

B) Identification by Dither Signal In general, in the servo valve 12, a minute signal d having a relatively high frequency is used as a position control signal in order to reduce the influence of static friction of the actuator 14 and the spool 19 and improve responsiveness. It is superimposed on v (see FIG. 1A). This signal d is called a dither signal.

By superimposing the dither signal d on the position control signal v, a minute vibration can be constantly excited on the spool 19, and the influence of static friction can be reduced to improve the responsiveness. The dither signal d is generally a periodic signal and has a fundamental frequency and its harmonic components. The dither signal d is always superposed regardless of whether the valve 12 is operated or not, that is, regardless of whether the operation lever 1 is operated.

Therefore, in FIG. 1A, the operation lever 1 is not operated and only the dither signal d is controlled by the control target 40.
, The position control signal v in which the dither signal d is superimposed is input to the controlled object 40, the operation lever 1 is operated,
The identification operation may be performed by operating the parameter identifier 50.

C) Identification by External Signal Before Operation Start In each of the embodiments a) and b) described above, it is assumed that the parameters are constantly identified during the operation of the operation lever 1, that is, during the operation of the valve 12. ing. However, before the operation of the valve 12, the identification calculation processing is performed by inputting a predetermined signal suitable for parameter identification in advance, and basically the parameter identification calculation processing is not performed during the operation of the valve 12, It is also possible to carry out identification calculation processing as necessary.

In this case, a signal generator 90 is added as shown in FIG. The signal generator 90 generates a signal S for parameter identification.

Therefore, before the operation of the valve 12 is started, the signal generator 90 generates the signal S and inputs it to the position controller 20. Therefore, the position control unit 20 generates the position control signal v corresponding to the signal S and outputs it to the controlled object 40. As a result, before the operation of the valve 12 is started, the input v and the output y of the controlled object 40 are obtained, and the parameter can be identified by the parameter identifier 50 based on these.

It is also possible to use the dither signal d.

That is, as shown in FIG. 1A, the dither signal d is superimposed on the position control signal v input to the controlled object 40 during the operation of the valve 12. Therefore, before starting the operation of the valve 12, only the dither signal d is controlled by the control target 40.
Is input as a signal for parameter identification. As a result, before the operation of the valve 12 starts, the input d of the controlled object 40
(V) and the output y are obtained, and the parameter can be identified by the parameter identifier 50.

The three embodiments shown in a), b), and c) have been described above, but these may be combined.

D) Updating the control parameters The parameters in the control model Pm in the disturbance observer 30 are updated by the parameters identified by the timing parameter identifier 50. There are various possible timings for updating the parameters.

For example, the parameters can be sequentially updated based on the identification result obtained momentarily during the identification calculation.

Further, it is possible to set a fixed time, average each parameter identified within this fixed time, and update using the averaged parameter.

It is also possible to average each parameter identified by the time when one lever operation is completed and update the averaged parameters.

In the embodiment described above, the control model Pm
It is assumed that the parameters of will be identified.

However, the present invention is not limited to this, and can be applied to the case of updating the control parameters in the position control unit 20, for example, the control gains of P, I and D.

FIG. 2 shows an embodiment in which the control gain in the position controller 20 is identified by the parameter identifier 50.

In FIG. 2, the same elements as those of FIG. 1 are the same and the description thereof will be omitted. In the embodiment shown in FIG. 2, a configuration corresponding to the disturbance observer 30 of FIG. 1A is incorporated in the position control unit 20. That is, as shown in FIG. 2, a spool position command S is input to the position control unit 20, a position control signal u is generated based on the position command S and a control parameter such as a control gain, and the position control signal u is generated. u is output to the actual controlled object 10 including hydraulic equipment such as the valve 12.

The position control signal u is input to the actual controlled object 10, the valve 12 is operated, and the spool 19 is positioned at the position y.

Disturbance observer 3 in position control unit 20
0 indicates the position control signal u and the valve 12 of the actual control target 10.
A signal for canceling the disturbance applied to the actual controlled object 10 is generated based on the error e from the position y at which the spool 19 is operated, and is output to the actual controlled object 10 as a position control signal u.

The parameter identifier 50 is used for the actual controlled object 10
The system composed of the disturbance observer 30 and the disturbance observer 30 is newly regarded as the apparent controlled object 40, and the signal v input to this apparent controlled object 40 and the signal y output from the apparent controlled object 40. Based on, the control parameters such as the control gain in the position control unit 20 are identified.

As described above, according to this embodiment, various control gains set in the position control section 20 are identified,
Since the optimum control is performed according to the environment such as the oil temperature, the response of the actual controlled object 10 can be always optimized.

FIG. 3 shows an embodiment in which a structure corresponding to FIG. 1B is added to the structure of FIG.

According to the configuration shown in FIG. 3, the parameter identifier 50 can be operated every time the operating lever 1 is operated, as described in the above a), and in this way, the parameter identifier 50 can be operated in real time during operation. The parameters can be identified.

Further, each of the above examples b), c), and d) can be appropriately applied to the embodiment shown in FIG.

The present invention also relates to the PID described in the embodiment.
The present invention is not limited to the control and the control using the disturbance observer, and can be applied to a control method in which a control system is configured based on a control model, such as H ∞ control or sliding mode control.

[Brief description of drawings]

FIG. 1A is a diagram showing a configuration of a control device of an embodiment, and FIG. 1B is a diagram illustrating a modified example of the control device of FIG. 1A.

FIG. 2 is a diagram showing a configuration of another control device according to the embodiment.

3 is a diagram showing a modification of the control device of FIG. 1, and is a diagram corresponding to FIG. 1 (b).

FIG. 4 is a diagram showing a hydraulic circuit including the valve of the embodiment.

5 is a diagram showing a structure of the valve shown in FIG.

6 (a) and 6 (b) are diagrams for explaining the conventional problems.

FIG. 7 is a diagram illustrating an internal configuration of a position control unit and an actual control target.

FIG. 8 is a diagram illustrating an internal configuration of a disturbance observer.

[Explanation of symbols]

1 control lever 10 Actual control target 12 valves 20 Position controller 30 Disturbance observer 31 Control model part 40 Apparent controlled object as seen from the identifier 50 parameter identifier 90 signal generator 92 Switch

Claims (6)

[Claims] 1. An actual control target including a hydraulic device that operates by inputting a control signal, the actual control target is described by control parameters, and an output when a control signal is input to the actual control target is calculated. And a disturbance observer that corrects the control signal by generating a signal that cancels a disturbance applied to the actual control object, based on an error between the output of the actual control object and the output of the actual control object. A parameter identifier that identifies the control parameter of the control model based on the input signal and the output signal of the apparent controlled object, with the actual controlled object and the disturbance observer as the apparent controlled object. A parameter identification device in a control device for hydraulic equipment, characterized in that 2. An actual control target including a hydraulic device that operates by inputting a control signal, and a command signal, and the control signal is generated based on this command signal, the output of the actual control target, and a control parameter. A control unit that generates the control signal and outputs the control signal to the actual control target, and generates a signal that cancels a disturbance applied to the actual control target based on the input signal and the output signal of the actual control target. A disturbance observer that corrects a control signal, and the actual control target and the disturbance observer are set as apparent control targets, and the control parameters of the control unit are identified based on the input signal and the output signal of the apparent control target. A parameter identifying device in a control device for hydraulic equipment, comprising: 3. The parameter identification device in a control device for hydraulic equipment according to claim 1, wherein the control parameter is a friction coefficient of the control model. 4. The hydraulic device is an operation valve operated by an operation means, and the parameter identifier is detected each time the operation means is detected and the operation means is detected. 3. The parameter identification device in a control device for hydraulic equipment according to claim 1, wherein the control parameter is identified by operating the. 5. A signal generator that generates a signal for identifying a control parameter is prepared, and a signal is generated by the signal generator and input to the actual control target or the control unit before the operation of the hydraulic equipment is started. The parameter identification device in the control device for hydraulic equipment according to claim 1 or 2, wherein the control parameter is identified thereby. 6. A dither signal is superimposed on a control signal input to the actual control target during operation of the hydraulic device, and by inputting the dither signal to the actual control target,
The parameter identification device in a control device for hydraulic equipment according to claim 1 or 2, wherein a control parameter is identified.