JPH06250706A - Fuzzy feedback controller - Google Patents

Abstract

PURPOSE:To effectively control a system which increases in overshoot since its process delay time and dead time are long. CONSTITUTION:A fuzzy arithmetic part 21 once inputting a set value SV and a process value PV of the controlled system finds their deviation and the quantity of variation in the deviation and also calculates the speed type manipulated variable of PI type control by fuzzy inference while referring to variation in PV, and outputs it to a manipulated variable conversion part 22. The manipulated variable conversion part 22 converts the manipulated variable into a position type. Further, a differential operation part 23 similarly finds the deviation and the quantity of variation in the deviation from the SV and PV and calculates a D manipulated variable by position type differential operation while referring to even variation in the PV. The calculated PI manipulated variable and D manipulated variable are added and the sum is sent to a process. Further, the quantity of variation in the process value is also used for control. Thus, the manipulated variables are determined on the basis of variation in process variation to improve the control characteristics.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fuzzy feedback control device used for process control such as a temperature controller, which has a large delay time and dead time.

[0002]

2. Description of the Related Art In process control, the present applicant has filed Japanese Patent Application No. 3-250450 as a fuzzy feedback controller that determines a control output by fuzzy inference so that a process value matches a target value.

This controller feeds back a process value to be controlled, determines a deviation between a set value and this process value, and a control output for matching the process value with the set value by using a variation of this deviation by fuzzy inference. And output it. The fuzzy control rule for that is to consider a reference straight line where the velocity control output change amount of PI control is zero on the phase plane with the deviation and the change amount of the deviation as coordinate axes, and the vertices are positioned on this straight line and a parallel line to this. In order to be proportional to the absolute value of the deviation of the same sign and the amount of change in the deviation, the phase plane is divided into rectangles of equal size, according to the distance between the reference straight line and the parallel straight line. Specifies the magnitude of control output change.

Further, in the antecedent part membership function, the deviation component and the deviation change component of the central coordinates of the rectangle are the isosceles sides of the antecedent part membership function relating to the deviation and the deviation change amount as fuzzy variables. The center of each triangle is specified. Further, the consequent part membership function is specified such that the distance between the reference straight line and the parallel straight line is the center of the isosceles triangle of the consequent part membership function regarding the control output change amount as a fuzzy variable. To do.

[0005]

However, when this fuzzy feedback controller is used for a control target with long process delay time and dead time, overshooting caused by a change in set value is less suppressed. Therefore, this controller cannot be used for a control target that is extremely reluctant to overshoot. Further, the suppression effect when a disturbance is applied is also inferior to the conventional PID control in terms of quick response. The present invention has been made to solve the above problems, and an object of the present invention is to effectively control even a control target having a large process delay time, a large dead time, and a large overshoot. An object of the present invention is to provide a fuzzy feedback control device that can be used.

[0006]

In order to achieve the above object, a first aspect of the present invention is directed to a phase plane whose coordinate axis is a deviation between a process value fed back from a controlled object and a set value, and a change amount of this deviation. In a fuzzy feedback control device for determining a control output for matching a process value with a set value by fuzzy inference, based on the PI control method in, a means for differentiating a change amount of a process value to calculate an operation amount is provided. Characterize.

According to a second aspect of the present invention, in the first aspect, an upper and lower limit limiter for limiting the calculated value calculated by the differential operation and an upper and lower limit range of the upper and lower limiter are determined by the manipulated variable obtained from the differential operation within the evaluation time. And means for changing.

According to a third aspect of the invention, the process value is made to match the set value based on the deviation between the process value fed back from the controlled object and the set value, and the PI control method on the phase plane with the change amount of the deviation as the coordinate axis. In a fuzzy feedback control device that determines a control output by fuzzy inference, a process parameter is identified by a process value and a set value, and a means for calculating a normalization parameter from the process parameter and a deviation and a variation amount of the deviation are calculated by the normalization parameter. And a means for normalizing.

According to a fourth aspect of the present invention, the process value is made to match the set value based on the deviation between the process value fed back from the controlled object and the set value, and the PI control method on the phase plane whose coordinate axis is the amount of change in the deviation. A fuzzy feedback control device for determining a control output by fuzzy inference is characterized by including means for calculating a correction value when changing a set value from a control gain and a current process value and set value.

[0010]

In the first aspect of the invention, since the differential operation for the change amount of the process value is added, the dead time,
Even for a process with a long delay time, the responsiveness at the time of changing the set value and suppressing disturbance is improved.

In the second aspect of the invention, the upper and lower limiters for limiting the calculated value of the differential operation are set, and the range of the upper and lower limiter is changed by the manipulated variable obtained from the differential operation within the evaluation time. As a result, for example, in a process with a short dead time, the manipulated variable is reduced when the process value starts to follow the set value when the set value is changed, but the range of the upper and lower limiters is expanded, and quick response is maintained.

In the third invention, the process parameter is identified by the process value and the set value, and the normalization parameter is calculated from the process parameter. further,
The deviation and the amount of change in the deviation are normalized by the normalization parameter. Thereby, the reference fluctuation range is determined according to the characteristics of the target process.

In the fourth invention, the correction value for changing the set value is calculated from the control gain, the process value and the set value. As a result, it is possible to prevent the operation amount from becoming excessive or slow.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First, an outline of a fuzzy feedback control device according to the first aspect of the present invention will be described in comparison with a conventional PI type fuzzy feedback control device. In the conventional PI type fuzzy control, when the delay time and dead time of the process increase, it becomes vibrational and the response becomes slow. Therefore, in the present invention, the D operation (differential operation) in the PID control is added to make fuzzy control by the PID type.
That is, the arithmetic expression of the conventional PI type fuzzy control is as follows.

ΔU _i = F (e _i , Δe _i ) U _i = U _i-1 + ΔU _i where U is the control output, ΔU is the control output change amount, F is a fuzzy function, e is the deviation, and Δe is the deviation. The amount of change and the subscript i are sampling numbers.

When the position type D motion is added to the above equation, a PID position type arithmetic expression can be obtained as in the following equation. ΔU _i = F (e _i , Δe _i ) U _i = U _i-1 + ΔU _i U _i = U _i + U _di where U _di is the operation amount by the D movement.

The manipulated variable U _di is a differential calculation formula, for example, an incomplete differential calculation formula shown below is used.

[0018]

[Equation 1]

As described above, the differential preceding PID equation is used as the arithmetic expression of the position type D operation.

FIG. 1 is a block diagram showing a schematic configuration of a fuzzy feedback control device according to the first invention.
In the figure, a fuzzy operation unit 21 determines a deviation and a change amount of the deviation when a set value SV of a controlled object and a process value PV are input, and also refers to a fluctuation of the PV, to perform a PI type / speed type PI operation. The amount is calculated by fuzzy inference and output to the manipulated variable conversion unit 22. Operation amount conversion unit 22
Converts the velocity type PI operation amount into the position type.

Further, the differential operation section 23 similarly obtains the deviation and the variation amount of the deviation from the input SV and PV, and also performs the differential operation of the position type while referring to the variation of PV to obtain the D operation amount. calculate. The calculated PI operation amount and D operation amount are added and then output as an operation amount.
In the PID fuzzy control configured in this way, the amount of change in the process value is also used. Therefore, even when the set value is changed or the disturbance is suppressed in a process having a long delay time and a long dead time, the manipulated variable is determined by the change in the process fluctuation, and therefore the quick response is improved as compared with the conventional PI fuzzy control.

Next, an outline of the fuzzy feedback control device according to the second invention will be described. In the conventional PI type fuzzy control, both the control calculation internal limiter and the control output limiter are set to 0-100%. PID according to the first invention
In the fuzzy control, D operation is performed, so delay time,
When changing the set value for a process with a short dead time, D
The reverse operation of the operation impairs the quick response of the rising.
Therefore, the present invention uses 0-100% of the control calculation internal limiter.
The range has been expanded to secure a margin for lowering the reverse operation of D motion. The amount of lowering is determined by back-calculating the value of D operation.

When the lowering margin is α, the upper limit is 1
00% + α, and the lower limit limiter is 0% −α. Here, the amount of change in PV is used as the ratio of the process gain Kp and the process delay time Tp in the operation amount calculation formula based on the D operation, and the operation amount of the D operation within the evaluation time is back-calculated by the following equation to obtain the lower amount α. .

[0024]

[Equation 2]

Γ is a constant.

Thus, in the PID fuzzy control, when the internal operation limiter of 0-100% is used as in the conventional case, if the process delay time and dead time are short, the PV follows the SV when the set value is changed. The amount of operation will be reduced when starting, but the internal calculation limiter is 0-α%
By increasing the amount to 100 + α%, the rapid response is not impaired by the initial weight reduction.

Next, an outline of the fuzzy feedback control device according to the third invention will be described. In the conventional PI type fuzzy control, in order to judge the magnitude of the process fluctuation, the disturbance applied to the process is measured and judged by the reference fluctuation width. That is, the input data on the reference fluctuation range of the process is manually set. Therefore, if you do not know how much the process variation is, 1
%, 10%, etc. could only be set roughly.

Therefore, in the present invention, it is assumed that the normal fluctuation width of the process in PI control corresponds to the proportional band, and the reference fluctuation width of the process is β% (for example, about 50%) of the proportional band, and the process is identified by the limit cycle method. Calculate using the results. Here, assuming that the reference fluctuation width is A, A is calculated by the following equation.

[0029]

[Equation 3]

Note that L is the dead time of the process, and Kp
/ Tp and L are determined by the limit cycle method. As described above, in the third aspect of the invention, for the calculation of the PID control parameter, the reference fluctuation width is calculated based on the process identification result, assuming that the reference fluctuation width depends on the proportional band width of the P operation. By outputting the reference variation width as, for example, 50% of the P width, the reference variation width can be determined according to the characteristics of the process.

Next, an outline of the fuzzy feedback control device according to the fourth invention will be described. In the conventional PI type fuzzy control, the correction amount ΔU _ref for changing the set value is
As shown in the following equation, it is determined so as to be proportional to the proportional gain and the set value change width. ΔU _ref = K _c × _ΔSV × K _{cEX In} this case, when the set value is changed when PV and SV do not match, the operation may become excessive or slow.

Therefore, in the fourth invention, the correction amount is determined so as to be proportional to the proportional gain, the set value change range, and the deviation between SV and PV, respectively. That is, the correction amount ΔU _ref is calculated by the following equation.

[0033]

[Equation 4]

The above equation is the same as the conventional method when PV and SV match. When PV and SV do not match, the operation amount proportional to the deviation between the new SV and the current PV is corrected, so that the operation amount does not become excessive or slow.

Next, examples of the present invention will be described in more detail. FIG. 2 is a block diagram of an embodiment of the fuzzy feedback control device according to the present invention. In the figure, the set value SV of the control target process and the process value PV of the control target
Is input to the controller. Process parameter identification device 3
Identifies the process parameters using PV and SV, and sends the obtained process parameters to the control parameter calculation device 4 and the control parameter conversion device 5.

The control parameter calculation device 4 calculates the PID control parameter from the process identification parameter and sends it to the control parameter conversion device 5. The control parameter conversion device 5 obtains a fuzzy control parameter from the process parameter and the PID control parameter, and sends the fuzzy control parameter to the PV value deviation normalization device 11, the PV value change amount normalization device 12, and the speed type manipulated variable conversion device 14.

The parameter A sent to the normalization device 11 is calculated by the following equation.

[0038]

[Equation 5]

Similarly, the parameter B sent to the normalization device 12 is calculated by the following equation. B = (τ / T _i ) · 66 Similarly, the parameter C sent to the conversion device 14 is calculated by the following equation. C = K _c · B

The SV value change amount calculation device 6 calculates the change amount of the SV value from the SV value of the previous time and this time and sends it to the set value change correction operation amount calculation device 9. The PV value deviation calculation device 7 calculates the deviation between the SV value and the PV value and sends it to the PV value deviation normalization device 11. The PV value change amount calculation device 8 calculates the change amount of the PV value from the PV value of the previous time and the PV value of this time, and calculates the differential operation manipulated variable calculation device 1.
0, sent to the PV value change amount normalizing device 12. The set value change correction operation amount calculation device 9 calculates a correction value of the control output for the change of the set value and sends it to the speed type operation amount conversion device 14.

The differential operation manipulated variable calculator 10 calculates a differential operation manipulated variable from the amount of change in the PV value and the previous D operation manipulated variable and sends it to the internal calculation upper / lower limit device 16. P
The V value deviation normalization device 11 normalizes the deviation of the PV value by the fuzzy control parameter to generate the fuzzy inference operation device 1.
Send to 3. The PV value change amount normalization device 12 normalizes the change amount of the PV value by the fuzzy control parameter and sends it to the fuzzy inference operation device 13.

The fuzzy inference operation device 13 performs fuzzy inference based on the normalized deviation of the PV value and its change amount, calculates PI type and speed type operation amounts, and sends them to the speed type operation amount conversion device 14. The velocity type manipulated variable conversion device 14 transforms (denormalizes) the velocity type manipulated variable according to the fuzzy control parameter, and further adds a correction value to the manipulated variable converter 15.
Send to. The manipulated variable conversion device 15 converts the velocity type manipulated variable into a position type manipulated variable and sends it to the internal calculation upper / lower limit device 16.

The internal calculation upper and lower limit device 16 performs an upper and lower limit check on the position type manipulated variable, further adds the differential operation manipulated variable, and sends it to the control output upper and lower limit device 17. The control output upper / lower limit device 17 performs an upper / lower limit check of the manipulated variable, sends it to the process parameter identification device 3, and outputs it to the controlled object.

FIG. 3 is a graph comparing the embodiment of the present invention with the conventional PID control. In the figure, the solid line is the control result according to the embodiment of the present invention, and the broken line is the conventional PID.
The control result by control is shown. As is clear from the figure,
In the embodiment of the present invention, almost no overshoot at the time of rising occurs, and the quick response to disturbance is excellent.

As described above, in the embodiment, the D operation is added to the conventional PI fuzzy control, the differential operation by the upper and lower limiters is adjusted, the normalization parameter is changed by the process parameter, and the set value is changed and corrected. By calculating the value from the process value and the set value, it is possible to obtain a fuzzy feedback control device with less overshoot, quick response to disturbance and change of the set value and good control characteristics.

[0046]

As described above, according to the first invention,
The addition of the differential operation with respect to the change amount of the process value improves the quick response and the settling property when the set value is changed and the disturbance is suppressed, even for a process having a long dead time and a long delay time. This makes it possible to obtain control characteristics equivalent to or better than those of PID control.

According to the second aspect of the present invention, the upper and lower limiters for limiting the calculated value of the differential operation are set, and the range of the upper and lower limiter is changed from the manipulated variable obtained from the differential operation within the evaluation time. . As a result, for example, in a process with a short dead time, the manipulated variable is reduced when the process value starts to follow the set value when the set value is changed, but the range of the upper and lower limiters is expanded, and quick response is maintained, which is good. It is possible to obtain various control characteristics.

According to the third invention, the process parameter is identified by the process value and the set value, and the normalization parameter is calculated from the process parameter. Further, the deviation and the variation amount of the deviation are normalized by the normalization parameter. As a result, the reference fluctuation range is determined according to the characteristics of the target process, and good control characteristics can be obtained.

According to the fourth invention, the correction value for changing the set value is calculated from the control gain, the process value and the set value. As a result, the amount of operation becomes excessive,
Slowness can be prevented and good control characteristics can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment according to the first invention.

FIG. 2 is a block diagram showing an embodiment of the present invention in detail.

FIG. 3 is a graph comparing control results of an example of the present invention and conventional PID control.

[Explanation of symbols]

 3 process parameter identification device 4 control parameter calculation device 5 control parameter conversion device 6 SV value change amount calculation device 7 PV value deviation calculation device 8 PV value change amount calculation device 9 set value change correction operation amount calculation device 10 differential operation operation amount calculation Device 11 PV value deviation normalization device 12 PV value change amount normalization device 13 Fuzzy inference calculation device 14 Speed type manipulated variable conversion device 15 Manipulation amount conversion device 16 Internal calculation upper and lower limit device 17 Control output upper and lower limit device 21 Fuzzy calculation Unit 22 manipulated variable conversion unit 23 differential operation unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masato Kawaura 1-1, Tanabe Nitta, Kawasaki-ku, Kawasaki-shi, Kanagawa Fuji Electric Co., Ltd.

Claims (4)

[Claims] 1. A control output for matching a process value with a set value based on a deviation between a process value fed back from a controlled object and a set value, and a PI control method on a phase plane whose coordinate axis is a variation amount of this deviation. A fuzzy feedback control device which is determined by fuzzy inference, comprising means for differentiating a process value change amount to calculate an operation amount. 2. The fuzzy feedback control device according to claim 1, wherein an upper and lower limit limiter for limiting an operation value calculated by a differential action, and an upper and lower limit limiter range by an operation amount obtained from the differential action within the evaluation time. A fuzzy feedback control device comprising: 3. A control output for matching a process value to a set value based on a deviation between a process value fed back from a controlled object and a set value, and a PI control method on a phase plane having a coordinate change axis of the deviation. In a fuzzy feedback controller determined by fuzzy inference, a process parameter is identified by a process value and a set value, a means for calculating a normalized parameter from the process parameter, and a deviation and a variation amount of the deviation are normalized by the normalized parameter. A fuzzy feedback control device comprising: 4. A control output for matching a process value with a set value based on a deviation between a process value fed back from a controlled object and a set value, and a PI control method in a phase plane having a coordinate change axis of the deviation. A fuzzy feedback control device that is determined by fuzzy inference, comprising a means for calculating a correction value when changing a set value from a control gain and a current process value and set value.