JP2010238259A - Pid controller - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To tune a PID controller without giving a signal for identification to a control object in an integration process in which a controlled variable presents a periodical fluctuation. <P>SOLUTION: A period analyzing means 31 calculates a period from a manipulated variable added to the control object 12 from an operation end 14 and a controlled variable detected by a detector 15. A simple integrator 32 being an integration characteristic calculating means calculates an integration gain, and generates a process model 30. A control model 20 and an evaluator 21 for adjusting a PID parameter of the control model are combined to constitute a control simulator 16, calculating the optimum PID parameter by simulation. The calculated PID parameter of the control model is finely adjusted in response to a request of a process, and can be set as a control parameter of a controller 13. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a PID control device for an integration process.

  Many PID control devices having an auto-tuning function and a self-tuning function related to PID control are provided. FIG. 8 shows a configuration related to the PID control device 1 including a general auto-tuning function. The control object 2 of the plant is controlled by the controller 3 based on the PID parameter, the operation for controlling the control object 2 is performed by the operation end 4, and the control amount of the control object 2 reflecting the operation result is the detector 5. Detected by. The controller 3 drives the operation end 4 so as to eliminate the deviation between the control amount and the target value. The tuning mechanism 6 includes an identifier 8 and an adjuster 7. The identifier 8 gives an identification signal to the operation terminal 4 to identify the characteristic of the controlled object 2, and the adjuster 7 determines the PID of the controller 3 based on the result. Adjust the parameters to be optimal.

  However, not all of the PID controllers provided in the plant are adjusted using the above-mentioned tuning function, and the PID controller does not necessarily have an auto-tuning or self-tuning function. Most of them are adjusted based on the experience and intuition of the coordinator, because they do not want to give an identification signal and there are many things that can be adjusted without using the tuning function.

  As described above, many plant PID controllers are adjusted based on the experience and intuition of the adjuster. Therefore, there is a problem that the adjustment tends to be insufficient if the delay is large such as an integration process. is there. In addition, many controllers having an auto-tuning or self-tuning function give an identification signal to an object to be controlled. In particular, there are many tuning failures in the integration process, and most of them exhibit periodic fluctuations. Therefore, the integration process is identified without giving an identification signal to the controlled object using this periodic fluctuation, and the PID controller Therefore, a PID control device that can eliminate the tuning failure is required. In addition, the plant operator is reluctant to directly change the control parameters of the actual process by the self-tuning function, and there is a demand for confirming the control performance by simulation in advance.

  The object of the present invention is to identify an integration process, which tends to be insufficiently tuned, without identifying the process by giving an identification signal to the plant, obtain the optimal PID parameters in advance by simulation, and then tune the actual process. It is to provide a PID control device that can be reflected.

The present invention is a PID control device that is used for PID control for an integration process, and that can be tuned to optimize the PID parameters.
Calculation that obtains the periodicity of the controlled variable or manipulated variable that exhibits periodic fluctuations from periodic analysis, and calculates the integral characteristics of the process by performing a predetermined algebraic operation from the period and the fluctuation range of the controlled variable and manipulated variable Means,
Based on the integration characteristic calculated by the calculation means, a PID control simulator is configured using a process model that identifies the integration process and a control model that simulates the PID control function, and a test input signal is given to the simulator and its response Tuning means for tuning the PID parameters of the control model so that the waveform is optimal;
And a controller capable of setting a PID parameter used when controlling the integration process based on the PID parameter tuned by the tuning means.

  According to the present invention, the PID control device includes a calculation unit, a tuning unit, and a controller, compares the control amount detected from the integration process with the target value, is used for control based on the comparison result, and the PID parameter is It can be tuned to be optimal.

  The calculation means analyzes the control amount or the operation amount exhibiting periodic fluctuations to obtain the cycle. As a calculation method, an autocorrelation function of the control amount, a cross-correlation function of the control amount and the manipulated variable may be used, or Fourier analysis or spectrum analysis may be used. The calculation means regards the process as an integrator and calculates an integral gain from the fluctuation range of the control amount and the fluctuation area of the manipulated variable. Therefore, the process model is expressed as a simple integrator, and a parameter expressing the integral characteristic of the process corresponds to the integral gain. The fluctuation area of the manipulated variable is obtained by performing an approximate calculation by calculating the area of a triangle with the period obtained by the calculation means as the base and the fluctuation width of the manipulated variable as the height. The control model uses a PID arithmetic expression of the controller, and a control simulator is configured using the identified process model and control model.

  The tuning means performs a simulation on the test input signal and adjusts the PID parameter of the control model so that the response waveform is optimized. As the test input signal, a step signal or an impulse signal is used, and the response waveform is evaluated according to a predetermined evaluation function such as a minimum overshoot area in the step response. In the tuning process, a series of operations such as control simulation, response waveform evaluation, and control model PID parameter adjustment are repeated to obtain an optimum PID parameter.

  Optimal PID parameters can be obtained with the tuning means described above, but the response waveform for the test input signal is displayed on the screen so that requirements such as checking the control performance in advance by simulation and fine adjustments utilizing the know-how of the practitioner can be added. And a function that can change the PID parameter of the control model. Further, it is possible to set the control parameter of the controller that controls the integration process based on the PID parameter tuned using the built-in control simulator.

  As described above, according to the present invention, it is possible to optimally tune a PID controller in which the control amount and the operation amount of the integration process exhibit periodic fluctuations. By using the period variation data, it is not necessary to give an identification signal to the actual plant, and there is no possibility that the plant will fluctuate. In addition, since the model structure is simplified by setting the process model as an integrator, and the integral calculation is approximated by the area calculation of the triangle, the process model can be generated quickly.

  Furthermore, by providing a control model that simulates the PID control function inside, a control simulator can be configured, and PID tuning by the transient response method that is easy to understand is possible. Since the response waveform is displayed on the display screen, it is possible to confirm the control performance in advance.

  In addition, the present invention is equipped with an environment in which the practitioner can finely adjust the PID parameters in the simulation stage in consideration of adjustments according to process requirements that only the plant operator can know. With this function, adjustment know-how can be incorporated in advance, and safer control parameters can be set for the actual plant.

  According to the present invention, it is possible to perform PID tuning of liquid level control and pressure control exhibiting periodic fluctuations.

It is a block diagram which shows the schematic structure of the PID control apparatus 11 as one Embodiment of this invention. It is the piping system diagram which simplified and showed the plant model which performs liquid level control using the PID control apparatus 11 of FIG. It is a graph of the time series data of the controlled variable and the manipulated variable obtained from the plant model of FIG. It is a graph of the cross correlation function obtained from the time series data of FIG. It is a graph which shows the concept of the simple integration calculation based on the time series data of FIG. It is a graph which shows the transient response of a control simulator which gave the step input signal to the process model 30 of FIG. It is a flowchart which shows the procedure which sets a control parameter to the PID control apparatus 11 of FIG. It is the block diagram which showed the schematic structure of the PID control apparatus containing a general auto tuning function.

  FIG. 1 shows a configuration diagram of a PID control device 11 as an embodiment of the present invention. Similar to the PID control device 1 of FIG. 8, the controller 13 controls the control target 12 of the plant based on the PID parameter, and the operation for controlling the control target 12 is performed by the operation end 14 and reflects the operation result. The control amount of the controlled object 12 is detected by the detector 15. The controller 13 drives the operation end 14 so as to eliminate the deviation between the control amount and the target value.

  The PID parameter of the controller 13 can be set based on the calculation result of the control simulator 16, and the control simulator 16 is composed of a regulator 17 and a part of the identifier 18. A general auto-tuning or self-tuning mechanism includes the adjuster 7 and the identifier 8 shown in FIG. 8, but does not have a control model because it does not have a control model. According to the present invention, the control simulator 16 can be configured by providing the control model 20, and offline control simulation is possible. Accordingly, the identifier 18 includes a period analysis unit 31 and a simple integrator 32 that is an integral characteristic calculation unit, a process model 30, and an adjuster 17, which includes a control model 20 and an evaluator 21, of which the control model 20 and the evaluator 21, The process model 30 constitutes the control simulator 16. An object of the present invention is to optimally tune a PID controller in which a control amount and an operation amount exhibit periodic fluctuations, and an identification method for an integration process and a tuning method based on a control simulation It is characterized by being.

  The identification method of the integration process of the present invention will be described based on the tank liquid level control model shown in FIG. The tank 40 is provided with an inlet 41 and an outlet 42, and stores a liquid 43 therein. The PID control device 11 controls the liquid level of the liquid 43 by adjusting the opening degree of the control valve that is the operation end 14 provided at the outlet 42. The height of the liquid level that is the control amount is detected by a liquid level meter that is the detector 15. When the valve opening degree of the control valve is changed, the outlet flow rate changes, and if the inlet flow rate is constant, the change in the liquid amount in the tank 40 is equal to the integration of the outlet flow rate change from the law of conservation of mass. If the tuning of the liquid level control is poor, in many cases, the control amount and the operation amount as shown in FIG. By using this periodic variation data, it is possible to identify the process without giving an identification signal to the plant and finally tune the PID controller.

  In the process identification, since the process is an integral system, it is necessary to associate the control amount with the integral of the operation amount. In the process of integrating the manipulated variable, it is necessary to obtain the fluctuation period and fluctuation range of the manipulated variable, and the cross-correlation function [Equation 1] is used as one means for obtaining the period. FIG. 4 shows a cross-correlation function between the controlled variable and the manipulated variable in FIG. When the phases overlap, a positive correlation is shown, and when the phase is reversed, a negative correlation is shown. A half period can be obtained from the positive and negative extreme values, and the period is obtained by doubling.

  In the example of FIG. 4, the ½ period is 21 minutes, so that one period is 42 minutes. Next, the fluctuation range of the control amount and the operation amount is obtained. As shown in FIG. 5, the time series data is divided into strips at a time slightly longer than one cycle, preferably 1.2 cycles, so that one cycle is included in the divided data. The fluctuation range is obtained from the difference between the maximum value and the minimum value of this data. The variation widths for the divided numbers are averaged to obtain a typical variation width for the control amount and the operation amount. By averaging, random disturbance is removed, and there is an effect that a more accurate fluctuation range is required. In the example of FIG. 5, the liquid level fluctuation range was 6%, and the opening degree change of the control valve was 6%.

  Although the control amount and the operation amount include an instrumental delay, they are ignored because they are sufficiently small compared to the process delay, and are considered to have no dead time. Furthermore, the valve opening degree of the control valve, which is the operating end, is also considered to be locally linear within the fluctuation range, and the process model is set as a linear integrator. By setting the process model as an integrator, the number of model parameters to be identified becomes one, and the prospects are good in numerical calculation.

  Next, a fluctuation area is calculated from the cycle and fluctuation width of the manipulated variable, and an integral gain that is a model parameter is calculated by comparison with the fluctuation width of the controlled variable. The integral of the manipulated variable is approximately calculated using a triangular area calculation shown in [Equation 2].

  From the cycle and fluctuation range of the manipulated variable, the integral value of the manipulated variable with periodic fluctuation was obtained. An integral gain representing the integral characteristic of the process is obtained from the ratio of the integral value of the manipulated variable and the controlled variable [Formula 3]. The residence time of the liquid in the tank is calculated by [Equation 4].

  In the example of FIG. 5, the control amount average 35 [%] and the operation amount average 53 [%], and the integral gain and the residence time were obtained as follows.

  Therefore, if the process model is expressed by Laplace transform, it is 0.0008 / s. The above is the process model calculation process and the function of the identifier 18 of FIG.

  As the control model, the PID arithmetic expression described in the functional specification of the PID controller is used as it is. The PID arithmetic expression of the embodiment is an expression [Expression 6] called an I-PD expression, and the PID parameters are a proportional band, an integration time, and a derivative time.

  Since the most frequently used method in the transient response method is the step response method, the embodiment is also a tuning means based on the step response. In general, it is said that tuning by step response is good when the control amount is 20%, but conservative tuning without overshooting is preferred in an actual plant. In accordance with the tuning method that has been implemented, [Equation 7] that minimizes the overshoot area without overshooting is taken as the evaluation function. An example of a response waveform with respect to a step input is shown in FIG.

  Since there are three PID parameters to be adjusted, various adjustment methods are proposed. As an example, consider a method of obtaining an integration time that does not cause periodic fluctuations with a fixed proportional band and a differential time of 0 seconds. For the proportional band, the most conservative value that does not cause the tank to become empty or overflow only by proportional control is calculated from [Equation 8], and this value is used as a fixed value. However, when the value actually set in the PID controller is smaller than the obtained value, it is fixed at the value on the controller side.

  Since the differential time is also determined to be 0 seconds, an integration time that does not cause periodic fluctuations is obtained by control simulation. As an initial value of the integration time, twice the residence time is given, and an integration time satisfying the evaluation function of [Equation 7] is obtained. A control simulator is composed of a plant model, a control model, and an evaluator, and a step signal of 3% is given as an input signal. Step response simulation is executed. If the evaluation function value is 0 or less, the integration time value is 0.75 times the current value, and if the evaluation function is greater than 0, the integration time value is 1.25 of the current value. Adjust to double. The step response simulation and the integration time adjustment are repeatedly performed, and the process ends when the evaluation function value becomes positive and close to zero. As described above, an optimum PID parameter having a fast control operation without overshooting has been obtained. A flowchart describing a series of procedures is shown in FIG.

  The example of the optimum value in FIG. 6 is a step response waveform in which the initial value of the integration time is set to 1650 seconds, and the proportional band of 66% and the integration time of 656 seconds are obtained by the tuning means. The above is the tuning process, and functions related to the adjuster 17 and the control simulator 16 of FIG.

  In FIG. 7, the procedure starts from step s0, and in step s1, as shown in FIG. 3, the periodic component of the controlled variable is extracted. In step s2, as shown in FIG. 4, the period is calculated from the period component. In step s3, as shown in FIG. 5, a simple integration calculation is performed to obtain the residence time and the integral gain. In step s4, a process model is generated based on the integral gain. In step s5, initial setting of PID parameters is performed.

  In step s6, a simulation with a step input as shown in FIG. 6 is performed. In step s7, evaluation is performed using transient response characteristics. If the evaluation is not satisfactory, the PID parameter is adjusted in step s8, and the process returns to step s6. The procedure from step s6 to step s8 is repeated until a satisfactory transient response is obtained. Since the evaluation of the transient response can be routinely performed with respect to the step input, the iterative procedure from step s6 to step s8 can be automated.

  If a satisfactory evaluation is obtained in step s7, the actual control parameter of the controller 13 is adjusted by looking at the obtained parameter in step s9. As described above, the actual control parameters can be fine-tuned according to human demand. The procedure ends at step s10.

  Next, a fine adjustment function using a control simulator in which the practitioner can change the PID parameter before applying the actual plant will be described. Depending on the process requirements, there are cases where it is desired to make the control operation slightly faster or slower than the control operation obtained by the tuning means. However, general auto-tuning or self-tuning can be finely adjusted only after automatic tuning is applied. Therefore, even if the PID parameter obtained by the tuning means is optimal in terms of control, it does not meet the requirements of the process, or even if it is optimally tuned, there is a risk that improper fine adjustment will be performed later. is there. In consideration of this point, the present invention enables a fine adjustment in advance without affecting the process by using a control simulator. Based on the parameters changed by the practitioner, a control simulation is executed, and the step response waveform is displayed on the display screen to facilitate understanding. The practitioner can evaluate the response waveform and evaluate it, and if a satisfactory result is obtained, it can be set as a control parameter of the PID controller.

The following embodiments are possible for the present invention.
(1) A PID control device that is used for PID control for an integration process and that can be tuned so as to optimize the PID parameters.
A calculation means that calculates the period of a controlled variable or manipulated variable that exhibits periodic fluctuations from cycle analysis, and calculates the integral characteristics of the process by performing simple algebraic calculation from the period and the fluctuation range of the controlled variable and manipulated variable When,
Based on the integral characteristic calculated by the calculation means, a PID control simulator is configured using a modeled process model and a control model simulating a PID control function, and a test input signal is given and the response waveform is optimal. Tuning means for tuning the PID parameter of the control model so that
And a controller capable of setting a control parameter for controlling the integration process based on the PID parameter tuned by the tuning means.

(2) A method of tuning a PID control device for an integration process so that PID parameters are optimized,
The periodicity of the controlled variable or manipulated variable that exhibits periodic fluctuations is obtained from periodic analysis, and a simple algebraic operation is performed from the period and the fluctuation range of the controlled variable and manipulated variable to calculate the integral characteristics of the process.
Based on the calculated integral characteristics, a PID control simulator is configured using a modeled process model and a control model simulating a PID control function, and a test input signal is given so that the response waveform is optimized. Tune and
A tuning method for a PID control device, wherein a control parameter of a PID controller is adjusted based on a PID parameter obtained by tuning.

(3) The PID control simulator can change the PID parameter of the obtained control model,
A tuning method for a PID control device, characterized in that a simulation based on a changed parameter is performed, and an environment is provided in which fine adjustment according to process requirements can be considered in the simulation stage.

  (4) In the process of integrating the operation amount in the generation of the plant model based on the periodic analysis result, the integral operation is approximately calculated by the area calculation of a triangle having the period as the base and the fluctuation amount of the operation amount as the height. A characteristic tuning method for a PID control device.

  (5) The PID control device tuning method, wherein the integration process is intended for at least one of liquid level control and pressure control.

DESCRIPTION OF SYMBOLS 1 PID control apparatus 2 Control object 3 Controller 4 Operation end 5 Detector 6 Tuning mechanism 7 Adjuster 8 Identifier 11 PID control apparatus 12 Control object 13 Controller 14 Operation end 15 Detector 16 Simulator 17 Adjuster 18 Identifier 20 Control model 21 Evaluator 30 Process model 31 Period analysis means 32 Integral characteristic calculation means 40 Tank 41 Inlet 42 Outlet 43 Liquid

Claims (4)

A PID control device that is used for PID control for an integration process and that can be tuned to optimize PID parameters,
Calculation that obtains the periodicity of the controlled variable or manipulated variable that exhibits periodic fluctuations from periodic analysis, and calculates the integral characteristics of the process by performing a predetermined algebraic operation from the period and the fluctuation range of the controlled variable and manipulated variable Means,
Based on the integration characteristic calculated by the calculation means, a PID control simulator is configured using a process model that identifies the integration process and a control model that simulates the PID control function, and a test input signal is given to the simulator and its response Tuning means for tuning the PID parameters of the control model so that the waveform is optimal;
And a controller capable of setting a PID parameter used when controlling the integration process based on the PID parameter tuned by the tuning means. The PID control simulator can change the PID parameter of the obtained control model,
2. The PID control apparatus according to claim 1, wherein a simulation based on the changed PID parameter is performed by a simulator, the response waveform displayed on the display means is confirmed, and the PID parameter obtained by tuning is adjusted. .   3. The PID control apparatus according to claim 1, wherein the algebraic calculation is approximately calculated as a triangular area having a period as a base and a fluctuation range of an operation amount as a height.   The PID control device according to any one of claims 1 to 3, wherein the integration process targets at least one of liquid level control and pressure control.

Images