KR20110101813A - Modeling Techniques and Automatic Tuning PID Control Method of Industrial Heater System Using UDEAS - Google Patents

Abstract

The present invention is an autotuning PID control system that automatically performs the tuning process of a PID controller applied to an industrial heater plant control system. Based on a PID controller that is frequently used by field users of a heater plant, uDEAS (high speed arithmetic optimization algorithm) The present invention relates to a modeling technique and an automatic tuning PID control method of an industrial heater system using uDEAS, comprising a system for automatically tuning a PID controller using univariate Dynamic Encoding Algorithm for Searches. By using the method of estimating parameters, it enables fast modeling, and also can be applied to similar linear system in addition to heater system for accurate and faster modeling, and also auto-tuning PID using IMC-PID proposed in the present invention. The control method is made of conventional PID The possible complement unstable elements of an industrial control system used, and the automatic tuning so easily possible to facilitate the control gain change for an industrial environment and disturbances, and there is an effect that can greatly increase the user's convenience.

Description

Modeling of industrial heater systems and Auto-tuning PID control using uDEAS

The present invention relates to a modeling technique and an automatic tuning PID control method of an industrial heater system using uDEAS applied to an industrial heater plant control system used in a fiber twist process.

In general, the PID controller has been applied to many industrial sites and controllers requiring various controls due to its simple structure and ease of understanding, and is a control system most familiar and easy to use for users in the field. However, the process of obtaining new control gain value is complicated and time consuming and burdens the field users when the PID controller is changed in process environment or when the controller parameter is changed due to disturbance. Therefore, most field users perform tuning work in the process environment that they use the most, and then, even if the process environment changes, operate without any new tuning work or perform new tuning work using empirical gain values. In many cases, the process can be rapidly changed and used manually. Therefore, the efficiency of the controlling plant is reduced, and when the plant is controlled in a transient state, it is a obstacle to normal operation.

Field users and related experts detect the response signal waveform of the system output stage when tuning the PID controller and compare it with the control reference value. Through this, the system diagnosis and control parameters such as proportional coefficient ( K _P ), integral coefficient ( T _I ) and derivative The coefficient T _D is adjusted. In general, the automatic tuning technique of PID controller is a modelless adjustment technique that is used when the model of the plant is not clear or the model changes due to working conditions or disturbances, but it is difficult to apply and control the above problems. exist.

In the prior art, in Korea Patent Publication No. 1993-0006519, a tuning method of a PID controller and an automatic tuning method using an integration of a relay feedback response in Korean Patent Publication No. 846554 have been proposed. In order to improve the performance of the PID controller, the unit controller can be upgraded or the critical information and frequency model of the process can be obtained more accurately. However, there are limitations in retrieving the parameters of the process at high speed.

The present invention for solving the above problems is created in view of the above problems with the prior art control method of the industrial heater plant, to solve this, a method for automatically performing the tuning process of a complex and difficult general PID controller The modeling technique of the industrial heater system using uDEAS is characterized by the automatic tuning PID controller algorithm using the operation optimization technique and the system for operation and monitoring in the existing PID controller so that users can easily understand and apply it. And an automatic tuning PID control method.

The tuning system using the operation optimization technique according to the present invention includes the steps of obtaining an initial PID coefficient K _P , T _D , T _I in a normal PID controller tuning state; Calculating a parameter value of each heater plant through uDEAS after initial tuning; Applying the parameter value calculated in the step to the PID controller; Applying a current control signal when the error error is less than the tolerance; An object of the present invention is to provide a modeling technique and an automatic tuning PID control method for an industrial heater system using uDEAS, comprising adjusting PID coefficients by using uDEAS when an error occurs above a reference value.

In the present invention for achieving the above object in the method for modeling the industrial heater system using uDEAS,

Using the transfer function of the 3-parameter model due to the nature of the linear heater system; Using uDEAS, an operation optimization technique, to estimate unknown parameters in the model; Applying an IAE (sum of absolute error) to verify a parameter estimated by the uDEAS; The method of modeling an industrial heater system using uDEAS includes a step of determining appropriateness based on the parameter and repeatedly performing uDEAS, selecting the appropriate parameter, and applying the same.

The cost function of uDEAS is set as a result function of IAE performance index using static variable K , time constant T , and invalid time L ,

The parameter static variable K , time constant T , and invalid time L of the cost function of uDEAS are characterized by estimating the optimal parameter by repeatedly performing uDEAS according to whether the performance index of IAE is minimized and well identified in the system.

And the present invention relates to an automatic tuning method of a heater system modeled using IMC-PID,

Applying to a changing heater system using the IMC-PID technique; Finding the applied IMC-PID coefficient with uDEAS; Applying an IAE (sum of absolute error) to verify the PID coefficient obtained by the uDEAS; Automatically tuning PID coefficients of a control system by repeatedly performing uDEAS by determining whether the titration coefficients are appropriate based on the PID coefficients, and an automatic tuning method of a heater system modeled using an IMC-PID. It is done.

By applying the IMC-PID, only the filter coefficient T _F is adjusted, not the PID technique for finding the proportional coefficient ( T _P ), the integral coefficient ( K _I ), and the differential coefficient ( K _P ).

The filter coefficient T _F is adjusted using uDEAS,

The found filter coefficient T _F is modeled using an IMC-PID, characterized by determining whether the titration is appropriate using a result function of an IAE performance index.

The filter coefficients may be automatically found using uDEAS according to whether the filter coefficients are appropriate and the error of the corresponding plant.

Therefore, the plant modeling method of the industrial heater system proposed by the present invention by the above-mentioned problem solving means does not require a complicated calculation equation, and enables fast modeling by using a method of estimating a parameter using uDEAS, which is an operation optimization technique. In addition, it can be applied to similar linear system in addition to heater system for accurate and faster modeling.

In addition, the automatic tuning PID control method using the IMC-PID proposed in the present invention can compensate for the unstable elements of the industrial control system using the conventional PID control, and the automatic tuning is easy to control the industrial environment or disturbance. The gain can be easily changed and the user's convenience can be greatly increased.

1 is a flowchart illustrating a modeling technique and an automatic tuning PID control method of an industrial heater system using the present invention uDEAS.
Figure 2 is a graph showing the step response of the heater system estimated using uDEAS in the present invention.
3 is a block diagram showing the structure of an IMC-PID applied in the present invention.

Hereinafter, with reference to the accompanying drawings, the preferred technical configuration of the present invention for achieving the above effects in detail as follows. In the following description, only the parts necessary for understanding the modeling technique and the automatic tuning PID control method of the industrial heater system using the uDEAS according to the present invention will be described. Note that it will be omitted from.

Referring to the present invention with reference to the accompanying drawings, Figures 1 to 3 as follows.

1 is a structure and flowchart of an autotuning PID controller according to the present invention. In an initial tuning step, a heater plant is modeled to determine an area of an initial PID coefficient. Based on this, an initial search for tuning using uDEAS is performed. Create a point. Decode it to measure the coefficient of performance index and calculate the cost function. If the solution for calculating the cost function is not improved, the above process is repeated. If the solution is improved, the coefficient is determined as the initial PID coefficient.

In the online tuning step, the error error is checked using coefficient values obtained in the initial tuning. The method of checking the error error is obtained by simulation of the heater system. If the error error is less than the tolerance, apply this value to the plant controller to limit the plant. However, if the error error is larger than the allowable error, uDEAS is used to determine the PID coefficients again. The controlled plant returns the temperature as an output value and adjusts the system diagnostics and control coefficients by comparing the output temperature with the control reference.

In the present invention, the parameter is detected by modeling a heater plant to be controlled for tuning. A 3-parameter model is used to model the heater plant, and the transfer function of the 3-parameter model is shown in Equation 1 below.

[Equation 1]

In the above equation, T is a time constant, K is a static gain, and L is a dead time. These values were often found graphically from the measured unit response data. In the present invention, the equation is expressed as a response to the step input to estimate the parameter using an operation optimization technique. When the equation is expressed as a response to the step input, Equation 2 below is given.

[Equation 2]

In the above equation, y _o represents room temperature and can be obtained directly from the measured value. Here, the measured value is the step response to the roll heater, and FIG. 2 shows the step response of the roll heater. In the above equation , three parameters , K, T, and L, to be found by the operation optimization technique, are totaled.

In the present invention, the parameters are estimated by using an operation optimization technique. The parameter should be determined so that the figure of merit is minimized. In the present invention, the performance index, which is the absolute value of the estimated error (Integral Absolute Error, IAE), should be minimized, and the cost function to be minimized is defined as the following equation.

&Quot; (3) "

In the above equation, N is the number of temperature measurement points of the heater. T _f represents the total time the simulation was performed, which is used to represent the integral discretely. FIG. 2 compares Equation 3 with the rise time of each section of the heater plant temperature estimated using the operation optimization technique and the actual heater plant temperature. The present invention proposes a modeling method for a heater system by applying a parameter estimated in Equation 1 to Equation 1.

로 분리가능하며 여기서,

은 최소 위상,

은 크기가 1인 비최소위상이라고 가정한다. 필터

는 공칭폐루프 전달함수를 설계자가 원하는 형태로 구성할 수 있으며, 해당 필터는 공칭모델의 최소 위상 부분을 보상할 수 있도록 선정한다. 필터

를 최소 위상 부분을 보상할 수 있도록 선정하면 다음 수학식 4와 같이 표현할 수 있다.
In addition, the PID controller used to control the model for the estimated heater system in the present invention is an IMC-PID controller structure is shown in FIG. The IMC-PID has internal stability compared to the conventional PID controller structure, and has a complete controller structure in which the output completely follows the input, assuming that there is little modeling error. The general IMC-PID structure uses a primary model including a time delay as a nominal model as shown in FIG. 3. The nominal model (model of the plant)

Can be separated from

Is the minimum phase,

Assumes a non-minimum phase of size 1. filter

The nominal closed-loop transfer function can be configured in the form desired by the designer, and the filter is selected to compensate for the minimum phase portion of the nominal model. filter

If to select to compensate for the minimum phase portion can be expressed as shown in Equation 4.

[Equation 4]

는 필터 계수이며, n은

의 차수가 된다. 이를 이용하여 제어기 P를 나타내면 다음 수학식 5와 같이 표현할 수 있다.
In the above equation

Is the filter coefficient and n is

Is the order of. If controller P is represented using this, it can be expressed as Equation 5 below.

&Quot; (5) "

를 3-파라메터 1차 근사 모델이라고 가정한다. 1차 근사화

되는 다음 수학식 6과 같이 표현할 수 있다.
Also, to approximate the nominal model to the actual plant

Is assumed to be a 3-parameter first approximation model. 1st approximation

It can be expressed as Equation 6 below.

&Quot; (6) "

이며, 정상상태이득(steady state gain)은

는 , 무효시간은

으로 나타낸다. 상기 수학식을 수학식 4와 5를 이용해 정리하면 다음 수학식 7처럼 나타낼 수 있다.
The above equation may be expressed similarly to Equation 1, and a time constant may be expressed to approximate the Equation 1 with an approximated model.

And steady state gain is

, Invalid time is

Represented by When the equations are arranged using Equations 4 and 5, they can be expressed as Equation 7 below.

&Quot; (7) "

Since the non-minimum phase of the above equation is in the form of an exponential function, it is very complicated to calculate, so it is represented by first-order Pade-Approximation for simplicity of operation. Equation 8 shows the non-minimum phase by approximating the first Pade.

[Equation 8]

When the equations (7) and (8) are substituted into equation (5), the equations (7) and (8) may be expressed as in equation (9).

[Equation 9]

The PID transfer function applied to the IMC-PID is shown in Equation 10 below.

[Equation 10]

In the above equation, K (= K _P ), T _d , and T _i represent the proportional coefficient, the integral time coefficient, and the derivative time coefficient, respectively. In relation to Equation 9, the relationship between the controller design variables may be derived as in Equation 11 below.

[Equation 11]

이며 이 값을 이용하여 PID 계수를 결정하게 된다. uDEAS를 이용하여 상기 설계 변수인

를 추정하게 되며 이에 대한검증은 역시 상기 내용에서 설명한 IAE 성능지수를 최소화 하는 방법을 이용한다.
In the above equation, the design variable of the IMC-PID controller is

This value is used to determine PID coefficients. Using uDEAS, the design variable

The verification is also performed using a method of minimizing the IAE performance index described above.

Q (S): Filter model with toughness filter (F)
G (S): Plant to be controlled
G _O (S): nominal model of the plant

Claims (8)

In the method for modeling an industrial heater system using uDEAS,
Using the transfer function of the 3-parameter model due to the nature of the linear heater system; Using uDEAS, an operation optimization technique, to estimate unknown parameters in the model; Applying an IAE (sum of absolute error) to verify a parameter estimated by the uDEAS; Determining whether the titration is appropriate based on the parameter, and repeatedly performing uDEAS, selecting an appropriate parameter and applying the same. The method of modeling an industrial heater system using uDEAS includes;
The method of claim 1,
The cost function of the uDEAS is set as a function of the result of the IAE performance index using the static variable K , time constant T , invalid time L.
The method of claim 1,
The parameter static variables K , time constant T , and invalid time L of the cost function of uDEAS are minimized by the performance index of the IAE, and the uDEAS is repeatedly estimated according to whether they are well identified in the system. Modeling method of industrial heater system using.
Auto-tuning method of heater system modeled using IMC-PID
Applying to a changing heater system using the IMC-PID technique; Finding the applied IMC-PID coefficient with uDEAS; Applying an absolute integration error (IAE) to verify the PID coefficient obtained with the uDEAS; Automatically tuning PID coefficients of a control system by repeatedly performing uDEAS by determining whether the titration coefficient is appropriate based on the PID coefficients. The automatic tuning method of the heater system modeled using the IMC-PID.
The method of claim 4, wherein
The applying IMC-PID to the IMC-PID, characterized in that for adjusting the filter coefficients T _F not the PID technique to find the conventional proportional coefficient (T _P) and the integral coefficient (K _I), the differential coefficient (K _P) only Auto-tuning method of heater system modeled using.
The method of claim 4, wherein
Autotuning method of the heater system modeled using IMC-PID, characterized in that the filter coefficient T _F is adjusted by using uDEAS.
The method of claim 4, wherein
Autotuning method of the heater system modeled using the IMC-PID characterized in that determining the appropriate filter coefficient T _F by using a result function of the IAE performance index.
The method of claim 4, wherein
Autotuning method of a heater system modeled using IMC-PID characterized in that the filter coefficients are automatically found using uDEAS according to whether the filter coefficients are appropriate and the error of the corresponding plant.

Images