Crystallizer liquid level control method based on improved Smith pre-estimation compensation

Abstract

According to the crystallizer liquid level control method based on improved Smith pre-estimation compensation, a Smith pre-estimation compensation link is arranged in a liquid level control system, and pre-estimation and compensation for dynamic response of the control process are established, so that the adjusted quantity of time lag time in the system is reflected to a liquid level regulator in advance, advanced action of the liquid level regulator is achieved, and the crystallizer liquid level is controlled. Comprising the following steps: 1, determining the time lag time of a liquid level control system according to a rod punching signal; 2, completing the structure setting of the Smith predictor in the form of setting response and disturbance response decoupling, and forming a Smith predictor structure with a tracking controller and a disturbance controller; and 3, respectively completing the determination of each parameter in the tracking controller and the disturbance controller, and completing the prediction compensation based on the Smith according to the determination. According to the crystallizer liquid level control method based on improved Smith pre-estimation compensation, time delay time can be determined, and system steady-state errors can be eliminated and controlled.

Description

Technical Field

[0001] The invention belongs to the technical field of steelmaking and continuous casting, and in particular relates to a crystallizer liquid level control method based on improved Smith predictive compensation. Background Art

[0002] Stable control of the crystallizer liquid level is crucial for the safe production of continuous casting and ensuring product quality. The molten steel is injected into the crystallizer through the submerged water nozzle through the adjustment of the stopper mechanism in the tundish. Under the action of the cold water in the crystallizer, the molten steel forms a cast billet with a certain shell thickness. The continuous casting machine pulls the gradually solidified cast billet out of the crystallizer at the speed set by the process. In the continuous casting production process, the crystallizer liquid level is required to be controlled within a certain range. If the liquid level fluctuates too much, it is easy to cause slag inclusions and cracks on the surface of the cast billet, resulting in quality loss, or even steel leakage, causing more serious production accidents. Therefore, the high-precision control method of the continuous casting crystallizer liquid level has received great attention. Like most industrial processes, the crystallizer liquid level control and adjustment has a time lag problem, which will lead to the deterioration of the crystallizer liquid level control quality, and Smith predictive control is an effective method to solve pure time lag process objects. However, there are two difficult problems to solve when applying the classic Smith predictive control to the crystallizer liquid level control system. 1. The crystallizer liquid level control system is a time-varying lag system, and the lag time varies with different working conditions and steel types, making it difficult to obtain accurately. 2. Since the crystallizer object is an integral proportional link, the steady-state error of the disturbance is not zero.

[0003] For problem 1, the hysteresis characteristics of the liquid level can be accurately identified in real time based on the data of the crystallizer liquid level control system, which can be used to improve the monitoring of liquid level flow characteristics and improve the control effect. Conventional system identification methods are mainly divided into non-parametric model identification methods and parametric model identification methods. The non-parametric model identification method assumes that the system is linear, but there is no need to determine the specific structure of the model in advance. According to the input and output of continuous time, the parameter model of the system is calculated by the response characteristics. This type of method can be applied to any complex system and is often used in engineering. This type of method mainly includes step response method, frequency method and correlation method; the parametric model identification method must first assume the model structure and determine the model parameters by minimizing the error function between the model output and the actual output of the system. This type of method can be divided into least squares method, gradient correction method and maximum likelihood method according to different basic principles.

[0004] Since there are many factors that affect the liquid level of the crystallizer, they are coupled with each other and the signals are mixed together. It is difficult to identify the lag time of the crystallizer adjustment system using conventional system identification methods based solely on the existing time domain signals.

[0005] For question 2, existing technologies include identifying the object model and time lag size through the collected step response data, and then correcting and compensating the accumulated error caused by the process disturbance by setting the prediction function. However, in the actual continuous casting production process, the step response cannot be set for the crystallizer liquid level control system, which will have a great impact on the liquid level control system and may cause serious production accidents. Moreover, the method of reducing the disturbance steady-state error through prediction function control is more difficult to select the prediction function, and it is difficult to ensure long-term stability.

[0006] How to identify the pure lag time of the crystallizer liquid level control process in real time during the production process under Smith predictive compensation and ensure that no disturbance steady-state error occurs has become an urgent problem to be solved in the existing crystallizer liquid level control based on improved Smith predictive compensation.

[0007] The invention application with application number: CN201310647691X discloses "a servo system identification method based on time domain characteristics of relay position feedback", and its implementation steps are: setting parameters of the relay module and experimental parameters; conducting a relay position feedback experiment and recording the corresponding actual displacement oscillation curve; determining the oscillation amplitude and oscillation period; determining the servo system model and its parameters; and accurately calculating the controlled servo system gain and time constant.

[0008] The invention application with application number: CN2019111061938 discloses "a linear quadratic optimal dynamic feedforward-feedback PID control system based on a closed-loop identification model and its control method", including sampling on-site closed-loop operation data, closed-loop identification of the ARX model, converting it into an observable normalized discrete-time state space model, and obtaining a measurable disturbance quantity feedforward control model; designing quadratic performance indicators; minimizing the performance indicators to obtain a linear state feedback control matrix; calculating the state quantity estimate of the measurable disturbance quantity, and then obtaining the feedforward control law of the linear optimal quadratic dynamic feedforward controller; combining the linear quadratic optimal dynamic feedforward controller with the PID controller to design a feedforward-feedback PID control system.

[0009] The invention application with application number: CN2022105346458 discloses "a predictive function control method for a time-delay integral process based on Smith predictive correction". First, a model is built based on the step response data collected in real time to find out the basic characteristics of the object, and then a new error correction and compensation is performed on the predictive function control based on Smith predictive correction. Finally, the obtained optimal control law is implemented on the controlled integral process. Summary of the invention

[0010] In order to solve the above problems, the present invention provides a method for controlling the liquid level of a crystallizer based on improved Smith prediction compensation, and the technical solution thereof is as follows:

[0011] A method for controlling the liquid level of a crystallizer based on improved Smith predictive compensation is provided. By setting a Smith predictive compensation link in a liquid level control system, a pre-estimation and compensation of the dynamic response of a control process is established, so that the controlled variable of the time lag time in the system is advancedly reflected to the liquid level regulator, and the advance action of the liquid level regulator is realized. The method is characterized in that it comprises the following steps:

[0012] S1: Determine the time delay of the liquid level control system according to the punch signal;

[0013] S2: Complete the structural setting of the Smith predictor in the form of decoupling the set response and the disturbance response, forming a Smith predictor structure with dual controllers of a tracking controller and a disturbance controller;

[0014] S3: Complete the determination of each parameter in the tracking controller and the disturbance controller respectively, and complete the Smith-based estimated compensation according to the determination.

[0015] Further,

[0016] Step S1 is specifically as follows:

[0017] S11: According to the set sampling time, the actual stopper opening signal and the measured liquid level signal are periodically obtained to form respective time series signals;

[0018] S12: performing deviation processing on the time series signal of the actual stopper rod opening and the time series signal of the measured liquid level, respectively, to obtain the stopper rod opening fluctuation signal and the liquid level fluctuation signal;

[0019] S13: performing time-frequency conversion on the stopper rod opening fluctuation signal and the liquid level fluctuation signal, and determining the stopper rod opening phase and the liquid level phase at the frequency according to the set frequency of the plunger rod signal; and determining the actual phase difference of the plunger rod signal accordingly;

[0020] S14: determining a theoretical lag phase of the punch signal according to the set frequency of the punch signal and the structural characteristics of the mold object;

[0021] S15: According to the actual phase difference of the punch signal determined in step S13 and the theoretical lag phase of the punch signal determined in step S14, the time delay time is determined.

[0022] Further,

[0023] Step S14 is specifically as follows: according to the proportional differential link structure of the crystallizer object and the set frequency of the punch signal, the theoretical lag phase of the punch signal is determined as follows:

[0024]

[0025] In the formula,

[0026] Theoretical lag phase of the punch signal, unit: degree;

[0027] f0: The set frequency of the punch signal, unit: Hz;

[0028] L: lag time, unit: S.

[0029] Further,

[0030] The disturbance controller in step S2 is structured based on a basic control method that uses proportional control to characterize the steady-state error and integral control to reduce the steady-state error.

[0031] Further,

[0032] The disturbance controller has the following specific structure:

[0033]

[0034] In the formula,

[0035] k1, k2, k3: optimal parameters.

[0036] Further,

[0037] The determination of the parameters in the disturbance controller in step S3 is completed based on the constraint that the steady-state error is zero.

[0038] The present invention discloses a method for controlling a crystallizer liquid level based on improved Smith predictor compensation, in which the phase lag of a punch signal is characterized by a model lag, and accordingly, the time lag is determined by the punch signal; and by setting the Smith predictor to a dual controller structure with a tracking controller and a disturbance controller, the tracking controller and the disturbance controller are optimized separately, thereby achieving control of the steady-state error of the disturbance response under the premise that the set response and the disturbance response do not interfere with each other, so as to overcome the deterioration effect of the introduction of Smith predictor compensation on the disturbance response characteristics of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the steps of Smith estimation compensation of the present invention;

[0040] Figure 2 A schematic diagram of the step of determining the time lag time in the present invention;

[0041] Figure 3 It is a block diagram of the control system based on improved Smith predictive compensation of the present invention;

[0042] Figure 4It is a block diagram of a generally conventional Smith predictive compensated crystallizer liquid level control system in the explanation of the principle and setting process of the present invention;

[0043] Figure 5 A flow chart of the lag time identification process in the principle and setting process of the present invention;

[0044] Figure 6 It is a simulation model based on the improved Smith predictive compensation control system in an embodiment of the present invention;

[0045] Figure 7 It is a time series schematic diagram of the actual opening and the measured liquid level in an embodiment of the present invention;

[0046] Figure 8 The phase spectrum of the plug rod opening and the measured liquid level plug rod component of the simulation model in the embodiment of the present invention;

[0047] Fig. 9 It is a schematic diagram of the dynamic lag time identification result in an embodiment of the present invention;

[0048] Fig.10 It is a schematic diagram of comparison results of measuring liquid level following set liquid level under different control conditions in an embodiment of the present invention;

[0049] Fig.11 It is a schematic diagram of the comparison results of the measured liquid level following the set liquid level under different control conditions in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] Below, a crystallizer liquid level control method based on improved Smith predictive compensation of the present invention is further described in detail according to the drawings and specific implementation methods of the specification.

[0051] Principle and setting process description:

[0052] like Figure 1 A method for controlling the liquid level of a crystallizer based on improved Smith predictive compensation is shown in FIG. Figure 4 The conventional Smith predictive compensation crystallizer level control system shown in FIG. 1 is improved, and the block diagram of the improved control system is shown in FIG. Figure 3 As shown, after the improvement, the set response and the disturbance response can be completely decoupled, and by setting the specific structural parameters of H(s), the problem of the traditional Smith predictive control that the integral link cannot eliminate the steady-state error of the disturbance for the control object model is solved. At the same time, the real data such as the stopper opening and the measured liquid level in the actual production process of the crystallizer liquid level control are used to realize the pure delay time online identification of the crystallizer liquid level control system based on the punch signal of this technical solution.

[0053] In the process of crystallizer liquid level control, there is a pure lag process in the controlled process object, which leads to poor control effect due to untimely feedback. Smith predictor compensation is an effective means to solve the above problem. The output result is predicted by the mathematical model of the prior object process and the lag time, and then the input of the system is adjusted in time according to the predicted result. In an ideal situation, the regulation performance of the system can be the same as that of the non-lag process, except that the output is delayed in time. The above original PI control system is improved by Smith predictor to obtain the crystallizer liquid level control system with Smith predictor compensation, as shown in Figure 1 shown.

[0054] Where s is the Laplace transform operator, Y(s) is the Laplace transform of the measured liquid level, R(s) is the Laplace transform of the set liquid level, K(s) is the set of PI controllers and actuators Ga(s), Gp(s) includes the flow coefficient and the crystallizer object, e -Ls is a pure lag function, e -L′s is the lag function for identification.

[0055] The closed-loop transfer function of the original crystallizer liquid level PI control system is:

[0056]

[0057] The closed-loop transfer function of the Smith estimated compensation crystallizer level control system is:

[0058]

[0059] When L′=L, the closed-loop transfer function is:

[0060]

[0061] The pure lag link of the original PI crystallizer liquid level control system exists in the closed-loop characteristic equation, which will have an adverse effect on the stability and dynamic performance of the system. Through the crystallizer liquid level control system compensated by Smith predictor, the pure lag link is no longer included in the characteristic equation of the system and is transferred outside the closed-loop control loop, so it will no longer have an adverse effect on the system. The introduction of Smith predictor can well compensate the system containing pure lag, which can effectively improve the stability and dynamic performance of the system.

[0062] However, the introduction of Smith prediction compensation leads to a deterioration in the disturbance response characteristics. Figure 4 According to the Mason formula, the transfer function of the system from process disturbance D(s) to measured liquid level Y(s) is:

[0063] In PI control, the disturbance response is:

[0064]

[0065] When Smith predictive compensation control is used, the disturbance response is:

[0066]

[0067] When L′=L, the disturbance response is:

[0068]

[0069] The transient response of the disturbance depends on the object model and whether the pole of the object is close to the imaginary axis. The crystallizer object process is a proportional integral model. If there is a constant process disturbance, the impact of the disturbance may exist for a long time in this case. According to the above equation, it is obvious that the Smith predictive compensation crystallizer liquid level control has poor anti-interference ability, which can be further verified according to the final value theorem. The final value theorem of the disturbance response is:

[0070]

[0071]

[0072] When D(s) is a unit step disturbance, it will cause a constant steady-state error in the disturbance response. This has an adverse effect on accurate mold level control.

[0073] By changing the Smith predictor structure, the set response and disturbance response are decoupled, so the set response and disturbance response can be optimized separately. And the lag time is identified by the stopper rod opening and the measured liquid level data, providing real-time parameter updates for the time lag of the improved Smith predictor. The crystallization liquid level control system with Smith predictor compensation set by this technology is as follows: Figure 3 The designed control system modules include PI controller, liquid level actuator, flow coefficient, dynamic lag time, crystallizer object, improved Smith predictor, and lag time identification module.

[0074] Set the response to:

[0075]

[0076] The disturbance response is:

[0077]

[0078] When L′=L,

[0079]

[0080]

[0081] Since the control object Gp(s) is unchangeable, the set response can optimize K(s) separately to achieve a better follow-up effect on the set liquid level without affecting the disturbance response; the disturbance response can optimize H(s) separately to achieve a better suppression effect on the disturbance without affecting the set response.

[0082] The optimization method of K(s) can be determined by conventional PID parameter tuning methods in engineering, including: critical proportion method, reaction curve method, attenuation method, expanded critical proportion tuning method, etc. The selection of H(s) is more complicated. First of all, it is necessary to ensure that the disturbance response does not have a steady-state error and to reduce overshoot as much as possible.

[0083] H(s) selects a specific structure, the structure is as follows:

[0084]

[0085] Among them, k1, k2, k3 are optimal parameters, and L is the system lag time.

[0086] The disturbance response at this time is:

[0087]

[0088] According to the final value theorem, the final stable liquid level output of the unit step disturbance is:

[0089]

[0090] To ensure that the steady-state error is 0, let:

[0091] k4=k2+k3*L

[0092] This ensures that:

[0093]

[0094] By selecting different k1, k2, and k3, different control effects on disturbance responses can be obtained, and they can be used as tuning parameters.

[0095] This technology not only improves the structure of Smith predictive control, and eliminates the steady-state error of disturbance through a specific structure, but also can realize lag time identification according to the stopper opening and measured liquid level data under different working conditions and steel grade changes, and perform online adjustment of Smith predictor lag parameters. The technical solution of real-time lag time identification of this technology, such as Figure 5As shown, the process is: 1) Collect the discrete actual stopper rod opening data and measured liquid level data collected by the PLC in S7-400 according to the scanning cycle; 2) Preprocess the actual stopper rod opening and measured liquid level data obtained; 3) Perform fast Fourier transform FFT and phase angle conversion on the actual opening data and measured liquid level data preprocessed in step 2 to obtain the phase of each frequency of the actual opening and measured liquid level; 4) Through the phase obtained in step 3, find the phase difference corresponding to the plunger frequency in the actual opening and measured liquid level; 5) According to the phase-frequency characteristics of the process model from the stopper rod opening to the measured liquid level, calculate the theoretical lag phase of the plunger signal from the actual opening to the measured liquid level; 6) Calculate the initial lag time L' based on the actual phase difference of the plunger signal from the actual opening to the liquid level calculated in step 4 and the theoretical lag phase calculated according to the phase-frequency characteristics of the process model in step 5; 7) Update the time lag parameter in the parameter model of the Smith predictor. The specific process of lag time identification is as follows (can be combined with Figure 2 To understand):

[0096] (1) Collect the discrete actual stopper rod opening data and measured liquid level data collected by the PLC in S7-400 according to the scanning cycle. Each time, take the actual stopper rod opening and measured liquid level time series data with the current adjacent time span of TN, and the data update window is 1s.

[0097] (2) Preprocess the actual stopper rod opening data and measured liquid level data taken out, and obtain the actual stopper rod opening and measured liquid level by subtracting the average value to obtain the fluctuation period signal. The processing method is as follows:

[0098] Liquid level fluctuation signal y(n) = measured liquid level signal - average value of measured liquid level signal Opening fluctuation signal sn(n) = actual opening signal - average value of actual opening signal

[0099] The above-mentioned liquid level fluctuation signal y(n) and opening fluctuation signal sn(n) are discrete time signals, the data length is N, and the sampling period is Ts.

[0100] (3) The liquid level fluctuation signal y(n) and the actual opening signal sn(n) obtained in step 2 are subjected to fast Fourier transform FFT to obtain a complex sequence in the frequency domain. The complex sequence is subjected to angle calculation in four quadrants to obtain the phase of the final sampling frequency interval of the liquid level fluctuation signal and the actual opening signal. The specific process is as follows:

[0101] Perform discrete Fourier transform DFT on the liquid level fluctuation signal y(n) and the opening fluctuation signal sn(n) to obtain the discrete Fourier transform sequence:

[0102]

[0103] In the above formula:

[0104]

[0105] The basic idea of ​​the FFT algorithm is to use the fact that some items in the DFT operation can be combined and to decompose the long sequence DFT into a short sequence DFT. It is not a new transformation, but a fast algorithm for simplifying the discrete Fourier transform operation, so the transformed frequency domain sequence is the same as above.

[0106] According to Euler's formula, we know that:

[0107]

[0108] Through the above process, we can know that the final result obtained by FFT is a complex sequence:

[0109]

[0110] The complex sequence performs angle calculations in four quadrants, and the phase of the complex sequence is:

[0111]

[0112]

[0113] There is a corresponding relationship between discrete k and frequency f, as follows:

[0114]

[0115]

[0116] Where fs is the sampling frequency of the actual opening signal and the measured liquid level. Through the above process, the phase of the actual opening and the measured liquid level at each sampling interval frequency can be obtained, that is, ∠Y(f) and ∠SN(f). When k = 0, the corresponding frequency f = 0, and ∠SN(f) = 0 and ∠Y(f) = 0. Obviously, the DC component has no phase.

[0117] (4) According to the frequency f of the anti-blocking punch signal o , find the corresponding punch component in the fast Fourier transform FFT results of the actual opening signal and the measured liquid level signal, and according to step 3, obtain the phase ∠SN (f o )=ph1 and ∠Y(f o )=ph2, and the phase difference Δph is calculated as follows:

[0118]

[0119] (5) According to step 4, the actual phase difference Δph of the punch signal is obtained, and the process model from the actual opening of the tundish plug to the crystallizer liquid level data is known as:

[0120]

[0121] Where Gc is the flow coefficient, A is the cross-sectional area of ​​the crystallizer, s is the Lagrange differential operator, and Gp(s) is the crystallizer object. It can be seen that the crystallizer object is similar to a container and can be approximated as an integral proportional link. What needs to be identified is the lag time L.

[0122] The process model from actual opening to measured liquid level data can obtain the theoretical phase-frequency characteristics as follows:

[0123]

[0124] The frequency of the punch signal is known to be f o According to the theoretical phase-frequency characteristics, the theoretical lag phase of the punch signal from the actual opening to the measured liquid level can be obtained as:

[0125]

[0126] (6) The actual stopper rod opening action is basically to offset the up and down fluctuations caused by the process disturbance D(s) of the measured liquid level. Due to the existence of process disturbance D(s), the time domain or frequency domain identification results using the actual stopper rod opening and the measured liquid level directly are not good. Process disturbance D(s) mainly includes low-frequency interference: bulging, standing waves, etc., as well as some high-frequency noise. However, the plunger rod signal is different: 1) The plunger rod signal frequency is higher, while the high-frequency noise energy is more dispersed, and the amplitude is small near the plunger rod frequency, so the external disturbance has little effect on the plunger rod signal phase; 2) The plunger rod signal is the active action of the plunger rod according to a specific set frequency and amplitude, which is different from the passive action of the plunger rod caused by the feedback of the liquid level fluctuation. The above two points are the main reasons why the plunger rod signal can be used alone for lag time identification.

[0127] According to step 4, the actual phase difference Δph of the punch signal is obtained, and the theoretical lag phase of the punch signal is obtained in step 5. To calculate the lag time L', the expression is:

[0128]

[0129] Simplify the lag time calculation expression to get:

[0130]

[0131] L′ is the identification delay time, Δph is the actual phase difference, f o is the frequency of the punch signal, L minis the minimum lag time, L max is the maximum lag time.

[0132] (7) According to the lag time L′ calculated in step 6, the Smith predictor parameter, i.e., the lag time L′, is updated online, and the stability and dynamic characteristics of the original crystallizer liquid level control system are improved by the Smith predictor.

[0133] Example

[0134] 1. Specific identification method of time lag parameters

[0135] The simulation model of the crystallizer liquid level control system of this embodiment is as follows: Figure 6 As shown:

[0136] a) Liquid level setting value: When 0 to 250 seconds, the liquid level setting value is 64; when 251 to 1000 seconds, the liquid level setting value is 84.

[0137] b) The expression of PI controller is:

[0138]

[0139] c) The expression of liquid level actuator is:

[0140]

[0141] d) The anti-blocking punch signal is set to a sinusoidal periodic signal with a frequency of 1 Hz and an amplitude of 1.

[0142] e) The process disturbance d(t) includes standing wave signal, bulging signal, and Gaussian white noise. The standing wave signal is set to have a frequency of 0.6 Hz and an amplitude of 0.2; the bulging signal is set to have a frequency of 0.055 Hz and an amplitude of 0.2.

[0143] f) The controlled object Gp(s) includes the flow coefficient Gc of the stopper opening into flow, and the crystallizer process object The expression is:

[0144]

[0145] The flow coefficient Gc=0.35, and the cross-sectional area of ​​the crystallizer A=0.264.

[0146] g) Dynamic lag time e -Ls , where L changes with time t, the process is as follows:

[0147] L=L0+sin(2*pi*f1*t)

[0148] L0 is the lag time under normal working conditions, L0=0.5; f1 is the delay time change frequency, f1=0.001Hz; t is time; the lag time range is L∈[0,1].

[0149] h) Specific structure H(s), the structural parameters are as follows:

[0150]

[0151] Assume a set of parameters k1=4, k2=3, k3=0.6, k4=k2+k3*L′, L′ is the lag time identification result. The final H(s) is:

[0152]

[0153] i) Delay time identification process

[0154] According to step 1, take the simulated actual stopper opening and measured liquid level data for 200 seconds, such as Figure 7 shown.

[0155] 1) After data preprocessing in step 2, the periodic fluctuation signal of the actual opening and the measured liquid level is obtained.

[0156] 2) The periodic fluctuation signal of the stopper rod opening and the measured liquid level obtained in step 2 is subjected to fast Fourier transform FFT according to the process of step 3, and angle calculation is performed according to four quadrants to obtain the FFT transformation phase spectrum, wherein the phase spectrum of the stopper rod opening and the punch rod signal in the measured liquid level is as follows: Figure 8 shown.

[0157] 3) According to Figure 6 It can be seen that the phase of the plug rod opening plunger signal component is ph1 = 264.7°, and the phase of the liquid level plunger signal component is ph2 = -6.531°.

[0158] 4) According to step 3, the phase ph1 of the plug rod opening plunger signal component and the phase ph2 of the liquid level plunger signal component are calculated, and the actual phase difference of the plunger signal is calculated according to the phase difference expression of step 4:

[0159] Δph=ph2-ph1=-271.231°

[0160] 5) According to the process model of the actual opening to the measured liquid level of the above simulation model, the theoretical lag phase of the punch signal from the actual opening to the measured liquid level is calculated according to step 5

[0161]

[0162] 6) Calculate the current lag time according to the lag time calculation expression obtained in step 6:

[0163]

[0164] 7) Update the Smith predictor lag parameter according to the lag time L′ obtained in step 6 above.

[0165] Simulation experiment

[0166] Within the simulation time of 1000s, the real lag time L changes dynamically around 0.5s. According to the real-time data of the actual opening and the measured liquid level at the corresponding moment, the lag time L' is repeatedly calculated according to the above process i). The comparison between the identified lag time result and the actual real lag time is as follows: Fig. 9 shown.

[0167] There are some errors between the identified lag time result and the real lag time result, but they are basically within the acceptable range. According to the above identified lag time L′ result, the model lag parameters of the Smith predictor are updated online in real time to realize the crystallizer liquid level control based on Smith predictor compensation of lag time identification.

[0168] When the liquid level is set at 250s, it changes from 64mm to 84mm. Figure 7 It can be seen that the real pure lag time of the system is 1s. PI control, Smith predictive compensation control (lag parameter is 0.5s), and improved Smith predictive compensation control based on lag time identification are compared respectively. The simulation results of the measured liquid level following the set liquid level under different control algorithms are shown in Figure 2. Fig.10 shown.

[0169] Compared with the traditional Smith predictive control, the improved crystallizer liquid level control system based on time identification of the technical solution can follow the set liquid level changes more quickly and in real time, with small overshoot and quickly reach a stable state.

[0170] At 500s, a step disturbance is added to the process disturbance. PI control, Smith predictive compensation control (delay parameter is 0.5s), and improved Smith predictive compensation control based on lag time identification are compared. The simulation results of the measured liquid level following the set liquid level under different control algorithms are shown in Figure 2. Fig.11 shown.

[0171] according to Fig.11 The simulation results show that the conventional Smith predictive control cannot eliminate the steady-state error of the disturbance response, resulting in the measured liquid level not being able to follow the set liquid level well. The crystallizer liquid level control system with improved Smith predictive compensation based on lag time identification can not only eliminate the steady-state error problem of the disturbance response, but also achieve better control accuracy.

[0172] This technology achieves complete decoupling of the set response and the disturbance response by improving the structure of the Smith predictor, and solves the problem that the Smith predictor control contains an integral link for the crystallizer object and cannot eliminate the stability error of the disturbance by designing a specific structure of H(s). In addition, a new lag time identification method is designed and implemented, which can adjust the lag parameters of the Smith predictor online. The above three simulation results show that the crystallizer liquid level control system with improved Smith predictor compensation based on lag time identification of this technical solution can effectively identify the lag time change online; when the set value changes, this technical solution can follow the set liquid level more quickly with a smaller overshoot than PI control and Smith predictor control; when sudden disturbances such as nodule / shedding occur, this technical solution will not have problems such as steady-state error.

Claims

1. A method for controlling the liquid level of a crystallizer based on improved Smith predictive compensation, by setting up a Smith predictive compensation link in the liquid level control system, establishing a pre-estimation and compensation for the dynamic response of the control process, so that the controlled variable of the time lag time in the system is advanced to the liquid level regulator, and the advance action of the liquid level regulator is realized, characterized in that: The steps include: S1: Determine the time delay of the liquid level control system according to the punch signal; S2: Complete the structural setting of the Smith predictor in the form of decoupling the set response and the disturbance response, forming a Smith predictor structure with dual controllers of a tracking controller and a disturbance controller; S3: Complete the determination of each parameter in the tracking controller and the disturbance controller respectively, and complete the Smith-based estimated compensation according to the determination.

2. A method for controlling a crystallizer liquid level based on improved Smith predictive compensation according to claim 1, characterized in that: Step S1 is specifically as follows: S11: According to the set sampling time, the actual stopper opening signal and the measured liquid level signal are periodically obtained to form respective time series signals; S12: performing deviation processing on the time series signal of the actual stopper rod opening and the time series signal of the measured liquid level, respectively, to obtain the stopper rod opening fluctuation signal and the liquid level fluctuation signal; S13: performing time-frequency conversion on the stopper rod opening fluctuation signal and the liquid level fluctuation signal, and determining the stopper rod opening phase and the liquid level phase at the frequency according to the set frequency of the plunger rod signal; and determining the actual phase difference of the plunger rod signal accordingly; S14: determining a theoretical lag phase of the punch signal according to the set frequency of the punch signal and the structural characteristics of the mold object; S15: According to the actual phase difference of the punch signal determined in step S13 and the theoretical lag phase of the punch signal determined in step S14, the time delay time is determined.

3. A method for controlling a crystallizer liquid level based on improved Smith predictive compensation according to claim 2, characterized in that: Step S14 is specifically as follows: according to the proportional differential link structure of the crystallizer object and the set frequency of the punch signal, the theoretical lag phase of the punch signal is determined as follows: In the formula, Theoretical lag phase of the punch signal, unit: degree; f0: The set frequency of the punch signal, unit: Hz; L: lag time, unit: S.

4. The method for controlling the mold level based on improved Smith predictive compensation according to claim 1, characterized in that: The disturbance controller in step S2 is structured based on a basic control method that uses proportional control to characterize the steady-state error and integral control to reduce the steady-state error.

5. The method for controlling the liquid level of a crystallizer based on improved Smith prediction compensation according to claim 4, characterized in that: The disturbance controller has the following specific structure: In the formula, k1, k2, k3: optimal parameters.

6. The method for controlling the crystallizer liquid level based on improved Smith predictive compensation according to claim 1, characterized in that: The determination of the parameters in the disturbance controller in step S3 is completed under the premise that the steady-state error is zero.

Images