Crystallizer liquid level control method with surface wave filtering structure and surface wave filtering structure

Abstract

According to the crystallizer liquid level control method with the surface wave filtering structure and the surface wave filtering structure, a surface wave monitoring and filtering structure based on a liquid level measurement signal is arranged in a feedback loop of crystallizer liquid level feedback control; periodic real-time monitoring of liquid level fluctuation signals is established through the arranged surface wave monitoring and filtering structure. When the monitoring result is larger than or equal to a set threshold value, surface wave filtering based on a wave trap is started, and liquid level measurement signal feedback based on surface wave filtering is formed; otherwise, direct feedback based on the liquid level measurement signal is formed. According to the crystallizer liquid level control method with the surface wave filtering structure and the surface wave filtering structure, through production and use after the technical setting, the harmful influence of surface waves on crystallizer liquid level control in the actual production process can be effectively reduced.

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 with a surface wave filtering structure and a surface wave filtering structure. 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 adjusted by the stopper mechanism in the tundish and injected into the crystallizer through the immersed water nozzle. Under the action of the cold water in the crystallizer, the molten steel forms a billet with a certain shell thickness. The continuous casting machine pulls the gradually solidified 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 billet, resulting in quality loss, or even steel leakage, which brings about more serious production accidents. Therefore, the high-precision control method of the continuous casting crystallizer liquid level has received great attention.

[0003] In the actual production process, we have noticed a flow asymmetry caused by crystallizer vibration or bifurcated outflow from the immersed water nozzle, which causes the molten steel inside the crystallizer to sway left and right. At this time, the internal liquid level value of the crystallizer is detected by the liquid level meter VUHZ, and the measured liquid level will show periodic fluctuations, but the total volume of the molten steel inside the crystallizer has not changed. Automatically adjusting the process through the liquid level PID will intensify this oscillation. This liquid level oscillation process is referred to as surface wave. Once the above problem occurs in actual working conditions, general operators will choose to switch the liquid level control to manual open-loop control, or choose to reduce the pulling speed to ensure the safety and stability of production. However, this method of suppressing surface waves by switching to manual open-loop control or reducing the pulling speed not only greatly reduces production efficiency, but also adds a large workload to on-site operators.

[0004] The invention application with application number: CN201310287655.7 discloses "a method for controlling liquid level fluctuation of a crystallizer", which utilizes the original continuous casting liquid level measurement system, communication control system and stopper rod control system, and adds a liquid level fluctuation analysis system, and utilizes the liquid level detection unit of the crystallizer liquid level measurement system to transmit the fluctuation data in real time to the liquid level fluctuation analysis system based on the spectrum analysis method, decomposes each frequency component by means of Fourier transform, and measures the relative importance of each component by means of spectral density function, finds out the main frequency components existing in the sequence, thereby grasps the periodic fluctuation characteristics of the sequence, analyzes the liquid level fluctuation frequency range of the main position of crystallizer slag curling under steady-state and non-steady-state casting conditions, thereby monitoring the liquid level fluctuation in real time, reducing the casting speed by controlling the stopper rod, and adjusting the protective slag and water outlet angle to reduce the possibility of slag curling.

[0005] The invention application with application number: CN 202210767436.8 discloses "a method and system for distinguishing abnormal fluctuations in the liquid level of a crystallizer", which can be applied to offline historical data analysis and online evaluation of the fluctuation of the liquid level of a crystallizer, reducing the impact of the fluctuation of the liquid level of the crystallizer on the quality of the ingot, reducing the loss of the quality of the ingot, and improving the efficiency of continuous casting production. It uses the crystallizer liquid level fluctuation data and an analysis method based on a combination of fast Fourier transform and wavelet entropy to comprehensively analyze the fluctuation of the molten steel in the crystallizer over a period of time (different furnaces or different castings), accurately locate the time when the abnormal fluctuation of the liquid level occurs, and quickly trace the cause of its occurrence. Summary of the invention

[0006] In order to solve the above problems, the present invention provides a crystallizer liquid level control method with a surface wave filtering structure and a surface wave filtering structure, wherein a crystallizer liquid level control method with a surface wave filtering structure has a technical solution as follows:

[0007] In the feedback loop of the crystallizer liquid level feedback control, a surface wave monitoring and filtering structure based on the liquid level measurement signal is set up.

[0008] The periodic real-time monitoring of the liquid level fluctuation signal is established through the surface wave monitoring and filtering structure;

[0009] When the monitoring result is greater than or equal to the set threshold, the surface wave monitoring and filtering structure based on the liquid level measurement signal is turned on to form a liquid level measurement signal feedback based on the surface wave filtering; otherwise, a direct feedback based on the liquid level measurement signal is formed.

[0010] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0011] The surface wave monitoring and filtering structure is composed of a surface wave identification unit and a notch filter unit.

[0012] The surface wave identification unit is used to receive the liquid level measurement signal and the mold width detection signal in real time, and determine the actual notch center frequency, notch depth, notch bandwidth and the fluctuation signal amplitude corresponding to the actual notch center frequency according to the two signals;

[0013] The notch filter unit is used to receive the amplitude of the fluctuation signal corresponding to the actual notch center frequency, and determine whether to start the notch filter according to the comparison result of the fluctuation signal amplitude and the set threshold.

[0014] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0015] The method of determining the actual notch center frequency according to the liquid level measurement signal and the mold width detection signal specifically includes the following steps:

[0016] S1: Determine the theoretical frequency of the surface wave according to the width of the crystallizer and the order of the trap;

[0017] S2: Zero-mean processing is performed on the liquid level measurement signal to obtain a liquid level fluctuation signal;

[0018] S3: Perform time-frequency conversion on the liquid level fluctuation signal and determine the actual notch center frequency in combination with the theoretical frequency of the surface wave.

[0019] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0020] Based on the notch depth and notch bandwidth, the specific determination steps are as follows:

[0021] SS1: Determine the phase margin of the current system with notch filter in real time;

[0022] SS2: Calculate the difference between the theoretical phase margin of the system and the phase margin of the current system with the notch filter. When the calculated result is greater than the set value, go to step SS3, otherwise go to step SS4;

[0023] SS3: Determine the notch depth to be 0.4 and the notch bandwidth to be 0.26;

[0024] SS4: The notch depth is determined according to the actual surface wave amplitude, and the notch bandwidth is determined according to the common constraints of the normal feedback of the liquid level and the robustness of the notch filter.

[0025] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0026] The theoretical frequency of the surface wave is determined according to the following formula:

[0027]

[0028] In the above formula,

[0029] f: theoretical frequency of surface wave, unit: Hz;

[0030] c: the order of the notch filter;

[0031] g: acceleration due to gravity, unit: m / s 2 ;

[0032] L: Crystallizer width, unit: m.

[0033] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0034] The time-frequency conversion of the liquid level fluctuation signal is completed by spectrum refinement analysis based on the frequency shift method.

[0035] Based on this and the theoretical frequency of the surface wave, the actual notch center frequency is determined. Specifically, a peak search is performed within a frequency range set near the theoretical frequency of the surface wave to obtain the maximum value of the amplitude spectrum of the liquid level fluctuation signal after time-frequency conversion within the frequency range. The actual notch center frequency is determined based on this maximum value.

[0036] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0037] The notch depth is determined according to the actual surface wave amplitude, specifically:

[0038] Set the target amplitude Af of the notch center frequency after filtering.

[0039] Determine the surface wave amplitude A in real time;

[0040] The notch depth is determined by the ratio of Af to A.

[0041] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0042] The notch bandwidth ranges from [0.05, 0.4].

[0043] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0044] The notch filter unit includes a first notch filter and a second notch filter which are serially connected in a feedback loop.

[0045] According to a crystallizer liquid level control method with a surface wave filtering structure of the present invention,

[0046] The first notch filter and the second notch filter are respectively a first-order notch filter and a second-order notch filter.

[0047] A crystallizer liquid surface wave filtering structure,

[0048] The surface wave filtering structure is arranged in the feedback loop of the crystallizer liquid level feedback control system. According to the surface wave filtering structure, periodic real-time monitoring of surface waves based on the liquid level measurement signal is established; according to the monitoring result, the surface wave filtering structure controls to form two feedback forms: feedback of the liquid level measurement signal after filtering the surface wave when the monitoring value exceeds the set threshold, or direct feedback based on the liquid level measurement signal when the monitoring value does not exceed the set threshold.

[0049] Further,

[0050] The surface wave filtering structure is composed of a surface wave identification unit and a notch filter unit.

[0051] The surface wave identification unit is used to receive the liquid level measurement signal and the mold width detection signal in real time, and determine the actual notch center frequency, notch depth, notch bandwidth and the fluctuation signal amplitude corresponding to the actual notch center frequency according to the two signals;

[0052] The notch filter unit is used to receive the amplitude of the fluctuation signal corresponding to the actual notch center frequency, and determine whether to start the notch filter according to the comparison result of the fluctuation signal amplitude and the set threshold.

[0053] Further,

[0054] The method of determining the actual notch center frequency according to the liquid level measurement signal and the mold width detection signal specifically includes the following steps:

[0055] S1: Determine the theoretical frequency of the surface wave according to the width of the crystallizer and the order of the trap;

[0056] S2: Zero-mean processing is performed on the liquid level measurement signal to obtain a liquid level fluctuation signal;

[0057] S3: Perform time-frequency conversion on the liquid level fluctuation signal and determine the actual notch center frequency in combination with the theoretical frequency of the surface wave.

[0058] Further,

[0059] Based on the notch depth and notch bandwidth, the specific determination steps are as follows:

[0060] SS1: Determine the phase margin of the current system with notch filter in real time;

[0061] SS2: Calculate the difference between the theoretical phase margin of the system and the phase margin of the current system with the notch filter. When the calculated result is greater than the set value, go to step SS3, otherwise go to step SS4;

[0062] SS3: Determine the notch depth to be 0.4 and the notch bandwidth to be 0.26;

[0063] SS4: The notch depth is determined according to the actual surface wave amplitude, and the notch bandwidth is determined according to the common constraints of the normal feedback of the liquid level and the robustness of the notch filter.

[0064] Further,

[0065] The theoretical frequency of the surface wave is determined according to the following formula:

[0066]

[0067] In the above formula,

[0068] f: theoretical frequency of surface wave, unit: Hz;

[0069] c: the order of the notch filter;

[0070] g: acceleration due to gravity, unit: m / s 2 ;

[0071] L: Crystallizer width, unit: m.

[0072] Further,

[0073] The time-frequency conversion of the liquid level fluctuation signal is completed by spectrum refinement analysis based on the frequency shift method.

[0074] Based on this and the theoretical frequency of the surface wave, the actual notch center frequency is determined. Specifically, a peak search is performed within a frequency range set near the theoretical frequency of the surface wave to obtain the maximum value of the amplitude spectrum of the liquid level fluctuation signal after time-frequency conversion within the frequency range. The actual notch center frequency is determined based on this maximum value.

[0075] Further,

[0076] The notch depth is determined according to the actual surface wave amplitude, specifically:

[0077] Set the target amplitude Af of the notch center frequency after filtering.

[0078] Determine the surface wave amplitude A in real time;

[0079] The notch depth is determined by the ratio of Af to A.

[0080] Further,

[0081] The notch bandwidth ranges from [0.05, 0.4].

[0082] Further,

[0083] The notch filter unit includes a first notch filter and a second notch filter which are serially connected in a feedback loop.

[0084] Further,

[0085] The first notch filter and the second notch filter are respectively a first-order notch filter and a second-order notch filter.

[0086] The present invention discloses a crystallizer liquid level control method and surface wave filtering structure with a surface wave filtering structure. By serially inserting a notch filter in a feedback loop, the narrowband characteristics of the notch filter are utilized to establish a good response to surface wave fluctuations in actual working conditions, so as to establish fully automatic adaptive real-time control of the surface wave, thereby improving the stability and control accuracy of the crystallizer liquid level control.

[0087] In the process of putting the notch filter into use, since the accuracy of the narrowband notch filter center frequency will affect the performance of the notch filter, how to accurately determine the center frequency of the notch filter in real time becomes the primary technical problem to be solved. The present technical scheme solves this problem from two aspects. On the one hand, based on the theoretical frequency of the surface wave and the actual liquid level, a method for determining the actual center frequency in real time is established. On the other hand, considering the fence effect, a spectrum refinement analysis based on the frequency shift method is adopted to complete the time-frequency conversion and obtain a more ideal spectrum.

[0088] At the same time, considering that the notch filter will cause certain phase margin damage, in order to ensure and maintain the stability of the system itself, the present technical scheme further optimizes the above technical scheme and establishes a notch filter parameter determination method based on discriminant constraints. When the difference between the theoretical phase margin of the system and the current phase margin with the notch filter exceeds the set value, a fixed notch depth and notch bandwidth are used to complete the determination of the notch filter parameters; otherwise, the method established in the present technical scheme is used to determine the notch depth and notch bandwidth, thereby completing the determination of the notch filter parameters.

[0089] The crystallizer liquid level control method and surface wave filtering structure with a surface wave filtering structure of the present invention can effectively reduce the harmful effects of surface waves on the crystallizer liquid level control in the actual production process through production and use after the above technical settings. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 It is a schematic diagram of the structure of the surface wave monitoring and filtering structure in the present invention;

[0091] Figure 2 It is a schematic diagram of a crystallizer liquid level control system with a surface wave monitoring and filtering structure in the present invention;

[0092] Figure 3 It is a schematic diagram of the crystallizer liquid level surface wave filtering process in the working principle and process of the present invention;

[0093] Figure 4 It is a schematic diagram of a simulation model of a crystallizer liquid level control with a surface wave monitoring and filtering structure in an embodiment of the present invention;

[0094] Figure 5 This is a schematic diagram showing the comparison of the liquid level signal before and after data preprocessing in an embodiment of the present invention;

[0095] Figure 6 It is a schematic diagram of a liquid level fluctuation signal in an embodiment of the present invention and a time-frequency conversion of the liquid level fluctuation signal by FFT and a time-frequency conversion of the refined FFT;

[0096] Figure 7Schematic diagram for comparing the liquid level measurement signal and the liquid level feedback signal in an embodiment of the present invention. DETAILED DESCRIPTION

[0097] Below, a crystallizer liquid level control method with a surface wave filtering structure and a surface wave filtering structure of the present invention are further specifically described according to the drawings and specific implementation methods of the specification.

[0098] The present invention combines the on-site process and proposes a crystallizer liquid level control method and a surface wave filtering structure with a surface wave filtering structure. In summary, the method can preliminarily determine the theoretical frequency of the surface wave by modeling the crystallizer structure and combining it with the crystallizer width in the actual production process; through advanced signal processing technology, the measured liquid level of the crystallizer liquid level control system is analyzed and combined with the above-calculated theoretical frequency of the surface wave to determine the opening and closing of the surface wave filtering structure and the internal parameter setting of the surface wave filtering structure. In the above manner, fully automatic adaptive control of the surface wave is achieved without human intervention, and the liquid level fluctuation components caused by these harmful interferences in the liquid level measurement data are automatically removed, thereby improving the stability and control accuracy of the crystallizer liquid level control.

[0099] The working principle and process of this technical solution are described below to provide a clearer understanding of this technical solution:

[0100] Since first-order and second-order surface waves are the most common in the actual production process, the notch filter unit in this technical solution is composed of first-order and second-order notch filters. Figure 1 .

[0101] The proposed technical solution is to combine the width change, theoretical frequency and liquid level fluctuation under actual working conditions, as well as signal processing for combined analysis. The actual process is to roughly determine the frequency band range of surface wave fluctuations through the width of the crystallizer, and automatically determine whether the surface wave exists and the actual fluctuation size through the liquid level fluctuation under actual working conditions through signal processing and other means. Through the above analysis, the opening and closing of the first and second notch filters are determined, and the notch filter parameters are set in real time. Through the above control means, the liquid level instability caused by the surface wave is realized, and fully automatic online adaptive control is achieved.

[0102] The crystallizer liquid level control system with the surface wave monitoring and filtering structure designed by this technical solution is as follows Figure 2 As shown, the modules include a PID controller, a liquid level actuator, a crystallizer object, a first notch filter, a second notch filter, and a surface wave identification module. The control method and steps based on this are as follows: Figure 3As shown in the figure, the method comprises: (1) calculating the theoretical frequency f of surface waves of different orders according to the width information of the crystallizer in the continuous casting process; (2) subtracting the average value from the liquid level measurement data to obtain the liquid level fluctuation signal; (3) obtaining the liquid level fluctuation signal by using step 2, and transforming the liquid level fluctuation signal from the time domain to the frequency domain by performing Zoom-FFT; (4) obtaining the liquid level fluctuation frequency domain result by using step 3, finding the surface wave component in the current actual process, and determining the notch center frequency and notch attenuation factor of the surface wave component distributed in the frequency domain; (5) using the notch center frequency determined by step 4 The rate and notch attenuation factor are calculated by parameters to set the parameters of the two notch filters; (6) Analyze the changes in the phase margin of the open-loop transfer function after using the first and second notch filters to infer the rationality of the notch filter parameter settings. If it is unreasonable, use a set of default parameters (7) According to the surface wave components determined in the frequency domain by Zoom-FFT in step 3, determine the enablement and disablement of the two notch filters, calculate the deviation of the filtered liquid level feedback value from the liquid level setting value, and calculate the control adjustment amount through the PID controller for normal closed-loop liquid level control. The specific process is as follows:

[0103] (1) Theoretically, according to the internal structure of the crystallizer, the theoretical frequency of the surface wave can be described by the following formula:

[0104]

[0105] Where c is the number of nodes (c = 1, 2, 3...); g is the acceleration due to gravity, in m / s 2 ; L is the width of the crystallizer (which can be determined based on the current ingot size information collected by PLC), unit: m. According to the different orders c, the theoretical value of the dominant frequency of the surface wave of the current crystallizer is calculated. The theoretical frequencies of the first two order surface waves are mainly considered, including: the first order surface wave frequency f1 (c = 1) and the second order surface wave frequency f2 (c = 2), and their relationship is generally f2>f1. However, under actual working conditions, the actual fluctuation frequency of the surface wave often deviates near the theoretical frequency of the surface wave due to errors in width measurement.

[0106] (2) Data preprocessing

[0107] In order to better analyze the actual working conditions, it is necessary to remove the DC component in the measured liquid level. The crystallizer liquid level measurement value is preprocessed to obtain the liquid level fluctuation periodic signal in the following way:

[0108] Liquid level fluctuation signal y(n) = liquid level measurement signal - average value of liquid level measurement signal

[0109] (3) Zoom-FFT

[0110] In the crystallizer liquid level control, the sampling frequency of data acquisition is fs. In order to ensure the real-time calculation of the analysis system, N data points are generally taken for frequency domain transformation, and the frequency resolution is Δf = fs / N. Due to the existence of the fence effect, the FFT of the liquid level fluctuation signal only has values ​​at certain quantiles, but the accuracy of the center frequency of the narrowband notch filter has a huge impact on the performance of the final notch filter. Based on the above considerations, a fast Fourier transform Zoom-FFT is used to increase the frequency band resolution near the theoretical frequency of the surface wave without increasing the number of transformation points. The process is:

[0111] First, perform Fourier transform on the liquid level fluctuation signal preprocessed in step 2:

[0112]

[0113]

[0114] Generally, Zoom-FFT only refines the frequency band within a certain bandwidth, but this technology takes into account the impact of the first two order surface waves at the same time. The refined frequency band bandwidth must at least include the frequency band between the first and second order surface wave theoretical frequencies, and leave a certain margin ∈. The center frequency of the frequency band for refinement analysis is determined as:

[0115] f c =[(f1-∈)+(f2+∈))] / 2,

[0116] Complex modulation of the original liquid level fluctuation signal

[0117]

[0118] According to the frequency shift characteristics of discrete Fourier transform, there is the following relationship

[0119] Y c (k) = Y(k + f c / Δf)

[0120] In order to better preserve the signal components within the refined frequency band, the frequency-shifted signal is low-pass filtered.

[0121]

[0122] N(k) is a low-pass filter. The original fluctuation signal is resampled using the sampling frequency f′s=fs / D, and the expression is:

[0123]

[0124] Where S = k / D, l = n / D. Perform discrete Fourier transform on the resampled g(l)

[0125]

[0126]

[0127] Finally, restore the spectrum to the actual frequency to get the final Zoom-FFT output:

[0128] Y(k)=Y(kf c / Δf)

[0129]

[0130] (4) Determine the actual notch center frequency and notch attenuation factor

[0131] According to the frequency domain result Y(f) obtained by Zoom-FFT of the liquid level fluctuation signal in step 3, The Y(f) peak search is performed in a certain frequency range fr1 and fr2 near the theoretical frequencies f1 and f2 of the first-order and second-order surface waves, and the Zoom-FFT maximum values ​​max(A(Y(fr1))) and max(A(Y(fr2))) of the liquid level fluctuation signals of fr1 and fr2 in this frequency range are obtained, and the notch center frequencies N_f1 and N_f2 of the first notch filter and the second notch filter are determined accordingly.

[0132] N_f2, the conditions met are:

[0133] fr1={f1-ε≤fr1≤f1+ε}

[0134] fr2={f2-ε≤fr2≤f3+ε}

[0135] A(Y(N_f1))=max(A(Y(fr1)))

[0136] A(Y(N_f2))=max(A(Y(fr2)))

[0137] The parameter notch attenuation factor σ should change dynamically with the actual fluctuation of the surface wave under actual working conditions. When the surface wave is more violent, the notch attenuation factor σ should be smaller, so that the surface wave can be effectively suppressed quickly; when the surface wave is relatively stable, the notch attenuation factor σ should be larger to reduce the impact of phase lag on the system. The frequency domain characteristics of the notch filter are analyzed. The notch attenuation factor of the center frequency is the largest, and the signal component of this frequency is attenuated to σ times the original. The notch attenuation factor σ is set to:

[0138]

[0139]

[0140] Among them Af It is expressed as the target amplitude of the signal at the center frequency of the notch after filtering; A1 is the first-order surface wave amplitude of the liquid level fluctuation signal identified online by Zoom-FFT; A2 is the second-order surface wave amplitude of the liquid level fluctuation signal identified online by Zoom-FFT.

[0141] (5) Parameters of the first and second notch filters

[0142] Using the notch center frequencies N_f1 and N_f2 of the first notch filter and the second notch filter determined in step 4, and the notch attenuation factors σ1 and σ2, set the parameters of the first and second notch filters, including: setting the notch center frequency N_f, the notch attenuation factor σ, and the notch frequency range wb. The setting of the notch frequency range wb needs to be fixed according to the empirical method: the notch frequency range wb is generally required to be in the range of 0.05≤wb≤0.4. If wb is too large, it will affect the normal feedback control of other liquid level fluctuation frequency bands. If wb is too small, the robustness of the notch filter is poor.

[0143] The parameters of the first notch filter are as follows: 1) notch center frequency N_f1; 2) notch attenuation factor σ1, that is, the attenuation multiple of the center frequency σ1; 3) notch frequency range wb1. The parameters of the second notch filter are as follows: 1) notch center frequency N_f2; 2) notch attenuation factor σ2; 3) notch frequency range wb2.

[0144] The general structure of the first and second notch filters is as follows (i.e. the basic transfer function of the notch filter):

[0145]

[0146] Where s is the Laplace transform operator; ω is the notch center frequency; m, n are notch coefficients, which are adjustable parameters according to the notch center frequency, notch attenuation factor, and notch frequency range.

[0147] First notch filter structure and parameters:

[0148]

[0149] ω1=2π*N_f1,

[0150]

[0151] m1=σ1*n1,

[0152] Second notch filter structure and parameters:

[0153]

[0154] ω2=2π*N_f2,

[0155]

[0156] m2=σ2*n2,

[0157] (6) Reasonableness of the first and second notch filter parameter settings

[0158] The open-loop transfer function of the original system is G(s), where s is the Laplace differential operator. The first and second notch filters are obtained in step 5 and then connected in series to the system. At this time, the open-loop transfer function of the system is Gn(s)=G(s)*N_F1(s)*N_F2(s).

[0159] Among them, the phase lag caused by the first-order notch filter to the system is

[0160]

[0161] The phase lag caused by using a second-order notch filter on the system is

[0162]

[0163] If the phase margin difference exceeds 10 0 It is believed that the notch filter has a greater impact on the stability of the liquid level control system, so the notch attenuation factor and notch width of the first-order and second-order notch filters use a set of default parameters, and the process is as follows:

[0164]

[0165] Where ω0 is the amplitude crossing frequency of the original liquid level control system, ω f is the amplitude crossing frequency after using the notch filter, but because the notch filter has little effect on the amplitude in other frequency bands, ω0≈ω f Then β=-∠N_F1(ω0)-∠N_F2(ω0). β>10°, then let the notch attenuation factor σ1=0.4, σ2=0.4, the notch width wb1=0.26, wb2=0.26, but do not change the notch center frequency, and recalculate the first and second notch filter parameters through step 5.

[0166] (7) First and second notch filter enable and disable conditions

[0167] Step 3 is used to determine the surface wave component in the liquid level fluctuation signal in the frequency domain, that is, the surface wave amplitude A(Y(N_f2)) of the first and second order surface waves determined by Zoom-FFT, and the activation and deactivation of the notch filter are determined by comparing with the set threshold δ.

[0168] To enable or disable the first notch filter:

[0169]

[0170] To enable or disable the second notch filter:

[0171]

[0172] When the amplitude of the surface wave exceeds a certain limit, the notch filter is turned on; when the amplitude of the surface wave is suppressed, the notch filter is turned off, thereby reducing the impact of the phase lag of the notch filter on the normal liquid level feedback. However, the start threshold and the shutdown threshold are not equal, that is, δ1≠δ2, which is a dead zone limit set to avoid frequent startup and shutdown of the notch filter.

[0173] Example

[0174] Build a simulation model, such as Figure 4 As shown:

[0175] a) The liquid level setting value is 64mm.

[0176] b) The expression of PI controller is:

[0177]

[0178] c) The expression of liquid level actuator is:

[0179]

[0180] d) Flow coefficient 0.35.

[0181] e) Time lag 0.24s.

[0182] f) Crystallizer process object:

[0183]

[0184] g) First-order surface wave model

[0185]

[0186] h) Second-order surface wave model

[0187]

[0188] i) The width of the crystallizer is L = 1.5 m.

[0189] Surface wave filtering process

[0190] 1) Calculation of theoretical frequency of surface waves

[0191] According to the theoretical formula of surface wave frequency:

[0192]

[0193] Theoretical frequency calculation of first-order surface wave (c = 1; L = 1.5m)

[0194] f1=0.721Hz

[0195] Theoretical frequency calculation of second-order surface wave (c = 2; L = 1.5m)

[0196] f2=1.02Hz

[0197] 2) Data preprocessing

[0198] Liquid level fluctuation signal = liquid level measurement signal - average value of liquid level measurement signal, get liquid level fluctuation signal, the result is as follows Figure 5 shown.

[0199] 3) Zoom-FFT

[0200] In order to improve the real-time performance of the system for surface wave detection, a sliding window is used to obtain data for the above liquid level fluctuation signal. The window size is 204.8s, the sliding step is 10s, and Zoom-FFT is performed on the data in the window. The results are as follows: Figure 6 As shown in Figure 2, Zoom-FFT significantly improves the frequency resolution near the surface wave frequency compared to ordinary FFT.

[0201] 4) Determine the notch center frequency and notch attenuation factor

[0202] Through a certain frequency range fr1, fr2 of the theoretical frequencies f1, f2 of the surface wave, the Zoom-FFT result Y(f) of the liquid level fluctuation signal is searched for peaks to determine the surface wave frequency offset ∈ = 0.05. Within this frequency range, the maximum value of the Zoom-FFT amplitude spectrum of the liquid level fluctuation signal corresponds to the frequency N_f1, N_f2.

[0203] A(Y(N_f1))=max(A(Y(fr1)))

[0204] A(Y(N_f2))=max(A(Y(fr2)))

[0205] It can be determined that the notch center frequency N_f1 of the first notch filter is 0.721 Hz; and the notch center frequency N_f2 of the second notch filter is 1.019 Hz.

[0206] The signal amplitude of the notch center frequency after filtering should be less than 0.1. f =0.1,

[0207] Set the target amplitude of the notch center frequency after filtering A f=0.1, the identification results of Zoom-FFT are A1=0.2007, A2=0.2736, so σ1=0.4982, σ2=0.3655.

[0208] 5) Determine the notch filter structure and parameters

[0209] For the first notch filter, according to the notch center frequency N_f1 = 0.721Hz, the notch attenuation factor σ1 = 0.4982, and the notch frequency range wb1 = 0.1Hz, the structural parameters m1 = 0.0560, ω1 = 3.9235, and n1 = 0.1125 can be calculated. Therefore, the structural parameters of the first notch filter are as follows:

[0210]

[0211] The second notch filter, according to the notch center frequency N_f2 = 1.019Hz, the notch attenuation factor σ2 = 0.3655, the notch frequency range wb2 = 0.1Hz; the structural parameters m2 = 0.0241, ω2 = 5.5487, n2 = 0.0660 can be calculated. Therefore, the structural parameters of the second notch filter are as follows:

[0212]

[0213] 6) Reasonableness of the first and second notch filter parameter settings

[0214] The open-loop transfer function of the original system

[0215]

[0216] The amplitude crossing frequency ω0 = 0.868 rad / s, the phase margin is γ0 = 53.7°, and after using the first and second notch filters, the amplitude crossing frequency ω f =0.868rad / s, phase margin γ0 =52°; therefore β =53.7°-52°=1.7°<10°. The parameters of the first and second notch filters are set within a reasonable range.

[0217] 7) Determine whether the notch filter is enabled or disabled

[0218] The start threshold δ1 = 0.5 mm, the closing threshold δ2 = 0.2 mm. Near the theoretical frequency of the surface wave, the liquid level fluctuation signal Zoom-FFT amplitude spectrum A(Y(N_f1)) and the thresholds δ1 and δ2 are used to determine whether the notch filter is turned on or off.

[0219] To enable or disable the first notch filter:

[0220]

[0221] To enable or disable the second notch filter:

[0222]

[0223] 7) Simulation results

[0224] according to Figure 7 From the simulation results of the liquid level feedback signal and liquid level measurement signal output by the simulation model, it can be seen that at 2 minutes, the first-order and second-order surface wave external signals are added. At this time, the liquid level control effect and the surface wave interact with each other, resulting in an increase in liquid level fluctuations; at 10 minutes, the first and second notch filters are enabled, which can effectively filter out the surface wave components in the liquid level feedback signal and reduce the crystallizer liquid level fluctuations caused by the interaction between the control effect and the surface wave. This technology proposes a crystallizer liquid level control and method with a surface wave filtering structure, which can timely and effectively remove the surface wave component in the original measured liquid level signal, prevent the interaction between the liquid level control and the surface wave, and successfully realize the fully automatic online adaptive control of the surface wave.

Claims

1. A method for controlling a crystallizer liquid level with a surface wave filtering structure, characterized in that: In the feedback loop of the crystallizer liquid level feedback control, a surface wave monitoring and filtering structure based on the liquid level measurement signal is set up. The periodic real-time monitoring of the liquid level fluctuation signal is established through the surface wave monitoring and filtering structure; When the monitoring result is greater than or equal to the set threshold, the surface wave monitoring and filtering structure based on the liquid level measurement signal is turned on to form a liquid level measurement signal feedback based on the surface wave filtering; otherwise, a direct feedback based on the liquid level measurement signal is formed.

2. A method for controlling a crystallizer liquid level with a surface wave filtering structure according to claim 1, characterized in that: The surface wave monitoring and filtering structure is composed of a surface wave identification unit and a notch filter unit. The surface wave identification unit is used to receive the liquid level measurement signal and the mold width detection signal in real time, and determine the actual notch center frequency, notch depth, notch bandwidth and the fluctuation signal amplitude corresponding to the actual notch center frequency according to the two signals; The notch filter unit is used to receive the amplitude of the fluctuation signal corresponding to the actual notch center frequency, and determine whether to start the notch filter according to the comparison result of the fluctuation signal amplitude and the set threshold.

3. A method for controlling a crystallizer liquid level with a surface wave filtering structure according to claim 2, characterized in that: The method of determining the actual notch center frequency according to the liquid level measurement signal and the mold width detection signal specifically includes the following steps: S1: Determine the theoretical frequency of the surface wave according to the width of the crystallizer and the order of the trap; S2: Zero-mean processing is performed on the liquid level measurement signal to obtain a liquid level fluctuation signal; S3: Perform time-frequency conversion on the liquid level fluctuation signal and determine the actual notch center frequency in combination with the theoretical frequency of the surface wave.

4. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 3, characterized in that: Based on the notch depth and notch bandwidth, the specific determination steps are as follows: SS1: Determine the phase margin of the current system with notch filter in real time; SS2: Calculate the difference between the theoretical phase margin of the system and the phase margin of the current system with the notch filter. When the calculated result is greater than the set value, go to step SS3, otherwise go to step SS4; SS3: Determine the notch depth to be 0.4 and the notch bandwidth to be 0.26; SS4: The notch depth is determined according to the actual surface wave amplitude, and the notch bandwidth is determined according to the common constraints of the normal feedback of the liquid level and the robustness of the notch filter.

5. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 3, characterized in that: The theoretical frequency of the surface wave is determined according to the following formula: In the above formula, f: theoretical frequency of surface wave, unit: Hz; c: the order of the notch filter; g: acceleration due to gravity, unit: m / s 2 ; L: Crystallizer width, unit: m.

6. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 3, characterized in that: The time-frequency conversion of the liquid level fluctuation signal is completed by spectrum refinement analysis based on the frequency shift method. Based on this and the theoretical frequency of the surface wave, the actual notch center frequency is determined. Specifically, a peak search is performed within a frequency range set near the theoretical frequency of the surface wave to obtain the maximum value of the amplitude spectrum of the liquid level fluctuation signal after time-frequency conversion within the frequency range. The actual notch center frequency is determined based on this maximum value.

7. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 4, characterized in that: The notch depth is determined according to the actual surface wave amplitude, specifically: Set the target amplitude Af of the notch center frequency after filtering. Determine the surface wave amplitude A in real time; The notch depth is determined by the ratio of Af to A.

8. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 4, characterized in that: The notch bandwidth ranges from [0.05, 0.4].

9. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 2, characterized in that: The notch filter unit includes a first notch filter and a second notch filter which are serially connected in a feedback loop.

10. The method for controlling the liquid level of a crystallizer with a surface wave filtering structure according to claim 9, characterized in that: The first notch filter and the second notch filter are respectively a first-order notch filter and a second-order notch filter.

11. A crystallizer liquid surface wave filtering structure, characterized in that: The surface wave filtering structure is arranged in the feedback loop of the crystallizer liquid level feedback control system. According to the surface wave filtering structure, periodic real-time monitoring of surface waves based on the liquid level measurement signal is established; according to the monitoring result, the surface wave filtering structure controls to form two feedback forms: feedback of the liquid level measurement signal after filtering the surface wave when the monitoring value exceeds the set threshold, or direct feedback based on the liquid level measurement signal when the monitoring value does not exceed the set threshold.

12. A crystallizer liquid surface wave filtering structure according to claim 11, characterized in that: The surface wave filtering structure is composed of a surface wave identification unit and a notch filter unit. The surface wave identification unit is used to receive the liquid level measurement signal and the mold width detection signal in real time, and determine the actual notch center frequency, notch depth, notch bandwidth and the fluctuation signal amplitude corresponding to the actual notch center frequency according to the two signals; The notch filter unit is used to receive the amplitude of the fluctuation signal corresponding to the actual notch center frequency, and determine whether to start the notch filter according to the comparison result of the fluctuation signal amplitude and the set threshold.

13. A crystallizer liquid surface wave filtering structure according to claim 12, characterized in that: The method of determining the actual notch center frequency according to the liquid level measurement signal and the mold width detection signal specifically includes the following steps: S1: Determine the theoretical frequency of the surface wave according to the width of the crystallizer and the order of the trap; S2: Zero-mean processing is performed on the liquid level measurement signal to obtain a liquid level fluctuation signal; S3: Perform time-frequency conversion on the liquid level fluctuation signal and determine the actual notch center frequency in combination with the theoretical frequency of the surface wave.

14. A crystallizer liquid surface wave filtering structure according to claim 13, characterized in that: Based on the notch depth and notch bandwidth, the specific determination steps are as follows: SS1: Determine the phase margin of the current system with notch filter in real time; SS2: Calculate the difference between the theoretical phase margin of the system and the phase margin of the current system with the notch filter. When the calculated result is greater than the set value, go to step SS3, otherwise go to step SS4; SS3: Determine the notch depth to be 0.4 and the notch bandwidth to be 0.26; SS4: The notch depth is determined according to the actual surface wave amplitude, and the notch bandwidth is determined according to the common constraints of the normal feedback of the liquid level and the robustness of the notch filter.

15. The crystallizer liquid surface wave filtering structure according to claim 13, characterized in that: The theoretical frequency of the surface wave is determined according to the following formula: In the above formula, f: theoretical frequency of surface wave, unit: Hz; c: the order of the notch filter; g: acceleration due to gravity, unit: m / s 2 ; L: Crystallizer width, unit: m.

16. The crystallizer liquid surface wave filtering structure according to claim 13, characterized in that: The time-frequency conversion of the liquid level fluctuation signal is completed by spectrum refinement analysis based on the frequency shift method. Based on this and the theoretical frequency of the surface wave, the actual notch center frequency is determined. Specifically, a peak search is performed within a frequency range set near the theoretical frequency of the surface wave to obtain the maximum value of the amplitude spectrum of the liquid level fluctuation signal after time-frequency conversion within the frequency range. The actual notch center frequency is determined based on this maximum value.

17. A crystallizer liquid surface wave filtering structure according to claim 14, characterized in that: The notch depth is determined according to the actual surface wave amplitude, specifically: Set the target amplitude Af of the notch center frequency after filtering. Determine the surface wave amplitude A in real time; The notch depth is determined by the ratio of Af to A.

18. The crystallizer liquid surface wave filtering structure according to claim 14, characterized in that: The notch bandwidth ranges from [0.05, 0.4].

19. The crystallizer liquid surface wave filtering structure according to claim 12, characterized in that: The notch filter unit includes a first notch filter and a second notch filter which are serially connected in a feedback loop.

20. The crystallizer liquid surface wave filtering structure according to claim 19, characterized in that: The first notch filter and the second notch filter are respectively a first-order notch filter and a second-order notch filter.

Images