A method for eliminating slurry noise in electromagnetic flowmeter based on adaptive filtering

Abstract

The present invention discloses a method for eliminating slurry noise of an electromagnetic flowmeter based on adaptive filtering. The method first collects the original voltage signal in the slurry flow measurement process based on a single-frequency square wave excitation method, then constructs a sine reference signal and a cosine reference signal according to the excitation signal frequency and the sampling frequency, and then reconstructs the signal components related to the flow in the original measurement voltage signal point by point based on the sine and cosine reference signals with a certain step length and weight coefficient, and iteratively calculates the weight coefficients of the two reference signals through the residual between the partial signal components and the original measurement voltage signal, and uses this as the input of the next iteration until the last sampling point is calculated. The difference between the original measurement voltage signal and the residual sequence is the voltage signal after noise reduction, and the noise-reduced flow value can be calculated by combining the current signal. The present invention can realize the extraction of the original measurement voltage signal under the interference of slurry noise, and improve the slurry measurement stability of the electromagnetic flowmeter.

Description

Technical Field

[0001] The invention discloses a method for eliminating slurry noise of an electromagnetic flowmeter based on adaptive filtering. Background Art

[0002] The demand for measuring slurry (solid-liquid two-phase flow) flow is widely present in many industries such as mining, energy, chemical industry, papermaking and metallurgy. The hydraulic transportation, processing and sales of industrial raw materials such as coal, ore and silt are inseparable from flow measurement; pulp is the basic raw material for papermaking. Through the accurate measurement of pulp flow, the automatic control of the papermaking production process can be realized, which improves production efficiency and reduces production costs. In addition, by measuring the pulp flow, the pulp concentration can be calculated and controlled to ensure the stability of paper quality; water-coal slurry is a new type of clean material widely used in thermal power generation and other industrial production. The accurate measurement of water-coal slurry flow is the prerequisite for ensuring the safe and efficient operation of related equipment. Otherwise, the equipment may stop frequently, causing economic losses and even causing production accidents. It can be seen that high-precision and accurate measurement of slurry flow is crucial.

[0003] Instruments commonly used for slurry flow measurement include ultrasonic flowmeters, vortex flowmeters, differential pressure flowmeters and electromagnetic flowmeters. Among them, the electromagnetic flowmeter has a simple structure, is resistant to corrosion and friction, has no flow-blocking components in the measuring tube, is not easily affected by factors such as fluid density, temperature, pressure, conductivity and viscosity during the measurement process, has a large range ratio and high measurement accuracy, is reliable to use and easy to maintain, and is therefore more suitable for slurry flow measurement than other flowmeters. However, during the measurement process, the solid particles in the slurry will randomly scratch the measuring electrodes of the electromagnetic flowmeter, destroying the ionization balance on the electrode surface. The non-stationary noise generated by the continuous cycle of ionization balance recovery and being broken by solid particles is superimposed on the original voltage signal of the flow measurement, causing signal distortion, which ultimately leads to excessive fluctuations in the measurement results of the electromagnetic flowmeter and unstable flow output. Therefore, it is necessary to take measures to eliminate the slurry noise in the electromagnetic flowmeter process.

[0004] In order to reduce the interference of slurry noise, the existing means mainly include improving sensors and signal processing. The interference of slurry noise can be effectively reduced by processing special measuring electrodes to reduce their contact area with the slurry and using polishing technology to reduce the surface roughness of the electrode, but this method will significantly increase the internal resistance of the sensor measurement, resulting in the loss of the original voltage signal, which is not conducive to the measurement of low conductivity fluids. The power spectrum of slurry noise is inversely proportional to the frequency, so some studies have proposed to use high-frequency excitation to reduce the interference of slurry noise. The document "Development of Slurry Electromagnetic Flowmeter Based on DSP" uses statistical analysis and signal reconstruction methods in conjunction with high-frequency excitation to suppress slurry noise and achieve a steady-state fluctuation rate of less than 4% in slurry measurement, but this method will lead to a decrease in the zero-point stability of the measurement signal. The invention patent "Electromagnetic Flowmeter" with patent number CN87101677A proposes a dual-frequency rectangular wave excitation method, which takes into account the slurry noise suppression and zero-point stability problems by adjusting the weight between the low-frequency signal and the high-frequency signal. However, this method has a high hardware implementation difficulty and high cost. The invention patent with patent number CN109489747A, "A Signal Processing Method for Electromagnetic Flowmeter Based on Harmonic Analysis", proposes a noise reduction method based on harmonic signal screening, which selects odd harmonics with less interference from slurry noise through the amplitude ratio of adjacent odd harmonics to reflect the flow rate. This method can reduce the slurry noise interference to a certain extent, but it requires the selection of appropriate parameters according to the specific scenario, and it is difficult to universally apply to complex and diverse field conditions. Summary of the invention

[0005] The present invention aims to overcome the problem of low detection signal-to-noise ratio of existing electromagnetic flowmeters during slurry measurement, realize the extraction of original measurement voltage signals under slurry noise interference, and improve the measurement stability of the electromagnetic flowmeter.

[0006] The present invention is implemented by the following technical solutions:

[0007] A method for eliminating slurry noise based on adaptive filtering. The method first collects the original voltage signal in the slurry flow measurement process based on a single-frequency square wave excitation method, then constructs sine and cosine reference signals according to the excitation signal frequency and the sampling frequency, and then reconstructs the flow-related signal components in the original voltage signal point by point with a certain step size and weight coefficient based on the sine and cosine reference signals. The weight coefficients of the sine and cosine reference signals are iteratively calculated through the residual between the signal component and the original voltage signal, and are used as the input for the next iteration until the last sampling point is calculated. The difference between the original voltage signal and the residual sequence is the denoised voltage signal, and the denoised flow value can be calculated in combination with the current signal.

[0008] Furthermore, a method for eliminating slurry noise based on adaptive filtering of the present invention specifically comprises the following steps:

[0009] Step 1: Flow signal acquisition, using single-frequency square wave excitation, the excitation frequency is f; during the sampling process, the time series t=[t 1 , t 2 , ..., t i , ...], the sampling frequency is f s , the original sampled signal is Data processing and sampling are carried out synchronously to ensure the continuity of the collected signals;

[0010] Step 2: Construct a reference signal, a sinusoidal reference signal in Cosine reference signal in In the above expression, N is the approximate order, which is generally an integer greater than 5;

[0011] Step 3: Coefficient setting: set the iteration step size μ and the magnification M during the adaptive filtering algorithm operation. The magnification should be greater than the number of sampling signal points in one cycle, that is, Set the sinusoidal reference signal R sin The weight coefficient w cos =0; Set the cosine reference signal R cos The weight coefficient w sin =0;

[0012] Step 4: Calculate the residuals.

[0013] Step 5: Coefficient update, the sine weight coefficient update formula is The cosine weight coefficient update formula is: Step 6: Residual iteration, repeating steps 4 and 5 until all sampling points in the original sampling signal are calculated, and the calculated residuals constitute the residual sequence E = [e 1 , e 2 , ..., e i , ...];

[0014] Step 7: Signal denoising. The residual sequence is the slurry noise based on adaptive filtering fitting. Subtracting the residual sequence from the original sampling signal can obtain the denoised flow detection signal S Denoised =S raw -E; At this point, the calculation ends.

[0015] The beneficial effects of the present invention are:

[0016] The algorithm of the present invention can extract the flow signal waveform from the original measurement signal under the interference of slurry noise, thereby achieving measurement signal noise reduction and improving the measurement stability of the electromagnetic flowmeter. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of the signal processing method.

[0018] Figure 2 This is the excitation current signal diagram.

[0019] Figure 3 This is the measured voltage signal diagram before adaptive filtering.

[0020] Figure 4 This is the measured voltage signal after adaptive filtering.

[0021] Figure 5 Schematic diagram of sinusoidal reference signal.

[0022] Figure 6 Schematic diagram of cosine reference signal.

[0023] Figure 7 Schematic diagram of flow measurement results before adaptive filtering processing.

[0024] Figure 8 Schematic diagram of flow measurement results after adaptive filtering processing. DETAILED DESCRIPTION

[0025] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0026] In the embodiment, a DN50 electromagnetic flowmeter is used to measure the flow of slurry containing sediment. The electromagnetic flowmeter uses square wave excitation with a frequency of 12.5 Hz and a single-sided amplitude of 100 mA. The waveform is as follows: Figure 2 As shown in the figure, the original voltage signal related to the flow is sampled after a certain method multiple, and the sampling frequency is 1000Hz. The waveform of the original voltage signal is as follows Figure 3 As shown in the figure, due to the interference of slurry noise, there are large fluctuations in the square wave flat section of the original voltage signal. Therefore, directly selecting sampling points from these flat sections of the original voltage signal and the excitation current signal to calculate the peak-to-peak value and convert it into flow value, the result has a large degree of instability. The original flow measurement result is shown in Figure 7 shown.

[0027] The present invention is used to process the original voltage signal, and the algorithm flow is as follows: Figure 1 As shown, the steps are as follows:

[0028] Step 1: Flow signal acquisition, using single-frequency square wave excitation, the excitation frequency is f; during the sampling process, the time series t=[t 1 , t 2 , ..., t i , ...], the sampling frequency is f s, the original sampled signal is Data processing and sampling are performed synchronously to ensure the continuity of the collected signal. In this embodiment, the excitation frequency f = 12.5 Hz, the sampling frequency f s = 1000 Hz, so the time series t = [0, 0.001, 0.002, 0.003, ...], the original sampling signal S raw Some waveforms are as follows Figure 3 As shown, the number of signal sampling cycles is 1000, so the total number of sampling points is 80000;

[0029] Step 2: Construct a reference signal, a sinusoidal reference signal in Cosine reference signal in In the above expression, N is an approximate order, which is generally an integer greater than 5. In this embodiment, the waveforms of the sine reference signal and the cosine reference signal are respectively as follows: Figure 5 , 6 As shown, the approximate order N = 10. Step 3: Coefficient setting, set the iteration step size μ and the magnification M during the adaptive filtering algorithm operation process. The magnification should be greater than the number of sampling signal points in one cycle, that is, Set the sinusoidal reference signal R sin The weight coefficient w sin =0; Set the cosine reference signal R cos The weight coefficient w cos =0, in this embodiment, the iteration step size μ is set to 5×10 -6 ; Magnification M = 800;

[0030] Step 4: Calculate the residuals.

[0031] Step 5: Coefficient update, the sine weight coefficient update formula is The cosine weight coefficient update formula is:

[0032] Step 6: Residual iteration, repeating steps 4 and 5 until all sampling points in the original sampling signal are calculated, and the calculated residuals constitute the residual sequence E = [e 1 , e 2 , ..., e i , ...]. In this embodiment, when i=80000, the iteration stops;

[0033] Step 7: Signal denoising. The residual sequence is the slurry noise based on adaptive filtering fitting. Subtracting the residual sequence from the original sampling signal can obtain the denoised voltage signal S Denoised =S raw -E; At this point, the calculation ends.

[0034] In this embodiment, the waveform of the voltage signal after noise reduction is as follows: Figure 4 As shown in Figure 2, it can be observed from the processed signal that the signal stability in the flat section is significantly improved. The flow signal is calculated from the voltage signal and current signal after noise reduction as shown in Figure 2. Figure 8 As shown in the figure. At the beginning of the calculation, the reference signal weight coefficient has not converged, resulting in a large deviation between the flow and the actual flow. After the reference signal weight coefficient converges, the output flow gradually stabilizes. Figure 7 It can be found that after being processed by this method, the flow noise has been significantly reduced and the flow measurement stability has been significantly improved.

[0035] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A slurry noise elimination method based on adaptive filtering, It is characterized in that The method first collects the original voltage signal in the slurry flow measurement process based on a single-frequency square wave excitation method, then constructs sine and cosine reference signals according to the excitation signal frequency and the sampling frequency, and then reconstructs the flow-related signal components in the original voltage signal point by point with a certain step size and weight coefficient based on the sine and cosine reference signals. The weight coefficients of the sine and cosine reference signals are iteratively calculated through the residual between the signal component and the original voltage signal, and used as the input for the next iteration until the last sampling point is calculated. The difference between the original voltage signal and the residual sequence is the denoised voltage signal, and the denoised flow value can be calculated in combination with the current signal.

2. According to the method for eliminating slurry noise based on adaptive filtering according to claim 1, It is characterized in that The specific steps include: Step 1: Flow signal acquisition, using single-frequency square wave excitation, the excitation frequency is f; during the sampling process, the time series t=[t 1 , t 2 , ..., t i , ...], the sampling frequency is f s , the original sampled signal is Data processing and sampling are carried out synchronously to ensure the continuity of the collected signals; Step 2: Construct a reference signal, a sinusoidal reference signal in Cosine reference signal in In the above expression, N is the approximate order, which is generally an integer greater than 5; Step 3: Coefficient setting: set the iteration step size μ and the magnification M during the adaptive filtering algorithm operation. The magnification should be greater than the number of sampling signal points in one cycle, that is, Set the sinusoidal reference signal R sin The weight coefficient w cos =0; Set the cosine reference signal R cos The weight coefficient w sin =0; Step 4: Calculate the residuals. Step 5: Coefficient update, the sine weight coefficient update formula is The cosine weight coefficient update formula is: Step 6: Residual iteration, repeat steps 4 and 5 until all sampling points of the original sampling signal are calculated, and the calculated residuals constitute the residual sequence E = [e 1 , e 2 , ..., e i , ...]; Step 7: Signal denoising. The residual sequence is the slurry noise based on adaptive filtering fitting. Subtracting the residual sequence from the original sampling signal can obtain the denoised flow detection signal S Denoised =S raw -E; At this point, the calculation ends.

Images