Ultrasonic signal filtering method and system for internal defect of roller

Abstract

The invention belongs to the technical field of signal filtering, and discloses a roller internal defect ultrasonic signal filtering method, which adopts improved median filtering as a primary filter, can effectively eliminate high-frequency noise in ultrasonic signals, and greatly improves the parameter optimization convergence speed of a subsequent secondary filter. According to the method, the multi-stage filter is constructed as the second-stage filter, the intelligent optimization algorithm is adopted to replace a traditional least square algorithm to obtain the parameters of the filter, operation of a related parameter matrix and an inverse matrix of the related parameter matrix is not needed, and the filter parameter solving process is greatly simplified. The two-stage filter designed by the invention is simple in structure and easy to realize, does not need complex matrix operation, can effectively filter background clutters and interference signals in the roller ultrasonic signals, and provides a basis for subsequent roller internal defect automatic identification research.

Description

Technical Field

[0001] The invention belongs to the technical field of signal filtering, and in particular relates to a method and system for filtering ultrasonic signals of internal defects of a rolling mill. Background Art

[0002] Internal defects of rolling mill rolls determine the service life of rolling mill rolls and are important factors that cause rolling mill roll failure and affect safe production. Ultrasonic testing is a widely used detection method in the field of non-destructive testing. The changes in ultrasonic waves during propagation are used as the basis for judging the material properties and structural changes of the object being tested. Ultrasonic testing has a relatively wide detection range and good directionality and sensitivity. It is a commonly used method for detecting surface and internal defects of metal devices. Therefore, ultrasonic testing is a commonly used detection method for internal defects of rolling mill rolls. However, during the ultrasonic testing process, in addition to the surface echo, bottom echo and defect echo, there are often some noise interference signals in the ultrasonic echo, resulting in inaccurate detection results. Especially when the defect is close to the surface of the rolling mill roll, the defect echo and the surface echo are coupled. Since the amplitude of the defect echo is often smaller than that of the surface echo, it is not conducive to the accurate extraction of the defect echo with a small amplitude.

[0003] Generally, it is necessary to filter the ultrasonic signal to eliminate or reduce the noise signal. Common signal filtering methods include low-pass filtering, high-pass filtering, band-pass filtering, mean filtering, median filtering, Wiener filtering, etc. Frequency filtering can only remove noise signals outside the cutoff frequency or signal bandwidth. For different noise signals, the cutoff frequency, center frequency and bandwidth also vary greatly, and this method has poor adaptability. Mean filtering smoothes the data sequence by calculating the local average value of the data points. The degree of smoothing depends directly on the size of the local data interval. The larger the data interval, the higher the smoothness, but the sensitivity is low at this time. Median filtering is a nonlinear smoothing filtering technology that uses the middle value in the adjacent area of ​​the current sampling moment as the current sampling value. It can effectively avoid the influence of accidental factors on the results and is suitable for data that changes relatively slowly. The selection of the adjacent area of ​​the sampling point will also affect the actual filtering effect. The Wiener filter uses the least squares method to make the square difference between the filter output value and the expected output value the optimal criterion for the linear filter. The parameters of the filter are fixed and it is suitable for optimal filtering under stable random conditions.

[0004] 201910249006.5 A data filtering method for ultrasonic partial discharge detection: In view of the possible pulse interference in ultrasonic partial discharge detection, a median filter is used. When the pulse width is less than half of the window width, the median filter is very effective for narrow pulse interference. Then, for the processing of random noise, after comparing the wavelet packet threshold function denoising and the singular value decomposition algorithm, a singular value decomposition filtering algorithm based on the singular value energy difference spectrum is proposed. This method requires the construction of a Hankel matrix, the singular value decomposition of the Hankel matrix constructed by the ultrasonic signal sequence, and the reconstruction of the matrix to achieve the best approximation of the Hankel matrix. The filtering process is cumbersome, the amount of calculation is large, and there are multiple process parameters. The actual effect needs to be tested.

[0005] Wang Jun et al. Near-surface defect detection of ultrasonic TOFD based on adaptive filtering [J] Mechanical and Electrical Engineering, 2018, 35(6): In order to solve the problem of aliasing of direct wave and defect signal in the near-surface blind area of ​​ultrasonic diffraction time-of-flight method image, a near-surface defect signal processing method based on adaptive filtering is proposed. This method uses the recursive least squares method to construct an adaptive FIR filter and the gradient search method to solve the minimum value of the error measurement function. The calculation is cumbersome, and there are many traps in the convergence curve that are minimum values ​​but not global minimum values. It cannot guarantee that the adaptive filter weight coefficients will converge to the optimal value quickly, and the filtering effect is poor.

[0006] Through the above analysis, the problems and defects of the prior art are as follows:

[0007] In the ultrasonic testing process, in addition to the surface echo, bottom echo and defect echo, there are often some noise interference signals in the ultrasonic echo, which leads to inaccurate test results. Especially when the defect is close to the roller surface, the defect echo and the surface echo are coupled. Since the amplitude of the defect echo is often smaller than that of the surface echo, it is not conducive to the accurate extraction of the defect echo with a small amplitude. Summary of the invention

[0008] In view of the problems existing in the prior art, the present invention provides a method for filtering ultrasonic signals of internal defects of a rolling mill roll.

[0009] The present invention is implemented as follows: a method for filtering ultrasonic signals of internal defects of a roller comprises:

[0010] Step 1, using an ultrasonic digital flaw detector to obtain the internal defect ultrasonic signal P(n) of the roller and the background clutter reference signal R(n);

[0011] Step 2, performing median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) to remove signal burrs, thereby obtaining a first-level median filtered ultrasonic signal P1(n) of the roller internal defect and a first-level median filtered reference signal R1(n) of the background clutter;

[0012] Step 3: construct a multi-stage filter Fs, obtain the filter core parameter Ks through the intelligent optimization algorithm, and obtain the background clutter estimation signal

[0013] Step 4: Compare the first-level median filtered ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal The open operation is performed to obtain the roller internal defect ultrasonic signal Po(n) with background clutter suppressed.

[0014] Further, the background clutter reference signal R(n) is obtained by:

[0015] In the defect-free roll area, multiple A-scan signals are measured by an ultrasonic digital flaw detector, and their average value is taken as the background noise reference signal.

[0016] Further, the median filtering of the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) comprises:

[0017] The median filter window radius is m, the window width is 2m+1, and the ultrasonic signal X(n)={x1,x2,......,xn};

[0018] Improved median filter formula:

[0019]

[0020] Referring to formula (1), the burr clutter is filtered out to obtain the first-level median filtering ultrasonic signal P1(n)=[p11, p12, ..., p1n] of the internal defect of the roller and the first-level median filtering reference signal R1(n)=[r11, r12, ..., r1n] of the background clutter.

[0021] Further, the construction of the multi-stage filter Fs includes:

[0022] The filter order is d, and the filter weight vector is ω=[ω0,ω1,......,ωd-1];

[0023] The filter output is the background clutter estimation signal, which can be expressed as:

[0024]

[0025] Furthermore, the method of obtaining the filter core parameter Ks by using an intelligent optimization algorithm includes:

[0026] The error signal e(n) between the first-level median filtering ultrasonic signal P1(n) of the internal defect of the roller and the output of the multi-level filter Fs is:

[0027]

[0028] Select the intelligent optimization algorithm indicator function as:

[0029]

[0030] Intelligent optimization algorithms can be particle swarm, genetic, differential evolution, simulated annealing, etc., or composite optimization algorithms;

[0031] An intelligent optimization algorithm is selected, initial parameters are set, and parameter iterative optimization is performed to obtain the optimal filter core parameters Ks = [Ks1, Ks2, ..., Ksd], and ω = Ks.

[0032] Furthermore, the first-level median filtered ultrasonic signal P1(n) of the internal defect of the roller is combined with the background clutter estimation signal Open operations include:

[0033] Substituting the background clutter primary median filter reference signal R1(n) and the optimal filter core parameter Ks into formula (2) to obtain the background clutter estimation signal

[0034] The background clutter estimation signal is subtracted from the first-order median filtered ultrasonic signal P1(n) of the internal defect of the roller The ultrasonic signal Po(n) of internal defects of the roller with background clutter suppressed is obtained.

[0035] Another object of the present invention is to provide a roller internal defect ultrasonic signal filtering system comprising:

[0036] A signal acquisition module, used to respectively acquire the ultrasonic signal P(n) of the internal defect of the roller and the background clutter reference signal R(n) by using an ultrasonic digital flaw detector;

[0037] A filtering module is used to perform median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) to remove signal burrs, thereby obtaining a first-level median filtering ultrasonic signal P1(n) of the roller internal defect and a first-level median filtering reference signal R1(n) of the background clutter;

[0038] The background clutter estimation signal acquisition module is used to construct a multi-stage filter Fs, obtain the filter core parameter Ks through an intelligent optimization algorithm, and obtain the background clutter estimation signal

[0039] The defect ultrasonic signal acquisition module is used to combine the first-level median filtered ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal The open operation is performed to obtain the roller internal defect ultrasonic signal Po(n) with background clutter suppressed.

[0040] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for filtering ultrasonic signals of internal defects of a rolling mill.

[0041] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the method for filtering ultrasonic signals of internal defects of a rolling mill roll.

[0042] Another object of the present invention is to provide an information data processing terminal, which is used to implement the roller internal defect ultrasonic signal filtering system.

[0043] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0044] First, the purpose of the present invention is to overcome the interference of the background clutter signal of the existing ultrasonic A-scan of internal defects of the roller, and further provide a method for filtering the ultrasonic signal of internal defects of the roller. The ultrasonic detection signal containing background clutter and noise and the background clutter reference signal are filtered once by median filtering to eliminate the peak pulse interference in the signal. A multi-stage filter is constructed, and the core parameters of the filter are obtained by an intelligent optimization algorithm to estimate the background clutter signal, thereby separating the ultrasonic A-scan signal after filtering out the background clutter.

[0045] The present invention adopts improved median filtering as a first-stage filter, which can effectively eliminate high-frequency noise in ultrasonic signals and greatly improve the convergence speed of subsequent second-stage filter parameter optimization.

[0046] The present invention constructs a multi-stage filter as a secondary filter, adopts an intelligent optimization algorithm to replace the traditional least squares algorithm to obtain the filter parameters, does not need to perform related parameter matrix and its inverse matrix operations, and greatly simplifies the filter parameter acquisition process.

[0047] The secondary filter designed in the present invention has a simple structure and is easy to implement without complex matrix operations. It can effectively filter out background clutter and interference signals in the ultrasonic signal of the roller, and provides a basis for subsequent research on automatic identification of internal defects of the roller.

[0048] The advantages of the present invention also include:

[0049] 1. Multi-stage filtering: the first-stage median filter removes high-frequency interference signals, and the second-stage filter achieves effective estimation of background clutter.

[0050] 2. The second-stage filter does not require a large number of matrix operations and uses intelligent algorithms to estimate filter parameters.

[0051] The present invention mainly adopts a two-stage filtering structure, combines median filtering and FIR filtering, and uses an intelligent optimization algorithm to obtain filter parameters. The filter structure and parameter iterative optimization process are simple, the calculation amount is small, the adaptability is strong, and the filtering effect is obvious.

[0052] Second, the ultrasonic signal filtering method of the present invention mainly solves the following technical problems in the prior art in industrial applications and has achieved significant technical progress:

[0053] Technical issues solved

[0054] 1. Background clutter interference: Traditional ultrasonic flaw detection methods are easily interfered by background clutter, resulting in low signal-to-noise ratio and affecting the accuracy of defect detection. Especially in the steel and oil industries, due to the complex surface and internal structure of the rolls and strong background noise, it is often difficult to effectively extract defect signals, resulting in misjudgment or missed judgment.

[0055] 2. Signal burr influence: Roller ultrasonic testing signals usually contain burr signals, which will interfere with the analysis and processing of defect signals. Traditional methods are difficult to effectively remove these burrs, thus affecting the stability of the test results.

[0056] 3. Difficulty in optimizing filter parameters: In existing filter designs, it is difficult to obtain optimized parameters, resulting in poor filtering effects and even a weakening effect on target defect signals. This problem is particularly prominent in complex signal environments.

[0057] Technological advancements:

[0058] 1. Improve signal-to-noise ratio and detection accuracy: By introducing a multi-stage filter and using an intelligent optimization algorithm to determine the core parameters of the filter, this method can significantly suppress background clutter interference and optimize the signal-to-noise ratio. Thanks to the combined use of median filtering and background clutter estimation signals, the present invention significantly improves the clarity of ultrasonic signals, making defect signals easier to identify, thereby improving the accuracy of defect detection.

[0059] 2. Enhance the robustness of signal processing: The improved median filtering method is used to remove burr signals, avoiding the interference of signal mutation on the detection results. This progress improves the stability of ultrasonic detection and greatly enhances the reliability of detection signals, which is suitable for roller detection of different materials and sizes.

[0060] 3. Intelligent filter parameter optimization: The optimal filter parameters are selected through intelligent optimization algorithms, which reduces manual intervention and improves the adaptability and processing effect of the filter. This method not only reduces the cost of equipment debugging and maintenance in actual detection, but also shortens the detection time and improves the detection efficiency.

[0061] In summary, this technical solution has achieved remarkable technical progress in background clutter suppression, signal stability improvement, and intelligent filter design, providing a more efficient and accurate technical means for non-destructive testing of rolling mill rolls, and meeting the needs of modern industry for high-precision testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 is a flow chart of a method for filtering ultrasonic signals of internal defects of a rolling mill provided by an embodiment of the present invention;

[0063] Figure 2 It is a control block diagram of a method for filtering ultrasonic signals of internal defects of a rolling mill provided by an embodiment of the present invention;

[0064] Figure 3 1 is a graph of a background clutter reference signal R(n) provided by the method for filtering ultrasonic signals of internal defects of a rolling mill roll according to Example 1 of the present invention;

[0065] Figure 4 1 is a graph of a background clutter primary median filtering reference signal R1(n) after being processed by the ultrasonic signal filtering method for internal defects of a roller according to Example 1 of the present invention;

[0066] Figure 5 is a graph of an ultrasonic signal P(n) of an internal defect of a roller provided by the method for filtering ultrasonic signals of an internal defect of a roller in Example 1 of the present invention;

[0067] Figure 6 is a graph of the first-level median filtered ultrasonic signal P1(n) of the internal defect of a roller after being processed by the ultrasonic signal filtering method of the internal defect of a roller according to Example 1 of the present invention;

[0068] Figure 7 It is a convergence curve diagram of the optimal index J optimized by the PSO intelligent algorithm in the method for filtering ultrasonic signals of internal defects of a rolling mill roll according to Example 1 of the present invention;

[0069] Figure 8 It is a background clutter estimation signal curve diagram after secondary intelligent filtering of the ultrasonic signal filtering method for internal defects of a rolling mill roll according to Example 1 of the present invention;

[0070] Fig. 9 It is a graph of the Po(n) of the ultrasonic signal of the internal defect of the roller with suppressed background clutter after the secondary intelligent filtering of the ultrasonic signal filtering method of the internal defect of the roller in Example 1 of the present invention;

[0071] Fig.10 It is a structural block diagram of the roller internal defect ultrasonic signal filtering system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0073] The ultrasonic signal filtering technology based on this method is widely used in the field of roller manufacturing and detection, especially in the following two specific industrial scenarios:

[0074] 1. Roller inspection in the steel industry

[0075] In the steel industry, rolls are important consumables, and their internal defects will affect the durability of the rolls and the surface quality of the finished steel. Through this ultrasonic signal filtering method, internal defects can be accurately detected before the rolls are put into use to avoid misjudgments caused by clutter signals. For example, during the factory inspection of rolls, technicians use ultrasonic flaw detectors to obtain ultrasonic signals inside the rolls, and use the above filtering method to remove background clutter, accurately locate possible defects such as cracks and pores, and improve detection accuracy. After implementing this method, not only can the service life of the rolls be extended, but maintenance costs can also be reduced and production efficiency can be improved.

[0076] 2. Maintenance of rolling mill equipment in the oil industry

[0077] Roller equipment in the oil industry is mainly used to transport oil pipes and other heavy steel products. The equipment is prone to internal defects after long-term high-intensity operation. Using this ultrasonic signal filtering method, the health of the roller can be detected during regular maintenance, and internal potential defects can be quickly discovered. In actual applications, operators obtain roller ultrasonic signals through non-destructive testing methods, and then perform background clutter suppression processing to obtain clean defect signals, avoiding ignoring potential problems due to clutter interference. The application of this technology effectively ensures the safe operation of the equipment and improves the quality control level during the oil pipe processing process.

[0078] like Figure 1 As shown, the method for filtering ultrasonic signals of internal defects of a roller provided by an embodiment of the present invention comprises the following steps:

[0079] S101, using an ultrasonic digital flaw detector to obtain a roller internal defect ultrasonic signal P(n) and a background clutter reference signal R(n);

[0080] S102, performing median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) to remove signal burrs, thereby obtaining a first-level median filtered ultrasonic signal P1(n) of the roller internal defect and a first-level median filtered reference signal R1(n) of the background clutter;

[0081] S103, construct a multi-stage filter Fs, obtain the filter core parameter Ks through an intelligent optimization algorithm, and obtain the background clutter estimation signal

[0082] S104, comparing the first-order median filtered ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal The open operation is performed to obtain the roller internal defect ultrasonic signal Po(n) with background clutter suppressed.

[0083] The filtering method of the ultrasonic signal of the internal defect of the roller provided by the present invention accurately identifies and suppresses the background clutter through multi-step filtering and intelligent optimization algorithm, thereby enhancing the clarity of the defect signal. First, in step 1, the defect signal P(n) and the background clutter reference signal R(n) inside the roller are respectively obtained by using an ultrasonic digital flaw detector. These signals contain a large amount of clutter and noise from the roller material, and it is difficult to directly distinguish the defect signal from the background clutter, so further processing is required.

[0084] In step 2, the acquired defect signal P(n) and background clutter signal R(n) are subjected to preliminary median filtering. Median filtering can effectively remove burrs and sharp noise in the signal, preventing these transient signals from affecting subsequent analysis and processing. After median filtering, preliminary smoothed signals P1(n) and R1(n) are obtained. These signals are more stable than the original signals, laying the foundation for subsequent multi-stage filtering processing.

[0085] In step 3, a multi-stage filter Fs is constructed, and the core parameter Ks of the filter is determined by an intelligent optimization algorithm. The intelligent optimization algorithm adaptively adjusts the filter parameters according to the signal characteristics to more accurately separate the background clutter signal and the defect signal. The optimized multi-stage filter can generate an estimated signal of the background clutter, which is helpful for suppressing the background clutter in the subsequent steps.

[0086] In step 4, the defect signal P1(n) is calculated with the background clutter estimation signal to finally obtain the defect signal Po(n) that eliminates most of the background clutter. The key to this step is to use the previously optimized filter to accurately separate and suppress the background clutter, leaving a clearer defect signal. The final processed Po(n) signal can be used for accurate detection of internal defects of the roll, improving the reliability and accuracy of defect detection.

[0087] The method for obtaining the background clutter reference signal R(n) provided by the embodiment of the present invention is as follows:

[0088] In the defect-free roll area, multiple A-scan signals are measured by an ultrasonic digital flaw detector, and their average value is taken as the background noise reference signal.

[0089] The method of performing median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) provided in an embodiment of the present invention includes:

[0090] The median filter window radius is m, the window width is 2m+1, and the ultrasonic signal X(n)={x1,x2,......,xn};

[0091] Improved median filter formula:

[0092]

[0093] Referring to formula (1), the burr clutter is filtered out to obtain the first-level median filtering ultrasonic signal P1(n)=[p11, p12, ..., p1n] of the internal defect of the roller and the first-level median filtering reference signal R1(n)=[r11, r12, ..., r1n] of the background clutter.

[0094] The multi-stage filter Fs provided in the embodiment of the present invention comprises:

[0095] The filter order is d, and the filter weight vector is ω=[ω0,ω1,......,ωd-1];

[0096] The filter output is the background clutter estimation signal, which can be expressed as:

[0097]

[0098] The method for obtaining the filter core parameter Ks by using an intelligent optimization algorithm provided by an embodiment of the present invention includes:

[0099] The error signal e(n) between the first-level median filtering ultrasonic signal P1(n) of the internal defect of the roller and the output of the multi-level filter Fs is:

[0100]

[0101] Select the intelligent optimization algorithm indicator function as:

[0102]

[0103] Intelligent optimization algorithms can be particle swarm, genetic, differential evolution, simulated annealing, etc., or composite optimization algorithms;

[0104] An intelligent optimization algorithm is selected, initial parameters are set, and parameter iterative optimization is performed to obtain the optimal filter core parameters Ks = [Ks1, Ks2, ..., Ksd], and ω = Ks.

[0105] The embodiment of the present invention provides a method for combining the first-level median filtering ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal Open operations include:

[0106] Substituting the background clutter primary median filter reference signal R1(n) and the optimal filter core parameter Ks into formula (2) to obtain the background clutter estimation signal

[0107] The background clutter estimation signal is subtracted from the first-order median filtered ultrasonic signal P1(n) of the internal defect of the roller The ultrasonic signal Po(n) of internal defects of the roller with background clutter suppressed is obtained.

[0108] like Fig.10 As shown, an embodiment of the present invention provides a roller internal defect ultrasonic signal filtering system provided by the embodiment of the present invention, comprising:

[0109] A signal acquisition module, used to respectively acquire the ultrasonic signal P(n) of the internal defect of the roller and the background clutter reference signal R(n) by using an ultrasonic digital flaw detector;

[0110] A filtering module is used to perform median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) to remove signal burrs, thereby obtaining a first-level median filtering ultrasonic signal P1(n) of the roller internal defect and a first-level median filtering reference signal R1(n) of the background clutter;

[0111] The background clutter estimation signal acquisition module is used to construct a multi-stage filter Fs, obtain the filter core parameter Ks through an intelligent optimization algorithm, and obtain the background clutter estimation signal

[0112] The defect ultrasonic signal acquisition module is used to combine the first-level median filtered ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal The open operation is performed to obtain the roller internal defect ultrasonic signal Po(n) with background clutter suppressed.

[0113] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for filtering ultrasonic signals of internal defects of a rolling mill.

[0114] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the method for filtering ultrasonic signals of internal defects of a rolling mill roll.

[0115] Another object of the present invention is to provide an information data processing terminal, which is used to implement the roller internal defect ultrasonic signal filtering system.

[0116] The present invention is specifically implemented:

[0117] The present disclosure provides a method for filtering ultrasonic signals of internal defects of a roller. Figure 1 The present invention is a flow chart of a method for filtering ultrasonic signals of internal defects of a roller given in an embodiment, the method comprising the following steps:

[0118] In step 101, an ultrasonic digital flaw detector is used to obtain the internal defect ultrasonic signal P(n) of the roller and the background noise reference signal R(n). In the defect-free roller area, multiple A-scan signals are measured by the ultrasonic digital flaw detector, and their average value is taken as the background noise reference signal R(n).

[0119] In step 102, the median filter window radius is selected as m, the window width is selected as 2m+1, the ultrasonic signal X(n)={x1,x2,......,xn}, and the improved median filter formula is:

[0120]

[0121] Referring to formula (5), the burr clutter is filtered out, and the first-level median filter ultrasonic signal P1(n)=[p11, p12, ..., p1n] of the internal defect of the roller and the first-level median filter reference signal R1(n)=[r11, r12, ..., r1n] of the background clutter are obtained respectively.

[0122] In step 103, the filter order is selected as d, and the filter weight vector is ω=[ω0,ω1,......,ωd-1];

[0123] The filter output is the background clutter estimation signal, which can be expressed as:

[0124]

[0125] The error signal e(n) output by the first-level median filter ultrasonic signal P1(n) of the internal defect of the roller and the multi-level filter Fs is:

[0126]

[0127] Select the intelligent optimization algorithm indicator function as:

[0128]

[0129] Intelligent optimization algorithms can be particle swarm, genetic, differential evolution, simulated annealing, etc., or composite optimization algorithms;

[0130] An intelligent optimization algorithm is selected, initial parameters are set, and parameter iterative optimization is performed to obtain the optimal filter core parameters Ks = [Ks1, Ks2, ..., Ksd], and ω = Ks.

[0131] In step 104, the background clutter first-order median filter reference signal R1(n) and the optimal filter core parameter Ks are substituted into formula (2) to obtain the background clutter estimation signal The signal is estimated by subtracting the background clutter from the first-order median-filtered ultrasonic signal P1(n) of the internal defect of the roller The ultrasonic signal Po(n) of internal defects of the roller with suppressed background clutter can be obtained.

[0132] A specific embodiment 1 provided by the present invention is described below:

[0133] The ultrasonic digital flaw detector is used to obtain the internal defect ultrasonic signal P(n) of the roller and the background noise reference signal R(n), as shown in Figure 5 , Figure 3 shown.

[0134] When the median filter window radius is selected as 3 and the window width is 7, the improved median filter formula is:

[0135]

[0136] Referring to formula (9), the burr clutter is filtered out, and the first-level median filter ultrasonic signal P1(n) of the internal defect of the roller and the first-level median filter reference signal R1(n) of the background clutter are obtained respectively. The curves are shown as follows: Figure 6 , Figure 4 shown.

[0137] The filter order is selected as 5, and the filter weight vector is ω = [ω0, ω1, ω2.ω3, ω4];

[0138] The filter output is the background clutter estimation signal, which can be expressed as:

[0139]

[0140] The error signal e(n) output by the first-level median filter ultrasonic signal P1(n) of the internal defect of the roller and the multi-level filter Fs is:

[0141]

[0142] Select the intelligent optimization algorithm indicator function as:

[0143]

[0144] Here, the PSO intelligent optimization algorithm is selected, and the initial parameters are set as shown in Table 1. After multiple iterations of optimization such as population parameter initialization, particle velocity position update, and selection operation update of individual optimal and global optimal, the optimal filter core parameter Ks = [1.013, 0.595, 0.088, -0.037, -0.637] is obtained. The optimal index J iterative optimization process is as follows Figure 7 As shown in Figure 2, after 15 iterations, the performance index J reaches the optimal value.

[0145] Table 1 Particle Swarm Optimization Algorithm (PSO) Parameters

[0146] Parameter name Parameter Value Parameter name Parameter Value Population size 40 Iterations 60 Dimensions 5 Inertia Weight 0.8 Variable lower bound [-1,-1,-1,-1,-1] Variable upper bound [1,1,1,1,1] Individual learning factor 0.5 Collective learning factor 0.6

[0147] Substitute the background clutter first-order median filter reference signal R1(n) and the optimal filter core parameter Ks into formula (10) to obtain the background clutter estimation signal The estimated curve is as follows Figure 8 As shown, combined Figure 4 It can be seen that the background clutter estimation signal The effective estimation of the background clutter primary median filter reference signal R1(n) can be well achieved.

[0148] The signal is estimated by subtracting the background clutter from the first-order median-filtered ultrasonic signal P1(n) of the internal defect of the roller The ultrasonic signal Po(n) of the internal defect of the roller with suppressed background noise can be obtained. The final curve is as follows: Fig. 9 As shown, the influence of background clutter can be effectively suppressed, thereby effectively extracting the ultrasonic signal of the internal defect of the roll.

[0149] The present invention is applied to ultrasonic detection of internal defects of rolls in steel rolling, and is mainly used to eliminate background clutter in ultrasonic A-scan signals of rolls to obtain denoised defect ultrasonic signals for subsequent roll defect identification and positioning.

[0150] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0151] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for filtering ultrasonic signals of internal defects of a rolling mill roll, characterized in that: The following steps are involved: Step 1, using an ultrasonic digital flaw detector to obtain the internal defect ultrasonic signal P(n) of the roller and the background clutter reference signal R(n); Step 2, performing median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) to remove signal burrs, thereby obtaining a first-level median filtered ultrasonic signal P1(n) of the roller internal defect and a first-level median filtered reference signal R1(n) of the background clutter; Step 3: construct a multi-stage filter Fs, obtain the filter core parameter Ks through the intelligent optimization algorithm, and obtain the background clutter estimation signal Step 4: Compare the first-level median filtered ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal The open operation is performed to obtain the roller internal defect ultrasonic signal Po(n) with background clutter suppressed.

2. The method for filtering ultrasonic signals of internal defects of a rolling mill roll according to claim 1, characterized in that: The background clutter reference signal R(n) is obtained by: In the defect-free roll area, multiple A-scan signals are measured by an ultrasonic digital flaw detector, and their average value is taken as the background noise reference signal.

3. The method for filtering ultrasonic signals of internal defects of a rolling mill roll according to claim 1, characterized in that: The median filtering of the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) comprises: The median filter window radius is m, the window width is 2m+1, and the ultrasonic signal X(n)={x1,x2,......,xn}; Improved median filter formula: Referring to formula (1), the burr clutter is filtered out to obtain the first-level median filtering ultrasonic signal P1(n)=[p11, p12, ..., p1n] of the internal defect of the roller and the first-level median filtering reference signal R1(n)=[r11, r12, ..., r1n] of the background clutter.

4. The method for filtering ultrasonic signals of internal defects of a rolling mill roll according to claim 1, characterized in that: The multi-stage filter Fs is constructed, comprising: The filter order is d, and the filter weight vector is ω=[ω0,ω1,......,ωd-1]; The filter output is the background clutter estimation signal, which can be expressed as:

5. The method for filtering ultrasonic signals of internal defects of a rolling mill roll according to claim 1, characterized in that: The method of obtaining the filter core parameter Ks by using an intelligent optimization algorithm includes: The error signal e(n) between the first-level median filtering ultrasonic signal P1(n) of the internal defect of the roller and the output of the multi-level filter Fs is: Select the intelligent optimization algorithm indicator function as: Intelligent optimization algorithms can be particle swarm, genetic, differential evolution, simulated annealing, etc., or composite optimization algorithms; An intelligent optimization algorithm is selected, initial parameters are set, and parameter iterative optimization is performed to obtain the optimal filter core parameters Ks = [Ks1, Ks2, ..., Ksd], and ω = Ks.

6. The method for filtering ultrasonic signals of internal defects of a rolling mill roll according to claim 1, characterized in that: The first-order median filtering ultrasonic signal P1(n) of the internal defect of the roller is combined with the background clutter estimation signal Open operations include: Substituting the background clutter primary median filter reference signal R1(n) and the optimal filter core parameter Ks into formula (2) to obtain the background clutter estimation signal The background clutter estimation signal is subtracted from the first-order median filtered ultrasonic signal P1(n) of the internal defect of the roller The ultrasonic signal Po(n) of internal defects of the roller with background clutter suppressed is obtained.

7. A roller internal defect ultrasonic signal filtering system implementing the roller internal defect ultrasonic signal filtering method according to any one of claims 1 to 6, characterized in that: The roller internal defect ultrasonic signal filtering system comprises: A signal acquisition module, used to respectively acquire the ultrasonic signal P(n) of the internal defect of the roller and the background clutter reference signal R(n) by using an ultrasonic digital flaw detector; A filtering module is used to perform median filtering on the roller internal defect ultrasonic signal P(n) and the background clutter reference signal R(n) to remove signal burrs, thereby obtaining a first-level median filtering ultrasonic signal P1(n) of the roller internal defect and a first-level median filtering reference signal R1(n) of the background clutter; The background clutter estimation signal acquisition module is used to construct a multi-stage filter Fs, obtain the filter core parameter Ks through an intelligent optimization algorithm, and obtain the background clutter estimation signal The defect ultrasonic signal acquisition module is used to combine the first-level median filtered ultrasonic signal P1(n) of the internal defect of the roller with the background clutter estimation signal The open operation is performed to obtain the roller internal defect ultrasonic signal Po(n) with background clutter suppressed.

8. A computer device, characterized in that: The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for filtering ultrasonic signals of internal defects of a rolling mill as claimed in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the method for filtering ultrasonic signals of internal defects of a rolling mill as claimed in any one of claims 1 to 6.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the roller internal defect ultrasonic signal filtering system as described in claim 7.

Images