Defect inversion method for irregular pipelines based on genetic algorithm and improved taboo search

Abstract

The present invention provides an irregular pipeline defect inversion method based on genetic algorithm and improved taboo search, which belongs to the technical field of pipeline in-detection. Based on the forward modeling of pipeline defects, this method can obtain preliminary inversion results through step-by-step iteration of genetic algorithm, but in order to avoid falling into the dilemma of local optimum, further search is required; the taboo search method is adopted to find the global optimal solution in a short time while jumping out of the local optimum. This method reduces the number of iterations of defect inversion, makes the convergence ability of the algorithm stronger, and makes it easier to find the best defect inversion data; this method greatly reduces the calculation time, and has the advantages of simple calculation method, fast calculation speed, high precision, etc.; this method has strong robustness and strong anti-interference ability, so it can reduce the social harm caused by pipeline leakage, thereby achieving adaptability to different defect sample sets.

Description

Technical Field

[0001] The invention belongs to the field of pipeline in-line detection, and in particular relates to an irregular pipeline defect inversion method based on genetic algorithm and improved taboo search. Background Art

[0002] Ferromagnetic materials are widely used in industrial fields such as petroleum and chemical industry. At present, they have been widely used in many fields such as oil and gas pipelines, and have achieved good results. Because they usually work in high temperature and high pressure environment, defects usually appear on their surface. Magnetic flux leakage detection is an important non-destructive testing method. It measures the leakage magnetic field distribution on the magnetic element by magnetic sensors, and is mainly used to detect defects and cracks in ferromagnetic materials. Compared with X-ray detection technology and magnetic particle detection technology, magnetic flux leakage detection technology has faster detection speed, higher detection accuracy, lower cost requirements and lower environmental requirements, and is more effective in non-destructive testing and evaluation. In magnetic flux leakage detection technology, quantifying the defect contour is a key step. By quantifying the size and shape of the defect, the severity of the defect can be judged, providing an accurate basis for subsequent repair and maintenance. The quantitative identification process of magnetic flux leakage detection signals is to determine whether there are defects in the material under test based on the magnetic flux leakage detection signals, and calibrate the shape and position of the defects, so as to realize the visualization of defect detection, which is called the inversion problem of magnetic flux leakage detection.

[0003] In the process of inversion using magnetic flux leakage signals, the methods used can be divided into direct inversion and iterative inversion. The results obtained by direct inversion are of low accuracy and cannot accurately depict the specific shape of the defect. When the actual defect is far from the training sample, the accuracy of the model is low, especially for irregular defects, which cannot be accurately predicted, and most of the defects in practical applications are irregular. Therefore, the iterative inversion method is used in the present invention. Its principle is: to improve the quantification accuracy of the defect contour in the magnetic flux leakage signal by continuously iterating and correcting the calculation process. Iterative inversion technology is an important technical method used in the field of magnetic flux leakage detection. This technology continuously optimizes model parameters in an iterative manner to more accurately invert the defect information in the material. The iterative inversion technology of magnetic flux leakage detection mainly consists of a forward model and an iterative optimization algorithm. Compared with the forward model, the inversion optimization problem is relatively complex. Due to the uneven distribution of the defect leakage magnetic field, and different defects may produce similar leakage magnetic field distributions, the results of defect inversion lack uniqueness and continuity, which makes it difficult to accurately solve the corresponding problems. Therefore, how to further improve the inversion accuracy and speed based on the existing iterative inversion technology is a problem that we urgently need to solve. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide an irregular pipeline defect inversion method based on genetic algorithm and improved taboo search to achieve the inversion of irregular pipeline defects in view of the deficiencies of the above-mentioned prior art.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: an irregular pipeline defect inversion method based on genetic algorithm and improved taboo search, comprising the following steps:

[0006] Step 1: Use sensors to collect magnetic flux leakage detection signals from defective sections on the pipeline, and use a defect reconstruction method based on a three-dimensional magnetic dipole model to perform forward modeling of pipeline defects;

[0007] Step 1.1: Define the initial depth of the pipeline defect;

[0008] Step 1.2: Using the initial depth of the pipeline defect in step 1.1, a three-dimensional magnetic dipole model is established for the pipeline defect to simulate the magnetic flux leakage detection signal of the pipeline defect;

[0009] Divide the pipeline defect into equal distances within the range grid, and set each grid as a cube; take a cube as a pipeline sub-defect to form a complete pipeline defect; set the height of the cube to be approximately equal to the equivalent depth of the grid area, obtain the leakage magnetic detection signal estimation matrix formed by the cube, and then obtain the leakage magnetic detection signal estimation matrix of the complete pipeline defect;

[0010] Step 1.3: Using the magnetic flux leakage detection signal estimation matrix obtained in step 1.2, simulate the magnetic flux leakage detection signal of the pipeline defect;

[0011] Step 1.4: Calculate the residual between the magnetic flux leakage detection signal of the pipeline defect simulated by the three-dimensional magnetic dipole model and the magnetic flux leakage detection signal collected by the sensor to obtain a reconstructed pipeline defect;

[0012] Step 1.5: Check whether the accuracy of the reconstructed pipeline defect meets the requirements. If not, return to step 1.1 to update the pipeline depth and repeat steps 1.2 to 1.5. Otherwise, directly reconstruct the three-dimensional pipeline defect contour to obtain the forward modeling result of the pipeline defect, that is, the exact depth value of the pipeline defect.

[0013] Step 2: Divide the pipeline defects into equal distances , construct the depth matrix of pipeline defects, the data in the depth matrix is ​​the estimated value of pipeline depth, and initialize the population of genetic algorithm;

[0014] The pipeline defects are divided into equal distances using the method in step 1 again. Each grid corresponds to a pipeline sub-defect. The entire pipeline defect is composed of pipeline sub-defects. The depth of each sub-defect in the grid constitutes The matrix is ​​the depth matrix. The data in the depth matrix is ​​the estimated value of the pipeline depth. ,in, is the data value in the depth matrix, d is the diameter of the pipe cross section; the population of the genetic algorithm is set to the data in the depth matrix;

[0015] Step 3: Set a fitness function so that the estimated pipe depth obtained in step 2 can match the exact depth value obtained in step 1;

[0016] The fitness function will score the chromosome data generated in each iteration to judge the fitness of the chromosome data, that is, to judge the quality of each group of data in the population; according to step 1, the leakage magnetic detection signal at the pipeline defect is obtained. ; The magnetic flux leakage detection signal generated by the magnetic flux leakage in each set of chromosome data of the population is , then there exists a fitness function ;

[0017] Step 4: Cross-pair and mutate the chromosome data in the population, and further match the pipeline depth estimate with the exact depth value obtained in step 1;

[0018] The mutation process is to randomly add or subtract 1 to the chromosome data after cross-pairing, thereby changing the genes in the individual chromosome data to increase the diversity of each set of chromosome data in the population;

[0019] Step 5: Replace the chromosome data of the parent generation with the generated offspring chromosome data to determine whether the set requirements are met. If not, re-execute step 4 until the set requirements are met to obtain an updated depth matrix. Otherwise, execute step 6;

[0020] Replace the chromosome data of the parent generation with the chromosome data of the offspring whose fitness is less than the set threshold after pairing and mutation, and determine whether it satisfies If it does not meet the requirement, repeat the pairing and mutation operation in step 4, iterate layer by layer until each set of chromosome data in the population, that is, the last data in each depth matrix meets the requirement. , where and , and are the minimum and maximum values ​​of the data in the depth matrix respectively;

[0021] Step 6: Perform a taboo search on the data in the updated depth matrix obtained in step 5 to obtain the final global optimal solution for pipeline defects;

[0022] Step 6.1: Use the data in the updated depth matrix obtained in step 5 as the current solution and set the taboo table Set to empty, set the taboo length T=d, and set the termination condition of the taboo search;

[0023] Step 6.2: Starting from the current solution, for the solution after the current iteration , generate a neighborhood solution set ;

[0024] In generating the neighborhood solution set In the process, a sliding window strategy is adopted, that is, a The pane is divided into The pipeline defects of the grids are slid from the beginning, where =N-2, so The pipeline defects of each grid are added and subtracted to obtain the neighborhood solution set , In the front pane The calculation formula for the data in each pane is: ,in, For the The data in each pane, , k is the step length, random is the result of the multiplication congruential method;

[0025] Data in the last pane ;

[0026] The specific formula for the random result of the multiplication congruential method is: , where a and c are constants and positive integers, m is the modulus, mod is the modulus operation, is the result of the k+1th step, and is the result of the kth step;

[0027] Step 6.3: Set the solution in the neighborhood Select the candidate solution set that is not tabooed; then select the candidate solution set from the neighborhood solution set according to the magnetic flux leakage detection signal and taboo table obtained in step 1 Select a set of candidate solutions ;

[0028] Step 6.4: Update the tabu table , repeat steps 6.2 to 6.4 to obtain the global optimal solution;

[0029] From the candidate solution set Select a set of optimal solutions as the current solution; repeat steps 6.2-6.4. When the solution obtained after repeated multiple times remains unchanged, the result is considered to be locally optimal; at this time, increase the step size k to increase the range of the solution neighborhood and search again; until the total difference with the leakage magnetic detection signal obtained in step 1 is less than the set threshold, stop searching and obtain the global optimal solution, that is, the data value of the final depth matrix.

[0030] The beneficial effects of the above technical solution are as follows: compared with the prior art, the irregular pipeline defect inversion method based on genetic algorithm and improved taboo search provided by the present invention has the following beneficial effects:

[0031] (1) In this method, the genetic algorithm is combined with the taboo search, which greatly improves the accuracy and convergence ability of the method, and avoids the local optimum at the fastest convergence speed to find the global optimal solution;

[0032] (2) This method is more suitable for real complex environments and has a stronger ability to resist external interference;

[0033] (3) This method divides the iteration process into two stages according to the discrete degree of the population: genetic calculation and layer-by-layer calculation. In addition, the law of corrosion is considered in the inversion process, making the defects after inversion more consistent with the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flow chart of an irregular pipeline defect inversion method based on the fusion of genetic algorithm and taboo search provided by an embodiment of the present invention;

[0035] Figure 2 Defect forward modeling flow chart provided by an embodiment of the present invention;

[0036] Figure 3 A magnetic flux leakage signal diagram obtained by simulation provided by an embodiment of the present invention;

[0037] Figure 4 A schematic diagram of a reconstructed three-dimensional pipeline defect contour provided by an embodiment of the present invention;

[0038] Figure 5 A schematic diagram of a real defect provided by an embodiment of the present invention;

[0039] Figure 6 A flowchart of taboo search in an iterative inversion method provided by an embodiment of the present invention;

[0040] Figure 7 A schematic diagram of the average error and the deepest point average error when 5% Gaussian white noise is randomly added to the test sample provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0042] In this embodiment, an irregular pipeline defect inversion method based on genetic algorithm and improved taboo search is provided. Figure 1 As shown, the following steps are included:

[0043] Step 1: Use sensors to collect magnetic flux leakage detection signals from defective sections on the pipeline, and use the defect reconstruction method based on the three-dimensional magnetic dipole model to perform forward modeling of pipeline defects, such as Figure 2 As shown;

[0044] Step 1.1: Define the initial depth of the pipeline defect;

[0045] Step 1.2: Using the initial depth of the pipeline defect in step 1.1, a three-dimensional magnetic dipole model is established for the pipeline defect to simulate the magnetic flux leakage detection signal of the pipeline defect;

[0046] In order to adapt to the three-dimensional contour of the pipeline defect, the pipeline defect range is divided into grid, and set each grid as a cube; take each cube as a pipeline sub-defect to form a complete pipeline defect; set the height of the cube to be approximately equal to the equivalent depth of the grid area, obtain the leakage magnetic detection signal estimation matrix formed by the cube, and then obtain the leakage magnetic detection signal estimation matrix of the complete pipeline defect;

[0047] In this embodiment, the pipeline defects are divided into five equal distances. 5 grids, the entire pipeline defect is composed of pipeline sub-defects; then for each grid, the leakage magnetic detection signal estimation matrix formed by the divided cube is , the magnetic flux leakage detection signal estimation matrix Superposition of the complete pipeline defect magnetic flux leakage detection signal estimation matrix ; ,in, The magnetic flux leakage detection signal estimation matrix is ​​formed by the cube formed by dividing the pipeline defect grid. is the magnetic flux leakage detection signal estimation matrix of the complete pipeline defect, i is the row of the magnetic flux leakage detection signal estimation matrix, j is the column of the magnetic flux leakage detection signal estimation matrix;

[0048] Step 1.3: Estimate the matrix using the magnetic flux leakage detection signal obtained in step 1.2 Simulate the magnetic flux leakage detection signal of pipeline defects, such as Figure 3 As shown;

[0049] Step 1.4: Calculate the residual between the magnetic flux leakage detection signal of the pipeline defect simulated by the three-dimensional magnetic dipole model and the magnetic flux leakage detection signal collected by the sensor to obtain the reconstructed pipeline defect;

[0050] In this embodiment, the three-dimensional magnetic dipole model is used to analyze the defects of the cubic-shaped pipeline and calculate the pipeline leakage magnetic field intensity at a point outside the pipeline defect affected by the defect. The specific formula is:

[0051] , in, is the magnetic field strength in the x-axis direction of the magnetic flux leakage detection of the pipeline affected by defects, is half the length of the pipe defect, is half the width of the pipe defect, is the depth of the pipeline defect, is the equivalent magnetic charge density at the pipeline defect location;

[0052] Step 1.5: Check whether the reconstructed pipeline defect accuracy meets the requirements. If not, return to step 1.1 to update the pipeline depth and repeat steps 1.2 to 1.5; otherwise, directly reconstruct the 3D pipeline defect contour, such as Figure 4 As shown, the forward modeling result of the pipeline defect, that is, the exact depth value of the pipeline defect, is obtained;

[0053] In this embodiment, the actual pipeline defect is as follows Figure 5 As shown;

[0054] Step 2: Divide pipeline defects into equal distances Grid, construct the depth matrix of pipeline defects, the data in the depth matrix is ​​the estimated value of pipeline depth, and initialize the population of genetic algorithm;

[0055] The pipeline defects are divided into equal distances using the method in step 1 again. Each grid corresponds to a pipeline sub-defect. The entire pipeline defect is composed of pipeline sub-defects. The depth of each sub-defect in the grid constitutes The matrix is ​​the depth matrix. The data in the depth matrix is ​​the estimated value of the pipeline depth. , is the data value in the depth matrix, d is the diameter of the pipe cross section; the population of the genetic algorithm is set to the data in the depth matrix;

[0056] Step 3: Set a fitness function so that the estimated pipe depth obtained in step 2 can match the exact depth value obtained in step 1;

[0057] The fitness function will score the chromosome data generated in each iteration to judge the fitness of the chromosome data, that is, to judge the quality of each group of data in the population; according to step 1, the leakage magnetic detection signal at the pipeline defect is obtained. ; The magnetic flux leakage detection signal generated by the magnetic flux leakage in each set of chromosome data of the population is , then there exists a fitness function , fitness function The smaller it is, the better the adaptability of the pipeline depth value is;

[0058] Step 4: Cross-pair and mutate the chromosome data in the population, and further match the pipeline depth estimate with the exact depth value obtained in step 1;

[0059] The mutation process is to randomly add or subtract 1 to the chromosome data after cross-pairing, thereby changing the genes in the individual chromosome data to increase the diversity of each set of chromosome data in the population;

[0060] Step 5: Replace the chromosome data of the parent generation with the generated offspring chromosome data to determine whether the set requirements are met. If not, re-execute step 4 until the set requirements are met to obtain an updated depth matrix. Otherwise, execute step 6;

[0061] Replace the chromosome data of the parent generation with the chromosome data of the offspring whose fitness is less than the set threshold after pairing and mutation, and determine whether it satisfies If it does not meet the requirement, repeat the pairing and mutation operation in step 4, iterate layer by layer until each set of chromosome data in the population, that is, the last data in each depth matrix meets the requirement. ,in, and They are the minimum and maximum values ​​of the data in the depth matrix respectively;

[0062] Step 6: Perform taboo search on the data in the updated depth matrix obtained in step 5 to obtain the final global optimal solution for pipeline defects, such as Figure 6 As shown;

[0063] Step 6.1: Use the data in the updated depth matrix obtained in step 5 as the current solution, set the taboo table T=d to empty, set the taboo length, and set the termination condition of the taboo search;

[0064] Step 6.2: Starting from the current solution, for the solution after the current iteration , generate a neighborhood solution set ;

[0065] In generating the neighborhood solution set In the process, a sliding window strategy is adopted, that is, a The pane is divided into The pipeline defects of grids are slid from the beginning, where n=N-2, so as to The pipeline defects of each grid are appropriately added and subtracted to obtain the neighborhood solution set , In the front pane The calculation formula for the data in each pane is: ,in, For the The data in each pane, k is the step length, random is the result of the multiplication congruential method;

[0066] Data in the last pane ;

[0067] The specific formula for the random result of the multiplication congruential method is: , where a and c are constants and positive integers, m is the modulus, mod is the modulus operation, is the result of the k+1th step, and is the result of the kth step;

[0068] Step 6.3: Set the solution in the neighborhood Select the candidate solution set that is not tabooed; then select the candidate solution set from the neighborhood solution set according to the magnetic flux leakage detection signal and taboo table obtained in step 1 Select a set of candidate solutions ;

[0069] Step 6.4: Update the tabu table , repeat steps 6.2 to 6.4 to obtain the global optimal solution;

[0070] From the candidate solution set Select a set of optimal solutions as the current solution; repeat steps 6.2 to 6.4. When the solution obtained after multiple repetitions remains unchanged, the result is considered to be a local optimum. At this time, increase the step size k to increase the range of the solution neighborhood and search again. Stop searching until the total difference with the leakage magnetic detection signal obtained in step 1 is less than the set threshold, and obtain the global optimal solution, that is, the data value of the final depth matrix.

[0071] In this embodiment, Figure 7 As shown, it reflects the average error of the defect size when 5% Gaussian white noise is randomly added to the test sample; the average error is within 0.5mm, and the error at the deepest point is within 1mm; the above results can fully prove that the result of defect inversion using the method of the present invention still has strong accuracy and robustness;

[0072] In this embodiment, firstly, based on the forward modeling of pipeline defects, the depth matrix is ​​gradually iterated using a genetic algorithm to obtain a preliminary inversion result. However, in order to avoid falling into the dilemma of local optimality, further search is required. Considering that it will take a long time for a general search method to obtain a global optimal solution, if a time-saving search method is used, the globality of the final result is poor. Therefore, a taboo search method is used to find the global optimal solution in a short time while jumping out of the local optimality.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. An irregular pipeline defect inversion method based on genetic algorithm and improved taboo search, characterized by: The following steps are involved: Step 1: Use sensors to collect magnetic flux leakage detection signals of defective sections on the pipeline, and use the defect reconstruction method based on the three-dimensional magnetic dipole model to perform forward modeling of pipeline defects, including the following steps: Step 1.1: Define the initial depth of the pipeline defect; Step 1.2: Using the initial depth of the pipeline defect in step 1.1, a three-dimensional magnetic dipole model is established for the pipeline defect to simulate the magnetic flux leakage detection signal of the pipeline defect; Divide the pipeline defect into equal distances within the range grid, and set each grid as a cube; take a cube as a pipeline sub-defect to form a complete pipeline defect; set the height of the cube to be approximately equal to the equivalent depth of the grid area, obtain the leakage magnetic detection signal estimation matrix formed by the cube, and then obtain the leakage magnetic detection signal estimation matrix of the complete pipeline defect; Step 1.3: Using the magnetic flux leakage detection signal estimation matrix obtained in step 1.2, simulate the magnetic flux leakage detection signal of the pipeline defect; Step 1.4: Calculate the residual between the magnetic flux leakage detection signal of the pipeline defect simulated by the three-dimensional magnetic dipole model and the magnetic flux leakage detection signal collected by the sensor to obtain a reconstructed pipeline defect; Step 1.5: Check whether the accuracy of the reconstructed pipeline defect meets the requirements. If not, return to step 1.1 to update the pipeline depth and repeat steps 1.2 to 1.

5. Otherwise, directly reconstruct the three-dimensional pipeline defect contour to obtain the forward modeling result of the pipeline defect, that is, the exact depth value of the pipeline defect. Step 2: Divide the pipeline defects into equal distances , construct the depth matrix of pipeline defects, the data in the depth matrix is ​​the estimated value of pipeline depth, and initialize the population of genetic algorithm; The pipeline defects are divided into equal distances using the method in step 1 again. Each grid corresponds to a pipeline sub-defect. The entire pipeline defect is composed of pipeline sub-defects. The depth of each sub-defect in the grid constitutes The matrix is ​​the depth matrix. The data in the depth matrix is ​​the estimated value of the pipeline depth. ,in, is the data value in the depth matrix, d is the diameter of the pipe cross section; the population of the genetic algorithm is set to the data in the depth matrix; Step 3: Set a fitness function so that the estimated pipe depth obtained in step 2 can match the exact depth value obtained in step 1; Step 4: Cross-pair and mutate the chromosome data in the population, and further match the pipeline depth estimate with the exact depth value obtained in step 1; Step 5: Replace the chromosome data of the parent generation with the generated offspring chromosome data to determine whether the set requirements are met. If not, re-execute step 4 until the set requirements are met to obtain an updated depth matrix. Otherwise, execute step 6; Step 6: Perform a taboo search on the data in the updated depth matrix obtained in step 5 to obtain the final global optimal solution for pipeline defects, including the following steps: Step 6.1: Use the data in the updated depth matrix obtained in step 5 as the current solution and set the taboo table Set to empty, set the taboo length T=d, and set the termination condition of the taboo search; Step 6.2: Starting from the current solution, for the solution after the current iteration , generate a neighborhood solution set ; Step 6.3: Set the solution in the neighborhood Select the candidate solution set that is not tabooed; then select the candidate solution set from the neighborhood solution set according to the magnetic flux leakage detection signal and taboo table obtained in step 1 Select a set of candidate solutions ; Step 6.4: Update the tabu table , repeat steps 6.2-6.4 to obtain the global optimal solution.

2. According to claim 1, the irregular pipeline defect inversion method based on genetic algorithm and improved taboo search is characterized by: The specific method of step 3 is: The fitness function will score the chromosome data generated in each iteration to judge the fitness of the chromosome data, that is, to judge the quality of each group of data in the population; according to step 1, the leakage magnetic detection signal at the pipeline defect is obtained. ; The magnetic flux leakage detection signal generated by the magnetic flux leakage in each set of chromosome data of the population is , then there exists a fitness function .

3. The irregular pipeline defect inversion method based on genetic algorithm and improved taboo search according to claim 2 is characterized by: The mutation process is to randomly add 1 or subtract 1 to the chromosome data after cross-pairing, thereby changing the genes in the individual chromosome data to increase the diversity of each set of chromosome data in the population.

4. The irregular pipeline defect inversion method based on genetic algorithm and improved taboo search according to claim 3 is characterized by: The specific method of step 5 is: Replace the chromosome data of the parent generation with the chromosome data of the offspring whose fitness is less than the set threshold after pairing and mutation, and determine whether it satisfies If it does not meet the requirement, repeat the pairing and mutation operation in step 4, and iterate layer by layer until each set of chromosome data in the population, that is, the last data in each depth matrix, meets the requirement: ,in, and are the minimum and maximum values ​​of the data in the depth matrix respectively.

5. The irregular pipeline defect inversion method based on genetic algorithm and improved taboo search according to claim 4 is characterized by: The specific method of step 6.2 is: In generating the neighborhood solution set In the process, a sliding window strategy is adopted, that is, a The pane is divided into The pipeline defects of the grids are slid from the beginning, where , thus The pipeline defects of each grid are added and subtracted to obtain the neighborhood solution set , In the front pane The calculation formula for the data in each pane is: ,in, For the The data in each pane, , k is the step size, random is the result of the multiplication congruential method; the data in the last pane ; The specific formula for the result random of the multiplication congruential method is: , where a and c are constants and positive integers, m is the modulus, and mod is the modulus operation. is the result of step k+1, is the result of step k.

6. The irregular pipeline defect inversion method based on genetic algorithm and improved taboo search according to claim 5 is characterized by: The specific method of step 6.4 is: From the candidate solution set Select a set of optimal solutions as the current solution; Repeat steps 6.2 to 6.

4. If the solution remains unchanged after repeated execution, the result is considered to be a local optimum. Increase the step size k to increase the range of the solution neighborhood and search again. Stop searching until the total difference between the magnetic flux leakage detection signal obtained in step 1 is less than the set threshold and obtain the global optimal solution, i.e., the data value of the final depth matrix.

Images