Strength test method for boiler steel plates with weld defects

Abstract

The present invention discloses a strength testing method for boiler steel plates with weld defects, which relates to the technical field of steel plate data analysis. The method comprises the following steps: collecting steel plate test samples, performing strength tests on the test samples, obtaining tensile strength test results, numbering the test samples, obtaining weld characteristic data of the test samples, analyzing the fitness between test samples with the same weld defect type, analyzing the reference accuracy of the test data, establishing a steel plate test selection model, obtaining weld characteristic data of the current steel plate to be processed, confirming the weld defect type and number of the current steel plate to be processed, performing weld characteristic comparison with test samples of the same type, analyzing the fitness between the current steel plate to be processed and the test samples, inputting actual data of the current steel plate to be processed into the model, and planning a method for obtaining the tensile strength data of the current steel plate to be processed, thereby improving the accuracy of the steel plate data required in the actual processing process and reducing unnecessary strength testing work.

Description

Technical Field

[0001] The invention relates to the technical field of steel plate data analysis, in particular to a strength testing method for boiler steel plates with weld defects. Background Art

[0002] Boiler steel plate mainly refers to the hot-rolled medium and thick plate material used to manufacture superheaters, main steam pipes and boiler firebox heating surfaces. Boiler steel plate is a key material in boiler manufacturing. In order to ensure the safe operation of the boiler and extend its service life, it is necessary to conduct a boiler steel plate strength test. The strength test can ensure that the boiler steel plate has good physical properties and workability. The tensile strength of the steel plate refers to the maximum tensile force that the steel plate can withstand during the stretching process. When conducting a tensile strength test on the steel plate, there is a method in the prior art to conduct a tensile strength test on the complete component. However, in real application scenarios, some defects may occur in the weld seam during the welding process. The tensile strength of the steel plate with weld defects is different from that of the steel plate without weld defects. Therefore, it is necessary to conduct a tensile strength test on the steel plate with weld defects to obtain a more accurate tensile strength of the steel plate with different weld defects, so as to set appropriate working parameters in the later processing of the steel plate and achieve the purpose of steel plate quality control.

[0003] After multiple tests, tensile strength data of steel plates with different weld defects can be collected. The collected data can be used as reference data when setting working parameters for actual steel plate processing. If a steel plate identical to the test sample appears in actual processing later, the tensile strength of the test sample obtained can be used as the tensile strength data of the steel plate during actual processing. However, during the strength test, the tensile strength of steel plates with the same weld defect type but different weld profiles may also be different, and the test data collected in the early stage may not be perfect. During actual processing, steel plates with weld profiles significantly different from those of the test sample may appear. For steel plates with the same weld defect type but significantly different weld profiles as the test sample, if the tensile strength of the test sample is still used as the tensile strength of the current actual steel plate, the obtained data may not be accurate enough. The existing technology cannot use the optimal method to obtain the tensile strength of steel plates with weld defects in actual situations, and cannot improve the accuracy of the tensile strength data of steel plates obtained during actual processing. Summary of the Invention

[0004] The object of the present invention is to provide a strength testing method for boiler steel plates with weld defects, so as to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for testing the strength of a boiler steel plate with weld defects, characterized in that it comprises the following steps:

[0006] S1: Collect steel plate test samples, perform strength tests on the test samples, and obtain tensile strength test results;

[0007] S2: Number the test samples, obtain the weld characteristic data of the test samples, and analyze the compatibility between the test samples with the same weld defect type;

[0008] S3: Count the total number of test samples with the same fitness and the number of test samples with the same fitness and the same tensile strength obtained in the test, and analyze the reference accuracy of the test data;

[0009] S4: Fitting training is performed on the sample number, fitness and reference accuracy data to establish a steel plate test selection model;

[0010] S5: Acquire the weld feature data of the current steel plate to be processed, confirm the weld defect type and number of the current steel plate to be processed, compare the weld features with the test sample of the same type, and analyze the compatibility between the current steel plate to be processed and the test sample;

[0011] S6: Inputting the serial number and the suitability of the current steel plate to be processed into the steel plate test selection model, and planning a method for obtaining the tensile strength data of the current steel plate to be processed.

[0012] Further, step S1 includes: butt welding two steel plate specimens to form a weld with a width of d, preparing and collecting steel plate test specimens with weld defects and different weld defect types, wherein the weld defect types include porosity, slag inclusion, lack of fusion, lack of penetration, and cracks, collecting a total of m steel plate test specimens, each having a weld width of d, setting a test environment temperature to T, performing a tensile test on the test specimens under the corresponding test environment, and obtaining the tensile strength of the test specimens obtained by the test;

[0013] Taking into account that in actual situations, some defects may occur in the welds, and the strength of the steel plate has decreased at this time, it can no longer be processed according to the strength test results of the steel plate without weld defects. The relevant parameters of the working conditions, such as temperature and pressure, should be lowered. Therefore, the steel plate test sample with welding defects is placed in a high temperature environment for tensile testing to test the tensile strength of the steel plate. The tensile strength of the steel plates with different types of welding defects obtained by the test can help to reasonably adjust the working parameters according to the strength test results of the steel plate test samples with welding defects during the actual processing process. Taking into account the impact of weld defects on the strength of the steel plate, it is helpful to avoid the problem of affecting the quality and service life of the steel plate due to improper setting of working parameters.

[0014] Furthermore, step S2 includes: randomly extracting samples with different weld defect types from the m test samples as samples to be calibrated: each weld defect type corresponds to a sample to be calibrated, and the steel plate test samples remaining after the extraction process are numbered, and the steel plate test samples with the same weld defect type have the same number, and the number set of the remaining n steel plate test samples is obtained as A = {A1, A2, ..., An}, and the same number exists in the number set, that is, some of the n steel plate test samples have the same number, and samples representing a random type of weld defect are retrieved from the n steel plate test samples, and the number of retrieved samples is counted as k, and weld images of the k retrieved samples are photographed, and the Canny edge detection technology is used to identify k The weld profile of the sample is obtained by shooting weld images of the sample to be calibrated with the same weld defect type as the k samples, identifying the weld profile of the corresponding sample to be calibrated, and extracting weld profile feature data using OpenCV technology. The feature data includes the area of ​​the area surrounded by the weld profile, the boundary length of the weld profile, and the area of ​​the minimum circle that can surround the weld profile. The area set of the area surrounded by the weld profile of the k samples is {R1, R2, ..., Rk}, the boundary length set of the weld profile is {L1, L2, ..., Lk}, and the area set of the minimum circle that can surround the weld profile is {r1, r2, ..., rk}. The weld profile feature data of the sample to be calibrated with the same weld defect type as the k samples are R ’ 、L ’ and r ’ , analyze the degree of fit Wi between a random sample among the k samples and the sample to be calibrated:

[0015] Wi=1 / (|R ’ -Ri|+|L ’ -Li|+|r ’ -ri|);

[0016] Since there are high probability differences between weld profiles in actual situations, the default value is |R ’ -Ri|+|L ’ -Li|+|r ’ -ri|≠0, the fitness set between the k samples and the sample to be calibrated is W = {W1, W2, ..., Wi, ..., Wk}, and there are equal fitness in the set W, where Ri represents the area of ​​the region enclosed by the weld contour of the i-th sample among the k samples, Li represents the boundary length of the weld contour of the i-th sample among the k samples, and ri represents the minimum circle area that can enclose the weld contour of the i-th sample;

[0017] Considering that the test sample data collected in the early stage may not be perfect, steel plates with weld profiles different from those of the test samples may appear in the actual processing process. Therefore, the samples to be calibrated are randomly selected from the test samples collected in the early stage, and the remaining steel plate test samples with the same weld defect type as the samples to be calibrated are compared with the weld feature data of the samples to be calibrated to analyze the degree of fit between the two. The higher the degree of fit, the smaller the difference in weld characteristics between the two samples.

[0018] Furthermore, step S3 includes: counting the number of samples with the same fitness W1 as the sample to be calibrated among the k samples as a, the number of samples with the same tensile strength as the sample to be calibrated among the a samples as b, and calculating the reference accuracy P1 of the tensile strength test data of the sample when the fitness between the steel plate to be processed and the sample is W1 for a sample representing a random type of weld defect, P1=b / a, and obtaining the reference accuracy set {P1, P2, ..., Pk}, where Pk represents the tensile strength test data of the sample representing a random type of weld defect. The sample of the type weld defect, when the fitness between the steel plate to be processed and the sample is Wk, the reference accuracy of the tensile strength test data of the sample is obtained, the fitness set between the n test samples representing all types of weld defects and the corresponding type of the to-be-calibrated sample is w = {w1, w2, ..., wn}, where the set w contains the set W, and the reference accuracy set of the tensile strength test data of the sample when the fitness between the steel plate to be processed and the sample is respectively the fitness in the set w is p = {p1, p2, ..., pn};

[0019] After analyzing the degree of fit between the remaining samples and the sample to be calibrated, the number of samples with the same degree of fit as the sample to be calibrated and the number of samples with the same degree of fit and the same tensile strength are counted from the remaining samples. The ratio between the two sample numbers is calculated to analyze the reference accuracy of the reference sample test data as the tensile strength data of the actual steel plate to be processed when the degree of fit between the actual steel plate to be processed and the sample is different. For example, when the degree of fit is W1, the higher the proportion of the number of samples with the same tensile strength as the sample to be calibrated, the more likely it is that the tensile strength of the actual steel plate to be processed with a degree of fit W1 with the sample is the tensile strength of the corresponding sample. At this time, the sample test data is used as the tensile strength of the actual steel plate to be processed. The more accurate the obtained strength data of the actual steel plate to be processed is, the higher the reference accuracy is.

[0020] Furthermore, step S4 includes: forming a training data set {(A1, w1, p1), (A2, w2, p2), ..., (An, wn, pn)} with the sample number, fitness and reference accuracy data, performing fitting training on the training data, and establishing a steel plate test selection model:

[0021] z=ζ0*x+ζ1*y+ζ2;

[0022] Among them, x represents the first variable representing the serial number in the steel plate test selection model, y represents the second variable representing the fitness degree in the model, z represents the dependent variable representing the reference accuracy in the model, and ζ0, ζ1, and ζ2 represent the fitting coefficients.

[0023] Furthermore, step S5 includes: obtaining the weld defect type of the currently to-be-processed steel plate, and obtaining that the weld defect type of the currently to-be-processed steel plate is the same as that of the test sample numbered A ’ , A ’ ∈ A, and there are e test samples numbered A ’ . Calculate the fitness degrees between the currently to-be-processed steel plate and the e test samples respectively. The calculation method of the fitness degrees between the currently to-be-processed steel plate and the e test samples is the same as the calculation method of Wi.

[0024] Furthermore, step S6 includes: comparing the fitness degrees between the currently to-be-processed steel plate and the e test samples, and obtaining that the highest fitness degree among the e test samples with the currently to-be-processed steel plate is Q max , input A ’ and Q max into the steel plate test selection model: let x = A ’ , y = Q max , and predict that the reference accuracy of using the tensile strength test data of the sample with the highest fitness degree among the e test samples with the currently to-be-processed steel plate as the reference data for the tensile strength of the currently to-be-processed steel plate is: ζ0 * A ’ + ζ1 * Q max + ζ2. Set the reference accuracy threshold as J, and compare ζ0 * A ’ + ζ1 * Q max + ζ2 and J: If ζ0 * A ’ + ζ1 * Q max + ζ2 ≥ J, plan the method for obtaining the tensile strength data of the currently to-be-processed steel plate as: using the tensile strength of the test sample with the fitness degree Q max with the currently to-be-processed steel plate as the tensile strength of the currently to-be-processed steel plate; if ζ0 * A ’ + ζ1 * Q max + ζ2 < J, plan the method for obtaining the tensile strength data of the currently to-be-processed steel plate as: placing the currently to-be-processed steel plate in a test environment for tensile testing, using the actual test result as the tensile strength of the currently to-be-processed steel plate, and at the same time supplementing the tensile strength test result data of the currently to-be-processed steel plate into the test database after obtaining the actual test result to enrich the test sample data;

[0025] A steel plate test selection model is established by fitting and training the number data, fitness data and reference accuracy data obtained from the previous analysis. The number represents the weld defect type of the steel plate, the fitness represents the difference between the welds of the steel plate, and the reference accuracy represents the accuracy of using the tensile strength test data of the specified sample as the tensile strength data of the actual steel plate to be processed when the actual steel plate has the same weld defect type and the fitness between the sample is the corresponding fitness. The fitness between the current steel plate to be processed and the sample with the same type of weld defect is analyzed, and the number corresponding to the weld defect type and the highest fitness data are selected and input into the steel plate test selection model. According to the output results, the method of obtaining the tensile strength data of the current steel plate to be processed is selected and planned: when the predicted reference accuracy is not lower than the threshold, it means that the tensile strength of the sample with the highest fitness is most likely equal to the tensile strength of the current steel plate to be processed. Therefore, the tensile strength test data of the sample corresponding to the highest fitness is selected as The tensile strength of the current steel plate to be processed reduces unnecessary testing work while ensuring the accuracy of the tensile strength data of the actual steel plate; when the predicted reference accuracy is lower than the threshold, the current steel plate to be processed is selected for strength testing to obtain the tensile strength. The reason for selecting the highest fitness input model is that the higher the fitness, the smaller the difference between the current steel plate to be processed and the sample. When the accuracy of the sample with the highest fitness as the reference data of the tensile strength data of the current steel plate to be processed is lower than the threshold, it means that the accuracy of the remaining sample data as a reference will only be lower. The method of inputting the fitness one by one is not adopted, which reduces unnecessary data input operations. The above method is selected to plan the method of obtaining the actual strength data of the steel plate to be processed, which improves the accuracy of the steel plate tensile strength data required in the actual processing process while reducing unnecessary strength testing work. At the same time, it gradually enriches the steel plate test sample data to a certain extent and improves the perfection of the test sample data.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] The present invention places a steel plate test sample with welding defects in a high temperature environment for tensile testing to test the tensile strength of the steel plate. The tensile strength of the steel plate with different types of welding defects obtained by the test can help reasonably adjust the working parameters according to the strength test results of the steel plate test sample with welding defects during actual processing, taking into account the influence of weld defects on the strength of the steel plate, which is conducive to avoiding the problem of affecting the quality and service life of the steel plate due to improper setting of working parameters;

[0028] Taking into account the influence of weld profile characteristics on steel plate strength data, samples to be calibrated are randomly extracted from the test samples collected in the early stage, and the remaining steel plate test samples with the same weld defect type as the samples to be calibrated are compared with the weld feature data of the samples to be calibrated, and the fitness between the two is analyzed. After the fitness between the remaining samples and the samples to be calibrated is obtained through analysis, the number of samples with the same fitness as the samples to be calibrated and the number of samples with equal fitness and equal tensile strength are counted from the remaining samples. By calculating the ratio between the two sample numbers, the test data of the reference sample is used as the tensile strength of the actual steel plate to be processed when the fitness between the actual steel plate to be processed and the sample is different. The reference accuracy of the strength data is determined by fitting the number data, fitness data and reference accuracy data obtained from the previous analysis to establish a steel plate test selection model, analyze the fitness between the current steel plate to be processed and the sample with the same type of weld defect, and select the number corresponding to the weld defect type and the highest fitness data to be input into the steel plate test selection model. Based on the output results, the method of obtaining the tensile strength data of the current steel plate to be processed is selected and planned. While improving the accuracy of the steel plate tensile strength data required in the actual processing process, unnecessary strength testing work is reduced. At the same time, the steel plate test sample data is gradually enriched to a certain extent, and the perfection of the test sample data is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The figure is a flow chart of the strength testing method of boiler steel plates with weld defects according to the present invention. DETAILED DESCRIPTION

[0030] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0031] Example 1:

[0032] like Figure 1As shown, this embodiment provides a strength testing method for boiler steel plates with weld defects, including: S1: collecting steel plate test samples, performing strength tests on the test samples, and obtaining tensile strength test results: butt-welding two steel plate specimens to form a weld with a width of d, preparing and collecting steel plate test samples with weld defects and different weld defect types, wherein the weld defect types include pores, slag inclusions, lack of fusion, incomplete penetration, and cracks, collecting a total of m steel plate test samples, each having a weld width of d, setting a test environment temperature range to T, performing a tensile test on the test samples under the corresponding test environment, and obtaining the tensile strength of the test samples obtained by the test;

[0033] S2: Number the test samples, obtain the weld feature data of the test samples, and analyze the fitness between the test samples with the same weld defect type: randomly extract samples with different weld defect types from the m test samples as samples to be calibrated: each weld defect type corresponds to a sample to be calibrated, and number the remaining steel plate test samples collected after the extraction process. The steel plate test samples with the same weld defect type have the same number, and the number set of the remaining n steel plate test samples is A = {A1, A2, ..., An}. There are the same numbers in the number set, that is, some of the n steel plate test samples have the same number. Samples representing a random type of weld defect are retrieved from the n steel plate test samples, and the number of retrieved samples is counted to k. The k retrieved samples are photographed for weld images. The weld contours of k samples are identified by using Canny edge detection technology. The weld images of the samples to be calibrated with the same weld defect type as the k samples are taken to identify the weld contours of the corresponding samples to be calibrated. The weld contour feature data are extracted using OpenCV technology. The feature data include the area of ​​the area surrounded by the weld contour, the boundary length of the weld contour, and the area of ​​the minimum circle that can surround the weld contour. The area set of the area surrounded by the weld contours of the k samples is {R1, R2, ..., Rk}, the boundary length set of the weld contour is {L1, L2, ..., Lk}, and the area set of the minimum circle that can surround the weld contour is {r1, r2, ..., rk}. The weld contour feature data of the samples to be calibrated with the same weld defect type as the k samples are R ’ 、L ’ and r ’ , analyze the degree of fit Wi between a random sample among the k samples and the sample to be calibrated:

[0034] Wi=1 / (|R ’ -Ri|+|L ’ -Li|+|r ’ -ri|);

[0035] The fitness set between the k samples and the sample to be calibrated is obtained as W = {W1, W2, ..., Wi, ..., Wk}, and there are equal fitness in the set W, where Ri represents the area of ​​the region enclosed by the weld contour of the i-th sample among the k samples, Li represents the boundary length of the weld contour of the i-th sample among the k samples, and ri represents the minimum circle area that can enclose the weld contour of the i-th sample;

[0036] S3: Count the total number of test samples with the same fitness and the number of test samples with the same fitness and the same tensile strength, and analyze the reference accuracy of the test data: count the number of samples with the fitness between the sample to be calibrated and the fitness W1 equal to a among the k samples, and the number of samples with the same tensile strength as the sample to be calibrated among a samples is b. Calculate the reference accuracy P1 of the tensile strength test data of the sample representing a random type of weld defect when the fitness between the steel plate to be processed and the sample is W1, P1 = b / a, and the reference accuracy set is {P1, P2 , ..., Pk}, Pk represents the reference accuracy of the tensile strength test data of the sample when the degree of fit between the steel plate to be processed and the sample is Wk for a sample representing a random type of weld defect, the set of fits between the n test samples representing all types of weld defects and the corresponding types of samples to be calibrated is w = {w1, w2, ..., wn}, where the set w contains the set W, and the set of reference accuracy of the tensile strength test data of the sample when the degrees of fit between the steel plate to be processed and the sample are the degrees of fit in the set w is p = {p1, p2, ..., pn};

[0037] S4: Fitting training is performed on the sample number, fitness and reference accuracy data to establish a steel plate test selection model: the sample number, fitness and reference accuracy data are combined into a training data set {(A1, w1, p1), (A2, w2, p2), ..., (An, wn, pn)}, and fitting training is performed on the training data to establish a steel plate test selection model:

[0038] z=ζ0*x+ζ1*y+ζ2;

[0039] Where * represents a multiplication sign, x represents the first variable representing the number in the steel plate test selection model, y represents the second variable representing the fitness in the model, z represents the dependent variable representing the reference accuracy in the model, ζ0, ζ1, and ζ2 represent the fitting coefficients;

[0040] S5: Obtain the weld seam feature data of the currently to-be-processed steel plate, confirm the weld defect type and number of the currently to-be-processed steel plate, compare the weld seam features with the test samples of the same type, and analyze the fitness between the currently to-be-processed steel plate and the test samples: Obtain the weld defect type of the currently to-be-processed steel plate, and obtain that the weld defect type of the currently to-be-processed steel plate is the same as that of the test sample with the number A ’ ∈A, and there are e test samples numbered A ’ . Calculate the fitness between the currently to-be-processed steel plate and the e test samples respectively. The calculation method of the fitness between the currently to-be-processed steel plate and the e test samples is the same as that of Wi;

[0041] S6: Input the number and fitness of the currently to-be-processed steel plate into the steel plate test selection model to plan the method for obtaining the tensile strength data of the currently to-be-processed steel plate: Compare the fitness between the currently to-be-processed steel plate and the e test samples, and obtain that the highest fitness between the currently to-be-processed steel plate and the e test samples is Q max . Input A ’ and Q max into the steel plate test selection model, and predict that the reference accuracy of using the tensile strength test data of the sample with the highest fitness with the currently to-be-processed steel plate among the e test samples as the reference data for the tensile strength of the currently to-be-processed steel plate is: ζ0*A ’ +ζ1*Q max +ζ2. Set the reference accuracy threshold as J, and compare ζ0*A ’ +ζ1*Q max +ζ2 and J: If ζ0*A ’ +ζ1*Q max +ζ2≥J, plan the method for obtaining the tensile strength data of the currently to-be-processed steel plate as: Use the tensile strength of the test sample with the fitness Q max with the currently to-be-processed steel plate as the tensile strength of the currently to-be-processed steel plate; If ζ0*A ’ +ζ1*Q max [[ID=(33)]]+ζ2<J, plan the method for obtaining the tensile strength data of the currently to-be-processed steel plate as: Place the currently to-be-processed steel plate in the test environment for tensile testing, and use the actual test result as the tensile strength of the currently to-be-processed steel plate;

[0042] For example: Establish a steel plate test selection model: z = 0.15x + 1.89y - 0.15, and obtain that the weld defect type of the currently to-be-processed steel plate is the same as that of the test sample with the number A ’ = 1, and there are 5 test samples numbered 1. Obtain that the highest fitness between the 5 test samples and the currently to-be-processed steel plate is Q max ​​=0.05, let x=1, y=0.45, and predict that the reference accuracy of using the tensile strength test data of the sample with the highest adaptability to the current steel plate to be processed among the five test samples as the reference data of the tensile strength of the current steel plate to be processed is 0.85, and the reference accuracy threshold is set to J=0.80, 0.85>0.80, ζ0*A ’ +ζ1*Q max +ζ2>J, the method of planning to obtain the tensile strength data of the current steel plate to be processed is: the adaptation degree between the current steel plate to be processed is Q max The tensile strength of the test sample is taken as the tensile strength of the steel plate to be processed.

[0043] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for testing the strength of boiler steel plates with weld defects, characterized by: The following steps are involved: S1: Collect steel plate test samples, perform strength tests on the test samples, and obtain tensile strength test results; S2: Number the test samples, obtain the weld characteristic data of the test samples, and analyze the compatibility between the test samples with the same weld defect type; S3: Count the total number of test samples with the same fitness and the number of test samples with the same fitness and the same tensile strength obtained in the test, and analyze the reference accuracy of the test data; S4: Fitting training is performed on the sample number, fitness and reference accuracy data to establish a steel plate test selection model; S5: Acquire the weld feature data of the current steel plate to be processed, confirm the weld defect type and number of the current steel plate to be processed, compare the weld features with the test sample of the same type, and analyze the compatibility between the current steel plate to be processed and the test sample; S6: Inputting the serial number and suitability of the current steel plate to be processed into the steel plate test selection model, and planning a method for obtaining the tensile strength data of the current steel plate to be processed; Step S4 includes: forming a training data set with sample numbers, fitness and reference accuracy data, performing fitting training on the training data, and establishing a steel plate test selection model: z=ζ0*x+ζ1*y+ζ2; Where x represents the first variable representing the number in the steel plate test selection model, y represents the second variable representing the fitness in the model, z represents the dependent variable representing the reference accuracy in the model, ζ0, ζ1, and ζ2 represent the fitting coefficients; Step S5 includes: obtaining the weld defect type of the current steel plate to be processed, obtaining the weld defect type of the current steel plate to be processed and the weld defect type numbered A ’ The weld defect types of the test samples are the same, A ’ ∈A, there are e numbers numbered A ’ The test samples are calculated respectively to calculate the degree of fit between the current steel plate to be processed and the e test samples; Step S6 includes: comparing the degree of fit between the current steel plate to be processed and the e test samples, and obtaining the highest degree of fit between the e test samples and the current steel plate to be processed as Q max , A ’ and Q max Input to the steel plate test selection model: Let x = A ’ ,y=Q max , the predicted reference accuracy of taking the tensile strength test data of the sample with the highest adaptability to the current steel plate to be processed among the e test samples as the reference data of the tensile strength of the current steel plate to be processed is: ζ0*A ’ +ζ1*Q max +ζ2, set the reference accuracy threshold to J, compare ζ0*A ’ +ζ1*Q max +ζ2 and J, get the comparison result; If ζ0 * A ’ + ζ1 * Q max + ζ2 ≥ J, the method for planning to obtain the tensile strength data of the currently to-be-processed steel plate is: taking the tensile strength of the test sample with a fitness of Q max to the currently to-be-processed steel plate as the tensile strength of the currently to-be-processed steel plate; if ζ0 * A ’ + ζ1 * Q max + ζ2 < J, the method for planning to obtain the tensile strength data of the currently to-be-processed steel plate is: placing the currently to-be-processed steel plate in a test environment for tensile testing and taking the actual test result as the tensile strength of the currently to-be-processed steel plate.

2. The method for testing the strength of boiler steel plates with weld defects according to claim 1, wherein: Step S1 includes: butt welding two steel plate specimens to form a weld with a width of d, preparing and collecting steel plate test samples with weld defects and different weld defect types, collecting a total of m steel plate test samples, setting the test environment temperature range to T, performing a tensile test on the test samples under the corresponding test environment, and obtaining the tensile strength of the test samples obtained by the test.

3. The strength testing method for boiler steel plates with weld defects according to claim 2, characterized in that: Step S2 includes: randomly extracting samples with different weld defect types from m test samples as samples to be calibrated: each weld defect type corresponds to a sample to be calibrated, numbering the remaining steel plate test samples collected after the extraction process, and the steel plate test samples with the same weld defect type have the same number, and the number set of the remaining n steel plate test samples is obtained as A = {A1, A2, ..., An}, retrieving samples representing a random type of weld defect from the n steel plate test samples, counting the number of retrieved samples to be k, taking weld images of the k retrieved samples, using Canny edge detection technology to identify the weld contours of the k samples, taking weld images of the samples to be calibrated with the same weld defect type as the k samples, identifying the weld contours of the corresponding samples to be calibrated, and extracting weld contour feature data using OpenCV technology.

4. The strength testing method for boiler steel plates with weld defects according to claim 3, characterized in that: The characteristic data include the area of ​​the area surrounded by the weld contour, the boundary length of the weld contour and the area of ​​the smallest circle that can surround the weld contour. The area set of the area surrounded by the weld contour of k samples is {R1, R2, ..., Rk}, the boundary length set of the weld contour is {L1, L2, ..., Lk}, and the area set of the smallest circle that can surround the weld contour is {r1, r2, ..., rk}. The weld contour characteristic data of the samples to be calibrated with the same weld defect type as the k samples are R ’ , L ’ and r ’ , analyze the degree of fit Wi between a random sample among the k samples and the sample to be calibrated: Wi=1 / (|R ’ -Ri|+|L ’ -Li|+|r ’ -ri|); The obtained fitness set between the k samples and the sample to be calibrated is W = {W1, W2, ..., Wi, ..., Wk}, where Ri represents the area of ​​the region enclosed by the weld contour of the i-th sample among the k samples, Li represents the boundary length of the weld contour of the i-th sample among the k samples, and ri represents the minimum circle area that can enclose the weld contour of the i-th sample.

5. The strength testing method for boiler steel plates with weld defects according to claim 4, characterized in that: Step S3 includes: counting the number of samples with the same fitness W1 as the sample to be calibrated among the k samples as a, the number of samples with the same tensile strength as the sample to be calibrated among the a samples as b, calculating the reference accuracy P1 of the tensile strength test data of the sample when the fitness between the steel plate to be processed and the sample is W1 for a sample representing a random type of weld defect, P1=b / a, and obtaining the reference accuracy set {P1, P2, ..., Pk}, Pk represents the tensile strength test data of the sample representing a random type of weld defect The defect sample, when the fitness between the steel plate to be processed and the sample is Wk, the reference accuracy of the tensile strength test data of the sample is obtained, the fitness set between the n test samples representing all types of weld defects and the corresponding types of samples to be calibrated is w = {w1, w2, ..., wn}, where the set w contains the set W, and the reference accuracy set of the tensile strength test data of the sample when the fitness between the steel plate to be processed and the sample is respectively the fitness in the set w is p = {p1, p2, ..., pn}.

6. The strength testing method for boiler steel plates with weld defects according to claim 5, characterized in that: The training data set consists of {(A1, w1, p1), (A2, w2, p2), ..., (An, wn, pn)}.

Images