Method for Evaluating the Full-Position Forming Quality of the Outer-Ring Weld in Pipeline Welding

Abstract

The present invention relates to a method for evaluating the all-position forming quality of the outer circumferential weld of a welding consumable in pipeline welding. The method includes: calibrating a number of reference lines around the pipe joint. After completing the welding of one pass of the outer circumferential weld of the pipeline, measuring the weld reinforcement corresponding to each reference line, and constructing a training set and a test set with several groups of weld reinforcements and the qualified rate of the next pass of welding seam flaw detection. Based on the BP neural network, training and optimizing the weld reinforcement prediction model of the number of reference lines and the h-M relationship, inputting the minimum value of the weld qualified rate into the weld reinforcement prediction model to calculate the corresponding weld reinforcement as a criterion, comparing the weld reinforcement corresponding to each reference line before the next pass of welding under the same welding conditions, and determining the weld forming quality of the welding consumable at the full position of the outer circumference of the pipeline. Without cutting a large number of specimens to measure the weld width and penetration depth, it can be associated with the defects of the next pass of weld seam, complete the evaluation online, significantly reduce the inspection cost, labor intensity and cycle, and is beneficial to controlling and improving the welding quality.

Description

Technical Field

[0001] The present invention belongs to the technical field of evaluation of the forming quality of welding materials, and particularly relates to a method for evaluating the all-position forming quality of the outer circumferential weld of a welding material in pipeline welding. Background Art

[0002] At present, in the welding of the outer circumferential weld of oil and gas pipelines, the oil and gas steel pipeline is fixed, and the 100 pipe orifices of the pipeline are processed into relatively narrow mixed Y-shaped grooves 101 and butt-jointed. During the welding process, root welding is first carried out inside the pipeline to form an inner circumferential weld, as Figure 1 shown. The fully automatic welding special machine uses welding materials, and then root welding, successive pass filling and capping are carried out on the outer circumference of the butt-jointed pipeline to form an outer circumferential weld, and the weld bead distribution of its outer circumferential weld 300 is as Figure 2 shown. During the welding process, as the welding trolley 200 moves, the welding position changes gradually, from Figure 1 the flat welding at the 0 o'clock position to the vertical welding at the 3 o'clock position, and then to the overhead welding at the 6 o'clock position. Therefore, the welding parameters need to be adjusted in real time according to the welding position. If the weld forming is not good, it will affect the melting effect of the next weld on the current weld, and there is a high probability of causing defects such as lack of fusion. Therefore, this welding method has extremely high requirements for the all-position welding forming quality of welding materials. Under the same welding method, equipment and parameters, the weld forming quality of different welding materials is different. Therefore, it is necessary to evaluate the weld forming quality of welding materials in the all-position welding of the outer circumferential weld of pipelines to guide the selection and evaluation of welding materials.

[0003] Existing methods for evaluating weld forming quality include: after cutting the weld, grinding, polishing, and corroding it, measuring the weld reinforcement h, weld width B, and penetration depth H under a stereoscopic microscope at an appropriate magnification, and calculating the weld forming coefficient By examining the weld reinforcement and the size of the weld forming coefficient of a sufficient number of welds, the weld forming quality is judged. The following problems exist:

[0004] (1) For the multi-layer welding method, after each layer of the weld is welded, a large number of specimens need to be cut from the pipe joints at different positions, ground, polished, corroded, and then measured. It is impossible to measure the weld characteristic parameters online quickly, and there are deficiencies such as high cost, many measurement specimens, and long cycle, which are not conducive to on-site inspection and timely treatment in production.

[0005] (2) For the outer circumferential weld of the pipeline, affected by the assembly and processing accuracy of the groove width dimension, the width is different, which in turn affects the actual value of the weld width, making the forming coefficient calculated by measuring the weld width and penetration depth not have reference significance, and the measurement and calculation of redundant parameters affect the evaluation efficiency.

[0006] (3) It is not associated with the defects of the unfused weld in the next pass. Through empirical evaluation, the reliability and stability of measurement evaluation cannot be guaranteed, which is not conducive to evaluating the melting effect of the next-pass weld on the current-pass weld. Furthermore, it is difficult to control the forming quality of the next-pass weld, affecting the production rhythm. Summary of the Invention

[0007] The present invention aims to solve at least one of the above technical problems to some extent. The present invention provides a method for evaluating the all-position forming quality of the outer-ring weld in pipeline welding with welding materials, which can be associated with the defects of the next-pass weld, complete the evaluation online, significantly reduce the inspection cost, labor intensity and cycle, and is conducive to controlling and improving the welding quality.

[0008] The technical solution adopted by the present invention to solve its technical problems is as follows:

[0009] A method for evaluating the all-position forming quality of the outer-ring weld in pipeline welding with welding materials, the method comprising:

[0010] Calibrate a plurality of reference lines perpendicular to the edge of the pipeline welding groove at equal intervals around half of the pipe joint;

[0011] After welding a pass of the outer-ring weld on the welded pipeline with welding materials, measure the weld reinforcement h corresponding to each reference line i ;

[0012] After welding the next pass on the welded pipeline with welding materials according to the welding parameters, detect the weld flaw detection qualification rate M corresponding to each reference line i ;

[0013] With a plurality of groups of weld reinforcements h i and the corresponding weld flaw detection qualification rates M i Construct a training set and a test set, and optimize the weld reinforcement prediction model of the number of reference lines and the h-M relationship based on BP neural network training;

[0014] Input the minimum value M of the weld qualification rate min into the weld reinforcement prediction model to calculate the corresponding weld reinforcement h min ;

[0015] Before welding the next pass on the welded pipeline with welding materials under the same welding conditions as those for training the weld reinforcement prediction model, measure the weld reinforcement h corresponding to each reference line i ”;

[0016] Based on h i ”≤h min Judge that the weld forming quality of the welding materials at the all-positions of the outer ring of the pipeline is qualified.

[0017] In the above method, further, benchmark points are calibrated on each reference line, and the depth measurement tool uses the benchmark points as the measurement zero points, which can ensure that the measurement positions of each layer of welds are consistent. The heights H1 and H2 from the outer wall of the pipeline to the two side edges of the weld are measured, and the height H3 from the outer wall of the pipeline to the middle top of the weld is measured. The weld reinforcement h i =(H1 + H2) / 2 - H3. Considering the misalignment of the weld, the measurement of the weld reinforcement is made more reasonable and accurate.

[0018] In the above method, further, when detecting the weld flaw detection qualification rate corresponding to each reference line, flaw detection is carried out at the weld position corresponding to the reference line, and the weld flaw detection qualification rate is obtained by detecting the position and size of the defects.

[0019] In the above method, further, when constructing the training set and the test set, outliers are removed according to the 3σ criterion of the normal distribution, which can reduce the influence of outliers caused by detection errors on the accuracy of the weld reinforcement prediction model.

[0020] In the above method, further, taking the weld flaw detection qualification rate M i of the training set as the input layer of the BP neural network, and taking the weld reinforcement h i of the training set as the output layer of the BP neural network, using the ReLU function as the activation function to map the input layer to the output layer, backward propagation can be carried out, and the convergence speed is relatively fast. The hidden layer of the BP neural network is solved to construct a weld reinforcement prediction model of the h-M relationship.

[0021] In the above method, further, the weld reinforcement prediction model uses the gradient descent method to correct the connection weights between the hidden layer and the output layer and the connection weights between the input layer and the hidden layer, which can improve the convergence speed.

[0022] In the above method, further, the weld flaw detection qualification rate M i of the test set is input into the weld reinforcement prediction model to calculate the predicted weld reinforcement h i ' of the output layer. The mean square error between the predicted weld reinforcement h i ' and the weld reinforcement h i of the test set is calculated. Taking the minimization of the mean square error as the goal, the number of reference lines is adjusted and the weld reinforcement prediction model is iteratively optimized, which can continuously reduce the variation of the target value through process improvement, obtain a better number of reference lines and a weld reinforcement prediction model of the h-M relationship, and then obtain a better weld reinforcement h min as the evaluation criterion.

[0023] In the above method, further, the number of the reference lines is 33, and through practical verification, it can have better evaluation efficiency and effect on the basis of fewer detections.

[0024] In the above method, further, determine the minimum value M of the weld pass rate according to the technical requirements of the pipeline project min , so that the evaluation method can meet the technical requirements of the pipeline project, is easy to implement, and can effectively evaluate.

[0025] In the above method, further, h i ” > h min , it indicates that it will have an adverse effect on the full penetration of the next weld, and it is easy to cause incomplete fusion defects. Accumulate h i ” > h min The number of weld bead reinforcements, and evaluate the weld formation quality of the welding consumables at all positions on the outer ring of the pipeline. h i ” > h min The larger the number of weld bead reinforcements, the worse the weld formation quality of the welding consumables at all positions on the outer ring of the pipeline.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0027] (1) Measure the height from the outer wall of the pipeline to the two side parts and the middle top of the weld on both sides respectively to calculate the weld bead reinforcement, without cutting a large number of specimens to measure the weld width and weld depth, greatly reducing the waste of materials, reducing the production cost, making the inspection cycle shorter, and the evaluation efficiency higher.

[0028] (2) Use several groups of weld bead reinforcements h i and the corresponding weld flaw detection pass rate M i to construct a training set and a test set. Based on the BP neural network, train and optimize the weld bead reinforcement prediction model of the number of baseline and h-M relationship, which can associate the weld bead reinforcement with the defect formation of the next incomplete fusion weld, improve the reliability and stability, and be used to evaluate the influence of the next weld on the melting effect of the current weld.

[0029] (3) Input the minimum value M of the weld pass rate min into the weld bead reinforcement prediction model to solve the corresponding weld bead reinforcement h min as a criterion. By comparing h i ” with h min , the number of unqualified weld formations of a welding consumable in the whole weld can be measured, and the evaluation result of the weld formation quality of the welding consumable at all positions on the outer ring of the pipeline can be obtained quickly and effectively.

[0030] In summary, the present invention can complete the evaluation online, does not affect the welding and production of the next weld, has a low inspection cost and a low labor intensity, can be directly applied to on-site production inspection, discover problems and deal with them in time, so as to control and improve the welding quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The above and / or additional aspects and advantages of the present invention will become obvious and easy to understand from the description of the embodiments in conjunction with the following drawings, in which:

[0032] Figure 1 It is a schematic diagram of the automatic welding structure for the outer ring weld of the pipeline;

[0033] Figure 2 It is a schematic diagram of the circumferential weld bead;

[0034] Figure 3 It is a top view schematic diagram for measuring the weld reinforcement in an embodiment of the present invention;

[0035] Figure 4 It is a cross-sectional view schematic diagram for measuring the weld reinforcement when the weld is convex in an embodiment of the present invention;

[0036] Figure 5 It is a cross-sectional view schematic diagram for measuring the weld reinforcement when the weld is concave in an embodiment of the present invention;

[0037] Figure 6 It is a flowchart of the method of the present invention.

[0038] Markings in the figure: pipeline 100, Y-shaped groove 101, welding carriage 200, weld 300, support 401, measuring ruler 402, probe 403, reference line 500, reference point 501. Detailed implementation manners

[0039] The embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals indicate the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention, but should not be construed as limiting the present invention.

[0040] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "axial", "radial", "vertical", "horizontal", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.

[0041] As Figure 3 - 6 shown, it is a preferred embodiment of the method for evaluating the full-position forming quality of the outer ring weld of pipeline welding using the welding materials of the present invention, and the method includes the following steps:

[0042] S1: As Figure 3 - 5 shown, take several groups of pipelines to be welded. Since as Figure 1The two welding trolleys shown weld two half - circumferences of each pipeline respectively. Therefore, the welding quality of the two half - circumferences of the welds is consistent. Mark n reference lines 500 perpendicular to the edge of the welding groove of the pipeline at equal intervals around the half - circumference of the pipe joint, and mark reference points 501 on each reference line 500;

[0043] S2: After the welding consumables complete one - pass welding of the outer - ring weld on each group of welded pipelines according to the welding parameters, the depth - measuring tool uses the respective reference points 501 in step S1 as the measurement zero points, measures the heights H1 and H2 from the outer wall of the pipeline to the two side edges of the weld respectively, measures the height H3 from the outer wall of the pipeline to the top middle of the weld, and calculates the weld reinforcement h corresponding to the i - th reference line 500 among the n reference lines 500 i =(H1 + H2) / 2 - H3;

[0044] The measuring tool includes a bracket 401 and a measuring scale 402. The bracket 401 is erected above the reference line 500. The measuring scale 402 can move up, down, left, and right along the bracket 401. A probe 403 is provided at the bottom of the measuring scale 402. During measurement, the probe 403 contacts the reference point 501 for zeroing. After moving the measuring scale 402 horizontally along the bracket 401 to drive the probe 403 to move, the probe 403 descends to contact the weld, and the corresponding value of the measuring scale 402 is used as the measured value of the height from the outer wall of the pipeline to the weld. Similarly, move the measuring scale 402 in the reverse direction for resetting in order to perform re - measurement;

[0045] S3: The welding consumables perform the next - pass welding on the welded pipelines in step S2 according to the welding parameters, perform flaw detection on the weld positions corresponding to the reference lines 500, record the defect conditions of incomplete fusion at the edges, and obtain the weld flaw - detection qualification rate M corresponding to the i - th reference line 500 i ;

[0046] S4: Optimize the weld reinforcement prediction model of the relationship between the number n of reference lines 500 and h - M based on BP neural network training. Specifically:

[0047] S401: Use several h i and the corresponding M i in steps S2 and S3 to construct a training set and a test set, and eliminate outliers outside the 3σ data interval according to the 3σ criterion of the normal distribution;

[0048] S402: Use the weld flaw - detection qualification rate M i of the training set as the input layer of the BP neural network, use the weld reinforcement h i of the training set as the output layer of the BP neural network, use the ReLU function as the activation function to map the input layer to the output layer, solve the hidden layer of the BP neural network, and construct the weld reinforcement prediction model of the h - M relationship;

[0049] S403: The weld reinforcement prediction model in step S402 uses the gradient descent method to correct the connection weights between the hidden layer and the output layer, and the connection weights between the input layer and the hidden layer;

[0050] S404: Input the N weld flaw detection qualification rates M of the test set i into the weld reinforcement prediction model in step S403 to calculate the predicted weld reinforcement h of the output layer i ', and calculate the mean square error between the predicted weld reinforcement h i ' and the weld reinforcement h of the test set i ; Adjust the number n of reference lines 500 and iteratively optimize the weld reinforcement prediction model with the goal of minimizing the mean square error, save the parameters, and obtain the weld reinforcement prediction model based on the BP neural network training and optimization of the h-M relationship;

[0051] S5: Determine the minimum value M of the weld qualification rate according to the technical requirements of the pipeline project min , and input M min into the weld reinforcement prediction model in step S4 to solve the corresponding weld reinforcement h min ;

[0052] S6: Using welding consumables, under the same welding conditions as in step S2, that is, the same welding equipment, welding method, and welding parameters, after completing one pass of the outer ring weld of the pipeline and before proceeding to the next weld, measure the weld reinforcement h corresponding to each reference line 500 according to the methods in steps S1 and S2 i ”;

[0053] S7: If the next weld of the pipeline in step S6 is carried out under the same welding conditions as in step S3, that is, the same welding equipment, welding method, and welding parameters, then compare the weld reinforcement h in step 5 min with the weld reinforcement h in step S6 i ”:

[0054] If h i ” > h min , it is determined that the weld forming quality corresponding to the reference line 500 is unqualified, which will have an adverse impact on the complete penetration of the next weld and is likely to cause lack of fusion defects. The cumulative unqualified quantity is counted, and the more unqualified quantities, the worse the weld forming quality of the welding consumables at the full position of the outer ring of the pipeline.

[0055] If h i ” ≤ h min , it is determined that the weld forming quality corresponding to the reference line 500 is qualified; if all h i ” ≤ h min , it is determined that the weld forming quality of the welding consumables at the full position of the outer ring of the pipeline is qualified.

[0056] Example 1:

[0057] Use a solid cored shielding wire with a diameter of 1.0 mm as the shielding gas, and use a fully automatic pipe girth welding machine to weld a steel pipe with an outer diameter of 20 * 1219 mm. The torch position, motion parameters, and welding parameters for the next welding are shown in Table 1 below:

[0058] Table 1

[0059]

[0060] Solve for h according to the above steps S1 - S5 min to be 0.3 mm. The number of reference lines is 33. Number the measurement positions corresponding to each reference line starting from 1, and measure the weld reinforcement h corresponding to each reference line according to step S6 above i ”, and the results are shown in Table 2 below:

[0061] Table 2

[0062] Measurement position serial number 1 2 3 4 5 6 7 Weld reinforcement / mm -0.480 -0.540 -0.400 -0.370 -0.305 -0.340 -0.340 Measurement position serial number 8 9 10 11 12 13 14 Weld reinforcement / mm -0.425 -0.565 -0.515 -0.545 -0.575 -0.510 -0.765 Measurement position serial number 15 16 17 18 19 20 21 Weld reinforcement / mm -0.545 -0.720 -0.635 -0.545 -0.605 -0.500 -0.485 Measurement position serial number 22 23 24 25 26 27 28 Weld reinforcement / mm -0.445 -0.370 -0.365 -0.380 -0.315 -0.195 -0.130 Measurement position serial number 29 30 31 32 33 Weld reinforcement / mm -0.075 -0.130 0.040 0.255 0.275

[0063] The negative values in Table 2 indicate that the weld is concave. The maximum value of the weld reinforcement is 0.275 mm, all less than 0.3 mm, and it is determined that the forming quality of the starting weld is good. Perform the next pass of welding according to Table 1. After the final weld is inspected by X-ray flaw detection, no lack of fusion defects are found. The weld forming performance of this wire is good in the all-position welding of the outer girth weld of the pipe, and the verification shows that the evaluation method is effective.

[0064] Example 2:

[0065] Use a solid cored shielding wire with a diameter of 1.0 mm as the shielding gas, and use a fully automatic pipe girth welding machine to weld a steel pipe with an outer diameter of 20 * 1219 mm. The torch position, motion parameters, and welding parameters for the next welding are shown in Table 3 below:

[0066] Table 3

[0067]

[0068] Solve for h according to the above steps S1 - S5 min to be 0.3 mm. The number of reference lines is 33. Number the measurement positions corresponding to each reference line starting from 1, and measure the weld reinforcement h corresponding to each reference line according to step S6 above i ”, and the results are shown in Table 4 below:

[0069] Table 4

[0070] Measurement position serial number 1 2 3 4 5 6 7 Weld reinforcement / mm -0.190 -0.350 -0.310 -0.505 -0.455 -0.590 -0.190 Measurement position serial number 8 9 10 11 12 13 14 Weld reinforcement / mm -0.17 -0.585 -0.545 -0.620 -0.500 -0.435 -0.785 Measurement position serial number 15 16 17 18 19 20 21 Weld reinforcement / mm -0.525 -0.695 -0.650 -0.515 -0.575 -0.605 -0.640 Measurement position serial number 22 23 24 25 26 27 28 Weld reinforcement / mm -0.705 -0.780 -0.415 -0.545 -0.420 -0.010 -0.320 Measurement position serial number 29 30 31 32 33 Weld reinforcement / mm 0.175 0.300 0.260 0.375 0.465

[0071] The negative values in Table 4 indicate that the weld is concave. The weld reinforcement has been greater than 0.3 mm since Point 32, and it is determined that the forming quality of the starting weld is unqualified. The next pass of welding is carried out according to Table 3. After the final X-ray inspection of the weld, 3 unfused defects are found near Points 32 and 33. It is determined that the weld forming performance of this welding wire is poor in the all-position welding of the outer ring weld of the pipeline. The verification shows that the evaluation method is effective.

[0072] Example 3:

[0073] A solid cored shielding wire with a diameter of 1.0 mm is used, and a fully automatic pipe girth welder is used to weld a steel pipe with an outer diameter of 20 * 1219 mm. The torch position, motion parameters, and welding parameters for the next pass of welding are shown in Table 5 below:

[0074] Table 5

[0075]

[0076] Calculating h according to the above steps S1 - S5 min is 0.5 mm, the number of reference lines is 33, and the measuring positions corresponding to each reference line are numbered starting from 1. Measuring the weld reinforcement h corresponding to each reference line according to step S6 above i ”, and the results are shown in Table 6 below:

[0077] Table 6

[0078] Measurement position serial number 1 2 3 4 5 6 7 Weld reinforcement / mm -0.045 -0.09 0.035 -0.085 -0.31 -0.24 -0.13 Measurement position serial number 8 9 10 11 12 13 14 Weld reinforcement / mm -0.3 -0.42 -0.47 -0.59 -0.885 -0.825 -0.99 Measurement position serial number 15 16 17 18 19 20 21 Weld reinforcement / mm -0.785 -0.81 -0.575 -0.92 -0.805 -0.88 -0.595 Measurement position serial number 22 23 24 25 26 27 28 Weld reinforcement / mm -0.28 -0.415 -0.165 -0.28 0.245 0.245 0.4 Measurement position serial number 29 30 31 32 33 Weld reinforcement / mm 0.39 0.51 0.465 0.735 0.695

[0079] The negative values in Table 6 indicate that the weld is concave. The weld reinforcement is greater than 0.5 mm at Points 30, 32, and 33. It is determined that the forming quality of the starting weld is unqualified. The next pass of welding is carried out according to Table 5. After the final X-ray inspection of the weld, 3 unfused defects are found near Points 30, 32, and 33. It is determined that the weld forming performance of this welding wire is poor in the all-position welding of the outer ring weld of the pipeline. The verification shows that the evaluation method is effective.

[0080] The above evaluation method: Measuring the height from the outer wall of the pipeline to the two side edges and the middle top of the weld to calculate the weld reinforcement, without the need to cut a large number of specimens to measure the weld width and weld depth. The number of inspection samples can be reduced according to the actual situation, greatly reducing the waste of materials, lowering the production cost, making the inspection cycle shorter, and having higher evaluation efficiency; Using several groups of weld reinforcement h i and the corresponding weld flaw detection qualification rate M i to construct a training set and a test set, and training and optimizing the weld reinforcement prediction model of the number of reference lines and the h - M relationship based on the BP neural network. The minimum value M min of the weld qualification rate is input into the weld reinforcement prediction model to calculate the corresponding weld reinforcement h min as a criterion, which can associate the weld reinforcement with the formation of defects in the next unfused weld. By comparing hi ” With h min It is possible to measure the number of unqualified points in the weld formation of a welding consumable in the entire weld seam, quickly and effectively obtain the evaluation result of the weld formation quality of the welding consumable at all positions on the outer ring of the pipeline, and the evaluation can be completed online without affecting the welding of the next weld seam or production. The inspection cost is low and the labor intensity is low. It can be directly applied to on-site production inspection to discover problems and handle them in a timely manner, thereby controlling and improving the welding quality.

[0081] The series of detailed descriptions listed above are only specific descriptions of the feasible embodiments of the present invention, and they are not intended to limit the protection scope of the present invention. Any equivalent embodiments or changes made without departing from the technical spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. Method for evaluating all-position forming quality of outer circumferential weld of pipeline welding with welding consumables, characterized in that, The method includes: Calibrating a number of reference lines (500) perpendicular to the edge of the pipe welding groove at equal intervals around half of the pipe joint; After the welding consumables complete a circumferential weld of the welded pipe, measure the weld reinforcement h corresponding to each reference line (500). i Calibrate reference points (501) on each reference line (500). Using the reference points (501) as the measurement zero points, measure the heights H1 and H2 from the outer wall of the pipe to the two side edges of the weld respectively, and measure the height H3 from the outer wall of the pipe to the top middle of the weld. The weld reinforcement h i =(H1 + H2) / 2 - H3; After the welding consumables are used to perform the next welding on the welded pipe according to the welding parameters, the qualified rate M of the weld flaw detection corresponding to each reference line (500) is detected i ; With several groups of weld reinforcement heights h i and the corresponding qualified rate of weld flaw detection M i Construct a training set and a test set, and train and optimize the weld reinforcement height prediction model for the number of baseline (500) and the h-M relationship based on the BP neural network; Set the minimum value M of the weld pass rate min Input the weld reinforcement h corresponding to the solution of the weld reinforcement prediction model min ; Before the next welding of the welding consumables on the welded pipeline according to the welding parameters, measure the weld reinforcement h corresponding to each reference line (500). i ”; According to h i ”≤h min It is determined that the weld forming quality of the welding consumables at the full position of the outer ring of the pipeline is qualified.

2. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to claim 1, characterized in that, When detecting the qualified rate of weld flaw detection corresponding to each reference line (500), flaw detection is performed at the weld position corresponding to the reference line (500).

3. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to claim 1, wherein When constructing the training set and the test set, outliers are removed according to the 3σ criterion of the normal distribution.

4. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding material according to claim 1, characterized in that Taking the qualified rate M of weld flaw detection in the training set i as the input layer of the BP neural network, and taking the weld reinforcement h of the training set i as the output layer of the BP neural network, using the ReLU function as the activation function to map the input layer to the output layer, solving the hidden layer of the BP neural network, and constructing a weld reinforcement prediction model of the h-M relationship.

5. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to claim 4, wherein The weld reinforcement prediction model uses the gradient descent method to correct the connection weights between the hidden layer and the output layer, and the connection weights between the input layer and the hidden layer.

6. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to claim 5, characterized in that The qualified rate M of weld flaw detection in the test set i Input the predicted weld reinforcement h of the weld reinforcement prediction model in the calculation output layer i ', calculate the predicted weld reinforcement h i ' and the weld reinforcement h of the test set i The mean square error of is calculated, and the number of baseline lines (500) is adjusted with the goal of minimizing the mean square error, and the weld reinforcement prediction model is iteratively optimized.

7. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to claim 6, characterized in that The number of the reference lines (500) is 33.

8. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to claim 1, characterized in that, Determine the minimum value M of the weld qualification rate according to the technical requirements of the pipeline project min .

9. The method for evaluating the full-position forming quality of the outer circumferential weld in pipeline welding of the welding consumables according to any one of claims 1 to 8, characterized in that, Cumulative h i ” > h min quantity of weld reinforcement, and evaluate the weld forming quality of the welding consumables in the full position of the outer ring of the pipeline.

Images