Method and system for measuring weld penetration in lap welding based on grey prediction

Abstract

The present invention relates to a method and system for measuring weld penetration in lap welding based on gray prediction, and relates to the field of ultrasonic testing technology. The method comprises: obtaining an ultrasonic imaging width sequence of a test piece to be tested; establishing a gray prediction model based on sample data; the sample data comprises ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by ultrasonic phased array testing of multiple sample specimens; the ultrasonic imaging width sequence of the test piece to be tested is input into the gray prediction model to determine a weld lower layer penetration prediction sequence for the test piece to be tested. The present invention obtains raw data through an ultrasonic testing method, and ultimately achieves a prediction of the weld lower layer penetration of dissimilar metal lap welds, which can directly and effectively improve the reliability of weld quality testing.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic detection, and in particular to a method and system for measuring weld penetration in lap welding based on grey prediction. Background Art

[0002] Dissimilar metal welding is much more difficult than welding homogeneous materials due to the large differences in melting points, linear expansion coefficients, thermal conductivity and specific heat capacity, and complex microstructures of these metals. This can easily lead to insufficient or excessive penetration of the lower layer. The penetration of the lower layer of the weld refers to the depth of the weld in the lower layer of metal during electron beam welding of two dissimilar metal layers. It is generally the distance between the interface between the two layers of dissimilar metals and the weld tip.

[0003] For ultrasonic testing of dissimilar metal welds, since the two sides of the weld are made of different materials and their sound beams are different, it is difficult to detect the longitudinal depth penetration, and it is difficult to achieve effective detection of the weld penetration. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for measuring weld penetration in lap welding based on grey prediction, combining grey prediction algorithm with ultrasonic phased array detection to achieve effective and accurate prediction of the penetration depth of the lower layer of the weld in lap welding of dissimilar materials.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] The present invention provides a method for measuring weld penetration in lap welding based on grey prediction, comprising:

[0007] Acquiring an ultrasonic imaging width sequence of the test piece to be inspected;

[0008] Establishing a grey prediction model based on sample data; the sample data includes ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by performing ultrasonic phased array testing on multiple sample specimens;

[0009] The ultrasonic imaging width sequence of the test piece to be inspected is input into the grey prediction model to determine that the lower layer of the weld of the test piece to be inspected enters the penetration prediction sequence.

[0010] Optionally, the process of establishing the grey prediction model specifically includes:

[0011] Preprocessing the weld lower layer penetration sample data based on the ultrasonic imaging width sample data to determine an equidistant penetration sequence;

[0012] According to the equidistant penetration sequence, a grey GM(1,1) model is established;

[0013] According to the grey GM (1,1) model, an equidistant penetration prediction sequence is calculated;

[0014] Determining a non-equidistant penetration prediction sequence based on the equidistant penetration prediction sequence; wherein the data in the non-equidistant penetration prediction sequence corresponds one-to-one to the weld lower layer penetration sample data;

[0015] When the relative error rate is within a set range, the grey GM (1,1) model is determined as a grey prediction model; the relative error rate is determined based on the penetration prediction value in the non-equidistant penetration prediction sequence and the corresponding weld lower layer penetration sample data.

[0016] Optionally, preprocessing the weld lower layer penetration sample data based on the ultrasonic imaging width sample data to determine an equidistant penetration sequence specifically includes:

[0017] According to the piecewise linear difference formula

[0018]

[0019] Convert the weld lower layer penetration sample data into an equidistant penetration sequence;

[0020] Where Δt represents the sequence x (0) The average serial interval of (t), x (0) (t) represents the equidistant penetration sequence, t n represents the ultrasonic imaging width sample data of the nth test piece to be tested, t1 represents the ultrasonic imaging width sample data of the first test piece to be tested, and t i represents the ultrasonic imaging width sample data of the i-th test piece to be inspected, t i-1 represents the ultrasonic imaging width sample data of the i-1th test piece to be tested, x (0) (t1) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the first test piece to be inspected, x (0) (t i ) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the i-th test piece to be inspected, x (0) (t i-1 ) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the i-1th test piece to be inspected, x (0) (t n ) represents the equally spaced penetration depth data corresponding to the ultrasonic imaging width sample data of the nth test piece to be inspected.

[0021] Optionally, establishing a grey GM(1,1) model according to the equidistant penetration sequence specifically includes:

[0022] According to the formula

[0023]

[0024] Establish the grey GM(1,1) model;

[0025] Among them, X (0) is the equidistant penetration sequence function, X (1) Represents the cumulative generation sequence function, Z (1) Represents the adjacent value generation sequence function; x (0) (1) represents the first penetration depth data in the equally spaced penetration depth sequence function, x (0) (2) represents the second penetration depth data in the equally spaced penetration depth sequence function, x (0) (n) represents the nth penetration depth data in the equally spaced penetration depth sequence function, x (1) (1) represents the first penetration depth data in the cumulative generation sequence function, x (1) (2) represents the second penetration depth data in the cumulative generation sequence function, x (1) (n) represents the nth penetration depth data in the cumulative generation sequence function, z (1) (1) represents the first penetration depth data in the adjacent value generation sequence function, z (1) (2) represents the second penetration depth data in the adjacent value generation sequence function, z (1) (n) represents the nth penetration depth data in the adjacent value generation sequence function, x (1) (t) represents the t-th penetration data in the cumulative generation sequence function, z (1) (t) represents the t-th penetration depth data in the adjacent value generation sequence function.

[0026] Optionally, the calculating of the equally spaced penetration prediction sequence according to the grey GM(1,1) model specifically includes:

[0027] According to the formula Determining the grey differential equation of the grey GM(1,1) model;

[0028] According to the formula determining a response function of the grey differential equation;

[0029] According to the formula Restore and calculate the equally spaced penetration depth prediction sequence;

[0030] Among them, a and b are unknown parameters, and a and b are estimated according to the least squares method. Sure; x (0) (3) represents the third penetration depth data in the equally spaced penetration depth sequence function, z (1) (3) represents the third penetration depth data in the adjacent value generation sequence function, Indicates the t+1th penetration depth data of the cumulative generation sequence function in the response function, Indicates the tth penetration depth data in the equally spaced penetration depth sequence function obtained by restoration calculation.

[0031] Optionally, determining a non-equidistant penetration depth prediction sequence based on the equidistant penetration depth prediction sequence specifically includes:

[0032] According to the formula

[0033] k=2,3,...,n to determine the non-uniformly spaced penetration prediction sequence;

[0034] Where Δt represents the sequence x (0) The average serial interval of (t), x (0) (t) represents the equally spaced penetration sequence, xd(k) represents the kth ultrasonic imaging width sample value, Indicates the first penetration depth prediction value in the non-uniformly spaced penetration depth prediction sequence. represents the k-1th predicted value of the non-uniformly spaced penetration depth prediction sequence, represents the kth predicted value of the non-uniformly spaced penetration depth prediction sequence, x (2) (1) represents the first predicted value of the equidistant depth prediction sequence, x (2) (k) represents the kth penetration depth prediction value in the equally spaced penetration depth prediction sequence.

[0035] Optionally, the relative error rate is calculated as follows:

[0036] According to the formula

[0037] Calculate the relative error rate;

[0038] The predicted value represents the penetration prediction value in the non-equidistant penetration prediction sequence, and the actual value represents the penetration sample data of the lower layer of the weld corresponding to the penetration prediction value in the non-equidistant penetration prediction sequence.

[0039] Optionally, the measurement method further includes:

[0040] The non-uniformly spaced penetration depth prediction sequence and a plurality of the weld lower layer penetration depth sample data are displayed respectively using a broken line graph.

[0041] To achieve the above-mentioned object, the present invention further provides a weld penetration measurement system for lap welding based on grey prediction, comprising:

[0042] A test data determination module is used to obtain an ultrasonic imaging width sequence of a test piece to be tested;

[0043] a model building module for establishing a grey prediction model based on sample data; the sample data includes ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by performing ultrasonic phased array testing on multiple sample specimens;

[0044] The prediction module is used to input the ultrasonic imaging width sequence of the test piece to be inspected into the grey prediction model to determine whether the lower layer of the weld of the test piece to be inspected enters the penetration prediction sequence.

[0045] Optionally, in terms of establishing a grey prediction model, the model building module specifically includes:

[0046] An equidistant processing submodule is used to pre-process the weld lower layer penetration sample data based on the ultrasonic imaging width sample data to determine an equidistant penetration sequence;

[0047] A model building submodule is used to build a grey GM (1,1) model according to the equidistant penetration sequence;

[0048] An equidistant prediction submodule is used to calculate an equidistant penetration prediction sequence according to the grey GM (1,1) model;

[0049] a non-equally spaced prediction submodule, configured to determine a non-equally spaced penetration prediction sequence based on the equal-spaced penetration prediction sequence; wherein the data in the non-equally spaced penetration prediction sequence corresponds one-to-one to the weld lower layer penetration sample data;

[0050] The error calculation submodule is used to determine the gray GM (1, 1) model as the gray prediction model when the relative error rate is within a set range; the relative error rate is determined based on the penetration prediction value in the non-equidistant penetration prediction sequence and the corresponding weld lower layer penetration sample data.

[0051] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0052] This invention provides a method and system for measuring weld penetration in lap welds based on gray prediction. This method uses a sequence of ultrasonic imaging widths from the test piece to be inspected, inputting them into a gray prediction model to accurately predict the penetration depth of the weld's lower layer. By combining a gray prediction algorithm with ultrasonic imaging technology, this method directly and effectively improves the reliability and accuracy of weld quality inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 Schematic diagram of the process of the method for measuring weld penetration in lap welding based on grey prediction according to the present invention;

[0055] Figure 2 A line graph showing the non-equidistantly spaced penetration prediction sequence and a plurality of weld lower layer penetration sample data of the present invention;

[0056] Figure 3 The figure is a schematic structural diagram of the weld penetration measurement system in lap welding based on grey prediction according to the present invention. DETAILED DESCRIPTION

[0057] 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.

[0058] In order to make the objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] Example 1

[0060] like Figure 1 As shown, this embodiment provides a method for measuring weld penetration in lap welding based on grey prediction, comprising:

[0061] Step 100: Acquire an ultrasonic imaging width sequence of a test piece to be inspected.

[0062] Step 200: Establish a grey prediction model based on sample data; the sample data includes ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by performing ultrasonic phased array testing on multiple sample specimens.

[0063] Step 300: Inputting the ultrasonic imaging width sequence of the test piece to be inspected into the grey prediction model to determine whether the lower layer of the weld of the test piece to be inspected enters the penetration prediction sequence.

[0064] The establishment process of the grey prediction model specifically includes:

[0065] 1) Based on the ultrasonic imaging width sample data, the weld lower layer penetration sample data is preprocessed to determine an equidistant penetration sequence. Specifically, the surface of the sample specimen is preprocessed by grinding or other pretreatment to obtain a pretreated specimen; ultrasonic phased array testing is performed based on the parameters of the pretreated specimen to obtain ultrasonic imaging width sample data of the dissimilar metal lap weld; then, the dissimilar metal lap weld is cut, ground, polished, and metallographically observed and measured to obtain the dissimilar metal lap weld lower layer penetration sample data. The ultrasonic imaging width refers to the gap width between the interface echoes between the dissimilar metals measured when the dissimilar metal lap weld is tested using an ultrasonic phased array.

[0066] According to the piecewise linear difference formula

[0067]

[0068] Convert the weld lower layer penetration sample data into an equidistant penetration sequence.

[0069] Where Δt represents the sequence x (0) The average serial interval of (t), x (0) (t) represents the equidistant penetration sequence, t n represents the ultrasonic imaging width sample data of the nth test piece to be tested, t1 represents the ultrasonic imaging width sample data of the first test piece to be tested, and t i represents the ultrasonic imaging width sample data of the i-th test piece to be inspected, t i-1 represents the ultrasonic imaging width sample data of the i-1th test piece to be tested, x (0) (t1) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the first test piece to be inspected, x (0) (t i ) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the i-th test piece to be inspected, x (0) (t i-1) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the i-1th test piece to be inspected, x (0) (t n ) represents the equally spaced penetration depth data corresponding to the ultrasonic imaging width sample data of the nth test piece to be inspected.

[0070] 2) According to the equidistant penetration sequence, a grey GM(1,1) model is established. Specifically, according to the formula

[0071]

[0072] Establish the grey GM(1,1) model;

[0073] Among them, X (0) is the equidistant penetration sequence function, X (1) Represents the cumulative generation sequence function, Z (1) represents a sequence function for generating adjacent values; specifically, the function value of the equally spaced penetration sequence function corresponds to the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the test piece to be inspected calculated in 1) above. The first penetration data in the equally spaced penetration sequence function is equal to the equally spaced penetration data corresponding to the first ultrasonic imaging width sample data of the test piece to be inspected. Other data can be obtained by analogy.

[0074] x (0) (1) represents the first penetration depth data in the equally spaced penetration depth sequence function, x (0) (2) represents the second penetration depth data in the equally spaced penetration depth sequence function, x (0) (n) represents the nth penetration depth data in the equally spaced penetration depth sequence function, x (1) (1) represents the first penetration depth data in the cumulative generation sequence function, x (1) (2) represents the second penetration depth data in the cumulative generation sequence function, x (1) (n) represents the nth penetration depth data in the cumulative generation sequence function, z (1) (1) represents the first penetration depth data in the adjacent value generation sequence function, z (1) (2) represents the second penetration depth data in the adjacent value generation sequence function, z (1) (n) represents the nth penetration depth data in the adjacent value generation sequence function, x (1) (t) represents the t-th penetration data in the cumulative generation sequence function, z (1) (t) represents the t-th penetration depth data in the adjacent value generation sequence function.

[0075] 3) According to the grey GM(1,1) model, calculate the equally spaced penetration prediction sequence. Specifically:

[0076] According to the formula Determine the grey differential equation of the grey GM(1,1) model.

[0077] According to the formula A response function of the grey differential equation is determined.

[0078] According to the formula Restore the calculation of the equally spaced penetration prediction sequence.

[0079] Among them, a and b are unknown parameters, and a and b are estimated according to the least squares method. Sure; x (0) (3) represents the third penetration depth data in the equally spaced penetration depth sequence function, z (1) (3) represents the third penetration depth data in the adjacent value generation sequence function, Indicates the t+1th penetration depth data of the cumulative generation sequence function in the response function, represents the t-th penetration depth data in the equally spaced penetration depth sequence function obtained by restoration calculation, that is, the t-th penetration depth data in the equally spaced penetration depth prediction sequence. (2) (k) is the kth penetration prediction value in the equally spaced penetration prediction sequence. When k = t,

[0080] 4) Based on the equidistant penetration prediction sequence, a non-equidistant penetration prediction sequence is determined; the data in the non-equidistant penetration prediction sequence corresponds one-to-one with the weld lower layer penetration sample data. Specifically, according to the formula

[0081] k=2,3,...,n to determine the non-uniformly spaced penetration prediction sequence;

[0082] Where Δt represents the sequence x (0) The average serial interval of (t), x (0) (t) represents the equidistant penetration sequence, xd(k) represents the independent variable value input to the gray prediction GM(1,1) model of the kth weld penetration, that is, the ultrasonic imaging width sample value of the kth weld in the sample specimen, Indicates the first penetration depth prediction value in the non-uniformly spaced penetration depth prediction sequence. represents the k-1th predicted value of the non-uniformly spaced penetration depth prediction sequence, represents the kth predicted value of the non-uniformly spaced penetration depth prediction sequence, x (2) (1) represents the first predicted value of the equidistant depth prediction sequence, x (2) (k) represents the kth penetration depth prediction value in the equally spaced penetration depth prediction sequence.

[0083] 5) When the relative error rate is within the set range, the gray GM (1,1) model is determined as the gray prediction model; the relative error rate is determined based on the penetration prediction value in the non-uniformly spaced penetration prediction sequence and the corresponding weld lower layer penetration sample data. Specifically, according to the formula

[0084] Calculate the relative error rate.

[0085] The predicted value represents the penetration prediction value in the non-equidistant penetration prediction sequence, and the actual value represents the penetration sample data of the lower layer of the weld corresponding to the penetration prediction value in the non-equidistant penetration prediction sequence.

[0086] 6) When the relative error rate is not within the set range, the established grey GM(1,1) model and the collected data are discarded, the sample data are reacquired, and the model is established.

[0087] In one specific embodiment, ultrasonic phased array equipment was used to inspect and perform metallographic measurements on dissimilar metal lap welds. The test subjects were 100x100mm GH4169 and nickel plates, 4mm and 5mm thick, respectively. Vacuum electron beam welding was used to weld the upper GH4169 layer and the lower nickel plate together along the long edges of the plates at the center. The welding voltage was fixed at 60kV, and various welding currents and speeds were used.

[0088] In the first step, as described in step 100 above, the sample specimen is subjected to pretreatment, ultrasonic phased array testing, cutting, grinding, polishing, and metallographic observation and measurement in sequence, thereby obtaining ultrasonic imaging width sample data of the dissimilar metal lap weld and sample data of the lower layer penetration depth of the dissimilar metal lap weld.

[0089] The above data are organized into a table for display, and the table content is arranged from small to large according to the ultrasound imaging width sample data, as shown in Table 1.

[0090] Table 1 Original data of gray prediction model for weld penetration of dissimilar metal lap welds

[0091]

[0092] In order to determine whether the sequence composed of the ultrasonic imaging width sample data and the lower layer penetration depth sample data in Table 1 is an equidistant sequence, the origin moment of the ultrasonic imaging width data is taken, and the results are shown in Table 2.

[0093] Table 2 Data table of gray prediction model for weld penetration of dissimilar metal lap welds

[0094]

[0095] If there is Δt i =t i -t i-1 ≠const(i=2,3,…,n), it means that the sequence of the lower layer penetration depth sample data is a non-uniformly spaced sequence. According to Table 2, it is obvious that the data collected this time is a non-uniformly spaced sequence. Therefore, the sequence of the ultrasonic imaging relative width sample data and the lower layer penetration depth sample data is subjected to spacing transformation processing. The specific processing method is as follows:

[0096] 1) Calculate the average serial interval:

[0097]

[0098] 2) Perform data conversion; denote the above data sequence as x (0) , use piecewise linear interpolation to convert the data into equally spaced data; the conversion formula is as follows:

[0099]

[0100] From the above formula we can get: x (0) (1)=2.3390,x (0) (2) = 2.8426, x (0) (3) = 3.3461, x (0) (4) = 2.1327, x (0) (5) = 2.4834, x (0) (6)=2.3068,x (0) (7) = 2.0774, x (0) (8) = 1.9120;

[0101] The second step is to establish a grey GM(1,1) model as described in step 200 above:

[0102] Then the grey differential equation model is:

[0103]

[0104] Solve the grey differential equation: On both sides of the grey differential equation, integrate on the interval (k-1, k) to obtain (where k∈[2,3,…,n]): x (0) (k)+az (1) (k) = b.

[0105] remember is the parameter list, let:

[0106] Then the least squares estimation parameter in the grey differential equation is

[0107] satisfy

[0108] The response function for solving the grey differential equation is:

[0109]

[0110] Let x (1) (0) = x (1) (1), the above formula is transformed into:

[0111]

[0112] Furthermore, the predicted sequence is restored:

[0113]

[0114] Among them, x (1) (0) = 2.3390, a = 0.0756, b = 3.3215, t is the sample number.

[0115] The equidistant penetration prediction sequence obtained after grey differential prediction is recorded as x (2) , and the above calculations yield:

[0116] x (2) (1)=2.3390,x (2) (2) = 3.0289, x (2) (3)=2.8085,x (2) (4)=2.6040,x (2) (5) = 2.4145

[0117] x (2) (6) = 2.2388, x (2) (7)=2.0758,x (2) (8) = 1.9248;

[0118] The third step is to perform an equal-interval inverse transformation on the equal-interval penetration prediction sequence after the grey differential prediction. The non-equal-interval penetration prediction sequence obtained by the equal-interval inverse transformation is recorded as

[0119] The corresponding calculation formula is as follows:

[0120]

[0121] Calculation can be obtained:

[0122]

[0123] The fourth step is to calculate the relative error rate based on the penetration prediction value in the non-uniformly spaced penetration prediction sequence and the corresponding weld lower layer penetration sample data. The specific calculation formula is:

[0124]

[0125] The relative error sequence for the eight samples is denoted as error. Then, error1 = 0, error2 = 16.2792%, error3 = 5.8247%, error4 = 9.7533%, error5 = 5.3828%, error6 = 2.5263%, error7 = 0.3235%, and error8 = 0.6667%. Table 3 shows the weld penetration data (actual values), the predicted penetration values ​​(predicted values) in the non-uniformly spaced penetration prediction sequence, and the corresponding relative error rates for the eight samples.

[0126] Table 3. Actual values, predicted values ​​and corresponding relative error rates

[0127]

[0128]

[0129] Furthermore, the average relative error rate is calculated according to the formula: average relative error rate = sum of relative error rates / N, where N represents the number of data in the non-uniformly spaced penetration prediction sequence. Based on the data in Table 3, the average relative error rate can be calculated to be 5.0945%.

[0130] The fifth step is to use a broken line graph to display the non-uniformly spaced penetration prediction sequence and multiple weld bottom layer penetration sample data. The line graph is drawn using the Origin software with the ultrasonic imaging width of the dissimilar metal lap weld as the horizontal axis. Figure 2 .

[0131] Example 2

[0132] like Figure 3 As shown, this embodiment provides a weld penetration measurement system for lap welding based on grey prediction, comprising:

[0133] The test data determination module 101 is used to obtain an ultrasonic imaging width sequence of the test piece to be detected.

[0134] The model building module 201 is used to establish a gray prediction model based on sample data; the sample data includes ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by performing ultrasonic phased array testing on multiple sample specimens.

[0135] The prediction module 301 is configured to input the ultrasonic imaging width sequence of the test piece to be inspected into the grey prediction model to determine whether the lower layer of the weld of the test piece to be inspected enters the penetration prediction sequence.

[0136] In terms of establishing the grey prediction model, the model building module specifically includes:

[0137] The equidistant processing submodule is used to pre-process the weld lower layer penetration sample data according to the ultrasonic imaging width sample data to determine an equidistant penetration sequence.

[0138] The model building submodule is used to establish a grey GM (1,1) model according to the equidistant penetration sequence.

[0139] The equally spaced prediction submodule is used to calculate an equally spaced penetration prediction sequence according to the grey GM (1,1) model.

[0140] The non-equidistant-interval prediction submodule is used to determine a non-equidistant-interval penetration prediction sequence based on the equidistant-interval penetration prediction sequence; the data in the non-equidistant-interval penetration prediction sequence corresponds one-to-one to the weld lower layer penetration sample data.

[0141] The error calculation submodule is used to determine the gray GM (1, 1) model as the gray prediction model when the relative error rate is within a set range; the relative error rate is determined based on the penetration prediction value in the non-equidistant penetration prediction sequence and the corresponding weld lower layer penetration sample data.

[0142] Compared with the prior art, the present invention also has the following advantages:

[0143] The present invention obtains original data through an ultrasonic detection method, and ultimately realizes the prediction of the penetration depth of the lower layer of the dissimilar metal lap weld, which can directly and effectively improve the reliability of weld quality detection.

[0144] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0145] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for measuring weld penetration in lap welding based on grey prediction, characterized in that: The weld penetration measurement in the lap welding includes: Acquiring an ultrasonic imaging width sequence of the test piece to be inspected; A grey prediction model is established based on sample data; the sample data includes ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by performing ultrasonic phased array testing on multiple sample specimens; wherein the ultrasonic imaging width refers to the gap width between the interface echoes between the dissimilar metals measured when using ultrasonic phased array to detect the dissimilar metal lap weld; The ultrasonic imaging width sequence of the test piece to be inspected is input into the grey prediction model to determine that the lower layer of the weld of the test piece to be inspected enters the penetration prediction sequence.

2. The method for measuring weld penetration in lap welding based on grey prediction according to claim 1, characterized in that: The establishment process of the grey prediction model specifically includes: Preprocessing the weld lower layer penetration sample data based on the ultrasonic imaging width sample data to determine an equidistant penetration sequence; According to the equidistant penetration sequence, a grey GM(1,1) model is established; According to the grey GM (1,1) model, an equidistant penetration prediction sequence is calculated; Determining a non-equidistant penetration prediction sequence based on the equidistant penetration prediction sequence; wherein the data in the non-equidistant penetration prediction sequence corresponds one-to-one to the weld lower layer penetration sample data; When the relative error rate is within a set range, the grey GM (1,1) model is determined as a grey prediction model; the relative error rate is determined based on the penetration prediction value in the non-equidistant penetration prediction sequence and the corresponding weld lower layer penetration sample data.

3. The method for measuring weld penetration in lap welding based on grey prediction according to claim 2, characterized in that: Preprocessing the weld lower layer penetration sample data based on the ultrasonic imaging width sample data to determine an equidistant penetration sequence specifically includes: According to the piecewise linear difference formula Convert the weld lower layer penetration sample data into an equidistant penetration sequence; Where Δt represents the sequence x (0) The average serial interval of (t), x (0) (t) represents the equidistant penetration sequence, t n represents the ultrasonic imaging width sample data of the nth test piece to be tested, t1 represents the ultrasonic imaging width sample data of the first test piece to be tested, and t i represents the ultrasonic imaging width sample data of the i-th test piece to be inspected, t i-1 represents the ultrasonic imaging width sample data of the i-1th test piece to be tested, x (0) (t1) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the first test piece to be inspected, x (0) (t i ) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the i-th test piece to be inspected, x (0) (t i-1 ) represents the equally spaced penetration data corresponding to the ultrasonic imaging width sample data of the i-1th test piece to be inspected, x (0) (t n ) represents the equally spaced penetration depth data corresponding to the ultrasonic imaging width sample data of the nth test piece to be inspected.

4. The method for measuring weld penetration in lap welding based on grey prediction according to claim 2, characterized in that: The grey GM (1,1) model is established according to the equidistant penetration sequence, specifically comprising: According to the formula Establish the grey GM(1,1) model; Among them, X (0) is the equidistant penetration sequence function, X (1) Represents the cumulative generation sequence function, Z (1) Represents the adjacent value generation sequence function; x (0) (1) represents the first penetration depth data in the equally spaced penetration depth sequence function, x (0) (2) represents the second penetration depth data in the equally spaced penetration depth sequence function, x (0) (n) represents the nth penetration depth data in the equally spaced penetration depth sequence function, x (1) (1) represents the first penetration depth data in the cumulative generation sequence function, x (1) (2) represents the second penetration depth data in the cumulative generation sequence function, x (1) (n) represents the nth penetration depth data in the cumulative generation sequence function, z (1) (1) represents the first penetration depth data in the adjacent value generation sequence function, z (1) (2) represents the second penetration depth data in the adjacent value generation sequence function, z (1) (n) represents the nth penetration depth data in the adjacent value generation sequence function, x (1) (t) represents the t-th penetration data in the cumulative generation sequence function, z (1) (t) represents the t-th penetration depth data in the adjacent value generation sequence function.

5. The method for measuring weld penetration in lap welding based on grey prediction according to claim 4, characterized in that: The calculation of the equally spaced penetration prediction sequence according to the grey GM (1,1) model specifically includes: According to the formula Determining the grey differential equation of the grey GM(1,1) model; According to the formula determining a response function of the grey differential equation; According to the formula Restore and calculate the equally spaced penetration depth prediction sequence; Among them, a and b are unknown parameters, and a and b are estimated according to the least squares method. Sure; x (0) (3) represents the third penetration depth data in the equally spaced penetration depth sequence function, z (1) (3) represents the third penetration depth data in the adjacent value generation sequence function, Indicates the t+1th penetration depth data of the cumulative generation sequence function in the response function, Indicates the tth penetration depth data in the equally spaced penetration depth sequence function obtained by restoration calculation.

6. The method for measuring weld penetration in lap welding based on grey prediction according to claim 2, characterized in that: Determining a non-equidistant penetration depth prediction sequence based on the equidistant penetration depth prediction sequence specifically includes: According to the formula Determine the non-equidistant penetration depth prediction sequence; Where Δt represents the sequence x (0) The average serial interval of (t), x (0) (t) represents the equally spaced penetration sequence, xd(k) represents the kth ultrasonic imaging width sample value, Indicates the first penetration depth prediction value in the non-uniformly spaced penetration depth prediction sequence. represents the k-1th predicted value of the non-uniformly spaced penetration depth prediction sequence, represents the kth predicted value of the non-uniformly spaced penetration depth prediction sequence, x (2) (1) represents the first predicted value of the equidistant depth prediction sequence, x (2) (k) represents the kth penetration depth prediction value in the equally spaced penetration depth prediction sequence.

7. The method for measuring weld penetration in lap welding based on grey prediction according to claim 2, characterized in that: The calculation process of the relative error rate is: According to the formula Calculate the relative error rate; The predicted value represents the penetration prediction value in the non-equidistant penetration prediction sequence, and the actual value represents the penetration sample data of the lower layer of the weld corresponding to the penetration prediction value in the non-equidistant penetration prediction sequence.

8. The method for measuring weld penetration in lap welding based on grey prediction according to claim 2, characterized in that: The measuring method further comprises: The non-uniformly spaced penetration depth prediction sequence and a plurality of the weld lower layer penetration depth sample data are displayed respectively using a broken line graph.

9. A weld penetration measurement system for lap welding based on grey prediction, characterized in that: The system comprises: A test data determination module is used to obtain an ultrasonic imaging width sequence of a test piece to be tested; A model building module is configured to establish a grey prediction model based on sample data; the sample data includes ultrasonic imaging width sample data and weld lower layer penetration sample data corresponding to the ultrasonic imaging width sample data; the ultrasonic imaging width sample data is determined by ultrasonic phased array testing of multiple sample specimens; wherein the ultrasonic imaging width refers to the gap width between interface echoes between dissimilar metals measured when ultrasonic phased array testing is used to detect dissimilar metal lap welds; The prediction module is used to input the ultrasonic imaging width sequence of the test piece to be inspected into the grey prediction model to determine whether the lower layer of the weld of the test piece to be inspected enters the penetration prediction sequence.

10. The weld penetration measurement system in lap welding based on grey prediction according to claim 9, characterized in that: In terms of establishing the grey prediction model, the model building module specifically includes: An equidistant processing submodule is used to pre-process the weld lower layer penetration sample data based on the ultrasonic imaging width sample data to determine an equidistant penetration sequence; A model building submodule is used to build a grey GM (1,1) model according to the equidistant penetration sequence; An equidistant prediction submodule is used to calculate an equidistant penetration prediction sequence according to the grey GM (1,1) model; a non-equally spaced prediction submodule, configured to determine a non-equally spaced penetration prediction sequence based on the equal-spaced penetration prediction sequence; wherein the data in the non-equally spaced penetration prediction sequence corresponds one-to-one to the weld lower layer penetration sample data; The error calculation submodule is used to determine the gray GM (1, 1) model as the gray prediction model when the relative error rate is within a set range; the relative error rate is determined based on the penetration prediction value in the non-equidistant penetration prediction sequence and the corresponding weld lower layer penetration sample data.

Images