Welding spot structure parameter multi-objective optimization method based on Taguchi-grey fuzzy

Abstract

The invention discloses a welding spot structure parameter multi-objective optimization method based on Taguchi-grey fuzzy, which comprises the following steps: 1) establishing a welding spot finite element model, and applying any load to the model; 2) selecting welding spot structure parameters as influence factors, establishing an orthogonal test table, applying the same load to the finite element model, and processing an analysis result to obtain a target (including welding spot maximum stress strain, service life and the like) to be optimized; 3) performing signal-to-noise ratio calculation on a target to be optimized, and performing normalization processing on the signal-to-noise ratio by using grey correlation analysis; and 4) constructing a fuzzy reasoning system based on a grey correlation analysis result, performing fuzzy reasoning by using the system to obtain an MPCI of a target needing to be optimized in each group of welding spot structure parameter combination, and performing range variance analysis on the MPCI to obtain and verify an optimal welding spot structure parameter combination.

Description

TECHNICAL FIELD

[0001] The application relates to a multi-objective optimization method for solder joint structure parameters based on Taguchi-gray fuzzy, and belongs to the technical field of electronic packaging. BACKGROUND

[0002] Welding technology plays a crucial role in modern industrial manufacturing. From the connection of complex structural components in the aerospace field to the assembly of body parts in automobile manufacturing, and the fixation of small components in electronic devices, the performance of the welded structure directly affects the quality and reliability of the product. As a key component of the welded structure, the structural parameters of the solder joint (such as solder joint size, shape, welding process parameters, etc.) have a decisive influence on the mechanical properties (including stress and strain distribution) and service life of the welded structure.

[0003] Most of the previous solder joint optimization methods focus on single-objective optimization, such as only considering the maximum stress of the solder joint. However, the performance of the solder joint is multi-faceted, and in addition to stress, strain distribution, fatigue life, etc. are also important factors to consider. Single-objective optimization often cannot fully consider the comprehensive performance of the solder joint, which may lead to the deterioration of other performance indicators after optimizing one performance indicator. With the increasing demand for the performance of welded structures in modern industry, a multi-objective optimization method for solder joints is needed that can consider multiple key performance indicators such as maximum stress, strain distribution, fatigue life, etc. to achieve comprehensive optimization of solder joint structure parameters.

[0004] In recent years, with the development of computing technology, numerical simulation methods such as finite element analysis have provided new means for solder joint optimization. By establishing a finite element model of the solder joint, the stress and strain performance indicators under different structural parameter combinations can be quickly obtained through simulation analysis in a virtual environment. This provides a basis for the application of multi-objective optimization methods. At the same time, advanced optimization theories such as Taguchi method, gray system theory, and fuzzy logic also provide new ideas for multi-objective optimization. The Taguchi method can efficiently screen out key factors and optimization directions through orthogonal test tables; the gray system theory can handle uncertainty and fuzziness problems, which is suitable for situations where multiple factors interact and the relationship is complex in solder joint optimization; fuzzy logic can handle the fuzzy relationship between objectives in the optimization process, providing strong support for comprehensive evaluation of solder joint performance.

[0005] In summary, traditional solder joint optimization methods have been unable to meet the needs of modern industry for high-performance and efficient optimization of welded structures. Therefore, it is of great practical significance to develop a multi-objective optimization method for solder joint structure parameters based on Taguchi-gray fuzzy. SUMMARY

[0006] The technical problem solved by the present application is to provide a multi-objective optimization method for solder joint structure parameters based on Taguchi-gray fuzzy, which can realize multi-objective optimization of solder joint stress, strain and fatigue life and the like.

[0007] In order to solve the above technical problem, the present application is implemented by adopting the following technical scheme:

[0008] A multi-objective optimization method for solder joint structure parameters based on Taguchi-gray fuzzy, comprising the following steps:

[0009] Step 1, a reasonable finite element model of the solder joint is established, and an arbitrary load is applied to the model, and the analysis result is processed to obtain the target to be optimized;

[0010] Step 2, the solder joint structure parameters are selected as the influencing factors, an orthogonal test table is established, and the same load is applied, and the analysis result is processed to obtain the target to be optimized;

[0011] Step 3, the signal-to-noise ratio of the target to be optimized is calculated, and the signal-to-noise ratio is normalized by using gray correlation analysis;

[0012] Step 4, based on the gray correlation analysis result, a fuzzy inference system is constructed, and fuzzy inference is performed by using the system to obtain the MPCI of the target to be optimized for each group of solder joint structure parameter combinations, and range variance analysis is performed on the MPCI to obtain and verify the optimal solder joint structure parameter combination.

[0013] Step 1 includes the following specific methods:

[0014] Step 1-1, a geometric model of the solder joint interconnection structure is established by using ANSYS software, mainly including the following parts: chip, solder joint and PCB board;

[0015] Step 1-2, an arbitrary load is applied to the established finite element model, and the analysis result is processed to obtain the target to be optimized.

[0016] Step 3 includes the following specific methods:

[0017] Step 3-1, the reliability characteristics of the solder joint can be represented by a Gaussian distribution, wherein S is a signal, N is a noise, and the signal-to-noise ratio originally refers to the power amplifier of a broadcasting device, and here it is the response value of the optimization target, as shown in the following formula:

[0018]

[0019] In the formula: The greater the value is, The smaller the value is, that is, the smaller the noise is, indicating that the solder joint reliability is less sensitive to external noise factors, and the higher the reliability is;

[0020] According to the different optimization target characteristics, the signal-to-noise ratio can be divided into three kinds of small, medium and large, and the expressions are as follows:

[0021]

[0022]

[0023]

[0024] In the formula: y i The value of the optimization target; n is the number of optimization target values, and the number of the method is 1;

[0025] Step 3-2, according to the different response characteristics, the response characteristics are divided into three kinds of small, medium and large, and the signal-to-noise ratio of the optimization target is normalized to 0-1, and the expressions are as follows:

[0026]

[0027]

[0028]

[0029] In the formula: Y(k) is the normalized value of the signal-to-noise ratio target of the i-th test, x i (k) is the signal-to-noise ratio value of the i-th test, and A(k) is the target signal-to-noise ratio.

[0030] The specific method of step 4 comprises:

[0031] Step 4-1, the signal-to-noise ratio grey correlation coefficient is used as an input variable to define the semantic function of Gaussian membership;

[0032] Step 4-2, according to the division of the fuzzy subset of the input variable membership function and the division of the corresponding output response fuzzy subset, the relationship between them is constructed into a fuzzy rule base, and the fuzzy rule is set;

[0033] Step 4-3, the fuzzy inference calculation is carried out through the division of the given input variable and output response fuzzy subset and the fuzzy rule base, the fuzzy relationship is obtained, the inference result after the anti-fuzzification of the input variable is calculated, that is, the output response MPCI;

[0034] Step 4-4, the MPCI of the output is subjected to range variance analysis, and the optimal structure parameter combination of the welding spot can be obtained and verified.

[0035] The method of the present application can optimize the structure parameters of the welding spot, and the targets include welding spot stress and strain and fatigue life, etc., which can avoid the deterioration of other performance indicators after optimizing one performance indicator. Attached Figure Description

[0036] Figure 1 This is an overall flowchart of the multi-objective optimization method of the present invention;

[0037] Figure 2 Geometric model for finite element analysis of BGA solder joints;

[0038] Figure 3 This is a schematic diagram of a fuzzy design.

[0039] Figure 4 For thermal structural coupling stress (stress I) at the solder joint of the stacked BGA;

[0040] Figure 5 The thermal structural coupling strain of the BGA solder joint (Strain I);

[0041] Figure 6 For the bending-torsional coupling stress (stress II) of the BGA solder joints in the stacked BGA;

[0042] Figure 7 The bending-torsional coupling strain (strain II) of the BGA solder joints in the multilayer BGA is shown. Detailed Implementation

[0043] This invention relates to a multi-objective optimization method for weld joint structural parameters based on Taguchi-grey fuzzing, used for multi-objective optimization of weld joint stress, strain, and lifespan, comprising the following steps:

[0044] Step 1: This paper uses the MPC8245 chip manufactured by NXP Semiconductors in the Netherlands as the research object and establishes a finite element analysis model of the BGA solder joints. The dimensions of each part of the model are: chip 30mm×30mm×1.17mm, solder ball diameter 0.65mm, pad diameter 0.40mm, single-layer solder joint height 0.50mm, solder joint spacing 1.27mm, and PCB board length, width, and thickness dimensions 170mm×100mm×1.60mm. The model contains 4 chips, spaced 10mm apart, and evenly distributed in the center of the PCB board. Each chip has 4 rings of solder joints, for a total of 352×4 (1408) solder joints. The geometric model of the BGA solder joint finite element analysis is as follows: Figure 2 As shown, the chip layout and PCB board size design refer to the relevant provisions of standard IPC / JEDEC-9702.

[0045] Step 2: An orthogonal experimental design was established using solder ball diameter, pad diameter, and solder joint height as key factors in the BGA solder joint model. Four levels were selected for each of the three factors, as shown in Table 1. Li 16 (4) 5The orthogonal test table obtains a level combination table as shown in Table 2. According to 16 different structural parameter level combinations in Table 3, a corresponding laminated BGA solder joint model is respectively established. A heat load and a boundary condition are applied to the finite element analysis model of the laminated BGA solder joint: four chips with a power consumption of 2W are used as heat sources of the model, and the heat generation rate of each chip is 1.814*10 6 W / m 3 As a body load, a convection heat transfer coefficient (environmental temperature 25℃) of 10W / (m 2 ·K) is set on the outer surface of the model, and full constraint thermal structure coupling analysis is performed on the four corners at the bottom of the PCB. The maximum stress and strain of the 16 groups of laminated BGA solder joints are shown in columns 5 and 6 of Table 3.

[0046] The laminated BGA solder joint finite element model is loaded in an indirect coupling manner. First, the model is subjected to bending load, and then subjected to torsional load. Finally, the bending-torsional coupling analysis result of the laminated BGA solder joint is obtained. The bending load standard adopts IPC / JEDEC-9702. The specific loading steps and boundary conditions are as follows: (1) the chip of the model is downward and the PCB support point is fixed, the support point span is 120mm, a downward displacement load of 0.5mm is applied to the middle line of the back surface of the PCB, which makes the laminated BGA solder joint finite element model deform in three-point bending; (2) continue to apply a torsional load to the laminated BGA solder joint finite element model, the chip of the model is upward, the finite element analysis result of the bending load is used as the initial condition, and a 0.3mm displacement load in the same diagonal direction and opposite adjacent angle direction is applied to the four corner points of the back surface of the PCB, so that the PCB deforms in torsion, and the maximum bending-torsional coupling stress and strain are shown in columns 5 and 6 of Table 4.

[0047] Table 2 Structural parameter factor level table of laminated BGA solder joint

[0048] Factor / Level Level 1 Level 2 Level 3 Level 4 Solder ball diameter / mm 0.55 0.60 0.65 0.76 Pad diameter / mm 0.40 0.42 0.44 0.46 Solder joint height / mm 0.44 0.46 0.48 0.50

[0049] Table 3 Orthogonal test design and thermal structure coupling simulation results

[0050] Serial number / Factor Solder ball diameter / mm Pad diameter / mm Solder joint height / mm Maximum stress / MPa Maximum strain 1 0.55 0.40 0.44 17.507 0.00046091 2 0.55 0.42 0.46 17.261 0.00045455 3 0.55 0.44 0.48 16.996 0.00044768 4 0.55 0.46 0.50 16.689 0.00044133 5 0.60 0.40 0.46 17.472 0.00045987 6 0.60 0.42 0.44 17.442 0.00045908 7 0.60 0.44 0.50 17.025 0.00044825 8 0.60 0.46 0.48 16.996 0.00044752 9 0.65 0.40 0.48 17.435 0.00045886 10 0.65 0.42 0.50 17.213 0.00045305 11 0.65 0.44 0.44 17.300 0.00455280 12 0.65 0.46 0.46 17.055 0.00448880 13 0.76 0.40 0.50 17.420 0.00045845 14 0.76 0.42 0.48 17.418 0.00045838 15 0.76 0.44 0.46 17.406 0.00045806 16 0.76 0.46 0.44 17.385 0.00045750

[0051] Table 4 Orthogonal test design and bending-torsional coupling simulation results

[0052] Serial number / Factor Solder ball diameter / mm Pad diameter / mm Solder joint height / mm Maximum stress / MPa Maximum strain 1 0.55 0.40 0.44 50.884 0.0013394 2 0.55 0.42 0.46 47.782 0.0012581 3 0.55 0.44 0.48 45.291 0.0011929 4 0.55 0.46 0.50 43.193 0.0011382 5 0.60 0.40 0.46 51.006 0.0013424 6 0.60 0.42 0.44 49.656 0.0013068 7 0.60 0.44 0.50 45.072 0.0011865 8 0.60 0.46 0.48 43.871 0.0011548 9 0.65 0.40 0.48 50.872 0.0013388 10 0.65 0.42 0.50 47.754 0.0012567 11 0.65 0.44 0.44 48.825 0.0012849 12 0.65 0.46 0.46 45.761 0.0012043 13 0.76 0.40 0.50 50.593 0.0013315 14 0.76 0.42 0.48 49.969 0.0013151 15 0.76 0.44 0.46 49.458 0.0013017 16 0.76 0.46 0.44 49.075 0.0012916

[0053] Step 3, based on the orthogonal test table 3 and table 4 data, by looking at the small characteristic signal-to-noise ratio calculation formula, the stress and strain signal-to-noise ratio of each combination of thermal structure coupling and bending torsion coupling in table 5 is calculated as shown in the last four columns of table 5 (thermal structure coupling maximum stress and strain are referred to as stress I and strain I, bending torsion coupling maximum stress and strain are referred to as stress II and strain II). In table 5, the thermal structure coupling stress and strain and bending torsion coupling stress and strain signal-to-noise ratio corresponding to the fourth test group are the largest, which are-24.449, 67.105, -32.708 and 58.876, respectively; while the thermal structure coupling stress and strain signal-to-noise ratio corresponding to the first test group is the smallest, which is-24.864, 66.728, respectively, and the bending torsion coupling stress and strain signal-to-noise ratio corresponding to the fifth test group is the smallest, which is-34.152, 57.442, respectively, the higher the signal-to-noise ratio value indicates that the reliability of the laminated BGA solder joint under the structure parameter level combination is higher;

[0054] Table 5 thermal structure coupling and bending torsion coupling stress and strain signal-to-noise ratio calculation

[0055] Serial number / Factor Solder ball diameter / mm Pad diameter / mm Solder joint height / mm Stress I SNR Strain I SNR Stress II SNR Strain II SNR 1 0.55 0.40 0.44 -24.864 66.728 -34.132 57.462 2 0.55 0.42 0.46 -24.741 66.848 -33.585 58.006 3 0.55 0.44 0.48 -24.607 66.981 -33.120 58.468 4 0.55 0.46 0.50 -24.449 67.105 -32.708 58.876 5 0.60 0.40 0.46 -24.847 66.747 -34.152 57.442 6 0.60 0.42 0.44 -24.832 66.762 -33.919 57.676 7 0.60 0.44 0.50 -24.622 66.970 -33.078 58.515 8 0.60 0.46 0.48 -24.607 66.984 -32.844 58.750 9 0.65 0.40 0.48 -24.828 66.766 -34.130 57.466 10 0.65 0.42 0.50 -24.717 66.877 -33.580 58.015 11 0.65 0.44 0.44 -24.761 66.834 -33.773 57.823 12 0.65 0.46 0.46 -24.637 66.957 -33.210 58.385 13 0.76 0.40 0.50 -24.821 66.774 -34.082 57.513 14 0.76 0.42 0.48 -24.820 66.775 -33.974 57.621 15 0.76 0.44 0.46 -24.814 66.782 -33.885 57.710 16 0.76 0.46 0.44 -24.803 66.792 -33.817 57.777

[0056] The data of stress and strain signal-to-noise ratio after normalization in table 5 is taken as the comparison sequence, the target sequence is taken as the reference sequence, the target sequence is set to 1, the stress and strain signal-to-noise ratio gray correlation coefficient obtained by calculation is shown in table 6, and the calculation formula is as follows:

[0057]

[0058] In the formula: is the normalized stress and strain signal-to-noise ratio value of the first group test, which is taken as 1 here; is the gray correlation coefficient; is the resolution coefficient, which is taken as 0.5.

[0059] Table 6 stress and strain signal-to-noise ratio gray correlation coefficient

[0060] Serial number / Factor Stress I Strain I Stress II Strain II 1 0.333 0.333 0.336 0.337 2 0.415 0.424 0.452 0.452 3 0.568 0.603 0.637 0.637 4 1.000 1.000 1.000 1.000 5 0.343 0.345 0.333 0.333 6 0.352 0.355 0.374 0.374 7 0.546 0.582 0.665 0.661 8 0.568 0.609 0.851 0.842 9 0.354 0.358 0.337 0.337 10 0.436 0.453 0.454 0.453 11 0.400 0.411 0.405 0.404 12 0.524 0.561 0.594 0.590 13 0.358 0.363 0.345 0.345 14 0.359 0.364 0.364 0.363 15 0.363 0.368 0.381 0.380 16 0.369 0.376 0.395 0.394

[0061] Step 4, a fuzzy reasoning system is constructed, and the fuzzy design schematic diagram is as Figure 3The stress I, strain I, stress II and strain II signal-to-noise ratio grey correlation coefficients of the laminated BGA solder joint are shown; the Gaussian membership function semantics definition is performed on the input variables, which are defined as three fuzzy subsets of large (L), medium (M) and small (S); the MPCI is defined as the output response, which is defined as five fuzzy subsets of very large (VL), large (L), medium (M), small (S) and very small (VS). According to the division of the three fuzzy subsets of the input variable membership function and the division of the corresponding five fuzzy subsets of the output response membership function, the relationship between them is constructed into a fuzzy rule base, and a total of 81 fuzzy rules are set, as shown in Table 7.

[0062] Table 7 Fuzzy rule setting of input variables and output response MPCI

[0063] Fuzzy rule Stress I Strain I Stress II Strain II MPCI Fuzzy rule Stress I Strain I Stress II Strain II MPCI 1 S S S S VS 42 M M M L M 2 S S S M S 43 M M L S M 3 S S S L S 44 M M L M M 4 S S M S S 45 M M L L L 5 S S M M S 46 M L S S M 6 S S M L M 47 M L S M M 7 S S L S S 48 M L S L M 8 S S L M M 49 M L M S M 9 S S L L M 50 M L M M M 10 S M S S S 51 M L M L L 11 S M S M S 52 M L L S M 12 S M S L M 53 M L L M L 13 S M M S S 54 M L L L L 14 S M M M M 55 L S S S S 15 S M M L M 56 L S S M M 16 S M L S M 57 L S S L M 17 S M L M M 58 L S M S M 18 S M L L M 59 L S M M M 19 S L S S S 60 L S M L M 20 S L S M M 61 L S L S M 21 S L S L M 62 L S L M M 22 S L M S M 63 L S L L L 23 S L M M M 64 L M S S M 24 S L M L M 65 L M S M M 25 S L L S M 66 L M S L M 26 S L L M M 67 L M M S M 27 S L L L L 68 L M M M M 28 M S S S S 69 L M M L L 29 M S S M S 70 L M L S M 30 M S S L M 71 L M L M L 31 M S M S S 72 L M L L L 32 M S M M M 73 L L S S M 33 M S M L M 74 L L S M M 34 M S L S M 75 L L S L L 35 M S L M M 76 L L M S M 36 M S L L M 77 L L M M L 37 M M S S S 78 L L M L L 38 M M S M M 79 L L L S L 39 M M S L M 80 L L L M L 40 M M M S M 81 L L L L VL 41 M M M M M

[0064] Through the fuzzy inference calculation of the input variable and output response fuzzy subset division and fuzzy rule base given above, the fuzzy relationship is obtained, and the inference results after the anti-fuzzification of the four input variables, i.e. the output response MPCI, are calculated. The MPCI values corresponding to the stress I, strain I, stress II and strain II signal-to-noise ratio grey correlation coefficients are respectively inferred and calculated, and the MPCI results corresponding to 16 groups of laminated BGA solder joint structure parameters are shown in the last column of Table 8. The MPCI value corresponding to the 4th test is the largest (0.887), and the MPCI value corresponding to the 1st test is the smallest (0.439). The larger the MPCI is, the more stable the system is, and the higher the reliability of the laminated BGA solder joint is.

[0065] Table 8 MPCI calculation results

[0066] Serial number / Factor Solder ball diameter / mm Pad diameter / mm Solder joint height / mm MPCI 1 0.55 0.40 0.44 0.439 2 0.55 0.42 0.46 0.484 3 0.55 0.44 0.48 0.539 4 0.55 0.46 0.50 0.887 5 0.60 0.40 0.46 0.440 6 0.60 0.42 0.44 0.455 7 0.60 0.44 0.50 0.552 8 0.60 0.46 0.48 0.692 9 0.65 0.40 0.48 0.444 10 0.65 0.42 0.50 0.491 11 0.65 0.44 0.44 0.477 12 0.65 0.46 0.46 0.520 13 0.76 0.40 0.50 0.449 14 0.76 0.42 0.48 0.458 15 0.76 0.44 0.46 0.462 16 0.76 0.46 0.44 0.466

[0067] The range analysis of the MPCI corresponding to the stacked BGA solder joint under different structural parameter level combinations in Table 8 is performed, and the range analysis result based on the orthogonal test design is shown in Table 9. In Table 9, the mean values K1, K2, K3 and K4 are the average values of the MPCI cumulative sum corresponding to each influencing factor under the levels 1, 2, 3 and 4 respectively. The range R is the difference between the maximum value and the minimum value in the mean values of each influencing factor, and the greater the difference is, the higher the weight of the influencing factor is, and the more the MPCI corresponding to the stacked BGA solder joint is affected. It can be obtained from the range values in Table 9 that the influence degree of each factor on the MPCI value of the stacked BGA solder joint is: pad diameter > solder joint height > solder ball diameter, that is, among the three influencing factors, the pad diameter has the greatest influence on the solder joint MPCI, followed by the solder joint height, and finally the solder ball diameter. Since the MPCI reflects the robustness and reliability of the stacked BGA solder joint in various environments, therefore, when optimizing the stacked BGA solder joint, improving the pad can obtain better optimization results. According to Tables 5-7, the optimal structural parameter level combination of the stacked BGA solder joint is: solder ball diameter 0.55 mm, pad diameter 0.46 mm and solder joint height 0.50 mm.

[0068] Table 9 MPCI range analysis result

[0069] Factor Solder ball diameter Pad diameter Solder joint height [K1] 0.587 0.443 0.459 [K2] 0.535 0.472 0.476 [K3] 0.483 0.508 0.533 [K4] 0.459 0.641 0.595 R 0.128 0.198 0.136

[0070] The variance analysis of the MPCI corresponding to the stacked BGA solder joint under different structural parameter level combinations in Table 8 is performed, and the variance analysis result based on the orthogonal test design is shown in Table 10. It can be known from Table 10 that, among the three influencing factors considered, the F value corresponding to the pad diameter is 3.407, which is greater than the critical value F 0.1 = 3.29, therefore, at a confidence level of 90%, the influence of the pad diameter on the MPCI of the stacked BGA solder joint is significant. The F values corresponding to the solder ball diameter and the solder joint height are 1.444 and 1.667 respectively, both of which are less than the corresponding critical value F 0.1 = 3.29, therefore, at a confidence level of 90%, the solder ball diameter and the solder joint height have no significant influence on the MPCI of the stacked BGA solder joint. According to the F values corresponding to each factor, it can also be obtained that the influence degree of each factor on the MPCI is: pad diameter > solder joint height > solder ball diameter, that is, the pad diameter has the greatest influence on the solder joint MPCI, followed by the solder joint height, and finally the solder ball diameter. The size order is consistent with the order of the range analysis result, verifying the correctness of the range analysis result.

[0071] Table 10 MPCI variance analysis result

[0072] Factor Sum of square deviation Degree of freedom F ratio F critical value Significance Solder ball diameter 0.039 3 1.444 3.29 Pad diameter 0.092 3 3.407 3.29 Significant (*) Solder joint height 0.045 3 1.667 3.29 Error 0.05 6

[0073] The optimal structure parameter horizontal combination (solder ball diameter 0.55 mm, pad diameter 0.46 mm and solder joint height 0.50 mm) obtained in the last section is used to establish the finite element model of the stacked BGA solder joint, and the thermal-structure coupling analysis and the bending-torsion coupling loading analysis are respectively carried out to obtain the thermal-structure coupling and bending-torsion coupling stress-strain distribution cloud diagrams of the stacked BGA solder joint as shown in FIG. 11. Figures 4-7 The comparison results of the stress-strain before and after optimization are shown in Table 11.

[0074] Table 11 Comparison results of stress-strain before and after optimization

[0075] Before optimization After optimization Optimization ratio Stress I 17.507 MPa 16.689 MPa Strain I 4.67% Stress II 0.00046091 0.00044133 4.25% 51.006 MPa 43.193 MPa Strain II 15.32% ​ 0.0013424 0.0011382 15.21%

[0076] As shown in Table 11, the four response targets after optimization are improved, the thermal-structure coupling stress-strain of the stacked BGA solder joint after optimization is reduced by 4.67% and 4.25% respectively, and the bending-torsion coupling stress-strain is reduced by 15.32% and 15.21% respectively, and the stress-strain in both loading environments is reduced. Therefore, when optimizing the structure parameters of the stacked BGA solder joint, the Taguchi-gray fuzzy method can be used to optimize the thermal-structure coupling and bending-torsion coupling stress-strain at the same time, and the comprehensive optimization effect of both is obtained.

Claims

1. A multi-objective optimization method for solder joint structure parameters based on Taguchi-gray fuzzy, characterized in that, The method comprises the following steps: Step 1, a reasonable finite element model of the welding spot is established, an arbitrary load is applied to the model, and an analysis result is processed to obtain a target to be optimized; Step 2, welding spot structure parameters are selected as influencing factors, an orthogonal test table is established, the same load is applied, and an analysis result is processed to obtain the target to be optimized; Step 3, a signal-to-noise ratio of the target to be optimized is calculated, and the signal-to-noise ratio is normalized by using grey correlation analysis; Step 4, a fuzzy inference system is constructed based on the grey correlation analysis result, fuzzy inference is carried out by using the system, an MPCI of the target to be optimized of each group of welding spot structure parameter combinations is obtained, and range variance analysis is carried out on the MPCI to obtain and verify an optimal structure parameter combination of the welding spot.

2. The spot structure parameter multi-objective optimization method based on Taguchi-gray fuzzy according to claim 1, characterized in that: The arbitrary load in the step 1 comprises heat, electricity, vibration, bending and torsion and the like, and multi-physical field coupling.

3. The spot structure parameter multi-objective optimization method based on Taguchi-gray fuzzy according to claim 1, characterized in that: The optimization target in the step 1 comprises a maximum stress and strain of the welding spot and a fatigue life of the welding spot and the like.

4. The spot structure parameter multi-objective optimization method based on Taguchi-gray fuzzy according to claim 1, characterized in that: The welding spot structure parameters in the step 2 comprise a welding spot diameter, a pad diameter and a welding spot height and the like.

Images