A pin soldering cavity risk assessment method under abnormal tin paste

Abstract

This invention discloses a method for assessing the risk of PIN soldering voids under abnormal solder paste incoming conditions, relating to the field of semiconductor module packaging assessment. The method includes: collecting multi-source data from the sample to be tested, performing time alignment and data normalization processing to construct a multi-source original dataset for soldering; based on the original dataset, performing feature decoupling processing to obtain a basic feature set, and constructing a structural consistency deviation relationship between different features based on the basic feature set to form a deviation feature set; further combining the basic feature set and the structural consistency deviation relationship to construct solder joint structural stability evolution potential parameters; based on the solder joint structural stability evolution potential parameters and the deviation feature set, assessing the risk of PIN soldering voids under abnormal solder paste incoming conditions and outputting a corresponding assessment report. By fusing multi-source data to construct evolution potential parameters, hidden failure modes can be identified, and graded handling suggestions can be output, improving assessment accuracy and process guidance capabilities.

Description

Technical Field

[0001] This invention relates to the field of semiconductor module packaging evaluation, specifically a method for assessing the risk of PIN soldering voids under abnormal solder paste delivery conditions. Background Technology

[0002] In semiconductor packaging processes, solder paste is a critical material for PIN soldering, and its incoming quality directly determines soldering reliability. Current technologies typically rely on multiple independent tests to evaluate incoming solder paste, including temperature recovery verification, dispensing parameter verification, post-soldering pull testing, solder joint visual inspection, and X-ray void detection. Each of these tests has its own pass / fail threshold or judgment criteria, and in practice, a "step-by-step" approach is often used to make an overall judgment on the acceptance or rejection of the incoming material.

[0003] However, the aforementioned methods are essentially based on single-index or simple combination judgments, making it difficult to effectively handle situations where different testing indicators contradict each other. For example, in actual production, samples often appear that meet the post-weld tensile strength and appearance standards but fail to meet the void rate requirements. This "hidden failure mode" is easily overlooked in existing evaluation systems because individual compliance conditions are met, but void defects may gradually expand during use, leading to early failure. Furthermore, existing technologies lack quantitative means to assess the degree of inconsistency between multi-source data and cannot provide differentiated release strategies or process modification suggestions for hidden failure modes.

[0004] Therefore, there is an urgent need for an assessment method that can integrate multi-source detection data, identify hidden risks such as qualified tensile strength but abnormal voids, and output refined handling opinions accordingly. Summary of the Invention

[0005] Based on the shortcomings of the prior art described above, the purpose of this invention is to provide a method for assessing the risk of PIN soldering voids under abnormal solder paste incoming materials, so as to solve the above-mentioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for assessing the risk of PIN soldering voids under abnormal solder paste incoming materials, comprising: S1: Collect multi-source data of the sample to be tested, perform time alignment and data normalization processing, and construct the original multi-source welding dataset; S2: Based on the original dataset of welding multi-sources, feature decoupling processing is performed to obtain a basic feature set. Based on the basic feature set, the structural consistency deviation relationship between different features is constructed to form a deviation feature set. The basic feature set and the structural consistency deviation relationship are further combined to construct the structural stability evolution potential parameters of the weld point. S3: Based on the stable evolution potential parameters and deviation feature set of the solder joint structure, assess the risk of PIN soldering voids under abnormal solder paste conditions and output the corresponding assessment report.

[0007] The present invention is further configured such that the multi-source data includes: solder paste incoming material temperature recovery status data, welding process parameter data, post-soldering mechanical property data, solder joint appearance inspection data, and X-ray void detection data; The solder paste incoming material status data includes: opening time, warm-up time, storage temperature curve and packaging status; the soldering process parameter data includes: dispensing amount, delay time, gap adjustment amount, pressure, speed, needle size and printing thickness; the post-soldering mechanical property data includes: maximum tensile force and fracture mode; the solder joint appearance inspection data includes: wetting angle, spreading area and bridging condition; and the X-ray void detection data includes: void rate, void size distribution and void location.

[0008] The present invention is further configured such that the construction of the welding multi-source raw dataset in S1 includes: Based on preset data identification rules, a unified data association identifier is established for multi-source data, wherein the data association identifier includes at least one or more of the following: batch identifier, product identifier, and solder joint location identifier; Based on the data association identifier, data from different sources are associated and matched, and aligned according to a preset time window rule, so that data from different time scales form a corresponding relationship under the same association identifier; The aligned data is standardized according to a unified data format and dimensional standard to form a structurally consistent multi-source raw dataset for welding.

[0009] The present invention is further configured such that the basic feature set includes: wetting-driven state, structural forming state, interface bonding state, and internal defect state.

[0010] The present invention is further configured such that S2 includes: a feature decoupling step, a deviation feature construction step, and a potential parameter construction step.

[0011] The present invention is further configured such that the feature decoupling step includes: The data in the original multi-source welding dataset are processed by interval mapping based on the preset process constraint interval or statistical distribution characteristics to obtain feature quantities with uniform dimensions and comparability. Based on the process physical meaning of the data and its role in the weld formation process, the feature quantities are divided into dimensions, and the data in the multi-source original dataset of welding are mapped to four dimensions: wetting drive state, structural forming state, interface bonding state and internal defect state. The specific mapping process is as follows: Based on the coupled correlation characterization process of reheat time and dispensing parameters, a wetting driving state is constructed to characterize the wetting driving capability. The structural forming state is obtained by processing the wetting angle, spreading area and bridging status in the weld joint appearance inspection data through feature association and consistency judgment. The interface bonding state is obtained by matching the maximum tensile force value in the post-weld mechanical property data with the preset bonding strength range. The internal defect state is obtained by spatially analyzing the void ratio, void size distribution, and void location in X-ray void detection data. The state data of four dimensions—wetting-driven state, structural forming state, interface bonding state, and internal defect state—are fused in a multidimensional manner to form a basic feature set containing multidimensional state characteristics.

[0012] The present invention is further configured such that the deviation feature construction step includes: Based on the basic feature set, structural consistency deviation relationships are constructed to characterize the degree of inconsistency between states through coupling association and matching judgment processing between different states. Various structural consistency deviation relationships are then integrated to form a deviation feature set. Specifically, the construction of the structural consistency deviation relationship includes: Based on the coupling correlation analysis between the wetting-driven state and the structural forming state, the first type of deviation relationship is constructed by comparing and extracting the differences of the corresponding feature quantities of the two. Based on the matching determination process between the structural forming state and the interface bonding state, a second type of deviation relationship is constructed by performing consistency comparison and deviation identification processing on the corresponding state feature quantities. Based on the consistency comparison and judgment process between the interface bonding state and the internal defect state, a third type of deviation relationship is constructed by screening and anomaly identification of internal defect features in samples that meet the interface bonding conditions.

[0013] The present invention is further configured such that the stability evolution potential parameters of the weld joint structure include: initial driving force potential, forming realization potential, and defect disturbance potential; The potential parameter construction steps include: The feature quantities corresponding to the wetting drive state in the basic feature set are numerically mapped according to the preset wetting drive capability quantification rules to obtain the initial driving force potential. By analyzing the correlation consistency of the characteristic quantities corresponding to the structural forming state and interface bonding state in the basic feature set, the forming realization potential is obtained. The feature quantities corresponding to the internal defect states in the basic feature set are combined with the third type of deviation relationship in the deviation feature set, and then comprehensively quantified according to the void location weight allocation rule, void size distribution classification rule, and the inverse correlation correction rule with the interface performance characterization result to obtain the defect perturbation potential. The initial driving force potential, the forming realization potential, and the defect disturbance potential are fused through multi-level processing using evolution constraint rules to obtain the stable evolution potential parameters of the weld joint structure. The evolution constraint rules include: when the initial driving potential is higher than a first set threshold and the forming realization potential is lower than a second set threshold, a potential reduction process is performed on the fusion result of the initial driving potential and the forming realization potential; when the forming realization potential is higher than a third set threshold and the defect disturbance potential is higher than a fourth set threshold, a suppression process is performed on the fusion result of the forming realization potential and the defect disturbance potential; when the third type of deviation relationship is in an abnormal state, a false high correction process is performed on the fusion result.

[0014] The present invention is further configured such that the evaluation report includes: incoming material release level, welding repair restrictions, and process modification recommendations.

[0015] The present invention is further configured such that S3 includes: Based on the stability evolution potential parameters of the solder joint structure, the overall structural stability level of the solder paste under the current process conditions is classified and judged to form the corresponding stability evaluation results. Based on the deviation relationships of various structural consistency in the deviation feature set, the degree of abnormality of the weld joint in each stage of wetting drive, structural forming, interface bonding and internal defects is classified and identified to determine the corresponding risk source category and risk impact range. The stability evaluation results are correlated and matched with the risk source categories. Based on the preset risk classification rules, the risk of PIN soldering voids is classified into levels to form corresponding risk level results. Based on the risk level results, and combined with the distribution characteristics and impact of various deviation relationships, the availability of incoming materials is determined, and the corresponding material release level is generated. Based on the risk source categories and risk impact scope, the operational constraints in the welding process are extracted and limited to generate corresponding repair welding constraints. Based on the structural inconsistency features reflected in the deviation feature set, the direction of process parameter adjustment is identified and processed, and combined with the stability evolution potential parameter of the weld joint structure, corresponding process correction suggestions are generated. The incoming material release level, resoldering restrictions, and process modification suggestions are integrated and output to form a PIN soldering void risk assessment report under abnormal solder paste incoming conditions.

[0016] This invention provides a method for assessing the risk of PIN soldering voids under abnormal solder paste incoming conditions. The method comprises: S1: collecting multi-source data from the sample to be tested, performing time alignment and data normalization processing to construct a multi-source original dataset for soldering; S2: based on the multi-source original dataset, performing feature decoupling processing to obtain a basic feature set, and constructing a structural consistency deviation relationship between different features based on the basic feature set to form a deviation feature set; further, combining the basic feature set and the structural consistency deviation relationship to construct solder joint structural stability evolution potential parameters; S3: based on the solder joint structural stability evolution potential parameters and the deviation feature set, assessing the risk of PIN soldering voids under abnormal solder paste incoming conditions and outputting a corresponding assessment report. The beneficial effects include: Improved accuracy and comprehensiveness of assessment: A unified dataset is constructed by collecting multi-source data and performing time alignment and normalization processing; four basic features are obtained through physical and semantic feature decoupling: wetting drive, structural forming, interface bonding, and internal defects; three types of structural consistency deviation relationships are further constructed to quantify the degree of mismatch between features, significantly improving the accuracy and comprehensiveness of the assessment.

[0017] Enhanced ability to identify hidden failure modes: For the hidden failure mode of "tensile strength is qualified but void rate is not up to standard", a stable evolution potential parameter of weld joint structure is constructed, and evolution constraint rules such as potential reduction, suppression and false height correction are introduced. Combined with the deviation feature set, the structural unstable qualified state is accurately identified, filling the gap in the existing technology for identifying hidden risks.

[0018] The value of process guidance and decision support is outstanding: it outputs a detailed assessment report that includes incoming material release level, welding repair restrictions and process modification suggestions, enabling refined hierarchical management of incoming material availability, and providing actionable process adjustment basis for engineering, quality and manufacturing departments, reducing subsequent production risks caused by incoming material abnormalities.

[0019] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1This is a flowchart illustrating a method for assessing the risk of PIN soldering voids under abnormal solder paste incoming material conditions, as an exemplary embodiment of the present invention. Detailed Implementation

[0021] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0022] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0023] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0024] Example: A method for assessing the risk of PIN soldering voids under abnormal solder paste incoming materials, such as... Figure 1 As shown, it includes: S1: Collect multi-source data of the sample to be tested, perform time alignment and data normalization processing, and construct the original multi-source welding dataset; S2: Based on the original dataset of welding multi-sources, feature decoupling processing is performed to obtain a basic feature set. Based on the basic feature set, the structural consistency deviation relationship between different features is constructed to form a deviation feature set. The basic feature set and the structural consistency deviation relationship are further combined to construct the structural stability evolution potential parameters of the weld point. S3: Based on the stable evolution potential parameters and deviation feature set of the solder joint structure, assess the risk of PIN soldering voids under abnormal solder paste conditions and output the corresponding assessment report.

[0025] The present invention is further configured such that the multi-source data includes: solder paste incoming material temperature recovery status data, welding process parameter data, post-soldering mechanical property data, solder joint appearance inspection data, and X-ray void detection data; The solder paste incoming material status data includes: opening time, warm-up time, storage temperature profile, and packaging status; the soldering process parameter data includes: dispensing amount, delay time, gap adjustment amount, pressure, speed, needle size, and printing thickness; the post-soldering mechanical property data includes: maximum tensile force and fracture mode; the solder joint appearance inspection data includes: wetting angle, spreading area, and bridging condition; and the X-ray void detection data includes: void rate, void size distribution, and void location. The construction of the multi-source raw dataset for welding in S1 includes: Based on preset data identification rules, a unified data association identifier is established for multi-source data, wherein the data association identifier includes at least one or more of the following: batch identifier, product identifier, and solder joint location identifier; Based on the data association identifier, data from different sources are associated and matched, and aligned according to a preset time window rule, so that data from different time scales form a corresponding relationship under the same association identifier; The aligned data is standardized according to a unified data format and dimensional standard to form a consistent multi-source raw dataset for soldering. Specifically, the multi-source data covers five dimensions: solder paste arrival temperature recovery status, soldering process parameters, post-soldering mechanical properties, solder joint appearance, and X-ray void detection. These data are obtained from the material management system, process equipment, mechanical testing instruments, automated optical inspection equipment, and X-ray inspection equipment, respectively. Under unified data association constraints, the above multi-dimensional data is organized and integrated to construct a structured multi-source raw dataset for soldering with cross-process correlations. Data on the solder paste arrival temperature recovery status is typically collected automatically through barcode scanning combined with the MES system. This data includes: opening time, temperature recovery time, storage temperature profile, and packaging status. The opening time is recorded when the solder paste is removed from its sealed packaging to determine if it has been used within the specified exposure window. The warm-up time is recorded as the time it takes for the solder paste to reach room temperature after being removed from the refrigerated environment; insufficient warm-up will lead to a decrease in flux activity. The storage temperature profile is provided by a temperature recorder or MES system, reflecting the temperature fluctuations of the solder paste in the warehouse; deviations from the recommended range may cause solder paste deterioration. Packaging status is recorded through manual barcode scanning or visual inspection to identify whether the packaging is damaged or damp, preventing external contamination. Soldering process parameter data is collected automatically through the equipment control system (such as SPI / dispensing machine), and the data includes: dispensing volume, delay time, gap adjustment, pressure, speed, needle size, and printing thickness. The amount of solder paste dispensed is obtained from the dispensing machine's flow meter or by weighing, affecting the solder paste coverage area. Delay time refers to the waiting time from solder paste extrusion to the start of soldering; excessive delay time can lead to solvent evaporation. Gap adjustment is the distance between the dispensing needle and the substrate, determining the solder paste extrusion effect. Pressure and speed are directly read by the dispensing machine's control system, affecting the stability of solder paste extrusion. Needle size is pre-entered into the equipment formula. Printing thickness is measured by the SPI device, reflecting the uniformity of the solder paste layer. Post-soldering mechanical property data are mainly recorded directly in the test report, including: maximum tensile force and fracture mode. The maximum tensile force is recorded by a tensile tester pulling the PIN at a constant rate in a direction perpendicular to the soldering surface until it detaches, reflecting the interfacial metallurgical bonding strength. The fracture mode is classified by the operator or image recognition software based on the fracture morphology (e.g., solder cohesive fracture, interfacial fracture, DBC delamination, etc.) to assist in determining the root cause of failure. Solder joint appearance inspection data are automatically output by the AOI equipment after soldering, including: wetting angle, spreading area, and bridging status. Wetting angle is measured by taking a cross-sectional image of the solder joint using an automated optical inspection (AOI) device and measuring the angle between the solder and the substrate surface. The smaller the angle, the better the wettability. Spread area is calculated by an image segmentation algorithm to determine the area of ​​the two-dimensional region covered by the solder, reflecting the spreading ability of the solder paste. Bridging is identified by AOI to determine whether there are abnormal solder bridges between adjacent solder joints, which is a type of appearance defect.X-ray void detection data is directly output by the AXI equipment analysis software. The data includes void rate, void size distribution, and void location. The void rate is calculated by the AXI equipment using X-ray transmission imaging, representing the percentage of void area to the total solder joint area. The void size distribution statistics show the number or area percentage of voids in different diameter ranges (e.g., <50μm, 50-100μm, >100μm). The void location is indicated by image coordinates, indicating whether the void is located in the center or edge region of the solder joint. In specific implementation, the data identification rules are used to uniformly index and label data from different sources to support cross-data source association and retrieval. The time window rules are used to coordinate data collected at different times along the time dimension, enabling multi-source data to form a corresponding data sequence under a unified time reference system. The normalization process eliminates differences in expression and units between different source data, allowing data from all dimensions to be uniformly represented under the same data structure. All multi-source data are ultimately associated through batch identifiers, product identifiers, and solder joint location identifiers to form a unified dataset.

[0026] The present invention is further configured such that the basic feature set includes: wetting drive state, structural forming state, interface bonding state, and internal defect state. Specifically, the wetting drive state is a feature state constructed based on solder paste inlet temperature recovery state data and soldering process parameter data, used to characterize the ability of solder paste to flow and wet the substrate surface under temperature recovery and dispensing conditions, serving as the initial driving feature for solder joint formation and used to evaluate the influence of process input conditions on wetting behavior. The structural forming state is a feature state constructed based on solder joint appearance inspection data, including appearance features such as wetting angle, spreading area, and bridging, used to reflect the actual spreading and forming results of the solder joint during the soldering process. The interface bonding state is a feature state constructed based on post-soldering mechanical property data, used to characterize the bonding at the interface between the solder joint, PIN pin, and substrate, and the interface bonding quality is characterized by the maximum tensile force value and fracture mode. The internal defect state is a feature state constructed based on X-ray void detection data, including void ratio, void size, and its spatial location, used to characterize the density of the internal structure of the solder joint and the distribution of defects.

[0027] The present invention is further configured such that S2 includes: a feature decoupling step, a deviation feature construction step, and a potential parameter construction step. Specifically, firstly, the feature decoupling step converts the original multi-source dataset of soldering into a basic feature set, outputting four dimensions of features: wetting-driven state, structural forming state, interface bonding state, and internal defect state. By unifying the dimensions and classifying the process mechanisms of the multi-source data, the solder paste reheating and dispensing data are mapped to the wetting-driven state, the solder joint appearance data to the structural forming state, the post-soldering mechanical property data to the interface bonding state, and the X-ray void data to the internal defect state. This step aims to standardize the heterogeneous data, providing basic features for deviation relationship analysis and potential parameter construction. Secondly, the deviation feature construction step outputs a deviation feature set to characterize the inconsistencies between different states. Through coupling analysis and consistency comparison between each state, the first to third types of structural consistency deviation relationships are constructed, reflecting the degree of deviation between wetting-driven and structural forming, structural forming and interface bonding, and interface bonding and internal defects, respectively. This step aims to quantify solder joint anomalies, providing a basis for identifying void risk sources. Finally, the potential parameter construction step transforms the basic features and deviation features into stable evolution potential parameters of the solder joint structure, outputting the initial driving force potential, forming realization potential, and defect disturbance potential; the initial driving force potential is obtained by mapping the wetting driving state, the forming realization potential is obtained by analyzing the structural forming and interface combination state, and the defect disturbance potential is obtained by quantifying the internal defect state and deviation relationship, and finally, the final stable evolution potential parameters of the solder joint are obtained by fusing them through evolution constraint rules; the role of this step is to provide a quantitative index of solder joint stability, providing a basis for risk assessment of PIN soldering voids under abnormal solder paste conditions.

[0028] The present invention is further configured such that the feature decoupling step includes: The data in the original multi-source welding dataset are processed by interval mapping based on the preset process constraint interval or statistical distribution characteristics to obtain feature quantities with uniform dimensions and comparability. Based on the process physical meaning of the data and its role in the weld formation process, the feature quantities are divided into dimensions, and the data in the multi-source original dataset of welding are mapped to four dimensions: wetting drive state, structural forming state, interface bonding state and internal defect state. The specific mapping process is as follows: Based on the coupled correlation characterization process of reheat time and dispensing parameters, a wetting driving state is constructed to characterize the wetting driving capability. The structural forming state is obtained by processing the wetting angle, spreading area and bridging status in the weld joint appearance inspection data through feature association and consistency judgment. The interface bonding state is obtained by matching the maximum tensile force value in the post-weld mechanical property data with the preset bonding strength range. The internal defect state is obtained by spatially analyzing the void ratio, void size distribution, and void location in X-ray void detection data. The system integrates state data from four dimensions—wetting-driven state, structural forming state, interface bonding state, and internal defect state—to form a basic feature set containing multi-dimensional state characteristics. Specifically, the core principle of the feature decoupling step is to transform the original parameters with different dimensions and physical meanings from the multi-source welding dataset into dimensionless feature quantities between 0 and 1 through interval mapping based on preset process constraint intervals (e.g., reheat time 2-4 hours, dispensing amount 15-25g) or statistical distribution characteristics (e.g., mean ± 3σ of a normal distribution). This ensures comparability between data from different sources. The underlying logic is that welding quality is affected by multiple independent physical factors, and each factor's contribution to the final weld structure is non-linear and non-additive. Therefore, the original data must first be unified to the same scale, and then dimensionally divided according to its process physical meaning to accurately extract the basic features reflecting the four independent dimensions of wetting, forming, bonding, and defects, providing physically meaningful input for subsequent deviation analysis. Specifically, the construction process for the four states is as follows: The construction principle of the wetting-driven state is to map the feature quantities after mapping the original parameters such as reheat time, dispensing volume, dispensing pressure, and dispensing speed, and then fuse them through coupled correlation characterization processing (such as weighted geometric mean or least squares fitting) to form a comprehensive index. The underlying logic is that the wetting ability of solder paste depends on the synergistic effect of sufficient reheat (affecting flux activity) and dispensing stability (affecting solder paste volume and extrusion morphology), both of which are indispensable. In specific implementation, the reheat time is first mapped to a reheat influencing factor according to the process window, and the dispensing parameters are mapped to a dispensing stability factor. Then, the wetting-driven state is obtained through nonlinear weighted fusion. The weighting coefficients can be set based on historical yield statistics or process expert experience to reflect the relative contribution of each factor to the wetting ability. For data with missing parameters, nearest neighbor interpolation or historical averages are used to supplement the data to ensure calculation continuity. The construction principle of the structural forming state is to perform feature correlation and consistency judgment processing on the original parameters such as wetting angle, spreading area, and bridging status in the solder joint appearance inspection data. The underlying logic is that good weld formation should simultaneously satisfy the following conditions: the wetting angle is less than a preset threshold, the spreading area reaches the design value, and there are no bridging defects. Any abnormality in any of these conditions will compromise structural integrity. In specific implementation, the wetting angle, spreading area, and bridging condition are mapped to corresponding influencing factors. Then, the overall forming state is determined through logical AND or minimum value operations. That is, the structural forming state is only qualified when all appearance indicators are qualified; otherwise, the lower value is taken according to the severity of the non-conformity. At the same time, for some missing test data, the average value of adjacent welds or batches can be used for compensation to avoid interruption of the state calculation. The principle of constructing the interface bonding state is to match the maximum tensile force value in the post-weld mechanical property data with the preset bonding strength range. The underlying logic is that the tensile test directly reflects the metallurgical bonding strength of the interface and is a rigid standard set internally by the enterprise.In specific implementation, the preset bonding strength range includes "unqualified zone (tensile force < 10 kg)," "qualified zone (10 kg ≤ tensile force < 15 kg)," and "excellent zone (tensile force ≥ 15 kg)." The maximum tensile force value is mapped to the interface bonding state value (e.g., 0 for unqualified, 0.6 for qualified, and 1.0 for excellent) through interval comparison, or continuous quantization is performed within the interval using linear interpolation. For missing tensile force data, the average value of other weld points in the same batch or historical similar samples can be used for interpolation estimation. The principle for constructing the internal defect state is: spatial distribution analysis of the void rate, void size distribution, and void location in the X-ray void detection data. The underlying logic is that the harm of voids to reliability depends not only on the overall void rate but also on the void size (larger voids are more harmful) and location (voids at the interface are more likely to cause crack propagation than voids at the center). In practice, the system first maps the void rate to a basic void score, then performs weighted correction based on void size distribution (larger voids are assigned higher weights), and finally performs spatial distribution correction based on void location weights (weight at the interface > weight at the center). The three factors combined yield the internal defect state. Weight coefficients can be set using historical good product statistics or process experience. Missing or abnormal void data are supplemented using the average of the same batch or adjacent solder joints to ensure the completeness of the overall state calculation. Finally, when performing multi-dimensional fusion processing on the above four state data, a basic feature set is formed using vector concatenation or structured storage. Each state in this set is a dimensionless comprehensive index between 0 and 1, retaining its independent physical meaning for use in subsequent deviation feature construction steps. By supplementing weight sources and handling missing data, the basic feature set can be continuously and stably calculated in a real production environment, improving implementation feasibility and repeatability.

[0029] The present invention is further configured such that the deviation feature construction step includes: Based on the basic feature set, structural consistency deviation relationships are constructed to characterize the degree of inconsistency between states through coupling association and matching judgment processing between different states. Various structural consistency deviation relationships are then integrated to form a deviation feature set. Specifically, the construction of the structural consistency deviation relationship includes: Based on the coupling correlation analysis between the wetting-driven state and the structural forming state, the first type of deviation relationship is constructed by comparing and extracting the differences of the corresponding feature quantities of the two. Based on the matching determination process between the structural forming state and the interface bonding state, a second type of deviation relationship is constructed by performing consistency comparison and deviation identification processing on the corresponding state feature quantities. Based on the consistency comparison and judgment process between the interface bonding state and the internal defect state, a third type of deviation relationship is constructed by screening and anomaly identification of internal defect features in samples that meet the interface bonding conditions. Specifically, the core principle of the deviation feature construction step is: based on the already constructed basic feature set, through coupling correlation and matching judgment between different states, potential inconsistencies or abnormal manifestations between states are identified, thereby extracting key features that can characterize structural deviations in the welding process. The underlying logic is that the final performance of welding quality depends not only on the absolute indicators of each individual state, but also on the interaction and coupling effect between states. For example, insufficient wetting may lead to abnormal forming, and abnormal forming may reduce the interface bonding strength, ultimately manifesting as voids or defects. By quantifying the degree of inconsistency between states, more sensitive and accurate welding quality deviation information can be provided than that of a single state, providing a physically meaningful input for subsequent quality analysis or anomaly warning. Specifically, the construction process of the structural consistency deviation relationship is as follows: the principle of constructing the first type of deviation relationship is: based on the coupling correlation analysis between the wetting-driven state and the structural forming state, by comparing and extracting the differences of the corresponding feature quantities of the two, a deviation index reflecting the influence of insufficient or excessive wetting driving on the forming state is formed. The underlying logic is that the wetting drive state directly affects the solder paste spreading and wetting angle. If the wetting drive index is low and the structural forming index is abnormal, it indicates insufficient wetting ability leading to forming deviation. If the wetting drive index is high but the forming state is still abnormal, there may be external interference or dispensing abnormalities. In specific implementation, we first obtain the wetting drive state characteristic quantity (a dimensionless comprehensive index of 0~1, weighted and fused from the temperature return influence factor and dispensing stability factor) and the structural forming state characteristic quantity (a comprehensive index of 0~1, calculated from the wetting angle, spreading area, and bridging situation). By calculating the normalized difference between the two (the difference is divided by the state quantity range to ensure the result is still within the 0~1 range), the first type of deviation value is formed. The average deviation and standard deviation are calculated in conjunction with historical good product data to set a deviation threshold for anomaly judgment. The construction principle of the second type of deviation relationship is: based on the matching judgment processing between the structural forming state and the interface bonding state, by comparing the consistency of the corresponding state characteristic quantities and identifying deviations, a deviation index reflecting the influence of forming quality on interface bonding is formed. The underlying logic is as follows: even if the forming state is qualified, if the interface bonding state is substandard, it indicates that there may be interface abnormalities caused by welding temperature, solder activity, or process fluctuations; conversely, if the forming is abnormal but the interface bonding is still qualified, it suggests that there may be local wetting or mechanical redundancy. In specific implementation, the structural forming state characteristic quantity and the interface bonding state characteristic quantity (maximum tensile force is mapped to the 0~1 interval) are obtained. By calculating the normalized difference between the two and combining it with weighting coefficients (such as determining the weights based on historical data), a second type of deviation value is formed, thereby quantifying the degree of inconsistency between the forming and bonding states.The principle behind constructing the third type of deviation relationship is as follows: Based on the consistency comparison and judgment of the interface bonding state and the internal defect state, the deviation index of the influence of the interface bonding state on the performance of internal defects is formed by screening and anomaly identification of internal defect features in samples that meet the interface bonding conditions. The underlying logic is that a good interface bonding usually corresponds to a low void rate and small void size. If voids or large-sized defects still appear in samples with good interface bonding, it indicates a potential risk of internal defects. In specific implementation, firstly, the interface bonding state threshold is determined based on historical good product statistics and process specifications. For example, the average value of historical good products minus one standard deviation (or the minimum qualified value according to the process specifications) is selected as the threshold, and samples with an interface bonding state ≥ this threshold are screened. The internal defect state feature quantity (a comprehensive index of 0~1, obtained by weighting void rate, void size and location) is obtained. The deviation value is calculated by normalizing the difference (internal defect value minus the expected low value, then divided by the range of 0~1) to form the third type of deviation value. It is then combined with weighted correction based on abnormal void size or location to ensure that the deviation value is in the range of 0~1 and can be directly used for feature integration. When integrating the above three types of structural consistency deviations, methods such as vector concatenation, principal component analysis (PCA), or weighted averaging can be used to form a deviation feature set. Each feature in this set is a dimensionless index between 0 and 1, preserving the physical meaning of the inconsistencies in coupling between states, and can be directly used in subsequent welding anomaly analysis, risk assessment, or prediction models. By clarifying the sources of deviation thresholds, the normalization calculation method for deviation values, and the method for handling missing data, it is ensured that the deviation feature set is continuous, stable, and repeatable in the actual production environment.

[0030] The present invention is further configured such that the stability evolution potential parameters of the weld joint structure include: initial driving force potential, forming realization potential, and defect disturbance potential; The potential parameter construction steps include: The feature quantities corresponding to the wetting drive state in the basic feature set are numerically mapped according to the preset wetting drive capability quantification rules to obtain the initial driving force potential. By analyzing the correlation consistency of the characteristic quantities corresponding to the structural forming state and interface bonding state in the basic feature set, the forming realization potential is obtained. The feature quantities corresponding to the internal defect states in the basic feature set are combined with the third type of deviation relationship in the deviation feature set, and then comprehensively quantified according to the void location weight allocation rule, void size distribution classification rule, and the inverse correlation correction rule with the interface performance characterization result to obtain the defect perturbation potential. The initial driving force potential, the forming realization potential, and the defect disturbance potential are fused through multi-level processing using evolution constraint rules to obtain the stable evolution potential parameters of the weld joint structure. The evolution constraint rules include: when the initial driving force potential is higher than a first set threshold and the forming realization potential is lower than a second set threshold, a potential reduction process is performed on the fusion result of the initial driving force potential and the forming realization potential; when the forming realization potential is higher than a third set threshold and the defect disturbance potential is higher than a fourth set threshold, a suppression process is performed on the fusion result of the forming realization potential and the defect disturbance potential; when the third type of deviation relationship is in an abnormal state, a false high correction process is performed on the fusion result. Specifically, the core principle of the potential parameter construction step is: based on the basic feature set and deviation feature set of the weld joint, various state variables in the welding process are mapped into physically meaningful potential parameters to quantify the driving force, forming realization capability, and defect disturbance influence of the weld joint structure in the evolution process. The underlying logic is that the stability of the weld joint structure not only depends on a single state but is also affected by the coupling effect between states and the deviation characteristics. By quantifying different states, they are transformed into mutually independent yet coupled potential parameters, providing clear and operable physical indicators for weld joint stability modeling and evolution analysis. Specifically, the construction of each type of potential parameter is as follows: The principle for constructing the initial driving force potential is to transform the wetting driving state feature quantities in the basic feature set into an initial potential energy index representing the solder joint wetting driving capability. The underlying logic is that the wetting driving state comprehensively reflects the influence of factors such as the sufficiency of solder paste reheating, dispensing amount, dispensing pressure, and dispensing speed on the wetting process. In the specific implementation, the wetting driving state feature quantities (0~1 dimensionless index, formed by weighted fusion of reheating influence factor and dispensing stability factor) are obtained, and numerical mapping is performed according to the preset wetting driving capability quantification rules. For example, a nonlinear mapping function (such as the sigmoid function or power normalization function) can be used to map the state features to the initial driving force potential value range of 0~1 to ensure that high wetting driving state corresponds to high potential value and low wetting driving state corresponds to low potential value, while retaining the ability to sensitively distinguish intermediate states. The principle for constructing the forming realization potential is to form a potential parameter representing the solder joint forming and interface bonding realization capability through the correlation consistency analysis of the structural forming state and interface bonding state feature quantities. The underlying logic is that a well-formed weld joint may still be unstable if the interface bonding is insufficient; conversely, a weld joint with slight deviations in formation but excellent interface bonding may maintain a certain level of stability. In practice, structural forming state characteristic quantities and interface bonding state characteristic quantities (both comprehensive indices of 0-1) are obtained. Consistency analysis is performed by calculating normalized differences, weighted averages, or nonlinear coupling functions (such as weighted geometric averages) to obtain the forming realization potential value. This potential value reflects the degree of consistency between the two; a high potential value corresponds to good forming and interface bonding, while a low potential value corresponds to a large deviation in either state. The principle for constructing the defect disturbance potential is as follows: the internal defect state characteristic quantities in the basic characteristic set are combined with the third type of deviation relationship in the deviation characteristic set to quantify and form potential parameters characterizing the stability disturbance caused by weld joint defects.The underlying logic is that the existence, size, and location of voids directly affect the reliability of solder joints, while the interface bonding state has a moderating effect on the degree of harm caused by voids. In specific implementation, internal defect state characteristics (0-1 indicators, weighted by void rate, void size, and void location) are obtained, and samples with an interface bonding state ≥ a threshold set according to historical good product statistics or process specifications (such as the mean of interface bonding state minus one standard deviation or the minimum acceptable value) are selected. Subsequently, a weighted combination is performed according to the void location weight allocation rule (voids at the interface have a higher weight than those at the center), the void size distribution grading rule (larger voids have a higher weight), and the inverse correlation correction rule with the interface bonding state, to obtain the final defect disturbance potential value. A high value indicates that the defect may cause a significant disturbance to the stability of the solder joint, while a low value indicates that the disturbance is negligible. When multiple potential parameters are fused into the solder joint structural stability evolution potential parameters, evolutionary constraint rules are used to control the comprehensive influence of each potential. The underlying logic is that there are dependencies and constraints between different potentials; the level of a single potential value is insufficient to fully reflect the stability of the solder joint, and dynamic adjustment is required in conjunction with conditions. In specific implementation, the evolution constraint rules include: when the initial driving force potential is higher than the first set threshold and the forming realization potential is lower than the second set threshold, a potential reduction processing is performed on the fusion result of the two, reflecting the structural instability that may be caused by sufficient driving force but insufficient forming; when the forming realization potential is higher than the third set threshold and the defect disturbance potential is higher than the fourth set threshold, a suppression processing is performed on the fusion result of the two, reflecting that even if the forming is good, the influence of defects may still lead to a decrease in overall stability; when the third type of deviation relationship is in an abnormal state, a false high correction processing is performed on the fusion result to prevent the potential value from being overestimated due to abnormal deviation, ensuring the reliability of the evolution potential parameters. Finally, after processing by the evolution constraint rules, the initial driving force potential, forming realization potential, and defect disturbance potential form the weld joint structural stability evolution potential parameters (0~1 interval), which can be used as the core input features for weld joint stability modeling, anomaly identification, and subsequent evolution analysis, ensuring that each potential parameter retains its independent physical meaning while reflecting the combined effect of state coupling and defect influence.

[0031] The present invention is further configured such that the evaluation report includes: incoming material release level, welding repair restrictions, and process modification recommendations; S3 includes: Based on the stability evolution potential parameters of the solder joint structure, the overall structural stability level of the solder paste under the current process conditions is classified and judged to form the corresponding stability evaluation results. Based on the deviation relationships of various structural consistency in the deviation feature set, the degree of abnormality of the weld joint in each stage of wetting drive, structural forming, interface bonding and internal defects is classified and identified to determine the corresponding risk source category and risk impact range. The stability evaluation results are correlated and matched with the risk source categories. Based on the preset risk classification rules, the risk of PIN soldering voids is classified into levels to form corresponding risk level results. Based on the risk level results, and combined with the distribution characteristics and impact of various deviation relationships, the availability of incoming materials is determined, and the corresponding material release level is generated. Based on the risk source categories and risk impact scope, the operational constraints in the welding process are extracted and limited to generate corresponding repair welding constraints. Based on the structural inconsistency features reflected in the deviation feature set, the direction of process parameter adjustment is identified and processed, and combined with the stability evolution potential parameter of the weld joint structure, corresponding process correction suggestions are generated. The incoming material release level, resoldering restrictions, and process modification suggestions are integrated and output to form a PIN soldering void risk assessment report under abnormal solder paste incoming material conditions. Specifically, firstly, based on the solder joint structure stability evolution potential parameters, the overall structural stability level of the incoming solder paste under the current process conditions is graded and judged to form corresponding stability evaluation results. The underlying logic is that the solder joint structure stability evolution potential parameters comprehensively reflect the influence of wetting drive capability, forming achievement capability, and defect disturbance on the overall stability of the solder joint. Through numerical potential parameters, the stability level of the solder joint under process conditions can be intuitively quantified. In specific implementation, the initial driving force potential, forming achievement potential, and defect disturbance potential (all normalized indices of 0~1) are obtained. According to the preset stability thresholds (e.g., high stability region 0.7~1.0, medium stability region 0.4~0.7, low stability region 0~0.4, set by historical good product statistics or process experience), the level is judged to form the stability evaluation results. Secondly, based on the deviation relationships of various structural consistency categories in the deviation feature set, the degree of abnormality of solder joints in each stage of wetting drive, structural forming, interface bonding, and internal defects is classified and identified to determine the corresponding risk source categories and risk impact range. The underlying logic is that solder joint defects or process deviations usually have different effects in different stages. The deviation features can quantitatively characterize the degree of abnormality in each stage and distinguish the main risk sources. In specific implementation, the deviation relationships of the first category (wetting drive-structural forming), the second category (structural forming-interface bonding), and the third category (interface bonding-internal defects) in the deviation feature set are obtained. Each deviation value is normalized (0~1 range) and classified according to the deviation value threshold. A high deviation value indicates a serious abnormality in the corresponding stage, and a low deviation value indicates a minor abnormality. The risk source categories can include "wetting abnormality", "forming abnormality", "bonding abnormality", and "internal defect abnormality". The risk impact range is determined by the location or number of solder joints involved in the deviation relationship. Subsequently, the stability evaluation results are correlated and matched with the risk source categories. According to the preset risk classification rules, the risk of PIN soldering voids is classified into levels to form the corresponding risk level results. The underlying logic is that the overall stability of the solder joints and the specific sources of risk jointly determine the incoming material risk level, and correlation matching can achieve comprehensive evaluation. In specific implementation, the stability evaluation level (high / medium / low) is comprehensively calculated with various deviation indicators, and mapped to a risk level (e.g., low risk, medium risk, high risk) through weighted scoring or a decision matrix. The weights can be set according to historical defect statistics or process requirements. Based on the risk level results, combined with the distribution characteristics and impact of various deviation relationships, the availability of incoming materials is judged, generating a corresponding incoming material release level. The underlying logic is that different risk levels correspond to different incoming material release strategies: high-risk incoming materials may need to be returned or partially re-inspected, medium-risk incoming materials can be used within a quota, and low-risk incoming materials can be released directly.In the specific implementation, the risk level and deviation relationship distribution are comprehensively analyzed, and the final incoming material release level is generated using rule-based judgment or threshold mapping (e.g., if the risk level is ≥ high, the release level is "restricted release"; if the risk level is = medium, the release level is "conditional release"; if the risk level is = low, the release level is "direct release"). Furthermore, based on the risk source category and the scope of risk impact, the operational constraints in the welding process are extracted and limited to generate corresponding repair welding constraints. The underlying logic is that different types of anomalies impose different constraints on welding operations. For example, wetting anomalies may require adjustments to the reheating strategy, forming anomalies may limit the dispensing speed, and internal defect anomalies may require local repair welding or rework. In the specific implementation, corresponding operational constraint templates are matched according to the risk source category, and refined by combining the scope of deviation relationship impact, such as "increasing the frequency of repair welding inspections for welds with severe deviations" and "limiting the welding temperature range for specific welds," thus forming repair welding constraints. Subsequently, based on the structural inconsistency features reflected in the deviation feature set, the direction of process parameter adjustment is identified and processed to generate corresponding process correction suggestions. The underlying logic is as follows: Although this method is based on single-test data, there is a deterministic mapping relationship between different deviation features and potential parameters, which can be used to infer the direction of process adjustments. In specific implementation, for feature dimensions with high deviation values, the corresponding parameter adjustment direction is determined based on a pre-established "deviation feature-process parameter influence relationship model" or empirical rule library. For example, when the wetting-driven related deviation is high, it is determined that the wetting ability is insufficient, and it is recommended to increase the reheat temperature or increase the dispensing amount; when the structural forming deviation is high, it is recommended to optimize the spreading time or adjust the dispensing speed; when the internal defect deviation is high, it is recommended to reduce the peak temperature or optimize the venting path. At the same time, combined with the numerical level of the current solder joint structure stability evolution potential parameter, the adjustment priority is sorted (such as prioritizing the improvement of deviation items with higher contribution to the potential parameter), thereby generating executable process correction suggestions without relying on multi-time series variation data. Finally, the incoming material release level, resoldering restriction conditions, and process correction suggestions are integrated and output to form a PIN soldering void risk assessment report under abnormal solder paste incoming conditions, thereby achieving quantitative assessment of incoming material risks and process guidance under single-test conditions.

[0032] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for evaluating the risk of voids in pin bonding under abnormality of tin paste incoming material, characterized by, include: S1: Collect multi-source data of the sample to be tested, perform time alignment and data normalization processing, and construct the original multi-source welding dataset; S2: Based on the original dataset of welding multi-sources, feature decoupling processing is performed to obtain a basic feature set. Based on the basic feature set, the structural consistency deviation relationship between different features is constructed to form a deviation feature set. The basic feature set and the structural consistency deviation relationship are further combined to construct the structural stability evolution potential parameters of the weld point. S3: Based on the stable evolution potential parameters and deviation feature set of the solder joint structure, assess the risk of PIN soldering voids under abnormal solder paste conditions and output the corresponding assessment report.

2. The method according to claim 1, wherein the tin paste is abnormal, and the method is characterized in that, The multi-source data includes: solder paste incoming temperature recovery status data, soldering process parameter data, post-soldering mechanical property data, solder joint appearance inspection data, and X-ray void detection data; The solder paste incoming material status data includes: opening time, warm-up time, storage temperature curve and packaging status; the soldering process parameter data includes: dispensing amount, delay time, gap adjustment amount, pressure, speed, needle size and printing thickness; the post-soldering mechanical property data includes: maximum tensile force and fracture mode; the solder joint appearance inspection data includes: wetting angle, spreading area and bridging condition; and the X-ray void detection data includes: void rate, void size distribution and void location.

3. The method for assessing the risk of PIN soldering voids under abnormal solder paste incoming material conditions according to claim 2, characterized in that, The construction of the multi-source raw dataset for welding in S1 includes: Based on preset data identification rules, a unified data association identifier is established for multi-source data, wherein the data association identifier includes at least one or more of the following: batch identifier, product identifier, and solder joint location identifier; Based on the data association identifier, data from different sources are associated and matched, and aligned according to a preset time window rule, so that data from different time scales form a corresponding relationship under the same association identifier; The aligned data is standardized according to a unified data format and dimensional standard to form a structurally consistent multi-source raw dataset for welding.

4. The method according to claim 1, wherein the tin paste is abnormal, and the method further comprises: determining the risk of the pin hole based on the tin paste and the soldering parameters. The set of basic features includes: wetting-driven state, structural forming state, interface bonding state, and internal defect state.

5. The method according to claim 4, wherein the tin paste is abnormal, and the method further comprises: determining the risk of the pin hole based on the tin paste abnormality. The S2 includes: a feature decoupling step, a deviation feature construction step, and a potential parameter construction step.

6. The method according to claim 5, wherein the tin paste is abnormal, and the method further comprises: determining the risk of the pin hole based on the tin paste abnormality. The feature decoupling step includes: The data in the original multi-source welding dataset are processed by interval mapping based on the preset process constraint interval or statistical distribution characteristics to obtain feature quantities with uniform dimensions and comparability. Based on the process physical meaning of the data and its role in the weld formation process, the feature quantities are divided into dimensions, and the data in the multi-source original dataset of welding are mapped to four dimensions: wetting drive state, structural forming state, interface bonding state and internal defect state. The specific mapping process is as follows: Based on the coupled correlation characterization process of reheat time and dispensing parameters, a wetting driving state is constructed to characterize the wetting driving capability. The structural forming state is obtained by processing the wetting angle, spreading area and bridging status in the weld joint appearance inspection data through feature association and consistency judgment. The interface bonding state is obtained by matching the maximum tensile force value in the post-weld mechanical property data with the preset bonding strength range. The internal defect state is obtained by spatially analyzing the void ratio, void size distribution, and void location in X-ray void detection data. The state data of four dimensions—wetting-driven state, structural forming state, interface bonding state, and internal defect state—are fused in a multidimensional manner to form a basic feature set containing multidimensional state characteristics.

7. The method for assessing the risk of PIN soldering voids under abnormal solder paste incoming material conditions according to claim 6, characterized in that, The deviation feature construction step includes: Based on the basic feature set, structural consistency deviation relationships are constructed to characterize the degree of inconsistency between states through coupling association and matching judgment processing between different states. Various structural consistency deviation relationships are then integrated to form a deviation feature set. Specifically, the construction of the structural consistency deviation relationship includes: Based on the coupling correlation analysis between the wetting-driven state and the structural forming state, the first type of deviation relationship is constructed by comparing and extracting the differences of the corresponding feature quantities of the two. Based on the matching determination process between the structural forming state and the interface bonding state, a second type of deviation relationship is constructed by performing consistency comparison and deviation identification processing on the corresponding state feature quantities. Based on the consistency comparison and judgment process between the interface bonding state and the internal defect state, a third type of deviation relationship is constructed by screening and anomaly identification of internal defect features in samples that meet the interface bonding conditions.

8. The method for assessing the risk of PIN soldering voids under abnormal solder paste incoming material conditions according to claim 7, characterized in that, The stability evolution potential parameters of the weld joint structure include: initial driving force potential, forming realization potential, and defect disturbance potential. The potential parameter construction steps include: The feature quantities corresponding to the wetting drive state in the basic feature set are numerically mapped according to the preset wetting drive capability quantification rules to obtain the initial driving force potential. By analyzing the correlation consistency of the characteristic quantities corresponding to the structural forming state and interface bonding state in the basic feature set, the forming realization potential is obtained. The feature quantities corresponding to the internal defect states in the basic feature set are combined with the third type of deviation relationship in the deviation feature set, and then comprehensively quantified according to the void location weight allocation rule, void size distribution classification rule, and the inverse correlation correction rule with the interface performance characterization result to obtain the defect perturbation potential. The initial driving force potential, the forming realization potential, and the defect disturbance potential are fused through multi-level processing using evolution constraint rules to obtain the stable evolution potential parameters of the weld joint structure. The evolution constraint rules include: when the initial driving potential is higher than a first set threshold and the forming realization potential is lower than a second set threshold, a potential reduction process is performed on the fusion result of the initial driving potential and the forming realization potential; when the forming realization potential is higher than a third set threshold and the defect disturbance potential is higher than a fourth set threshold, a suppression process is performed on the fusion result of the forming realization potential and the defect disturbance potential; when the third type of deviation relationship is in an abnormal state, a false high correction process is performed on the fusion result.

9. The method for assessing the risk of PIN soldering voids under abnormal solder paste incoming materials according to claim 1, characterized in that, The assessment report includes: incoming material release level, welding repair restrictions, and process modification recommendations.

10. The method for assessing the risk of PIN soldering voids under abnormal solder paste incoming material conditions according to claim 9, characterized in that, S3 includes: Based on the stability evolution potential parameters of the solder joint structure, the overall structural stability level of the solder paste under the current process conditions is classified and judged to form the corresponding stability evaluation results. Based on the deviation relationships of various structural consistency in the deviation feature set, the degree of abnormality of the weld joint in each stage of wetting drive, structural forming, interface bonding and internal defects is classified and identified to determine the corresponding risk source category and risk impact range. The stability evaluation results are correlated and matched with the risk source categories. Based on the preset risk classification rules, the risk of PIN soldering voids is classified into levels to form corresponding risk level results. Based on the risk level results, and combined with the distribution characteristics and impact of various deviation relationships, the availability of incoming materials is determined and processed to generate the corresponding material release level. Based on the risk source categories and risk impact scope, the operational constraints in the welding process are extracted and limited to generate corresponding repair welding constraints. Based on the structural inconsistency features reflected in the deviation feature set, the direction of process parameter adjustment is identified and processed, and combined with the stability evolution potential parameter of the weld joint structure, corresponding process correction suggestions are generated. The incoming material release level, resoldering restrictions, and process modification suggestions are integrated and output to form a PIN soldering void risk assessment report under abnormal solder paste incoming conditions.

Images