Soldering tin material pouring processing optimization method based on molding anomaly detection

Abstract

The invention relates to the technical field of soldering tin material processing, in particular to a soldering tin material pouring processing optimization method based on molding anomaly detection. The method comprises the following steps of: calculating an actual state transition probability of forming abnormity by outputting a forming abnormity signal spectrum in a soldering tin material pouring processing process, and constructing a hidden Markov model corresponding to an expected observation probability of forming abnormity when pouring processing equipment is used for pouring and producing a soldering tin material according to a standard process at the same time; and the hidden Markov model is utilized to identify the actual state transition probability so as to obtain abnormal forming elements causing abnormal forming of the soldering tin material in actual pouring processing, singular decomposition optimization is carried out on actual process parameters with the aim of eliminating the influence brought by the abnormal forming elements to the maximum extent, and a pouring process optimization scheme is obtained. According to the method, the technological parameters and the technological execution opportunity which cause abnormal molding in the pouring processing of the soldering tin material can be optimized, and the rework rate and the defective rate of abnormal molding in the pouring process of the soldering tin material are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of solder material processing, and particularly to an optimization method for solder material pouring processing based on forming anomaly detection. Background Art

[0002] With the continuous development of electronic products, welding technology plays a crucial role in the assembly of electronic components. The use of solder materials not only ensures good connection between electronic components but also plays an important role in ensuring the performance and reliability of products. However, during the welding process, due to the influence of various factors (such as temperature, time, pressure, etc.), the forming quality of solder materials may have defects, which may lead to poor welding connections and affect the performance and lifespan of products. Solder pouring processing, as an important method in welding technology, is widely used in the production of electronic products. During the pouring process, factors such as the fluidity, adhesiveness, and pouring speed of solder may lead to inconsistencies in solder forming, which not only affects the welding quality but also increases the rejection rate. Therefore, how to precisely optimize the process performance during solder pouring to eliminate forming anomalies has become an important research direction for improving product quality and reducing production costs.

[0003] Currently, many methods for optimizing forming anomalies during solder material pouring processing still rely on manual optimization or simple automatic optimization systems. Manual optimization and traditional automatic optimization systems based on forming anomaly detection often rely on experience and simple algorithms, making it difficult to accurately capture the random evolution law of forming anomalies to trace the forming elements that cause them, and easily resulting in inaccurate subsequent process optimization; at the same time, traditional optimization methods are difficult to quickly and accurately locate the process parameter parts that need to be optimized for the traced forming elements, so global calculation is required for optimization, which wastes a lot of time and computing costs; moreover, traditional optimization methods are difficult to self-check and optimize the forming anomaly phenomena caused by the process execution timing of pouring processing equipment, greatly reducing the process performance of the equipment for solder material pouring, increasing the rejection rate and defective rate of solder material pouring processing. Summary of the Invention

[0004] The present invention overcomes the deficiencies of the prior art and provides an optimization method for solder material pouring processing based on forming anomaly detection.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] The first aspect of the present invention provides an optimization method for solder material pouring processing based on forming anomaly detection, including the following steps:

[0007] S102: Obtain the forming abnormal signal spectrum of the solder material pouring process, describe the direct dependence relationship of random state variables according to the forming abnormal signal spectrum, and construct a forming abnormal undirected graph model based on the node pattern diagram. Calculate the joint probability distribution of the forming abnormal undirected graph model based on the standard technological process of the pouring process, and obtain the actual state transition probability of the forming abnormality caused by the output of the forming abnormal signal spectrum during the solder material pouring process;

[0008] S104: Calculate the hidden energy distribution of the forming abnormal signal on the solder material with the benchmark critical threshold interval when each forming element in each forming processing link has a forming abnormality as the target. Calculate, update and iterate to build a hidden Markov model of the forming abnormality of the solder material based on the expected observation probability of the pouring processing forming of the hidden energy distribution. Use the hidden Markov model to decode the actual state transition probability to obtain the abnormal forming elements that cause the forming abnormality of the solder material during the actual pouring process;

[0009] S106: Establish a permitted forming abnormality definition model for the abnormal forming elements. Calculate and determine the elimination singular value of each actual forming abnormality model to the maximum extent to reach the permitted forming abnormality definition model based on the termination and elimination benchmark of the forming abnormality. Perform singular value decomposition on the diagonal matrix of each actual process parameter after elimination based on the elimination singular value to obtain the pouring process optimization plan;

[0010] S108: If a forming abnormality is still detected during the simulated pouring process of the solder material based on the pouring process optimization plan, obtain the simulated signal spectrum. Determine the start and end timing nodes of the abnormal execution according to the hash misalignment between the normal signal spectrum of the abnormal process index and the simulated signal spectrum. Calculate the deviation of the abnormal execution timing compared with the standard execution timing based on the start and end timing nodes of the abnormality to optimize the preset process control strategy of the pouring processing equipment.

[0011] More specifically, the step S102 specifically includes the following steps:

[0012] Perform real-time detection on the pouring process of the solder material through the forming abnormality monitoring Internet of Things to obtain the forming abnormal signal spectrum of the solder material and the residence generation point of each forming abnormal signal in the forming abnormal signal spectrum;

[0013] Define the random state variables of the forming abnormality according to the forming abnormal signal, extract the direct dependence relationship between each random state variable when the solder material has a forming abnormality based on the forming abnormal signal spectrum, and determine the node pattern diagram of the random state variables based on the residence generation point of each forming abnormal signal;

[0014] Connect and depict the adjacent undirected edges of each random state variable in the node pattern diagram according to the direct dependency relationship, and construct an undirected graph model of the forming anomaly of the random state variable when the solder material has a forming anomaly through the connected adjacent undirected edges;

[0015] Obtain the casting processing forming process of the solder material, and obtain the standard process flow of the casting processing forming process. Based on the standard process flow, formulate a casting evaluation criterion for no forming anomaly, and assign a Gaussian potential function to the undirected graph model of the forming anomaly based on the casting evaluation criterion;

[0016] After the assignment is completed, obtain the Gaussian potential function of each adjacent undirected edge in the undirected graph model of the forming anomaly, which is defined as the marginal Gaussian potential function;

[0017] Multiply the marginal Gaussian potential functions corresponding to all adjacent undirected edges to obtain the joint probability distribution of the undirected graph model of the forming anomaly, and determine the actual state transition probability of the forming anomaly caused by the output of the forming anomaly signal spectrum during the casting processing of the solder material according to the joint probability distribution.

[0018] More specifically, the step S104 specifically includes the following steps:

[0019] Based on the standard process flow, obtain the reference critical threshold interval when the forming elements of each forming processing link for the casting of the solder material have forming anomalies. Calculate the energy distribution based on the historical casting processing forming signals within the reference critical threshold interval to obtain the hidden energy distribution of the forming anomaly signals on the solder material;

[0020] According to the reference critical threshold interval, preset the expected observation state when each forming element has a forming anomaly. Calculate the probability that each historical casting processing forming signal generates the expected observation state based on the hidden energy distribution to obtain the expected observation probability;

[0021] Introduce the maximum likelihood algorithm, and update and iterate the state transition probability matrix and the observation probability matrix for the historical casting processing forming signals to generate the expected observation state in the maximum likelihood algorithm through the expected observation probability, and output the current log-likelihood function;

[0022] If the current log-likelihood function is greater than the preset log-likelihood function, it is determined that the updated state transition probability matrix and the observation probability matrix are in a convergent form. At this time, terminate the continuous update and iteration operation to obtain the hidden Markov parameter architecture for a single historical casting processing forming signal to trigger the forming anomaly in the reference critical threshold interval, which is marked as the single hidden Markov parameter architecture;

[0023] Continuously repeat the above steps for the state transition probability matrix and the observation probability matrix update iteration and convergence determination operation of the historical pouring processing and forming signal to generate the desired observation state until all historical pouring processing and forming signals are processed, and generate all single hidden Markov parameter architectures;

[0024] Build a hidden Markov model for the abnormal pouring processing and forming of solder materials through all single hidden Markov parameter architectures, introduce the Viterbi algorithm, and decode the actual state transition probability in the hidden Markov model through the Viterbi algorithm to obtain the actual hidden state sequence;

[0025] Obtain the preset hidden state sequence set of each forming element, construct an identification map of the forming element - preset hidden state sequence, and match and identify the actual hidden state sequence through the identification map to obtain one or more forming elements corresponding to the abnormal forming during the actual pouring processing of the solder material, which are defined as abnormal forming elements.

[0026] More specifically, based on the standard process flow, obtain the reference critical threshold interval when the forming elements of the solder material pouring in each forming processing link show forming abnormalities, and calculate the energy distribution based on the historical pouring processing and forming signals in the reference critical threshold interval to obtain the hidden energy distribution of the forming abnormal signals on the solder material. The specific steps are as follows:

[0027] Obtain the pouring processing log of the pouring processing equipment for the solder material and the standard design drawing of the target solder material, and construct a standard pouring processing model of the solder material in the SolidWorks model design software according to the standard design drawing;

[0028] Based on the standard process flow, strip out several forming processing links when the solder material performs the pouring processing and forming process, and divide the standard pouring processing model into N sub - standard pouring processing models according to the several forming processing links;

[0029] Obtain the forming elements of the solder material pouring in each forming processing link, and obtain the reference critical threshold interval for each forming element to cause forming abnormalities in the pouring of each sub - standard pouring processing model; among them, the forming elements include the quality, temperature control, pressure control, and pouring speed of the solder material;

[0030] Extract the historical pouring processing and forming signals detected by the pouring processing equipment when pouring the standard pouring processing model of the solder material within a preset time period and being in the reference critical threshold interval corresponding to each forming element one by one through the pouring processing log;

[0031] A preset wavelet function is used to perform multi-scale decomposition of discrete wavelet transform on the historical casting processing and forming signal, so as to obtain the hidden energy distribution of the discrete approximation signal and the discrete detail signal corresponding to the output of each sub-standard casting processing model.

[0032] More specifically, the step S106 specifically includes the following steps:

[0033] Obtain the allowable forming anomaly boundary values of one or more of the abnormal forming elements, and establish an allowable forming anomaly boundary model for one or more of the abnormal forming elements according to the allowable forming anomaly boundary values;

[0034] Construct an actual forming anomaly model for one or more of the abnormal forming elements through the forming anomaly signal spectrum of the solder material. Taking the allowable forming anomaly boundary model as the termination elimination reference, calculate the covariance matrix of the actual forming anomaly model compared with the allowable forming anomaly boundary model based on the termination elimination reference;

[0035] Introduce the singular value decomposition algorithm to calculate multiple forming anomaly elimination eigenvalues and multiple forming anomaly elimination eigenvectors of the covariance matrix, and determine the elimination singular values for each actual forming anomaly model to maximize the elimination to reach the allowable forming anomaly boundary model according to the forming anomaly elimination eigenvalues and the forming anomaly elimination eigenvectors;

[0036] Use multiple forming anomaly elimination eigenvalues to generate an elimination eigenvalue matrix, and simultaneously use multiple forming anomaly elimination eigenvectors to generate an elimination eigenvector matrix;

[0037] Obtain the actual process parameters of each actual forming anomaly model generated by one or more of the abnormal forming elements when the casting processing equipment has a forming anomaly during the actual production of the solder material through the casting processing log. Based on the elimination singular values, describe the diagonal elements of each actual process parameter after elimination, and construct a diagonal matrix through a number of the described diagonal elements;

[0038] Use the elimination eigenvalue matrix and the elimination eigenvector matrix to perform singular decomposition on the diagonal matrix in the singular value decomposition algorithm to generate the final decomposition result, and determine the process optimization parameters required for each actual process parameter when each actual forming anomaly model maximally eliminates to reach the allowable forming anomaly boundary model according to the final decomposition result, so as to obtain the casting process optimization plan.

[0039] More specifically, the step S108 specifically includes the following steps:

[0040] Construct a simulation model of the casting processing equipment, upload the casting process optimization plan to the control terminal of the casting processing equipment, and perform casting processing simulation on the solder material by executing the casting process optimization plan through the simulation model;

[0041] If abnormal forming during the casting process of the solder material is still detected, an abnormal forming simulation signal detected during the simulation process is obtained at this time. The Fourier transform algorithm is introduced to extract the features of the abnormal forming simulation signal, and the simulation signal spectrum is obtained.

[0042] A number of process test cases for the casting process of the solder material by the casting processing equipment are obtained. An abnormal forming simulation model is constructed according to the simulation signal spectrum. Synchronously, the detection landing area of the simulation signal spectrum is planned based on the position of the abnormal forming simulation model on the standard casting model of the solder material.

[0043] One or more process indicators associated with the occurrence of abnormal forming in the detection landing area are obtained through a number of process test cases, which are defined as abnormal process indicators, and the preset process control strategy executed during the simulation of the casting processing equipment is obtained.

[0044] The simulation process execution time line of each abnormal process indicator is extracted through the process control strategy, and the signal spectrum detected by the casting processing equipment without abnormal forming in the detection landing area is obtained, which is defined as the normal signal spectrum.

[0045] The hash error function of the signal spectrum error of the simulation signal spectrum relative to the normal signal spectrum is calculated to obtain the hash misalignment function byte. By analyzing the hash misalignment function byte, the start and end timing nodes when the detection landing area begins to have abnormal forming are determined, which are defined as the execution abnormal start and end timing nodes.

[0046] The corresponding time segment is intercepted on the simulation process execution time line through the execution abnormal start and end timing nodes, which is marked as the abnormal execution timing segment. At the same time, the correct process execution time line of each abnormal process indicator when the casting processing equipment executes the preset process control strategy is obtained.

[0047] The sub - segment corresponding to the abnormal execution timing segment is intercepted on the correct process execution time line, which is marked as the standard execution timing segment. The deviation between the abnormal execution timing segment and the standard execution timing segment is calculated to obtain the process execution timing deviation degree. Based on the process execution timing deviation degree, the preset process control strategy of the casting processing equipment is optimized.

[0048] The second aspect of the present invention provides a solder material casting processing optimization system based on abnormal forming detection. The solder material casting processing optimization system includes a memory and a processor. A solder material casting processing optimization method program based on abnormal forming detection is stored in the memory. When the solder material casting processing optimization method program is executed by the processor, the steps of any of the solder material casting processing optimization methods described above are implemented.

[0049] The present invention solves the technical defects existing in the background art, and the beneficial technical effects of the present invention are as follows:

[0050] Obtain the abnormal signal spectrum of the solder material casting process. Construct an undirected graph model of forming anomalies based on the random state variables described by the abnormal signal spectrum. Calculate the joint probability distribution of the undirected graph model of forming anomalies based on the standard process flow of casting processing to obtain the actual state transition probability of forming anomalies caused by the output of the abnormal signal spectrum during the solder material casting process. Calculate the hidden energy distribution of the abnormal signal on the solder material with the reference critical threshold interval when each forming element has a forming anomaly as the target. Calculate, update and iterate to build a hidden Markov model of the forming anomaly of the solder material based on the expected observation probability of the casting processing forming with the hidden energy distribution. Use the hidden Markov model to decode the actual state transition probability to obtain the abnormal forming elements that cause the forming anomaly of the solder material during the actual casting process. Establish an allowable forming anomaly definition model for the abnormal forming elements. Calculate and determine the elimination singular values that can maximize the elimination of each actual forming anomaly model to reach the allowable forming anomaly definition model based on the termination elimination criterion of the forming anomaly. Perform singular value decomposition on the diagonal matrix of each actual process parameter after elimination based on the elimination singular values to obtain the casting process optimization plan. If forming anomalies are still detected during the simulated casting process of the solder material based on the casting process optimization plan, obtain the simulated signal spectrum, and calculate the deviation of the abnormal execution timing compared with the standard execution timing based on the execution abnormal start and end timing nodes of the hash misalignment between the normal signal spectrum and the simulated signal spectrum of the abnormal process index to optimize the preset process control strategy of the casting processing equipment. The present invention can optimize the process parameters and process execution timing that cause forming anomalies in the solder material casting process to eliminate the forming anomaly phenomenon in the solder material casting process and reduce the rework rate and defective rate of the solder material casting. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0052] Figure 1 Shows the first method flow chart of an optimization method for solder material casting processing based on forming anomaly detection;

[0053] Figure 2 Shows the second method flow chart of an optimization method for solder material casting processing based on forming anomaly detection;

[0054] Figure 3Shows the system framework diagram of a solder material pouring processing optimization system based on forming anomaly detection. Detailed implementation manners

[0055] In order to be able to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below in conjunction with the drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0056] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0057] The first aspect of the present invention provides a solder material pouring processing optimization method based on forming anomaly detection, as Figure 1 shown, including the following steps:

[0058] S102: Obtain the forming anomaly signal spectrum of the solder material pouring processing, construct a forming anomaly undirected graph model according to the direct dependence relationship of the random state variables described by the forming anomaly signal spectrum and the node pattern graph, calculate the joint probability distribution of the forming anomaly undirected graph model based on the standard process flow of the pouring processing, and obtain the actual state transition probability of the forming anomaly caused by the output of the forming anomaly signal spectrum during the solder material pouring processing;

[0059] S104: Calculate the hidden energy distribution of the forming anomaly signal on the solder material with the target of the reference critical threshold interval when each forming element in each forming processing link has a forming anomaly. Calculate, update and iterate to build a hidden Markov model of the solder material forming anomaly based on the expected observation probability of the pouring processing forming of the hidden energy distribution, and use the hidden Markov model to decode the actual state transition probability to obtain the abnormal forming elements that cause the forming anomaly of the solder material in the actual pouring processing;

[0060] S106: Establish an allowable forming anomaly definition model for the abnormal forming elements, calculate and determine the elimination singular values of each actual forming anomaly model that can be maximally eliminated to reach the allowable forming anomaly definition model based on the termination and elimination reference of the forming anomaly, and perform singular value decomposition on the diagonal matrix of each actual process parameter after elimination based on the elimination singular values to obtain the pouring process optimization scheme;

[0061] S108: If molding anomalies are still detected during the simulated casting process of the solder material based on the optimized casting process plan, obtain the simulated signal spectrum. Determine the start and end timing nodes of the anomaly execution based on the hash misalignment between the normal signal spectrum of the abnormal process index and the simulated signal spectrum. Calculate the deviation of the anomaly execution timing compared to the standard execution timing based on the start and end timing nodes of the anomaly, so as to optimize the preset process control strategy of the casting processing equipment.

[0062] More specifically, the step S102 specifically includes the following steps:

[0063] Perform real-time detection on the casting process of the solder material through the molding anomaly monitoring Internet of Things to obtain the molding anomaly signal spectrum of the solder material and the residence generation point of each molding anomaly signal in the molding anomaly signal spectrum;

[0064] Define the random state variables of the molding anomaly according to the molding anomaly signal. Extract the direct dependence relationship between each random state variable when the solder material has a molding anomaly based on the molding anomaly signal spectrum, and determine the node pattern diagram of the random state variables based on the residence generation point of each molding anomaly signal;

[0065] Connect and depict the adjacent undirected edges of each random state variable in the node pattern diagram according to the direct dependence relationship, and construct an undirected graph model of the molding anomaly of the random state variables when the solder material has a molding anomaly through the connected adjacent undirected edges;

[0066] Obtain the casting processing molding process of the solder material, and obtain the standard process flow of the casting processing molding process. Formulate a casting evaluation criterion for no molding anomalies based on the standard process flow, and assign the Gaussian potential function of the undirected graph model of the molding anomaly based on the casting evaluation criterion;

[0067] After the assignment is completed, obtain the Gaussian potential function of each adjacent undirected edge in the undirected graph model of the molding anomaly, which is defined as the marginal Gaussian potential function;

[0068] Multiply and calculate the marginal Gaussian potential functions corresponding to all adjacent undirected edges to obtain the joint probability distribution of the undirected graph model of the molding anomaly. Determine the actual state transition probability of the molding anomaly caused by the output of the molding anomaly signal spectrum during the casting process of the solder material according to the joint probability distribution.

[0069] It should be noted that due to the characteristics of the solder material, some abnormal phenomena are likely to occur during the casting and processing forming process. For example, bubbles of uneven size appear on the surface of the solder material, cracks with different lengths and depths, and local material embrittlement caused by uneven heating. These forming abnormal phenomena do not have traceable regular characteristics. Their distribution and morphology on the solder material are determined by the rationality of the process during casting and processing, showing a certain randomness of forming abnormalities. Therefore, if we want to accurately optimize the casting and processing of the solder material, we need to explore the state transition probability of the actual forming abnormalities caused by the casting and processing technology based on the detection results of the forming abnormalities. In this regard, this method obtains the direct dependence relationship and node pattern diagram between the random state variables corresponding to each forming abnormal signal through each forming abnormal signal recorded in the forming abnormal signal spectrum of the solder material and its residence generation point. Then, an undirected graph model reflecting the variable dependence relationship is constructed by connecting relevant variables with undirected edges, making the dependence relationship of the forming abnormalities clearer and more visual, avoiding unnecessary calculations, improving the reasoning reliability of the state change of the forming abnormalities caused by the casting and processing technology, and revealing the evolution law of the casting and processing technology for the random state variables of the forming abnormalities. Then, according to the standard process flow of the solder material casting and processing without forming abnormalities, a Gaussian potential function, that is, a marginal Gaussian potential function, is assigned to each adjacent undirected edge of the forming abnormal undirected graph model, so as to quantify the interaction between random state variables, describe the local characteristics of the joint distribution of random state variables, and ensure that the probability distribution of the forming abnormal state evolution can express the dependence relationship with the actual casting and processing. Finally, the actual state transition probability of the forming abnormalities caused by the output of the forming abnormal signal spectrum during the casting and processing of the solder material is determined according to the joint probability distribution calculated by multiplying the marginal Gaussian potential functions of all adjacent undirected edges. Through this method, the transition probability of the forming abnormal state deduction of the casting and processing standard process can be calculated according to the forming abnormal signals detected during the casting and processing of the solder material, thereby providing accurate identification data for the subsequent identification of the elements causing the actual forming abnormalities and improving the optimization accuracy of the solder material casting and processing based on the actual forming abnormal conditions. Among them, the Chinese name of the Gaussian potential function is the Gaussian potential function.

[0070] More specifically, the step S104 specifically includes the following steps:

[0071] Based on the standard process flow, obtain the benchmark critical threshold interval when the forming elements of the solder material casting in each forming processing link show forming abnormalities, and calculate the energy distribution based on the historical casting and processing forming signals within the benchmark critical threshold interval to obtain the hidden energy distribution of the forming abnormal signals on the solder material;

[0072] Preset the expected observation state when forming anomalies occur for each forming element according to the reference critical threshold interval, calculate the probability of generating the expected observation state for each historical casting and processing forming signal based on the hidden energy distribution, and obtain the expected observation probability;

[0073] Introduce the maximum likelihood algorithm, update and iterate the state transition probability matrix and the observation probability matrix for generating the expected observation state in the maximum likelihood algorithm through the expected observation probability, and output the current log-likelihood function;

[0074] If the current log-likelihood function is greater than the preset log-likelihood function, determine that the updated state transition probability matrix and the observation probability matrix are in a convergent form. At this time, terminate the continuous update and iteration operation, and obtain the hidden Markov parameter architecture for a single historical casting and processing forming signal to trigger a forming anomaly in the reference critical threshold interval, which is marked as the single hidden Markov parameter architecture;

[0075] Continuously repeat the above steps for the state transition probability matrix and the observation probability matrix update iteration and convergence determination operations for generating the expected observation state of the historical casting and processing forming signal until all historical casting and processing forming signals are processed, and generate all single hidden Markov parameter architectures;

[0076] Build a hidden Markov model for the forming anomalies of the solder material casting and processing through all single hidden Markov parameter architectures, introduce the Viterbi algorithm, and decode the actual state transition probability in the hidden Markov model through the Viterbi algorithm to obtain the actual hidden state sequence;

[0077] Obtain the preset hidden state sequence set of each forming element, construct an identification map of the forming element - preset hidden state sequence, and match and identify the actual hidden state sequence through the identification map to obtain one or more forming elements corresponding to the forming anomalies that occur in the actual casting and processing of the solder material, which are defined as abnormal forming elements.

[0078] It should be noted that there are many factors that cause abnormal forming in the casting process of solder materials. For example, improper temperature control parameters, unreasonable pressure injection process, or incorrect pouring rate, etc. These are all decisive factors for different abnormal forming phenomena. Therefore, by analyzing and identifying the state transition probability of the evolution of abnormal forming during the actual casting process of solder materials, the key factors causing abnormal forming of solder materials in the casting process can be traced back. However, most existing methods still cannot achieve the state transition probability of abnormal forming, and for the small part of methods that can achieve this function, their traceability and identification accuracy are relatively low, unable to meet the process optimization requirements of solder material casting. In response to this, this method first calculates the hidden energy distribution of the abnormal forming signal on the solder material, then calculates the expected observation probability of each historical casting process forming signal generating the expected observation state based on the hidden energy distribution, and then updates and iterates the state transition probability matrix and the observation probability matrix of the historical casting process forming signal generating the expected observation state in the maximum likelihood algorithm until the current log-likelihood function converges, obtaining a single hidden Markov parameter architecture where the forming abnormality starts to appear at the benchmark critical threshold interval triggered by a historical casting process forming signal, thereby constructing a hidden Markov model with high adaptability for traceability and identification of the forming factors causing forming abnormalities. This hidden Markov model represents the hidden states where the solder material starts to show forming abnormalities during casting according to different forming factors in the standard process flow, that is, it reveals the random laws and patterns of the evolution of forming abnormalities under different forming factor constraints, can be applied to the identification of calculation results of different forms of state transition probabilities, and can solve the problems that existing methods still cannot adaptively trace, identify and analyze the state transition probability of forming abnormalities and have relatively low traceability and identification accuracy.

[0079] It should be noted that by introducing the Viterbi algorithm to decode the actual state transition probability in the hidden Markov model, an actual hidden state sequence can be obtained. This actual hidden state sequence is a unique label for tracing the forming factors that cause abnormal forming in the casting process of solder materials, which can improve the identification accuracy of abnormal forming factors. Through this method, a hidden Markov model can be constructed to describe the different forming factors of the solder material in the standard process flow starting to show forming abnormalities, so as to specifically identify the actual state transition probability, improve the traceability fineness of the corresponding forming factors causing abnormal forming in the casting process of solder materials, and help improve the reliability of subsequent process optimization. Compared with existing methods, the process optimization of solder material casting is more efficient, stable and accurate.

[0080] More specifically, the reference critical threshold range is obtained for each forming and processing link for the casting of solder materials when forming anomalies occur in the forming elements. Based on the historical casting processing forming signals within the reference critical threshold range, the energy distribution is calculated to obtain the hidden energy distribution of the forming anomaly signals on the solder materials. The specific steps are as follows:

[0081] Obtain the casting processing log of the casting processing equipment for the solder materials and the standard design drawing of the target solder materials. Construct a standard casting processing model of the solder materials in the SolidWorks model design software according to the standard design drawing;

[0082] Based on the standard process flow, strip out several forming and processing links when the solder materials perform the casting processing forming process. Divide the standard casting processing model into N sub-standard casting processing models according to the several forming and processing links;

[0083] Obtain the forming elements of each forming and processing link for the casting of solder materials, and obtain the reference critical threshold range for each forming element to cause forming anomalies in the casting of each sub-standard casting processing model; wherein, the forming elements include the quality, temperature control, pressure control and casting speed of the solder materials;

[0084] Extract, through the casting processing log, the historical casting processing forming signals detected by the casting processing equipment when casting the standard casting processing model of the solder materials within a preset time period and successively in the reference critical threshold range corresponding to each forming element;

[0085] Preset a wavelet function, and perform multi-scale decomposition of discrete wavelet transform on the historical casting processing forming signals according to the wavelet function to obtain the hidden energy distribution of the discrete approximation signals and discrete detail signals output corresponding to each sub-standard casting processing model.

[0086] It should be noted that since the detection of molding anomalies is mainly output in the form of signals, and the existing methods still have technical gaps in interpreting and analyzing the random laws of the evolution of molding anomalies based on the signal form, this may cause the subsequent hidden Markov model constructed to reveal the random deduction of the molding abnormal state to have a large recognition error, affecting the process optimization of the solder material pouring process. Therefore, this method refines the model area according to the pouring processing model of the solder material standard in the molding process link. In other words, it is to zoning plan the local solder material area that may be affected by each molding process link, that is, the sub-standard pouring processing model, which can improve the definition accuracy of the critical threshold when different molding abnormal defects appear in different solder material areas, and improve the rigor and credibility of the evolution of the random change law of the molding abnormal state under the influence of different molding factors by the hidden Markov. Then, by performing discrete wavelet transform on the different scale frequency components of the historical pouring and molding signals detected when the pouring and molding equipment pours the solder material within a preset time period and is in the benchmark critical threshold interval corresponding to each molding element one by one, the hidden energy distribution of the output discrete approximate signal and the discrete detail signal corresponding to each sub-standard pouring and molding model can be obtained. The hidden energy distribution is the energy value hidden in the detected molding abnormality signal frequency band under the influence of the molding element, which can help identify the pattern of the molding abnormality state transition probability output in the form of a signal, thereby ensuring the adaptability and accuracy of the constructed hidden Markov model for the traceability and identification of abnormal molding elements in the form of signals.

[0087] More specifically, the step S106 is as follows: Figure 2 As shown, the specific steps include:

[0088] S202: Obtaining the allowable forming abnormality limit values of one or more abnormal forming elements, and establishing one or more allowable forming abnormality limit models of the abnormal forming elements according to the allowable forming abnormality limit values;

[0089] S204: constructing an actual forming abnormality model of one or more abnormal forming elements through the forming abnormality signal spectrum of the solder material, taking the allowed forming abnormality definition model as a termination elimination benchmark, and calculating a covariance matrix of the actual forming abnormality model compared with the allowed forming abnormality definition model based on the termination elimination benchmark;

[0090] S206: introducing a singular value decomposition algorithm to calculate a plurality of molding anomaly elimination eigenvalues and a plurality of molding anomaly elimination eigenvectors of the covariance matrix, and determining, according to the molding anomaly elimination eigenvalues and the molding anomaly elimination eigenvectors, to eliminate the elimination singular values of each actual molding anomaly model to the maximum extent possible to achieve the allowed molding anomaly definition model;

[0091] S208: Generate an elimination eigenvalue matrix by using multiple forming anomaly elimination eigenvalues, and simultaneously generate an elimination eigenvector matrix by using multiple forming anomaly elimination eigenvectors;

[0092] S210: Obtain, from the pouring process log, the actual process parameters of one or more abnormal forming elements that generate each actual forming anomaly model when the pouring processing equipment has a forming anomaly during the actual production of the solder material. Based on the described elimination singular values, describe the diagonal elements of each actual process parameter after elimination, and construct a diagonal matrix through a number of described diagonal elements;

[0093] S212: Use the elimination eigenvalue matrix and the elimination eigenvector matrix to perform singular value decomposition on the diagonal matrix in the singular value decomposition algorithm to generate a final decomposition result. According to the final decomposition result, determine the process optimization parameters required for each actual process parameter when each actual forming anomaly model is maximally eliminated to reach the allowable forming anomaly definition model, and obtain a pouring process optimization plan.

[0094] It should be noted that after tracing the abnormal forming elements, the process of pouring and processing the solder material can be optimized. However, existing methods are difficult to accurately locate the process parameters to be optimized only based on the obtained abnormal forming elements, and the calculation amount is large, and the process optimization efficiency is greatly limited and reduced. In response to this, this method first establishes an allowable forming anomaly definition model for the abnormal forming elements through the allowable forming anomaly definition values of the abnormal forming elements. This allowable forming anomaly definition model is the minimum allowable forming error during the pouring and processing of the solder material and is the minimum goal for eliminating actual forming anomalies. Then, using the allowable forming anomaly definition model as the termination elimination benchmark, calculate the covariance matrix of the actual forming anomaly model compared to the allowable forming anomaly definition model. This covariance matrix can reveal the main process components or directions required for optimizing the elimination of actual forming anomalies, improving the accuracy of process optimization. Then, use the singular value decomposition algorithm to calculate the forming anomaly elimination eigenvalues and forming anomaly elimination eigenvectors of this covariance matrix. Solving the eigenvalues and eigenvectors of the covariance matrix means finding the main process component directions for pouring process optimization and the contribution degree of each direction. Thus, matrices can be generated respectively based on the directions and contribution degrees to quickly locate the optimization process parameters, that is, the elimination eigenvalue matrix and the elimination eigenvector matrix, improving the positioning accuracy and positioning optimization efficiency of the pouring and processing process optimization of the solder material.

[0095] It should be noted that since it is difficult for existing process optimization methods to achieve the effect of synchronously corresponding to optimize process parameters according to the elimination accuracy of maximizing the elimination of the actual forming abnormality model to the allowable forming abnormality definition model, the optimization of the casting process is inaccurate, and there may be certain optimization errors, resulting in the recurrence of forming abnormalities in the subsequent casting process of the solder material. In response to this, this method determines the elimination singular value of each actual forming abnormality model to maximize the elimination to reach the allowable forming abnormality definition model according to the forming abnormality elimination eigenvalue and the forming abnormality elimination eigenvector; based on the elimination singular value, the diagonal elements of each actual process parameter after elimination in the casting process are described, and a diagonal matrix is constructed in this way, which projects the maximum elimination progress of the actual forming abnormality model against the allowable forming abnormality definition model into a new actual process parameter optimization space, maximizing the variance or information volume of the data in each direction, and improving the optimization synchronization and timeliness of process parameters in the elimination of forming abnormalities. Finally, by using the elimination eigenvalue matrix and the elimination eigenvector matrix to decompose the singular value constructed by the diagonal matrix, the process optimization parameters required for each actual process parameter when each actual forming abnormality model maximally eliminates to reach the allowable forming abnormality definition model can be determined. On the one hand, this method can achieve the optimization target effect of maximizing the elimination of the actual forming abnormality to the allowable forming abnormality, improving the optimization effect; on the other hand, it can optimize the timeliness synchronization of process parameters during the process of maximizing the elimination of the actual forming abnormality, improve the accuracy and reliability of the casting process optimization, avoid the forming abnormality phenomenon, reduce the rework rate and defective rate of the solder material forming abnormality, and improve the economic benefits of the solder material casting process.

[0096] More specifically, the step S108 specifically includes the following steps:

[0097] Construct a simulation model of the casting processing equipment, upload the casting process optimization plan to the control terminal of the casting processing equipment, and perform the casting process optimization plan through the simulation model to simulate the casting process of the solder material;

[0098] If the forming abnormality in the casting process of the solder material is still detected, then obtain the forming abnormality simulation signal detected during the simulation process at this time, and introduce the Fourier transform algorithm to extract the characteristics of the forming abnormality simulation signal to obtain the simulation signal spectrum;

[0099] Obtain several process test cases of the casting processing equipment for the casting process of the solder material, construct a simulated forming abnormality model according to the simulation signal spectrum, and synchronously plan the detection landing area of the simulation signal spectrum according to the position of the simulated forming abnormality model on the standard casting model of the solder material;

[0100] One or more process indicators associated with forming anomalies in the detection landing area are obtained through a number of process test cases, defined as abnormal process indicators, and the preset process control strategy executed during the simulation of the casting processing equipment is obtained;

[0101] The simulated process execution timeline of each abnormal process indicator is extracted through the process control strategy, and the signal spectrum detected by the casting processing equipment without forming anomalies in the detection landing area is obtained, defined as the normal signal spectrum;

[0102] The hash error function of the signal spectrum error of the simulated signal spectrum relative to the normal signal spectrum is calculated to obtain the hash misalignment function byte, and the start and end timing nodes when the forming anomaly begins to occur in the detection landing area are determined by analyzing the hash misalignment function byte, defined as the execution abnormal start and end timing nodes;

[0103] The corresponding time segment is intercepted on the simulated process execution timeline through the execution abnormal start and end timing nodes, marked as the abnormal execution timing segment, and at the same time, the correct process execution timeline of each abnormal process indicator when the casting processing equipment executes the preset process control strategy is obtained;

[0104] The sub - segment corresponding to the abnormal execution timing segment is intercepted on the correct process execution timeline, marked as the standard execution timing segment, the deviation between the abnormal execution timing segment and the standard execution timing segment is calculated to obtain the process execution timing deviation degree, and the preset process control strategy of the casting processing equipment is optimized based on the process execution timing deviation degree.

[0105] It should be noted that the pouring process optimization scheme is used to optimize the pouring process of the solder material to eliminate the abnormal forming. However, if the abnormal forming signal can still be detected during the pouring process, it indicates that under the condition of process stability, the control deviation of the pouring processing equipment may cause the abnormal forming not to be eliminated. Since the abnormal forming of the solder material mainly depends on the execution timing of the pouring process, for example, at a certain time node, a process of increasing the pressure input amount needs to be executed to avoid abnormal forming. However, if the process execution timing is disordered or incorrect, the probability of abnormal forming will be greatly increased. Therefore, it is necessary to optimize the process execution timing of the pouring processing equipment. In this regard, this method simulates the pouring process by executing the pouring process optimization scheme on the simulation model of the pouring processing equipment, and extracts the features of the detected abnormal forming simulation signal through Fourier transform to obtain the simulation signal spectrum. The specific form of the abnormal forming in the detection landing area can be reflected by this simulation signal spectrum, and this form is the basis and basis for determining the error node of the process execution timing. Then, the start and end timing nodes of the execution abnormality where the abnormal forming starts to occur in the detection landing area are determined by calculating the hash misalignment function byte of the signal spectrum error between the simulation signal spectrum and the normal signal spectrum. The process execution timing included in the start and end timing nodes of the execution abnormality has deviated from the correct timing. Therefore, based on the start and end timing nodes of the execution abnormality, the corresponding incorrect process execution time lines on the simulated process execution timeline and the correct process execution timeline are respectively intercepted, that is, the abnormal execution timing paragraph and the standard execution timing paragraph. According to the process execution timing deviation degree between the two, the problem of process execution timing disorder or error caused by the preset process control strategy of the pouring processing equipment can be optimized. Through this method, the process execution timing deviation when the pouring processing equipment executes the preset process control strategy can be optimized to repair the abnormal forming phenomenon caused by the equipment, improve the process performance of the pouring processing equipment, and achieve the self-inspection and self-optimization effect of the pouring processing equipment for the abnormal forming of the solder material.

[0106] The second aspect of the present invention provides a solder material pouring processing optimization system based on abnormal forming detection, as Figure 3 shown. The solder material pouring processing optimization system includes a memory 31 and a processor 32. A program of a solder material pouring processing optimization method based on abnormal forming detection is stored in the memory 31. When the program of the solder material pouring processing optimization method is executed by the processor 32, the steps of any one of the solder material pouring processing optimization methods are implemented.

[0107] The above are only the specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. An optimization method for the casting process of solder materials based on forming anomaly detection, characterized in that It includes the following steps: S102: Obtain the abnormal signal spectrum of the solder material casting process. Based on the abnormal signal spectrum, describe the direct dependence of random state variables and construct an undirected graph model of abnormal forming. Calculate the joint probability distribution of the undirected graph model of abnormal forming based on the standard process flow of casting processing, and obtain the actual state transition probability of abnormal forming caused by the output of the abnormal signal spectrum during the solder material casting process; S104: Calculate the hidden energy distribution of the abnormal forming signal on the solder material with the target of the reference critical threshold interval when each forming element in each forming processing link has abnormal forming. Based on the expected observation probability of casting processing forming of the hidden energy distribution, calculate, update and iterate to build a hidden Markov model of the solder material abnormal forming. Use the hidden Markov model to decode the actual state transition probability to obtain the abnormal forming elements that cause the solder material to have abnormal forming during the actual casting process; S106: Establish a defined model for allowing abnormal forming of abnormal forming elements. Based on the termination and elimination reference of abnormal forming, calculate and determine the elimination singular value for each actual abnormal forming model to maximize the elimination to reach the defined model of allowing abnormal forming. Based on the elimination singular value, perform singular value decomposition on the diagonal matrix of each actual process parameter after elimination to obtain the casting process optimization plan; S108: If abnormal forming is still detected during the simulated casting process of the solder material based on the casting process optimization plan, obtain the simulated signal spectrum. Determine the start and end timing nodes of the abnormal execution based on the hash misalignment between the normal signal spectrum of the abnormal process index and the simulated signal spectrum. Calculate the deviation of the abnormal execution timing compared to the standard execution timing based on the start and end timing nodes of the abnormal execution to optimize the preset process control strategy of the casting processing equipment.

2. The optimized method for casting and processing solder materials based on forming anomaly detection according to claim 1, wherein The step S102 specifically includes the following steps: Real-time detect the casting process of the solder material through the Internet of Things for monitoring abnormal forming to obtain the abnormal signal spectrum of the solder material and the residence generation point of each abnormal forming signal in the abnormal signal spectrum; Define the random state variables of abnormal forming according to the abnormal forming signal. Based on the abnormal signal spectrum, extract the direct dependence between each random state variable when the solder material has abnormal forming, and determine the node pattern diagram of the random state variables based on the residence generation point of each abnormal forming signal; Connect and depict the adjacent undirected edges of each random state variable in the node pattern diagram according to the direct dependence, and construct an undirected graph model of abnormal forming of random state variables when the solder material has abnormal forming through the connected adjacent undirected edges; Obtain the casting process of the solder material and the standard process flow of the casting process. Based on the standard process flow, formulate a casting evaluation criterion for no abnormal forming, and assign a Gaussian potential function to the undirected graph model of abnormal forming based on the casting evaluation criterion; After the assignment is completed, obtain the Gaussian potential function of each adjacent undirected edge in the undirected graph model of abnormal forming, which is defined as the marginal Gaussian potential function; Multiply the marginal Gaussian potential functions corresponding to all adjacent undirected edges to obtain the joint probability distribution of the formed abnormal undirected graph model, and determine the actual state transition probability of forming an abnormality due to the output of the formed abnormal signal spectrum during the soldering material pouring process according to the joint probability distribution.

3. The optimized method for casting and processing solder materials based on forming anomaly detection according to claim 1, characterized in that, The step S104 specifically includes the following steps: Based on the standard process flow, obtain the reference critical threshold interval when the forming elements of the soldering material pouring in each forming processing link show forming abnormalities, and calculate the energy distribution based on the historical pouring processing forming signals within the reference critical threshold interval to obtain the hidden energy distribution of the forming abnormal signals on the soldering material. Preset the expected observation state when each forming element shows a forming abnormality according to the reference critical threshold interval, and calculate the probability of each historical pouring processing forming signal generating the expected observation state based on the hidden energy distribution to obtain the expected observation probability. Introduce the maximum likelihood algorithm, and update and iterate the state transition probability matrix and the observation probability matrix of the historical pouring processing forming signal generating the expected observation state in the maximum likelihood algorithm through the expected observation probability, and output the current log-likelihood function. If the current log-likelihood function is greater than the preset log-likelihood function, it is determined that the updated state transition probability matrix and the observation probability matrix are in a convergent form. At this time, terminate the continuous update and iteration operation to obtain the hidden Markov parameter architecture triggered by a single historical pouring processing forming signal when the reference critical threshold interval appears with a forming abnormality, which is marked as the single hidden Markov parameter architecture. Continuously repeat the above steps for the update, iteration, and convergence determination operations of the state transition probability matrix and the observation probability matrix of the historical pouring processing forming signal generating the expected observation state until all historical pouring processing forming signals are processed to generate all single hidden Markov parameter architectures. Build a hidden Markov model for the forming abnormality of the soldering material pouring processing through all single hidden Markov parameter architectures, introduce the Viterbi algorithm, and decode the actual state transition probability in the hidden Markov model through the Viterbi algorithm to obtain the actual hidden state sequence. Obtain the preset hidden state sequence set of each forming element, construct an identification map of the forming element - preset hidden state sequence, and perform matching identification on the actual hidden state sequence through the identification map to obtain one or more forming elements corresponding to the forming abnormality of the soldering material during the actual pouring process, which are defined as abnormal forming elements.

4. An optimization method for soldering material pouring processing based on forming anomaly detection according to claim 3, characterized in that The step of obtaining the reference critical threshold interval when the forming elements of the soldering material pouring in each forming processing link show forming abnormalities based on the standard process flow, and calculating the energy distribution based on the historical pouring processing forming signals within the reference critical threshold interval to obtain the hidden energy distribution of the forming abnormal signals on the soldering material specifically includes the following steps: Obtain the pouring processing log of the soldering material by the pouring processing equipment and the standard design drawing of the target soldering material, and construct a standard pouring processing model of the soldering material in the SolidWorks model design software according to the standard design drawing. Based on the standard process flow, when stripping the solder material and performing the casting processing and forming process, several forming processing links are separated, and according to the several forming processing links, the standard casting processing model is divided into N sub-standard casting processing models; Obtain the forming elements of the solder material casting for each forming processing link, and obtain the reference critical threshold interval for each forming element to cause forming anomalies in the casting of each sub-standard casting processing model; wherein, the forming elements include the quality, temperature control, pressure control, and casting speed of the solder material; Extract the standard casting processing model of the casting processing equipment for casting the solder material within a preset time period through the casting processing log, and the historical casting processing forming signals detected when each is in the reference critical threshold interval corresponding to each forming element one by one; Preset a wavelet function, and perform multi-scale decomposition of discrete wavelet transform on the historical casting processing forming signal according to the wavelet function to obtain the hidden energy distribution of the discrete approximation signal and the discrete detail signal corresponding to each sub-standard casting processing model.

5. The optimized method for casting and processing solder materials based on forming anomaly detection according to claim 1, wherein The step S106 specifically includes the following steps: Obtain the allowable forming anomaly definition value of one or more of the abnormal forming elements, and establish an allowable forming anomaly definition model for one or more of the abnormal forming elements according to the allowable forming anomaly definition value; Construct an actual forming anomaly model for one or more of the abnormal forming elements through the forming anomaly signal spectrum of the solder material. Taking the allowable forming anomaly definition model as the termination elimination reference, calculate the covariance matrix of the actual forming anomaly model compared with the allowable forming anomaly definition model based on the termination elimination reference; Introduce the singular value decomposition algorithm to calculate multiple forming anomaly elimination eigenvalues and multiple forming anomaly elimination eigenvectors of the covariance matrix, and determine the elimination singular values for each actual forming anomaly model to maximize the elimination to reach the allowable forming anomaly definition model according to the forming anomaly elimination eigenvalues and the forming anomaly elimination eigenvectors; Generate an elimination eigenvalue matrix using multiple forming anomaly elimination eigenvalues, and simultaneously generate an elimination eigenvector matrix using multiple forming anomaly elimination eigenvectors; Obtain the actual process parameters of each actual forming anomaly model generated by one or more abnormal forming elements when the casting processing equipment has forming anomalies during the actual production of the solder material through the casting processing log. Based on the elimination singular values, describe the diagonal elements of each actual process parameter after elimination, and construct a diagonal matrix through the described several diagonal elements; Use the elimination eigenvalue matrix and the elimination eigenvector matrix to perform singular decomposition on the diagonal matrix in the singular value decomposition algorithm to generate the final decomposition result, and determine the process optimization parameters required for each actual process parameter when each actual forming anomaly model maximally eliminates to reach the allowable forming anomaly definition model according to the final decomposition result, and obtain the casting process optimization plan.

6. The optimized method for casting and processing of solder materials based on forming anomaly detection according to claim 1, characterized in that The step S108 specifically includes the following steps: Construct a simulation model of the casting processing equipment, upload the casting process optimization plan to the control terminal of the casting processing equipment, and perform casting processing simulation on the solder material by executing the casting process optimization plan through the simulation model; If molding abnormalities in the solder material pouring process are still detected, the molding abnormality simulation signals detected during the simulation process are obtained at this time, and the Fourier transform algorithm is introduced to extract the features of the molding abnormality simulation signals, obtaining the simulation signal spectrum; Obtain several process test cases of the solder material pouring process by the pouring processing equipment, construct a simulated molding abnormality model based on the simulation signal spectrum, and synchronously plan the detection landing area of the simulation signal spectrum according to the position of the simulated molding abnormality model on the standard pouring processing model of the solder material; Obtain one or more process indicators associated with the generation of molding abnormalities in the detection landing area through several process test cases, define them as abnormal process indicators, and obtain the preset process control strategy executed during the simulation of the pouring processing equipment; Extract the simulated process execution timeline of each abnormal process indicator during the simulation process through the process control strategy, and obtain the signal spectrum detected by the pouring processing equipment without generating molding abnormalities in the detection landing area, defined as the normal signal spectrum; Calculate the hash error function of the signal spectrum error of the simulation signal spectrum relative to the normal signal spectrum, obtain the hash misalignment function byte, and determine the start and end timing nodes when the molding abnormality starts to occur in the detection landing area by analyzing the hash misalignment function byte, defined as the execution abnormal start and end timing nodes; Intercept the corresponding time segment on the simulated process execution timeline through the execution abnormal start and end timing nodes, marked as the abnormal execution timing segment, and at the same time obtain the correct process execution timeline of each abnormal process indicator when the pouring processing equipment executes the preset process control strategy; Intercept the sub-segment corresponding to the abnormal execution timing segment on the correct process execution timeline, marked as the standard execution timing segment, calculate the deviation between the abnormal execution timing segment and the standard execution timing segment, obtain the process execution timing deviation degree, and optimize the preset process control strategy of the pouring processing equipment based on the process execution timing deviation degree.

7. A soldering material pouring and processing optimization system based on forming anomaly detection, characterized in that The solder material pouring processing optimization system includes a memory and a processor. There is a program for the solder material pouring processing optimization method based on molding abnormality detection stored in the memory. When the program for the solder material pouring processing optimization method is executed by the processor, the steps of the solder material pouring processing optimization method described in any one of claims 1-6 are implemented.

Images