A fastener production combined process determination method based on a parameter matching algorithm

Abstract

The present application relates to the technical field of fastener production, and discloses a fastener production combined process determination method based on a parameter matching algorithm. The method comprises: classifying fastener parameters using a multi-dimensional parameter classification method; collecting sample data based on the classification results, and establishing an association matrix of parameters and processes; generating process labels using a constraint clustering algorithm; dynamically adjusting parameter weights for samples to be processed, and generating candidate schemes using a weighted matching algorithm; selecting the optimal process combination through multi-objective evaluation; and determining detailed production parameters. The system comprises parameter classification, data analysis, constraint clustering, weighted matching, multi-objective evaluation, and parameter determination modules. The present application solves the problems of fixed parameter weights, lack of systematic process induction, and ignoring process constraints in existing methods, realizes the intelligentization and optimization of process determination, and improves production efficiency and scheme accuracy.

Description

Technical Field

[0001] This invention belongs to the field of fastener manufacturing technology, specifically a method for determining fastener manufacturing assembly processes based on a parameter matching algorithm. Background Technology

[0002] Fasteners are key basic components in mechanical assembly, and the rational determination of their production processes directly affects product quality, production efficiency, and manufacturing costs. With the increasing diversification of fastener products and the personalization of customer needs, how to quickly and accurately determine the appropriate combination of production processes for fasteners of different specifications has become an important issue for fastener manufacturers.

[0003] Currently, the determination of fastener production processes mainly relies on manual experience, rule-based expert systems, or simple database query matching methods. Manual experience methods are highly dependent on the individual skill level of technicians, resulting in significant differences in solutions provided by different personnel, and lacking systematic decision support when facing new product models. While rule-based expert systems solidify experience into decision rules, the workload of establishing and maintaining these rules is substantial, and they struggle to handle complex situations involving multiple coupled parameters. Database query methods perform parameter matching by retrieving historical cases, but they have the following shortcomings: First, they do not consider the relative importance of different parameters, using fixed or equal weights for similarity calculations, leading to low matching accuracy; second, they lack a systematic summary of process modules, failing to establish explicit relationships between parameters and processes, making it difficult to flexibly combine and optimize process solutions; third, they ignore the dependencies between processes and equipment compatibility constraints, potentially outputting solutions with unreasonable order or equipment conflicts; and fourth, the solution evaluation stage often uses single-objective optimization, failing to comprehensively consider multiple interdependent objectives such as cost, cycle time, quality, and energy consumption.

[0004] Therefore, there is a need for a method for determining fastener production combination processes that can systematically process multi-dimensional parameters, dynamically adjust parameter weights, consider process constraints, and perform multi-objective optimization. Summary of the Invention

[0005] The purpose of this invention is to overcome the problems of fixed parameter weights, lack of systematic process summarization, neglect of process dependency constraints, and single-objective optimization in existing fastener production process determination methods. This invention provides a fastener production process combination determination method based on parameter matching algorithm. This method improves the accuracy and rationality of process schemes by establishing a correlation matrix between parameters and processes, dynamically adjusting parameter weights, and conducting multi-objective evaluation, thereby realizing the intelligent determination of fastener production processes.

[0006] To achieve the above objectives, the first aspect of the present invention provides a method for determining fastener production assembly processes based on a parameter matching algorithm, the method comprising the following steps:

[0007] S1) The fastener sample parameters are classified using a multi-dimensional parameter classification method. The multi-dimensional parameters include shape geometry parameters, material property parameters, process constraint parameters, and quality index parameters.

[0008] S2) Based on the multidimensional parameter classification results, sample data are collected from the product database and production requirement documents to summarize the set of fastener production process modules and establish a correlation matrix between parameters and processes.

[0009] S3) Based on the process dependency relationships in the association matrix, a constraint clustering algorithm is used to cluster the process module set to generate process labels that satisfy process dependency relationships and equipment compatibility constraints;

[0010] S4) For the fastener sample to be processed, extract its multi-dimensional parameters and dynamically adjust the weight of each dimension according to the fastener type. Use a weighted parameter matching algorithm to match and score the process label with the parameters of the sample to be processed. Based on the matching score, retrieve and generate multiple production combination candidate schemes from the association matrix.

[0011] S5) Based on production cost, production cycle, product yield and energy consumption indicators, the optimal fastener production combination process is selected from the candidate schemes through multi-objective evaluation.

[0012] S6) Based on the selected optimal process combination and its corresponding process requirements, determine the detailed production parameters for each process, including shape processing parameters, process handling parameters and quality control parameters.

[0013] And / or, in S2, the correlation matrix between the parameters and the process records the set of feasible processes corresponding to each parameter combination and its historical selection frequency.

[0014] And / or, in S3, the constrained clustering algorithm satisfies must-link constraints and cannot-link constraints, wherein the must-link constraint represents a pair of processes with a cooperative relationship, and the cannot-link constraint represents a pair of mutually exclusive processes.

[0015] And / or, in S4, the calculation formula for dynamically adjusting the weights of each dimension parameter is:

[0016] Wi = αi × (1 + β × Ci)

[0017] Where Wi is the weight of the i-th dimension parameter, αi is the basic weight, β is the adjustment coefficient, and Ci is the complexity coefficient of this dimension;

[0018] The weighted parameter matching algorithm adopts corresponding distance measurement methods for different types of parameters. Hamming distance is used for discrete parameters and normalized Euclidean distance is used for continuous parameters. After calculating the distance in each dimension, the algorithm performs weighted summation according to dynamic weights to obtain the comprehensive matching score.

[0019] And / or, in S5, the multi-objective evaluation adopts a weighted summation method; before selecting the optimal solution, process feasibility verification is performed, process conflicts are detected by process compatibility matrix, and when the compatibility is lower than a preset threshold, an automatic correction mechanism is triggered, which includes retrieving similar historical cases, generating alternative process solutions, and adjusting parameters for rematching.

[0020] And / or, in S6, the determination of the detailed production parameters is based on the process requirements of the selected process combination; the process parameters include heat treatment temperature profiles and surface treatment process types; the quality control parameters include inspection node locations and process capability index Cpk target values.

[0021] And / or, after S6, perform learning updates based on feedback from actual production data, adjust parameter weights, correlation matrices, and evaluation models; and insert quality control processes at high-risk process nodes based on risk assessment results.

[0022] A second aspect of the present invention provides a fastener production assembly process determination system based on a parameter matching algorithm. Using the above method, the system includes:

[0023] Parameter classification module: performs multi-dimensional classification of fastener sample parameters;

[0024] Data analysis module: Collects sample data, summarizes the process module set, and establishes a correlation matrix between parameters and processes;

[0025] Constrained clustering module: Performs constrained clustering on the set of process modules to generate process labels;

[0026] Weighted matching module: Extracts multidimensional parameters of the sample to be processed, dynamically adjusts the weights of each dimension and executes a weighted parameter matching algorithm to generate multiple production combination candidate schemes;

[0027] Multi-objective evaluation module: Evaluates candidate solutions based on multiple indicators to select the optimal combination of processes;

[0028] Parameter determination module: Determines detailed production parameters based on the optimal process combination.

[0029] And / or, the system further includes a database interface module and a visualization output module. The database interface module is used to interact with the product database and production requirement documents, and the visualization output module is used to graphically display the production combination process and process parameters. Beneficial effects

[0030] This invention systematically summarizes fastener production process knowledge by establishing a multi-dimensional parameter classification system and a parameter-process correlation matrix; it employs a constrained clustering algorithm to handle process dependencies, ensuring that the generated process labels meet actual production constraints; it improves the matching accuracy of process schemes through a weighted matching algorithm that dynamically adjusts parameter weights; and it achieves optimized selection of process schemes by using a multi-objective evaluation method that comprehensively considers indicators such as cost, cycle time, quality, and energy consumption. Thus, it realizes intelligent and systematic determination of fastener production processes, improving the accuracy of process schemes and production efficiency. Attached Figure Description

[0031] Figure 1 A flowchart illustrating a method for determining fastener production assembly steps based on a parameter matching algorithm provided by the present invention;

[0032] Figure 2 This is a schematic diagram of a fastener production assembly process determination system based on a parameter matching algorithm, provided by the present invention. Detailed Implementation

[0033] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0034] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0035] Terminology and Scope

[0036] For ease of understanding, those skilled in the art should understand that the term "fastener sample parameters" as used herein refers to a set of geometrical parameters, material property parameters, process constraint parameters, and quality index parameters used to characterize the fastener product to be processed. Unless otherwise stated, the endpoints of the numerical ranges described herein and any specific values ​​therein can be combined to form new sub-ranges, all of which are considered to be within the scope of this invention.

[0037] In the text, "must-link constraint" indicates that two processes have a collaborative relationship and should be executed in the same group or sequentially under the condition of meeting equipment / process requirements; "cannot-link constraint" indicates that two processes have a mutually exclusive relationship and should not coexist in the same combination or should be configured in an exclusive selection manner.

[0038] The following specific embodiments are given to illustrate the technical solution of this application in detail.

[0039] The first aspect of this invention provides a method for determining fastener manufacturing assembly processes based on a parameter matching algorithm, such as... Figure 1 As shown, the method includes the following steps:

[0040] S1) The fastener sample parameters are classified using a multi-dimensional parameter classification method. The multi-dimensional parameters include shape geometry parameters, material property parameters, process constraint parameters, and quality index parameters.

[0041] S2) Based on the multidimensional parameter classification results, sample data is collected from the product database and production requirement documents to summarize the set of fastener production process modules and establish the correlation between parameters and processes.

[0042] matrix;

[0043] S3) Based on the process dependency relationships in the association matrix, a constraint clustering algorithm is used to cluster the process module set to generate process labels that satisfy process dependency relationships and equipment compatibility constraints;

[0044] S4) For the fastener sample to be processed, extract its multi-dimensional parameters and dynamically adjust the weight of each dimension according to the fastener type. Use a weighted parameter matching algorithm to match and score the process label with the parameters of the sample to be processed. Based on the matching score, retrieve and generate multiple production combination candidate schemes from the association matrix.

[0045] S5) Based on production cost, production cycle, product yield and energy consumption indicators, the optimal fastener production combination process is selected from the candidate schemes through multi-objective evaluation.

[0046] S6) Based on the selected optimal process combination and its corresponding process requirements, determine the detailed production parameters for each process, including shape processing parameters, process handling parameters and quality control parameters.

[0047] In a specific implementation, the external geometric parameters include thread diameter, length, thread specification, and dimensional tolerance; the material property parameters include material grade, hardness range, and tensile strength; the process constraint parameters include available equipment type and machining accuracy level; and the quality index parameters include surface roughness requirements, mechanical performance indicators, and corrosion resistance requirements.

[0048] In a specific implementation, the multidimensional parameter classification is stored using a standardized data structure, with each parameter dimension corresponding to an independent data field, which facilitates subsequent retrieval and matching calculations.

[0049] In a specific implementation, step S2, the sample data collection includes extracting parameter information and corresponding process routes from historical production orders from the enterprise's ERP system, MES system, and quality management system. The production process module set includes, but is not limited to, basic process modules such as blanking, turning, milling, drilling, tapping, thread rolling, heat treatment, surface treatment, and inspection.

[0050] In a specific implementation, the correlation matrix between the parameters and the process is stored in a sparse matrix structure, and the matrix element values ​​represent the correlation strength (probability weight) of the parameter combination in selecting a specific process combination.

[0051] In a specific implementation, the process of establishing the correlation matrix includes: discretizing the continuous parameters in historical production orders, for example, segmenting the length dimension into 10mm intervals and the hardness into HRC5 ranges; statistically analyzing the process combinations used in all historical orders falling within the same parameter interval and their frequency of occurrence; and storing the normalized frequency (frequency of the combination / total number of orders) as the correlation strength in the matrix. For each parameter interval, the top 3-5 process combinations with the highest correlation strength are typically stored.

[0052] In a specific implementation, in step S3, the constraint-based clustering algorithm satisfies both must-link and cannot-link constraints. The must-link constraint represents a pair of processes with a cooperative relationship; the cannot-link constraint represents a pair of mutually exclusive processes.

[0053] In a specific implementation, a constrained K-medoids clustering algorithm is adopted, and the following improvements are made for the special constraints of the fastener process;

[0054] (1) Constraint transformation: Convert the dependencies between processes into a must-link constraint set ML={(pi,pj)|pi must precede pj}, and convert the mutual exclusion relationship between devices into a cannot-link constraint set CL={(pi,pj)|pi and pj cannot be executed simultaneously};

[0055] (2) Mixed Distance Measurement: For the mixed attribute characteristics of the process, a distance function is defined:

[0056] d(pi,cj)=w1·dattr(pi,cj)+w2·dfreq(pi,cj)+w3·dequip(pi,cj)

[0057] Where dattr is the Hamming distance of process attributes, dfreq is the Euclidean distance of historical frequency, and dequip is the equipment type distance;

[0058] (3) Constraint embedding mechanism: Add constraint checks during the allocation phase: If there is a must-link(pi,pj) and pi has been allocated to Ck, then pj is forced to be allocated to Ck as well; if there is a cannot-link(pi,pj) and pi has been allocated to Ck, then pj is prohibited from being allocated to Ck.

[0059] (4) Penalty function: Introduce a constraint violation penalty term.

[0060] P=λ1·|MLviolated|+λ2·|CLviolated|

[0061] Minimize the objective function: Objective = Σd(pi,ci) + P, where the first term is the sum of cluster distances and the second term is the penalty for constraint violation;

[0062] (5) Iteration termination: The iteration terminates when the cluster center no longer changes and P=0, or when the maximum number of iterations of 100 is reached.

[0063] Algorithm execution flow: Initialization → Distance calculation → Constraint check and allocation → Center update → Penalty function calculation → Termination condition determination → Label output.

[0064] In a specific implementation, the process labels are divided into basic machining, finishing machining, heat treatment, and post-treatment categories. Basic machining includes processes such as blanking, rough turning, and rough milling; finishing machining includes processes such as finish turning, finish milling, grinding, and tapping; heat treatment includes processes such as quenching, tempering, heat treatment, and carburizing; and post-treatment includes processes such as surface treatment, cleaning, inspection, and packaging.

[0065] In a specific implementation, the calculation formula for dynamically adjusting the weights of each dimension parameter in step S4 is as follows:

[0066] Wi = αi × (1 + β × Ci)

[0067] Where Wi is the weight of the i-th dimension parameter, αi is the basic weight, β is the adjustment coefficient, and Ci is the complexity coefficient of this dimension.

[0068] The basic weight αi is obtained by analyzing historical data and calculating the correlation between each parameter dimension and process selection using information gain or chi-square tests, followed by normalization. The system provides an interface for expert fine-tuning. Typical value ranges are: external geometry parameters 0.20-0.25, material property parameters 0.25-0.30, process constraint parameters 0.15-0.20, and quality index parameters 0.20-0.25.

[0069] The adjustment coefficient β ranges from 0.3 to 0.8, preferably 0.5. It is a system-level hyperparameter used to control the magnitude of the complexity impact.

[0070] The complexity coefficient Ci is calculated using information entropy to quantify the uniformity of the parameter distribution. The formula is as follows:

[0071] Ci=-Σ(j=1tom)pij·log2(pij) / log2(m)

[0072] Where pij is the historical probability distribution of the j-th value of the i-th parameter, and m is the number of values.

[0073] Ci close to 1 indicates uniform distribution and high complexity; Ci close to 0 indicates concentrated distribution and low complexity.

[0074] In a specific implementation, the weighted parameter matching algorithm employs corresponding distance metric methods for different types of parameters. For discrete parameters (such as material grade and thread specification), the Hamming distance metric is used, and the calculation formula is as follows:

[0075] d=Σ(xi≠yi),

[0076] Where xi and yi are the i-th discrete values ​​of the sample parameter and label parameter, respectively; for continuous parameters (such as hardness, length, and roughness), the normalized Euclidean distance metric is used, and the calculation formula is:

[0077] d=√[Σ((xi-yi) / ri)²],

[0078] Where ri is the normalized range of the i-th parameter. After calculating the distance for each dimension, the results are weighted and summed according to the dynamic weights Wi to obtain the comprehensive matching score.

[0079] S=1-Σ(Wi×di),

[0080] In a specific implementation, the matching and scoring process includes: comparing the multidimensional parameters of the sample to be processed with the typical parameters of each tag in the process tag library one by one, calculating the distance value of each dimension, and summing them according to dynamic weights to obtain the matching score of each tag; sorting the tags according to the matching scores and selecting the top 3-5 tags with the highest scores; querying the association matrix according to the selected tags to retrieve the corresponding process combinations and form multiple candidate solutions.

[0081] In a specific implementation, step S5 involves a multi-objective evaluation using either a weighted summation method or a Pareto optimization method. When using the weighted summation method, production costs, production cycle time, product yield, and energy consumption are normalized, and then a comprehensive score is calculated according to set weights. The calculation formula is as follows:

[0082] F=w1×C'+w2×T'+w3×Q'+w4×E',

[0083] Where C', T', Q', and E' are the normalized cost, cycle time, yield, and energy consumption indicators, respectively, and w1, w2, w3, and w4 are the corresponding weights that satisfy w1+w2+w3+w4=1.

[0084] In a specific implementation, the normalization process employs the minimum-maximum normalization method. For indicators such as cost and energy consumption, where smaller values ​​are preferred, the normalization formula is as follows:

[0085] X'=(Xmax-X) / (Xmax-Xmin);

[0086] For indicators such as yield, where higher values ​​are better, the normalization formula is:

[0087] X'=(X-Xmin) / (Xmax-Xmin).

[0088] In a specific implementation, process feasibility verification is required before selecting the optimal solution. Process conflicts are detected using a process compatibility matrix, which records the compatibility values ​​between each pair of processes. The compatibility value ranges from 0 to 1, where 1 indicates complete compatibility and 0 indicates complete conflict. When a candidate solution contains a pair of processes with a compatibility value lower than a preset threshold (e.g., 0.6), an automatic correction mechanism is triggered.

[0089] In a specific implementation, the automatic correction mechanism follows this process:

[0090] Search for similar cases: In the historical case library, using the same weighted parameter matching algorithm as S4, search for successful historical cases with a comprehensive matching score higher than 0.85 with the current sample to be processed, and extract their process combinations as alternative solutions for reference;

[0091] Generating alternative solutions: If no suitable case is found in the first level, for the detected conflicting process pair (OpA, OpB), an alternative process OpA′ with similar function to OpA is searched from the process module library. The functional similarity is measured by cosine similarity based on the process function description text, requiring a similarity > 0.7, and the compatibility of OpA′ with other processes in the current solution is higher than the threshold of 0.6.

[0092] Adjust parameters and rematch: If the above two correction methods are ineffective, temporarily increase the weight Wp of the "process constraint parameter" (e.g., increase by 50%), re-execute the matching process in step S4, and use the weight change to guide the system to avoid parameter combinations that cause conflicts and generate new candidate solutions.

[0093] In a specific implementation, in step S6, the determination of detailed production parameters is based on the process requirements of the selected process combination. For each process, a baseline value of the process parameter matching the sample parameters is retrieved from the process parameter database, and then fine-tuned and optimized according to the specific parameter characteristics of the sample.

[0094] In a specific implementation, the external machining parameters include cutting speed, spindle speed, feed rate, depth of cut, tool type, and coolant type; the process parameters include heat treatment temperature curve, holding time, cooling rate, and surface treatment process type; and the quality control parameters include inspection node location, inspection items, inspection methods, sampling ratio, and process capability index Cpk target value.

[0095] In a specific implementation, the shape machining parameters and process parameters are determined by querying a process parameter database based on material properties and quality requirements. For example, the cutting speed is determined based on the material hardness and is typically 40-100 m / min; the heat treatment temperature is determined based on the material grade and target hardness, and the quenching temperature is typically 800-900℃.

[0096] In a specific implementation, after step S6, learning updates are performed based on feedback from actual production data. Actual production costs, cycles, yields, and energy consumption data of executed process plans are collected and compared with predicted values; prediction errors are calculated, and when the error exceeds 10%, parameter weight coefficients and frequency weights in the correlation matrix are adjusted; the weight parameters in the multi-objective evaluation model are updated to improve the accuracy of subsequent predictions.

[0097] In a specific implementation, quality control procedures are inserted at high-risk process nodes based on risk assessment results. By analyzing historical production data, process nodes with high defect rates are identified. When the defect rate or defect cost exceeds a preset threshold, a special inspection procedure is added thereafter.

[0098] In the specific implementation of the above method, firstly, the operator inputs the basic information of the fastener sample to be processed into the system interface. The system's parameter classification module automatically extracts or guides the user to input four categories of parameters: shape geometry, material properties, process constraints, and quality indicators, and stores them according to the classification standards. Secondly, the data analysis module collects historical production records from the enterprise's product database and production requirement documents through the database interface module, statistically analyzes them to obtain a set of commonly used process modules, and establishes a correlation matrix between parameter combinations and processes. The matrix records the feasible processes corresponding to each parameter combination and their historical selection frequency. Then, approximately... The bundled clustering module performs constrained clustering on the set of process modules, considering both must-link and cannot-link relationships between processes, grouping similar processes under the same label to generate a process label library. Next, the weighted matching module dynamically calculates the weights Wi for each dimension of the sample to be processed based on its fastener type and the complexity of each dimension's parameters. A weighted parameter matching algorithm is then used to calculate the matching score between the sample parameters and the process labels. Hamming distance is used for discrete parameters, and normalized Euclidean distance is used for continuous parameters. The distances of each dimension are then weighted and summed to obtain the overall score. The system first calculates the overall score by retrieving the corresponding process combination from the association matrix based on the matching score, generating 3-5 candidate solutions. Then, the multi-objective evaluation module assesses the production cost, production cycle, product yield, and energy consumption of each candidate solution. After normalizing each indicator, a weighted sum is calculated based on set weights to obtain the overall score. Feasibility is verified using a process compatibility matrix. If process conflicts with compatibility below a threshold are found, an automatic correction mechanism is triggered. Finally, the solution with the highest score and feasibility is selected as the optimal process combination. Finally, the parameter determination module queries the process database corresponding to each process based on the selected optimal process combination to determine... Detailed production parameters for each process, including blanking length, cutting speed, spindle speed, feed rate, depth of cut, heat treatment temperature profile, surface treatment type, inspection node location, and quality control parameters, are compiled into a complete production process parameter table. This table is displayed to the user for confirmation in the form of a table or flowchart through a visualization output module. After confirmation, the data is sent to an external production execution information system (such as a MES system) for execution via a data interface. During production execution, the system continuously collects actual production data feedback, learns and updates, and continuously optimizes parameter weights, correlation matrices, and evaluation models to improve the system's intelligence level and prediction accuracy.

[0099] The above embodiments introduce a method for determining fastener production assembly processes based on parameter matching algorithms from the perspective of process flow. The following embodiments introduce a system for determining fastener production assembly processes based on parameter matching algorithms. For details, please refer to the following embodiments.

[0100] A second aspect of the present invention provides a system for determining fastener production assembly processes based on a parameter matching algorithm, such as... Figure 2 As shown, using the above method, the system includes:

[0101] Parameter classification module 101: performs multi-dimensional classification of fastener sample parameters;

[0102] Data analysis module 102: collects sample data, summarizes the process module set and establishes a correlation matrix between parameters and processes. The data analysis module 102 is connected to the product database and production requirement documents.

[0103] Constrained Clustering Module 103: Performs constrained clustering on the set of process modules to generate process labels;

[0104] Weighted matching module 104: Extracts multidimensional parameters of the sample to be processed, dynamically adjusts the weights of each dimension and executes a weighted parameter matching algorithm to generate multiple production combination candidate schemes;

[0105] Multi-objective evaluation module 105: Evaluates candidate solutions based on multiple indicators and selects the optimal combination of processes;

[0106] Parameter determination module 106: Determines detailed production parameters based on the optimal process combination.

[0107] In the specific implementation of the above method, the parameter classification module 101 may include a parameter input submodule and a parameter verification submodule. The parameter input submodule provides a user interface for operators to input or import parameter data, and the parameter verification submodule verifies the completeness, rationality, and format standardization of the input parameters. The data analysis module 102 may include a data acquisition submodule, a process summarization submodule, and a matrix construction submodule. The constraint clustering module 103 may include a constraint identification submodule and a clustering calculation submodule. The constraint identification submodule extracts process dependencies and equipment compatibility constraints from the association matrix, and the clustering calculation submodule executes the constraint clustering algorithm. The weighted matching module 104 may include a weight calculation submodule and a similarity calculation submodule. The multi-objective evaluation module 105 may include an index quantification submodule, a feasibility verification submodule, and a scheme ranking submodule. The parameter determination module 106 connects to the process parameter database and determines the specific process parameters of each process through querying.

[0108] In a specific implementation, the system further includes a database interface module 107 and a visualization output module 108. The database interface module 107 is used to interact with the product database and production requirement documents, and the visualization output module 108 is used to display the production combination process and process parameters in a graphical manner.

[0109] In a specific implementation, the system may further include a learning update module 109 and a risk assessment module 110. The learning update module 109 is used to collect feedback from actual production data, analyze prediction errors, and adjust parameter weights, correlation matrices, and evaluation models to achieve continuous system optimization. The risk assessment module 110 is used to identify high-risk process nodes, calculate the defect rate and defect cost of each process, and provide decision support for the insertion of quality control processes.

[0110] In this invention, the system's workflow includes: a parameter classification module 101 receives the original parameters of the fastener sample to be processed, inputs the data through a parameter input submodule, verifies the data through a parameter verification submodule, and classifies the parameters into four categories according to a multidimensional classification method: shape geometry, material properties, process constraints, and quality indicators, and stores them in a structured data table; a data analysis module 102 reads historical production data from a product database through a database interface module 107, a data acquisition submodule extracts parameters and process information from relevant orders, a process summarization submodule performs statistical analysis and summarizes a set of process modules, and a matrix construction submodule establishes the relationship between parameters and processes. The association matrix is ​​concatenated and stored in the system database; the constraint identification submodule of the constraint clustering module 103 reads the dependency data in the process module set and association matrix, extracts must-link and cannot-link constraints, and the clustering calculation submodule executes the constraint clustering algorithm to generate a process label library and stores the label data for later use; the weighted matching module 104 receives the classification parameters output by the parameter classification module 101, the weight calculation submodule dynamically calculates the weights of each dimension according to the fastener type and complexity coefficient, and the similarity calculation submodule matches the sample parameters with the process label library, using Hamming distance and Euclidean distance respectively. The system measures discrete and continuous parameters, calculates a weighted sum to obtain a comprehensive matching score, and outputs 3-5 candidate process combination schemes based on the scores. The multi-objective evaluation module 105's index quantification submodule quantifies and normalizes the cost, cycle time, yield, and energy consumption indicators of each candidate scheme. The feasibility verification submodule detects process conflicts through a process compatibility matrix, triggering an automatic correction mechanism for schemes with compatibility below a threshold. The scheme ranking submodule calculates the comprehensive score using a weighted summation method and ranks the schemes, selecting the optimal scheme with the highest score. The parameter determination module 106 queries the process parameter database based on the optimal process combination, and performs parameter determination for each process. The system determines detailed production parameters such as cutting speed, feed rate, heat treatment temperature, and inspection points to form a complete process parameter table. The visualization output module 108 displays the process flow and parameters to the user in the form of tables, flowcharts, etc. After user confirmation, the system sends the process parameters to the production execution system (MES) through an interface. The learning and updating module 109 collects actual data after production execution, compares and analyzes the error between the predicted value and the actual value, and adjusts relevant parameters and models when the error exceeds the threshold. The risk assessment module 110 analyzes the historical defect rate of each process, identifies high-risk nodes, and suggests inserting quality control processes after key processes to form closed-loop management.

[0111] The present invention will be described in detail below through embodiments, but the scope of protection of the present invention is not limited thereto.

[0112] The following examples are as follows Figures 1-2The system shown is an implementation of a fastener production assembly process determination system based on a parameter matching algorithm. The system includes a parameter classification module, a data analysis module, a constrained clustering module, a weighted matching module, a multi-objective evaluation module, a parameter determination module, a database interface module, and a visualization output module. These modules are connected via a data bus to enable real-time data interaction and processing. Example

[0113] To determine the production process requirements for M10×50 hexagonal head bolts for a certain automobile manufacturing company, the method of this invention is applied:

[0114] S1) First, the multidimensional parameters of the bolt sample are extracted through the parameter classification module: the external geometric parameters include thread diameter M10, length 50mm, pitch 1.5mm, and dimensional tolerance IT7 grade; the material property parameters include material grade 45# steel, target hardness HRC28-32, and tensile strength ≥600MPa; the process constraint parameters include the availability of CNC lathes, tapping machines, heat treatment furnaces, and electroplating production lines, machining accuracy grade IT7, and batch size of 1000 pieces; the quality index parameters include surface roughness Ra≤3.2μm, strength grade 8.8, and the requirement for rust prevention treatment.

[0115] S2) The data analysis module retrieves 500 historical production records of M8-M12 bolts from the enterprise's product database over the past three years, summarizing them into eight process modules: blanking, turning, tapping, thread rolling, heat treatment, surface treatment, and inspection. A correlation matrix between parameters and processes is then established. In this matrix, for the parameter combination of 45# steel, M10 thread, strength grade 8.8, and batch size of 500-2000 pieces, the historical frequency of the process combination "blanking → turning → tapping → heat treatment → surface treatment → inspection" was 89 times, accounting for 89% of the total records for this parameter combination. Therefore, the correlation strength weight for this process combination is set to 0.89.

[0116] S3) The constraint clustering module performs constraint clustering on the set of process modules, identifying a must-link constraint between blanking and turning (must be performed sequentially), a cannot-link constraint between heat treatment and drilling (mutually exclusive), and a sequence constraint between tapping and heat treatment (heat treatment must be performed after tapping). Under these constraints, four process labels are generated: Label 1 - Basic Machining (Blanking, Rough Turning, Rough Milling), Label 2 - Finishing (Finish Turning, Tapping, Grinding), Label 3 - Heat Treatment (Quenching, Tempering, Heat Treatment), and Label 4 - Post-Processing (Surface Treatment, Inspection, Packaging).

[0117] S4) The weighted matching module determines that the bolt sample is a standard bolt based on its fastener type and calculates the complexity coefficients for each dimension: shape complexity Cs=0.3, material complexity Cm=0.5, process complexity Cp=0.2, and quality complexity Cq=0.4. Basic weights are set as αs=0.22, αm=0.28, αp=0.18, and αq=0.22, with an adjustment coefficient β=0.5. Dynamic weights are calculated using the formula Wi=αi×(1+β×Ci), and after normalization, Ws=0.24, Wm=0.31, Wp=0.18, and Wq=0.27. A weighted parameter matching algorithm was used to match sample parameters with process labels. Label 1 (basic machining) scored 0.995, Label 2 (finishing) scored 0.88, Label 3 (heat treatment) scored 0.75, and Label 4 (post-processing) scored 0.72. Based on the matching scores, the corresponding process combinations were retrieved from the association matrix, generating three candidate schemes: Scheme 1 is "material preparation → turning → tapping → heat treatment → surface treatment → inspection", Scheme 2 is "material preparation → turning → thread rolling → heat treatment → surface treatment → inspection", and Scheme 3 is "material preparation → turning → tapping → surface treatment → inspection" (heat treatment omitted).

[0118] S5) The multi-objective evaluation module comprehensively evaluates the three candidate solutions: Solution 1 has a production cost of 8500 yuan, a production cycle of 5 days, a product yield of 98%, and energy consumption of 320 kWh. After normalization and weighted summation, its comprehensive score is 0.85. Solution 2 has a cost of 9200 yuan, a cycle of 6 days, a yield of 99%, and energy consumption of 380 kWh, with a comprehensive score of 0.78. Solution 3 has a cost of 7800 yuan, a cycle of 4 days, a yield of 95%, and energy consumption of 280 kWh, with a comprehensive score of 0.72. Feasibility verification is performed using a process compatibility matrix. No process pairs with a compatibility below the threshold of 0.6 were found. Therefore, Solution 1, with the highest comprehensive score, is selected as the optimal process combination.

[0119] S6) The parameter determination module determines detailed production parameters from the process parameter database based on each process in Scheme 1: For turning, the spindle speed is 1200-1500 rpm, feed rate is 0.15-0.25 mm / r, and cutting speed is 90-100 m / min; for heat treatment, the quenching temperature is 840-860℃, holding time is 30 minutes, and tempering temperature is 400℃; for surface treatment, electro-galvanizing is used with a coating thickness of 8-12 μm; quality control parameters include hardness testing after heat treatment (10% sampling rate) and full-size inspection of finished products (20% sampling rate), with a process capability index (Cpk) target value set at ≥1.33. After the system displays the process parameters to the user for confirmation via the visualization output module, it is then sent to the production execution system.

[0120] After production execution, the learning and updating module collected actual data for feedback: the actual cost was 8650 yuan (an error of 1.8% compared to the predicted value of 8500 yuan), the actual cycle time was 5.2 days (an error of 4% compared to the predicted value of 5 days), the actual yield was 98.5% (an error of 0.5% compared to the predicted value of 98%), and the actual energy consumption was 325 kWh (an error of 1.6% compared to the predicted value of 320 kWh). All errors were below the 10% threshold, verifying the accuracy and effectiveness of the method of this invention.

[0121] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0122] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for determining fastener production assembly processes based on a parameter matching algorithm, characterized in that, Includes the following steps: S1) The fastener sample parameters are classified using a multi-dimensional parameter classification method. The multi-dimensional parameters include shape geometry parameters, material property parameters, process constraint parameters, and quality index parameters. S2) Based on the multidimensional parameter classification results, sample data is collected by retrieving product databases and analyzing production demand documents, and a set of fastener production process modules is summarized, and a correlation matrix between parameters and processes is established. S3) Based on the process dependency relationships in the association matrix, a constraint clustering algorithm is used to cluster the process module set to generate process labels that satisfy process dependency relationships and equipment compatibility constraints; S4) For the fastener sample to be processed, extract its multi-dimensional parameters and dynamically adjust the weight of each dimension according to the fastener type. Use a weighted parameter matching algorithm to match and score the process label with the parameters of the sample to be processed. Based on the matching score, retrieve and generate multiple production combination candidate schemes from the association matrix. S5) Based on production cost, production cycle, product yield and energy consumption indicators, the optimal fastener production combination process is selected from the candidate schemes through multi-objective evaluation. S6) Based on the selected optimal process combination and its corresponding process requirements, determine the detailed production parameters for each process, including shape processing parameters, process handling parameters and quality control parameters.

2. The method according to claim 1, characterized in that: The external geometric parameters include dimensions, dimensional tolerances, and thread specifications; The material property parameters include material grade, hardness range, and tensile strength; The process constraint parameters include available equipment types and machining accuracy levels; The quality indicators include surface roughness requirements, mechanical performance indicators, and corrosion resistance requirements.

3. The method according to claim 1, characterized in that: In S2, the correlation matrix between parameters and processes records the set of feasible processes corresponding to each parameter combination and their historical selection frequency.

4. The method according to claim 1, characterized in that: In S3, the constrained clustering algorithm satisfies must-link and cannot-link constraints, where the must-link constraint represents a pair of processes with a cooperative relationship, and the cannot-link constraint represents a pair of mutually exclusive processes.

5. The method according to claim 1, characterized in that: In S4, the calculation formula for dynamically adjusting the weights of each dimension parameter is as follows: Wi = αi × (1 + β × Ci) Where Wi is the weight of the i-th dimension parameter, αi is the base weight, β is the adjustment coefficient, and Ci is the complexity coefficient of this dimension; the weighted parameter matching algorithm adopts the corresponding distance measurement method for different types of parameters, calculates the distance of each dimension respectively, and then performs weighted summation according to dynamic weights to obtain the comprehensive matching score.

6. The method according to claim 1, characterized in that: In S5, the multi-objective evaluation adopts a weighted summation method; before selecting the optimal solution, the feasibility of the process is verified, and process conflicts are detected by the process compatibility matrix. When the compatibility is lower than a preset threshold, an automatic correction mechanism is triggered. The automatic correction mechanism includes retrieving similar historical cases, generating alternative process solutions, and adjusting parameters for rematching.

7. The method according to claim 1, characterized in that: In S6, the determination of the detailed production parameters is based on the process requirements of the selected process combination; the process parameters include the heat treatment temperature curve and the surface treatment process type; the quality control parameters include the detection node location and the process capability index Cpk target value.

8. The method according to claim 1, characterized in that: After S6, learning updates are performed based on feedback from actual production data, adjusting parameter weights, correlation matrices, and evaluation models; and quality control procedures are inserted at high-risk process nodes based on risk assessment results.

9. A fastener production assembly process determination system based on a parameter matching algorithm, used to execute the fastener production assembly process determination method based on a parameter matching algorithm as described in any one of claims 1 to 8, characterized in that, The system includes the following modules: Multidimensional parameter classification module: performs multidimensional classification of fastener sample parameters; Data analysis module: Collects sample data, summarizes the process module set, and establishes a correlation matrix between parameters and processes; Constrained clustering module: Performs constrained clustering on the set of process modules to generate process labels; Weighted matching module: Extracts multidimensional parameters of the sample to be processed, dynamically adjusts the weights of each dimension and executes a weighted parameter matching algorithm to generate multiple production combination candidate schemes; Multi-objective evaluation module: Evaluates candidate solutions based on multiple indicators to select the optimal combination of processes; Parameter determination module: Determines detailed production parameters based on the optimal process combination.

10. The system according to claim 9, characterized in that: The system also includes a database interface module and a visualization output module. The database interface module is used to interact with the product database and production requirement documents, and the visualization output module is used to display the production combination process and process parameters in a graphical manner.

Images