Server manufacturing process management system capable of automatically generating operation guidance based on AI

Abstract

The invention relates to the technical field of intelligent manufacturing control, in particular to a server manufacturing process management system for automatically generating operation guidance based on AI, which comprises a process time sequence analysis module, an operation arrangement optimization module, a guidance generation integration module, a knowledge base iteration updating module and an operation guidance distribution module. According to the invention, by combining the employee skill level and the equipment state and optimizing the process time sequence distribution, the process operation time can be accurately calculated and standardized, so that the production arrangement is dynamically adjusted, the load balance of the production line is ensured, the station configuration is optimized, the technical terms and matters needing attention are embedded, the production error and inconsistency are reduced, and the production efficiency is improved. By means of a knowledge base incremental updating mechanism, operation guidance files are optimized in real time, the timeliness and accuracy of operation standards are improved, rapid file distribution is achieved, the flexibility and customization capacity of a production line is enhanced, the production efficiency and the product quality are remarkably improved, and the real-time response capacity of the production process is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing control, and particularly relates to a server manufacturing process management system based on AI automatic generation of work guidance. BACKGROUND

[0002] The technical field of intelligent manufacturing control mainly involves optimizing and monitoring production processes, quality control, equipment management and other links through information technology and automation means in the manufacturing process. Its core tasks include production data collection and analysis, production scheduling and execution, equipment state monitoring and maintenance, quality traceability and management, material management and logistics control, etc. In this technical field, the focus is to achieve close cooperation between information systems and manufacturing equipment, personnel and products to improve production efficiency, reduce human errors, ensure product quality and achieve dynamic optimization. Intelligent manufacturing control supports various decisions in the production process through real-time data collection, analysis and feedback, which helps to achieve flexibility, customization and optimal allocation of resources in the production line.

[0003] Among them, the traditional server manufacturing process management system refers to that in the server manufacturing process, employees operate according to the static work guidance written by engineers. The work guidance needs to go through a long period of writing, auditing, approval and publishing. This static file cannot be updated in time according to the actual data in production, thus having certain limitations and hysteresis. The writing of work guidance needs to rely on the experience and knowledge of manufacturing process engineers and the organization assets of the team, and lacks a development platform, which leads to the inability to achieve quick response and adjustment when facing problems in production. Although the existing server manufacturing MES system can carry technical documents and record production scheduling and process, it cannot combine new technical standards, personnel conditions and line problems in production with key data for real-time guidance and updating, and lacks a dynamic optimization mechanism. Especially when facing customer customization requirements, it fails to update the work guidance in time, which easily leads to production quality problems.

[0004] The existing technology relies on static work guidance, which cannot be updated according to real-time data in the production process, leading to the inability to adjust or respond to changes in the manufacturing process execution process. The writing and approval period of work guidance is long, which relies on the experience and team knowledge of manufacturing process engineers, leading to the inability to quickly adjust when facing sudden problems or technical upgrades, and making it difficult to meet the growing customization requirements in the production environment. Although the existing MES system can record the production process, it cannot update in real time by combining new technical standards, personnel conditions and line problems with key data, and fails to achieve dynamic optimization in the production process. It also lacks sufficient flexibility and adaptability, which easily leads to production quality problems. Especially when facing customer customization requirements, the control of production efficiency and quality faces great challenges. SUMMARY

[0005] The object of the present application is to solve the shortcomings existing in the prior art, and the server manufacturing process management system based on AI automatic generation of work guidance is proposed.

[0006] In order to achieve the above object, the present application adopts the following technical scheme: the server manufacturing process management system based on AI automatic generation of work guidance comprises:

[0007] The process timing analysis module identifies and calculates the standard working hour interval of the process based on the operation time record of the process in the original order, combines the skill level of the employee and the running state of the equipment, and generates a process timing distribution diagram;

[0008] The work arrangement optimization module calls the process timing distribution diagram, extracts the logical relationship and operation dependency between processes, filters the optimal work station arrangement mode according to the production line configuration quantity and load balancing principle, and establishes a dynamic process distribution model;

[0009] The guidance generation and integration module calls the dynamic process distribution model, matches the work station operation content and operation requirements, generates the corresponding work guidance file by using the natural language processing algorithm, and embeds the technical terms and matters needing attention to form a standardized work guidance set;

[0010] The knowledge base iterative updating module calls the standardized work guidance set, combines the uploaded abnormal solution in the production process, performs incremental updating on the basic work guidance text library, extracts high-frequency problems and key operation points, and constructs the optimized technical knowledge system.

[0011] As a further scheme of the present application, the process timing distribution diagram comprises process number, standard working hour interval, operation dependency relationship, equipment state identifier, skill level mapping value, the dynamic process distribution model comprises work station number, process logic chain, load balancing parameter, arrangement priority label, working hour floating range, the standardized work guidance set comprises work station operation content, technical term index, matters needing attention list, guidance file number, operation step sequence, and the optimized technical knowledge system comprises problem handling scheme library, high-frequency problem label, key operation point index, knowledge base updating frequency, and technical term standardization table.

[0012] As a further scheme of the present application, the process timing analysis module comprises:

[0013] The working hour statistical submodule removes the equipment failure and unplanned downtime based on the operation time record of the process in the original order, calculates the average operation time of each employee under the same process, and generates a process working hour benchmark value;

[0014] The working time interval calculation submodule identifies the standard working time interval of the process based on the working time benchmark value of the process, combined with the differentiated employee skill level and equipment operation status, and marks the upper and lower limits of the interval to obtain the working time distribution characteristics of the process;

[0015] The time series distribution construction submodule calls the process working time distribution characteristics, identifies the logical relationship and operation dependency between processes, formulates the time series structure and sets the associated weight according to the process number and working time interval, and establishes the process time series distribution diagram.

[0016] As a further solution of the present invention, the job arrangement optimization module includes:

[0017] The process logic extraction submodule calls the process timing distribution diagram, extracts the logic chain and operation dependency between processes, analyzes the dependency strength and logical sequence, and obtains the process logic relationship matrix;

[0018] The load balancing calculation submodule calculates the load balancing value of the workstation according to the process logic relationship matrix, combined with the production line configuration quantity and load balancing principle, and screens out the optimal workstation arrangement;

[0019] The dynamic arrangement generation submodule identifies the workstation number and process logic chain based on the optimal workstation arrangement mode, determines the arrangement priority and the working time floating range, and generates a dynamic process allocation model.

[0020] As a further embodiment of the present invention, the guidance generation integration module includes:

[0021] The operation content matching submodule calls the dynamic process allocation model, extracts the workstation operation content and manufacturing process requirements, analyzes the operation complexity and technology dependency, and generates an operation content matching table;

[0022] The term embedding processing submodule uses a natural language processing algorithm to standardize technical terms based on the operation content matching table, embeds the operation content in the matching table, and forms a technical term index;

[0023] The instruction document generation submodule calls the technical term index, combines the precaution list and the operation step sequence, generates a standardized operation instruction document for the corresponding work station, and marks the instruction document number to form a standardized operation instruction set.

[0024] As a further solution of the present invention, the knowledge base iterative update module includes:

[0025] The exception solution collection submodule calls the standardized work instruction set, extracts the exception solutions uploaded during the production process, analyzes the applicable scenarios and solution efficiency of the solutions, and classifies them to form a problem handling solution library;

[0026] The high-frequency problem extraction submodule counts high-frequency problems and key operation points based on the problem solution library, calculates the weight value of high-frequency problems, and obtains high-frequency problem labels;

[0027] The knowledge base update submodule incrementally updates the basic operation instruction text library based on the high-frequency problem tags, combined with the key operation point index and the technical terminology standardization table, records the knowledge base update frequency, and constructs an optimized technical knowledge system.

[0028] As a further solution of the present invention, the system further includes:

[0029] The work instruction distribution module calls the optimized technical knowledge system, automatically distributes the work instruction file to the corresponding work station based on the employee login account and work station number, marks the file version number and update time, and outputs a distribution path mapping table;

[0030] The distribution path mapping table includes a workstation number, an operation instruction file number, a file version number, an update timestamp, and a distribution path identifier.

[0031] As a further solution of the present invention, the operation instruction distribution module includes:

[0032] The file version management submodule calls the tuned technical knowledge system, identifies the version number and update time of the work instruction file, compares the current version with the current version, selects the current version file, and generates a file version mapping table;

[0033] The distribution path generation submodule identifies the distribution path corresponding to the workstation based on the file version mapping table, in combination with the employee login account and the workstation number, records the path identifier and the update timestamp, and generates a distribution path mapping table.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are:

[0035] In this invention, by combining employee skill levels with equipment status to optimize the timing distribution of the manufacturing process, it is possible to accurately calculate and standardize process operation times, thereby dynamically adjusting production arrangements and ensuring that the configuration of workstations is optimized based on the load balancing of the production line. By using natural language processing algorithms to generate work instructions that meet production requirements and accurately embed technical terms and precautions, errors and inconsistencies in production are reduced. With the help of the incremental update mechanism of the knowledge base, the solution can perform real-time optimization based on anomalies generated during the production process, improving the timeliness and accuracy of operating standards, enabling work instructions to be quickly distributed according to actual needs, and enhancing the flexibility and customization capabilities of the production line, thereby significantly improving production efficiency and product quality, and enhancing the real-time responsiveness of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a system flow chart of the present invention;

[0037] Figure 2 This is a flow chart of the process timing analysis module in the present invention;

[0038] Figure 3 This is a flow chart of the job arrangement optimization module in the present invention;

[0039] Figure 4 To guide the generation of an integrated module flow chart in the present invention;

[0040] Figure 5 This is a flow chart of the knowledge base iterative update module in the present invention;

[0041] Figure 6 This is a flow chart of the operation instruction distribution module in the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0043] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0044] See also Figure 1 The server manufacturing process management system that automatically generates work instructions based on AI includes:

[0045] The process timing analysis module identifies and calculates the standard working time interval of the process based on the operation time records of the process in the original order, combined with the employee skill level and equipment operating status, and generates a process timing distribution diagram;

[0046] The job arrangement optimization module calls the process time sequence distribution diagram, extracts the logical relationship and operation dependency between processes, selects the optimal workstation arrangement method based on the production line configuration quantity and load balancing principle, and establishes a dynamic process allocation model;

[0047] The instruction generation and integration module calls the dynamic process allocation model to match the workstation operation content and operation requirements, uses the natural language processing algorithm to generate the corresponding work instruction documents, and embeds technical terms and precautions to form a standardized work instruction set;

[0048] The knowledge base iterative update module calls on standardized work instructions and combines them with exception solutions uploaded during the production process to incrementally update the basic work instruction text library, extracting high-frequency issues and key operating points to build a tuned technical knowledge system.

[0049] The work instruction distribution module calls the optimized technical knowledge system, automatically allocates the work instruction files to the corresponding work stations based on the employee login account and work station number, marks the file version number and update time, and outputs the distribution path mapping table.

[0050] The process timing distribution diagram includes the process number, standard working time interval, operation dependency, equipment status identification, and skill level mapping value. The dynamic process allocation model includes the workstation number, process logic chain, load balancing parameters, arrangement priority label, and working time floating range. The standardized operation instruction set includes the workstation operation content, technical term index, precaution list, instruction document number, and operation step sequence. The optimized technical knowledge system includes a problem handling solution library, high-frequency problem labels, key operation point index, knowledge base update frequency, and technical term standardization table. The distribution path mapping table includes the workstation number, operation instruction document number, file version number, update timestamp, and distribution path identification.

[0051] See also Figure 2 , the process timing analysis module includes:

[0052] The working time statistics submodule calculates the average operating time of each employee in the same process based on the operating time records of the original order, removes equipment failures and unplanned downtime, and generates a process working time benchmark value;

[0053] Based on the operation time records of the processes in the original order, when the production data of "Product A" in the original order "Order No.: 20240615001" is obtained, first identify and retrieve all the operation time records of the processes related to the production of "Product A". For example, for the process "drilling", the original record shows that the operation time includes "Employee Zhang San: 8.5 minutes (normal operation), 1.2 minutes (equipment failure shutdown), Employee Li Si: 9.0 minutes (normal operation), 0.5 minutes (unplanned shutdown), Employee Wang Wu: 8.8 minutes (normal operation)", etc. Then, automatically identify and eliminate the equipment failure time and unplanned shutdown time in the record, for example, the 1.2 minute equipment failure shutdown time and unplanned shutdown time of employee Zhang San in the "drilling" process The 0.5 minutes of unplanned downtime of employee Li Si is removed, and the actual effective operating time is 8.5 minutes for Zhang San and 9.0 minutes for Li Si. At the same time, for the same process "drilling", the effective operating time records of all employees are collected. For example, the effective time of employee Zhang San is 8.5 minutes, the effective time of employee Li Si is 9.0 minutes, and the effective time of employee Wang Wu is 8.8 minutes. The effective operating time is summarized and the average value is calculated, that is, (8.5+9.0+8.8) / 3=8.77 minutes. This average value is used as the working time benchmark value for the "drilling" process. This working time benchmark value is entered into the process working time benchmark database together with the benchmark values ​​of other processes (such as "milling", "grinding", etc.) to generate the process working time benchmark value.

[0054] The working time interval calculation submodule identifies the standard working time interval of the process based on the working time benchmark value of the process, combined with the differentiated employee skill levels and equipment operating status, and marks the upper and lower limits of the interval to obtain the working time distribution characteristics of the process;

[0055] Based on the process working time benchmark value, after obtaining the working time benchmark value of 8.77 minutes for the process "drilling", the preset employee skill level parameter is called. This parameter divides the employee skill level into "primary", "intermediate" and "senior", and corresponds to different operation time floating coefficients respectively. For example, the floating coefficient of "primary" employees is [+15%, +25%], the floating coefficient of "intermediate" employees is [+5%, +15%], and the floating coefficient of "senior" employees is [-5%, +5%]. At the same time, combined with the equipment operation status parameter, this parameter divides the equipment operation status into "good", "general" and "to be maintained" according to the equipment maintenance record and real-time sensor data, and corresponds to different time adjustment coefficients respectively. For example, the "good" status corresponds to [0%, +2%], the "general" status corresponds to [+2%, +5%], and the "to be maintained" status corresponds to [+5%, +10%]. The process working time benchmark value, employee skill level floating coefficient and equipment operation status adjustment coefficient are comprehensively calculated to identify the standard working time range of the process. For example, if a batch of tasks is performed by an employee with an intermediate skill level on equipment in the "normal" state, the working time interval is calculated as follows: the floating range of the intermediate employee is [8.77×(1+0.05), 8.77×(1+0.15)]=[9.21, 10.09] minutes, and the adjustment range of the equipment in the normal state is [8.77×(1+0.02), 8.77×(1+0.05)]=[8.94, 9.21] minutes. The final working time interval is the intersection of the two, that is, [9.21, 9.21]. ], if there is no intersection between the two, take the union and adjust it in steps of 5% until there is an intersection, so as to determine the final standard working time interval as [9.15, 9.35] minutes, where the lower limit of the interval 9.15 minutes represents the shortest operating time under ideal conditions, and the upper limit 9.35 minutes represents the longest operating time considering skills and equipment status. Associate the upper and lower limits of the interval with the process number, and mark and store the working time interval [9.15, 9.35] minutes of the "drilling" process to obtain the working time distribution characteristics of the process.

[0056] The time series distribution construction submodule calls the process time distribution characteristics, identifies the logical relationship and operation dependency between processes, formulates the time series structure and sets the associated weights based on the process number and time interval, and establishes the process time series distribution diagram;

[0057] The process procedure time distribution characteristics are called, and after the process procedure time distribution characteristics of "drilling", "milling", "polishing" and the like are obtained, for example, the standard process procedure time interval of the "drilling" procedure is [9.15, 9.35] minutes, the logical relationship and operation dependency between procedures are identified through a preset manufacturing process procedure and an engineer definition, for example, it is clear that "drilling" must be completed before "milling", and "polishing" must be performed after "drilling" and "milling" are both completed, which indicates that "drilling" is a preceding procedure of "milling", and "drilling" and "milling" are both preceding procedures of "polishing", a time sequence structure is formulated according to the procedure number and the determined process procedure time interval, for example, for the procedure A (drilling), the procedure B (milling) and the procedure C (polishing), the time sequence structure is set as A->B->C, wherein the arrow represents the logical sequence, meanwhile, the correlation weight is set according to the closeness of the operation dependency and the importance of the influence on the subsequent procedure, for example, if the "drilling" has a greater influence on the precision of the "milling", the correlation weight from "drilling" to "milling" is set as 0.8, and if the "milling" has a smaller influence on the surface quality of the "polishing", the correlation weight from "milling" to "polishing" is set as 0.5, the weight value is set in the range of 0 to 1, and the higher the value is, the stronger the dependency is, the procedure number, the process procedure time interval, the logical relationship and the correlation weight are integrated together, and the process procedure time distribution diagram is established.

[0058] Please refer to Figure 3 The operation arrangement optimization module comprises:

[0059] The procedure logic extraction submodule calls the process procedure time distribution diagram, extracts the logical chain and operation dependency between procedures, analyzes the dependency strength and logical sequence, and obtains a procedure logical relationship matrix.

[0060] The process procedure time distribution diagram is called, and after the process procedure time distribution diagram containing the procedures of "drilling", "milling", "polishing" and the like and the logical relationship and correlation weight thereof are obtained, the logical chain and operation dependency between procedures are identified and extracted, for example, it is identified from the distribution diagram that "drilling" is a preceding procedure of "milling", and "drilling" and "milling" are both preceding procedures of "polishing", the dependency strength and logical sequence of each logical chain are further analyzed, for example, the dependency strength between "drilling" and "milling" is 0.8, and the logical sequence is that "drilling" is prior to "milling", and the dependency strength between "milling" and "polishing" is 0.5, and the logical sequence is that "milling" is prior to "polishing", and the structure of the identified logical chain and dependency relationship is structured, and a procedure logical relationship matrix is generated.

[0061] The load balancing calculation submodule combines the procedure logical relationship matrix, the production line configuration number and the load balancing principle, and adopts the formula:

[0062]

[0063] The load balancing value of the workstations is calculated, and the optimal workstation arrangement is screened out;

[0064] Wherein, L represents the load balancing value of the workstations, T i is the standard man-hour of the i-th process, C i is the workstation configuration number of the i-th process, W i is the priority weight of the i-th process, and n represents the total number of processes;

[0065] According to the process logical relationship matrix, combined with the production line configuration number and the load balancing principle, the formula is used to calculate the load balancing value of the workstations;

[0066] Wherein, L represents the load balancing value of the workstations, represents the overall load level of a single workstation when it executes all the processes allocated to it, and the smaller the value is, the more balanced the load is;

[0067] T i is the standard man-hour of the i-th process, which represents the time required to complete the process, and the larger the value is, the longer the process takes;

[0068] C i is the workstation configuration number of the i-th process, which represents the number of workstations or devices used to execute the process, and the larger the value is, the stronger the processing capacity of the process is;

[0069] W i is the priority weight of the i-th process, which represents the importance or urgency of the process in the entire production process, and the larger the value is, the more important the process is, and the higher the priority is;

[0070] n represents the total number of processes, i.e., the number of all processes for which the load balancing value needs to be calculated;

[0071] The fraction represents the average time required for a single workstation to complete the i-th process under a given workstation configuration number, which reflects the processing efficiency of the workstation for the process, multiplied by the priority weight W i is to give higher weight to more important or more urgent processes when calculating the total load, so as to ensure that the load balancing consideration is more in line with actual production needs. The advantage of this formula is that by combining man-hour, workstation configuration number and priority weight, it can comprehensively evaluate the actual workload of each workstation and highlight the impact of high-priority processes on the load, so as to preferentially reduce the burden of key processes when optimizing the arrangement, thereby improving overall production efficiency. For example, for a production line that includes drilling, milling and polishing processes;

[0072] Assuming the drilling process (i = 1) standard time T1 = 9.25 minutes (take the process time interval median), the station configuration number C1 = 2, the priority weight W1 = 0.9;

[0073] The milling process (i = 2) standard time T2 = 12.5 minutes, the station configuration number C2 = 3, the priority weight W2 = 0.7;

[0074] The polishing process (i = 3) standard time T3 = 7.0 minutes, the station configuration number C3 = 1, the priority weight W3 = 0.8;

[0075] Priority weight W i The setting is based on the comprehensive evaluation of the influence degree of each process on the final product quality, delivery time and subsequent process, for example, the drilling process as the basic processing step, its precision directly affects the subsequent milling and polishing, so set a higher weight 0.9, the milling process second, set the weight 0.7, the polishing process is important, but its influence on the overall process is relatively low, so set the weight 0.8, the weight value is set according to historical production data analysis and expert experience, and verified by actual production test to ensure its rationality and effectiveness, for example, through A / B test, compare the influence of different weight settings on production efficiency and product pass rate, verify that the current weight setting can make the load balancing value L within the allowable fluctuation range, at the same time, the production efficiency is improved by 5, and the product pass rate is improved by 2, so as to select the optimal station arrangement mode, its calculation process is as follows:

[0076]

[0077] The calculated load balancing value L = 12.6794 represents the overall load level under the current station arrangement mode, compare this value with the preset load balancing threshold, if the L value is lower than the preset threshold, for example, the threshold is 15, then the current arrangement mode is acceptable, if the L value is significantly higher than the threshold, it means that the load is unbalanced, the station configuration or process allocation needs to be adjusted, through iterative optimization algorithm, adjust the station configuration number C i and process allocation, recalculate the L value, until a arrangement mode is found that makes the L value minimum or lower than the preset threshold, and the optimal station arrangement mode is selected.

[0078] The dynamic arrangement generation submodule is based on the optimal station arrangement mode, identifies the station number and process logic chain, determines the arrangement priority and time floating range, and generates a dynamic process allocation model;

[0079] Based on the optimal station arrangement mode, after obtaining the optimal station arrangement mode screened by the load balancing calculation submodule, the mode clearly indicates the station to which each process should be allocated and the number thereof, for example, the "drilling" process is allocated to station A, the "milling" process is allocated to station B, and the "polishing" process is allocated to station C, and the number of each station and the logical chain of the process carried thereby are identified, for example, station A carries the "drilling" process, which serves as the logical predecessor of the subsequent "milling" and "polishing", based on the priority weight and logical dependency of the process, the priority of each process in the arrangement is determined, for example, the high-priority "drilling" process is arranged first to avoid blocking the subsequent processes, according to the process time interval and equipment running state, the process time floating range of each process is determined, for example, the process time floating range of the "drilling" process is [9.15, 9.35] minutes, which indicates that the completion time of the process can float within this interval in actual operation, and the station number, process logical chain, arrangement priority and process time floating range are integrated to generate a dynamic process allocation model.

[0080] Please refer to Figure 4 The guidance generation module includes:

[0081] The operation content matching submodule calls the dynamic process allocation model, extracts the operation content and manufacturing process requirements of the station, analyzes the operation complexity and technical dependency, and generates an operation content matching table.

[0082] The dynamic process allocation model is called, after obtaining the dynamic allocation model of the "drilling" process, for example, the model indicates that the "drilling" process is allocated to station A, the operation content and manufacturing process requirements of station A are extracted from the preset operation database, for example, the drilling operation content of station A includes "workpiece clamping", "drill bit selection", "drilling depth setting" and the like, and the manufacturing process requirements thereof include "hole diameter accuracy ±0.02mm", "surface roughness Ra1.6" and the like, the complexity and technical dependency of the operation content are analyzed, for example, "drilling depth setting" involves numerical control machine tool parameter adjustment, which has high complexity and high dependency on the technical proficiency of the operator, while "workpiece clamping" is relatively simple and has low technical dependency, and the operation content, manufacturing process requirements, operation complexity analysis result and technical dependency evaluation result are integrated to generate an operation content matching table.

[0083] The term embedding processing submodule uses natural language processing algorithms to standardize technical terms based on the operation content matching table, embeds the operation content in the matching table, and forms a technical term index.

[0084] Based on the operation content matching table, after obtaining the "drilling depth setting", "hole diameter accuracy

[0085] \pm0.02mm” and technical requirements, and then using natural language processing algorithms, such as based on morphological analysis and entity recognition technology, to standardize technical terms, for example, standardizing “\pm0.02mm” to “tolerance range 0.02 millimeters”, and “Ra1.6” to “surface roughness Ra1.6 microns”, to ensure the uniformity and standardization of the term expression, and then embedding the standardized technical terms into the corresponding operation content description in the operation content matching table, for example, embedding “tolerance range 0.02 millimeters” in the operation description of “drilling depth setting”, to form a technical term index.

[0086] The guide file generation submodule calls the technical term index, combines the notes list and the operation step sequence, generates a standardized job guide file for the corresponding work station, and marks the guide file number to form a standardized job guide set;

[0087] After obtaining the index containing standardized technical terms, such as “tolerance range 0.02 millimeters” embedded in “drilling depth setting”, the technical term index is called, and the pre-set notes list and operation step sequence are combined, such as “notes list” including “drill wear check”, “cooling liquid use specification”, and “operation step sequence” including “step 1: check workpiece clamping, step 2: select corresponding drill, step 3: set drilling depth”, the information is integrated to generate a standardized job guide file for the corresponding work station, for example, generating a detailed “drilling procedure job guide” for work station A, which contains standardized terms, operation steps and notes, at the same time, the system will mark a unique guide file number for this guide file, for example, “GUIDE-DRL-001”, and archive it to form a standardized job guide set.

[0088] Please refer to Figure 5 The knowledge base iterative update module includes:

[0089] The abnormal solution collection submodule calls the standardized job guide set, extracts the abnormal solutions uploaded in the production process, analyzes the applicable scenarios and solution efficiency of the solutions, and classifies them to form a problem handling solution library;

[0090] The standardized work instruction set is called. After obtaining standardized work instruction documents such as the "Drilling Process Work Instruction (No.: GUIDE-DRL-001)", the abnormal solution data uploaded during the production process is monitored in real time. For example, when the operator encounters the abnormal situation of "drill bit breakage" when performing the "drilling" process and submits the solution of "replace the drill bit and adjust the feed speed", the abnormal solution will be extracted and the applicable scenarios and solution efficiency of the solution will be analyzed. For example, the solution of "drill bit breakage" "replace the drill bit and adjust the feed speed" is applicable to the situation where the drill bit is overloaded or the material is too hard in the "drilling" process, and the solution efficiency of this solution is recorded as high (for example, the average solution time is 5 minutes and the success rate is 95%). The solutions and their applicable scenarios and solution efficiency are classified. For example, all drill problem solutions related to the "drilling" process are grouped under the "drilling exception handling" category to form a problem handling solution library.

[0091] The high-frequency problem extraction submodule is based on the problem solution library, and counts high-frequency problems and key operation points using the formula:

[0092]

[0093] Calculate the weight value of high-frequency questions and obtain the high-frequency question labels;

[0094] Among them, F represents the weight value of high-frequency problems, P j is the frequency of occurrence of the j-th question, S j is the efficiency of solving the jth problem, m represents the total number of problems;

[0095] Based on the problem handling solution library, after obtaining the problem handling solution library containing various abnormal solutions, for example, including "drill bit breakage", "hole diameter tolerance", "surface roughness substandard" and other problems and their solutions, first count the high-frequency problems and key operation points in the library. For example, through data analysis, it was found that "drill bit breakage" occurred 50 times in the past month, which is a high-frequency problem, and its key operation points are "drill bit selection" and "feed speed setting". The formula is used Calculate the weight value of high-frequency questions;

[0096] Among them, F represents the weight value of high-frequency problems, which indicates the comprehensive importance of a certain problem among all problems. The larger the value, the more important the problem is and the more priority it needs to be paid attention to and solved.

[0097] P j is the frequency of occurrence of the jth problem, indicating the number of times the problem occurs in a specific time period. A larger value indicates a more common problem.

[0098] S jis the efficiency of solving the jth problem, which indicates the time or resources required to solve the problem, expressed as a percentage or fraction. A larger value indicates that the problem is easier or faster to solve;

[0099] m represents the total number of questions, that is, the number of all questions that need to calculate the weight value of high-frequency questions;

[0100] The summation symbol ∑ indicates the accumulation of the calculation results of all m problems;

[0101] Fraction molecules It represents the sum of the product of the frequency of occurrence of all problems and the efficiency of solving them, which can be understood as the sum of the impact of the problems, and the denominator is Represents the sum of the efficiency of solving all problems. This ratio measures the weighted average importance of high-frequency problems in the overall problem set.

[0102] The benefit of this formula is that it not only considers the frequency of problem occurrence, but also the efficiency of solving the problem. This gives more weight to problems that occur frequently and are easy to solve. This allows the knowledge base to be updated to prioritize solutions to "high-frequency, easy-to-solve" problems, thereby rapidly improving production efficiency.

[0103] For example, for the following three questions:

[0104] Problem 1 (drill bit breakage): frequency P1 = 50 times, solution efficiency S1 = 0.9 (representing a 90% solution success rate);

[0105] Problem 2 (aperture deviation): frequency of occurrence P2 = 20 times, resolution efficiency S2 = 0.7;

[0106] Problem 3 (surface roughness does not meet the standard): frequency of occurrence P3 = 30 times, resolution efficiency S3 = 0.8;

[0107] Solution efficiency S j The setting of S1 is derived from a comprehensive evaluation of historical resolution records, including average resolution time and first-time resolution success rate. For example, by analyzing the past 100 cases of resolving the "broken drill bit" problem, it was found that the average resolution time was 5 minutes and the first-time resolution success rate was 90%. Therefore, S1 was set to 0.9, and so on. The value is obtained through data statistics and analysis in an actual production environment to ensure its authenticity and representativeness. The calculation process is as follows:

[0108]

[0109] The calculated high-frequency problem weight value F is approximately 34.58, which represents the importance of the problem in the comprehensive consideration of frequency and solution efficiency. By comparing this weight value with other problems, the problems with F values higher than the preset threshold (for example, the threshold is set to 30) are marked as high-frequency problem labels, for example, the "drill bit fracture" problem is identified as a high-frequency problem because its F value is higher than 30. The high-frequency problem labels are stored to obtain the high-frequency problem labels.

[0110] The knowledge base updating submodule updates the basic job instruction text library incrementally according to the high-frequency problem labels, in combination with the key operation point index and the technical term standardization table, records the knowledge base update frequency, and constructs the optimized technical knowledge system;

[0111] According to the high-frequency problem labels, after obtaining the high-frequency problem labels such as "drill bit fracture", in combination with the key operation point index and the technical term standardization table, for example, the key operation point index shows that the key operation point of "drill bit fracture" is "drill bit selection" and "feed speed setting", and the technical term standardization table contains standardized terms, the basic job instruction text library is updated incrementally, for example, in the "drilling process job instruction book", for the high-frequency problem of "drill bit fracture", relevant chapters are added or modified, for example, in the "drill bit selection" part, the guidance content of "selecting a drill bit with appropriate hardness for different materials" is added, and in the "feed speed setting" part, the detailed explanation of "suggesting to adjust the feed speed according to the drill bit diameter and material hardness to avoid too fast or too slow" is supplemented, the updated content strictly uses the standardized terms in the technical term standardization table, and refers to the specific operation details in the key operation point index, at the same time, the frequency and specific update content of this knowledge base update are recorded, for example, "update frequency: update every week, this update date: June 16, update content: add "drill bit fracture" problem solving scheme and related operation point optimization", the optimized technical knowledge system is constructed.

[0112] Please refer to Figure 6 , the job instruction distribution module includes:

[0113] The file version management submodule calls the optimized technical knowledge system, identifies the version number and update time of the job instruction file, compares the current version with the current version, selects the current version file, and generates a file version mapping table;

[0114] The optimized technical knowledge system is called. After obtaining the optimized technical knowledge system including the updated "Drilling Process Operation Instructions", the version number and update time of each operation instruction document are identified. For example, the current version number of the "Drilling Process Operation Instructions" is V1.2, and the update time is 10:00 on June 16. The current version and the current version are compared regularly or on demand. For example, it is found that the version number of the old version of the "Drilling Process Operation Instructions" on the production line is V1.1, and its update time is 09:00 on May 10, which is inconsistent with the current version V1.2 in the knowledge system. The current version file, that is, the V1.2 version of the "Drilling Process Operation Instructions", will be automatically selected and associated with the old version to generate a file version mapping table. For example, the mapping table will record "old version V1.1 corresponds to the current version V1.2, and the file path is XXX".

[0115] The distribution path generation submodule identifies the distribution path of the corresponding workstation based on the file version mapping table, combines the employee login account and the workstation number, records the path identifier and update timestamp, and generates a distribution path mapping table;

[0116] Based on the file version mapping table, after obtaining the file version mapping table containing the current version file information such as "Drilling Process Operation Instructions V1.2", combined with the employee login account and workstation number, for example, it is identified that the employee "Zhang Ming" login account is "ZM001" and his current operation workstation number is "DRL-01"

[0117] (Drilling station), according to the workstation number of the employee, identify the distribution path of the corresponding workstation. For example, the distribution path of the work instruction file of the workstation "DRL-01" is set to " / / server / guidance / DRL-01 / ", and record this distribution path together with the file identifier and update timestamp. For example, record "GUIDE-DRL-001_V1.2: distribution path / / server / guidance / DRL-01 / , timestamp 202506161000" to generate a distribution path mapping table.

[0118] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A server manufacturing process management system based on AI that automatically generates work instructions, characterized by: The system comprises: The process timing analysis module identifies and calculates the standard working time interval of the process based on the operation time records of the process in the original order, combined with the employee skill level and equipment operating status, and generates a process timing distribution diagram; The job arrangement optimization module calls the process time sequence distribution diagram, extracts the logical relationship and operation dependency between processes, selects the optimal workstation arrangement method based on the production line configuration quantity and load balancing principle, and establishes a dynamic process allocation model; The instruction generation and integration module calls the dynamic process allocation model, matches the workstation operation content and operation requirements, uses the natural language processing algorithm to generate corresponding work instruction documents, and embeds technical terms and precautions to form a standardized work instruction set; The knowledge base iterative update module calls the standardized work instruction set, combines the exception solutions uploaded during the production process, incrementally updates the basic work instruction text library, extracts high-frequency problems and key operation points, and constructs an optimized technical knowledge system.

2. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 1 is characterized in that: The process timing distribution diagram includes process number, standard working time interval, operation dependency, equipment status identification, and skill level mapping value; the dynamic process allocation model includes workstation number, process logic chain, load balancing parameters, arrangement priority label, and working time floating range; the standardized operation instruction set includes workstation operation content, technical term index, precaution list, instruction document number, and operation step sequence; the optimized technical knowledge system includes problem handling solution library, high-frequency problem label, key operation point index, knowledge base update frequency, and technical term standardization table.

3. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 1 is characterized in that: The process timing analysis module includes: The working time statistics submodule calculates the average operating time of each employee in the same process based on the operating time records of the original order, removes equipment failures and unplanned downtime, and generates a process working time benchmark value; The working time interval calculation submodule identifies the standard working time interval of the process based on the working time benchmark value of the process, combined with the differentiated employee skill level and equipment operation status, and marks the upper and lower limits of the interval to obtain the working time distribution characteristics of the process; The time series distribution construction submodule calls the process working time distribution characteristics, identifies the logical relationship and operation dependency between processes, formulates the time series structure and sets the associated weight according to the process number and working time interval, and establishes the process time series distribution diagram.

4. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 3 is characterized in that: The job arrangement optimization module includes: The process logic extraction submodule calls the process timing distribution diagram, extracts the logic chain and operation dependency between processes, analyzes the dependency strength and logical sequence, and obtains the process logic relationship matrix; The load balancing calculation submodule calculates the load balancing value of the workstation according to the process logic relationship matrix, combined with the production line configuration quantity and load balancing principle, and screens out the optimal workstation arrangement; The dynamic arrangement generation submodule identifies the workstation number and process logic chain based on the optimal workstation arrangement mode, determines the arrangement priority and the working time floating range, and generates a dynamic process allocation model.

5. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 4 is characterized in that: The guidance generation integration module includes: The operation content matching submodule calls the dynamic process allocation model, extracts the workstation operation content and manufacturing process requirements, analyzes the operation complexity and technology dependency, and generates an operation content matching table; The term embedding processing submodule uses a natural language processing algorithm to standardize technical terms based on the operation content matching table, embeds the operation content in the matching table, and forms a technical term index; The instruction document generation submodule calls the technical term index, combines the precaution list and the operation step sequence, generates a standardized operation instruction document for the corresponding work station, and marks the instruction document number to form a standardized operation instruction set.

6. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 5 is characterized in that: The knowledge base iterative update module includes: The exception solution collection submodule calls the standardized work instruction set, extracts the exception solutions uploaded during the production process, analyzes the applicable scenarios and solution efficiency of the solutions, and classifies them to form a problem handling solution library; The high-frequency problem extraction submodule counts high-frequency problems and key operation points based on the problem solution library, calculates the weight value of high-frequency problems, and obtains high-frequency problem labels; The knowledge base update submodule incrementally updates the basic operation instruction text library based on the high-frequency problem tags, combined with the key operation point index and the technical terminology standardization table, records the knowledge base update frequency, and constructs an optimized technical knowledge system.

7. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 1 is characterized in that: The system also includes: The work instruction distribution module calls the optimized technical knowledge system, automatically distributes the work instruction file to the corresponding work station based on the employee login account and work station number, marks the file version number and update time, and outputs a distribution path mapping table; The distribution path mapping table includes a workstation number, an operation instruction file number, a file version number, an update timestamp, and a distribution path identifier.

8. The server manufacturing process management system based on AI automatic generation of work instructions according to claim 7 is characterized in that: The operation instruction distribution module includes: The file version management submodule calls the tuned technical knowledge system, identifies the version number and update time of the work instruction file, compares the current version with the current version, selects the current version file, and generates a file version mapping table; The distribution path generation submodule identifies the distribution path of the corresponding workstation based on the file version mapping table, in combination with the employee login account and the workstation number, records the path identifier and the update timestamp, and generates a distribution path mapping table.

Images