Method and device for supporting complex product manufacturing process model reconstruction

Abstract

The invention provides a method and device for supporting complex product manufacturing process model reconstruction. A user can input manufacturing process reconstruction requirements according to a language structure specified by a manufacturing process reconstruction template, wherein the manufacturing process reconstruction requirements comprise production units, manufacturing activities needing to be added, manufacturing targets and manufacturing constraints; the method comprises the following steps: analyzing a reconstruction demand according to a manufacturing process reconstruction dictionary, performing information extraction and element matching in a manufacturing process model library according to a production unit and manufacturing activities needing to be added to obtain a primary manufacturing process model scheme, and optimizing the primary manufacturing process model scheme according to a manufacturing target and manufacturing constraints to obtain a manufacturing process model; and a final manufacturing process scheme is reconstructed, and a model-driven self-adaptive reconfigurable manufacturing process function is realized. By using the method, rapid generation of the complex product manufacturing process model can be realized, and the efficiency and the quality of product design are improved.

Description

Technical Field

[0001] The present invention relates to the fields of system engineering and computer technology, and in particular to a method and device for supporting reconstruction of a complex product manufacturing process model. Background Art

[0002] As product complexity continues to increase, product manufacturing processes are also becoming increasingly complex. Model-based Systems Engineering (MBSE) supports formalization, modeling, design, analysis, and verification in the product manufacturing process through models, providing solutions to meet the manufacturing challenges of complex systems. Usually, in order to cope with different models of products or different batches of the same model, different manufacturing process models need to be designed to meet the needs of product manufacturing. This process involves the reconstruction of the manufacturing process model. Traditional manufacturing process model reconstruction methods mainly rely on manual experience and manual configuration, which have problems such as low efficiency and prone to errors. An automated method is needed to achieve adaptive reconstruction of the manufacturing process model to improve the efficiency and accuracy of reconstruction.

[0003] Automated adaptive reconstruction methods can efficiently reconstruct and optimize manufacturing processes by utilizing advanced algorithms and technologies, combined with artificial intelligence and algorithm optimization, thereby quickly responding to product changes and manufacturing needs.

[0004] CN202011214559.6 discloses a system and method for automatic generation and self-improvement of a dynamic flexible process, which decomposes each node in the process into tasks and data to establish a task model and a data model. The task model is associated with the data model to form an upstream and downstream relationship between tasks, and the position of each node in the task model in the process is determined by the generation and use of data, thereby automatically generating a process. When the input and output data model nodes of the task model nodes change, the process automatically changes, realizing the model-based dynamic process construction and self-improvement functions.

[0005] The method includes the following steps: S1: obtain the top-level task, select at least one task model node related to the top-level task according to multiple task model nodes in the task model module, and determine the task instance node after task decomposition in the task decomposition module; S2: establish a complete data instance node in the data storage module according to multiple data model nodes in the data model module; S3: determine the input-output relationship between the task instance node and multiple data model nodes according to the input-output relationship between multiple task model nodes and multiple data model nodes defined in the task data relationship module, and determine the input-output relationship between the task instance node and the complete data instance node; S4: obtain the execution order between the task instance nodes according to the input-output relationship between the task instance node and the complete data instance node. Although this method solves the problem of automatic process generation to a certain extent, it still has the following major disadvantages in terms of manufacturing process reconstruction:

[0006] 1. Lack of intelligent optimization and adaptive capabilities

[0007] In the process of process generation and self-improvement, existing technologies mainly rely on the association between predefined tasks and data models, and lack intelligent optimization and adaptive mechanisms. Since no optimization algorithm is introduced, the generated process may not be the optimal solution and cannot meet the optimization requirements of manufacturing goals.

[0008] 2. Insufficient processing capabilities for user input requirements

[0009] Existing methods mainly generate processes through task decomposition and data model association, and their ability to process user input requirements is relatively limited. For complex manufacturing process reconstruction requirements, it is difficult for the system to accurately understand and parse them, resulting in the generated process model being difficult to fully meet actual manufacturing requirements. This limitation significantly reduces the practicality of the system and user experience.

[0010] 3. Lack of support for model reconstruction of heterogeneous manufacturing process models

[0011] In the existing technology, the reconstruction of manufacturing process models is usually only for manufacturing process models in a specific modeling language, and it is difficult to support manufacturing process models in other modeling languages. However, the manufacturing process model can be constructed by multiple modeling languages ​​such as SysML, UML, BPMN, etc., and the semantic syntax of different modeling languages ​​is heterogeneous. The current technology has limited support for model reconstruction of heterogeneous manufacturing process models. Summary of the invention

[0012] In view of this, the present invention provides a method and device that supports the reconstruction of complex product manufacturing process models, which can construct a manufacturing process according to user needs, and solve the optimal model solution through an optimization algorithm, thereby realizing the rapid generation of complex product manufacturing process models and improving the efficiency and quality of product design.

[0013] In order to solve the above technical problems, the present invention is implemented as follows.

[0014] A method for supporting complex product manufacturing process model reconstruction, comprising:

[0015] Step S1: construct a manufacturing process model library, define a manufacturing process reconstruction dictionary and a manufacturing process reconstruction requirement template;

[0016] The manufacturing process model library includes the production units involved in the basic manufacturing process and the manufacturing activities they perform, as well as the execution sequence between the manufacturing activities;

[0017] The manufacturing process reconstruction requirement template is constructed based on the manufacturing process reconstruction dictionary, and specifies the sentence structure of the manufacturing process model reconstruction requirement text, including a production unit selection template, a manufacturing activity addition template, a manufacturing target template, and a manufacturing constraint template;

[0018] Step S2: Reconstruct the demand template according to the manufacturing process, input the manufacturing process model to reconstruct the demand text; parse the manufacturing process reconstruction demand according to the manufacturing process reconstruction dictionary, including production units, new manufacturing activities, manufacturing goals, and manufacturing constraints;

[0019] Step S3: According to the production units and newly added manufacturing activities analyzed in step S2, matching is performed in the manufacturing process model library, the manufacturing activities to be performed by the production units, as well as the production unit information, all manufacturing activity information and execution sequence are extracted, and a preliminary manufacturing process model solution is generated;

[0020] Step S4: Based on the manufacturing objectives and manufacturing constraints analyzed in step S2, multi-objective optimization is performed on the manufacturing activities in the preliminary manufacturing process model scheme to reconstruct the final manufacturing process scheme.

[0021] Preferably, step S4 uses a non-dominated sorting genetic algorithm II (NSGA-II) to perform multi-objective optimization.

[0022] Preferably, when performing multi-objective optimization in step S4, each manufacturing activity in the preliminary manufacturing process model scheme is taken as a node to construct a node set; a group of node subsets are generated as an initial population by randomly sampling in the node set, and each node in the node set is adaptively selected based on the resolved manufacturing goals and manufacturing constraints to obtain an optimal solution set.

[0023] Preferably, in step 1, the manufacturing process model library is constructed based on a manufacturing process metamodel library built on the basis of a multi-architecture modeling language KARMA; the metamodels included in the manufacturing process metamodel library include:

[0024] The graphic element models include: a manufacturing process graphic element model, which is used to describe the manufacturing process;

[0025] The object metamodel includes: a production unit object metamodel, a manufacturing activity object metamodel, a start object metamodel and an end object metamodel; the production unit object metamodel is used to describe the production unit in the manufacturing process and its related characteristics, including a name attribute; the manufacturing activity object metamodel is used to describe the manufacturing activity in the manufacturing process and its related characteristics, including a name attribute, a time attribute, a cost attribute, and a resource consumption rate attribute;

[0026] The relational metamodel includes: control flow relational metamodel, which is used to describe the sequence relationship between two manufacturing activities.

[0027] Preferably, the manufacturing process model library constructed in step S1 includes:

[0028] The production unit objects include: purchasing department, quality inspection department, production planning department, warehousing department, production workers, assembly workers, surface treatment workers, equipment maintenance workers, packaging workers, logistics department, data analysts;

[0029] Manufacturing activity objects include: purchasing raw materials constructed in the purchasing department object; accepting materials, quality inspection during production, product testing, and final product inspection constructed in the quality inspection department object; formulating production plans and arranging production tasks constructed in the production planning department object; managing raw material inventory, material distribution to production lines, and storing finished products constructed in the warehousing department object; pre-processing materials and parts manufacturing constructed in the production worker object; assembling products constructed in the assembly worker object; surface treatment and painting constructed in the surface treatment worker object; daily equipment inspection, preventive maintenance, and equipment failure repair constructed in the equipment maintenance worker object; packaging finished products constructed in the packaging worker object; arranging transportation constructed in the logistics department object; analyzing production data constructed in the data analyst object; Manufacturing activities also include: start objects and end objects;

[0030] After constructing the manufacturing activities, the control flow relational metamodel is used to connect the manufacturing activities in the required sequence to form a basic manufacturing process.

[0031] Preferably, the manufacturing process reconstruction dictionary includes {new manufacturing process model: new model}, {should: shall}, {include: contain}, {add: add}, {increase: add}, {name: name}, {is: is}, {production activity: object}, {maximum: max}, {greater than or equal to: more than}, {greater than: more than}, {less than or equal to: rather than}, {less than: rather than}, {manufacturing process model, model}.

[0032] Preferably, the manufacturing process reconstruction requirement template is specifically:

[0033] The production unit selection template is:<new model> <shall> <contain>…, …, …, ;

[0034] The template for adding manufacturing activities is:… <shall> <add> <name> <is> … <object>;

[0035] The manufacturing target template is: <model> … <max min>;

[0036] The manufacturing constraint template is:… <shall> <add> <name> <is> … <object>.

[0037] Preferably, in step S3, generating a preliminary manufacturing process model solution includes: manufacturing process model data conversion and preliminary manufacturing process solution generation and display:

[0038] The step of converting the manufacturing process model data is to convert the data to be displayed in the extracted data into form data;

[0039] The steps of generating and displaying the preliminary manufacturing process plan display the converted form data in a table format.

[0040] Preferably, the production unit objects and manufacturing activity objects in the manufacturing process model library also include the position and size of the object nodes on the layout;

[0041] When the final manufacturing process plan is reconstructed in step S4, the manufacturing activities and production units are matched in the manufacturing process model library according to the optimization plan, and the positions and sizes of these manufacturing activities and production units are obtained, and provided to the modeling tool to display icons of each object on the modeling canvas for users to view and modify.

[0042] The present invention also provides a device for supporting complex product manufacturing process model reconstruction, comprising: a manufacturing process metamodel library construction unit, a manufacturing process model library construction unit, and a manufacturing process model reconstruction unit;

[0043] The manufacturing process metamodel library construction unit is used to build a manufacturing process metamodel library based on the multi-architecture modeling language KARMA;

[0044] The manufacturing process model library construction unit is used to construct a manufacturing process model library based on the manufacturing process meta-model library; the manufacturing process model library includes the production units involved in the basic manufacturing process and the manufacturing activities performed by them, as well as the execution order between the manufacturing activities;

[0045] The manufacturing process model reconstruction unit includes an analysis module, a preliminary solution generation module, an optimization module and a reconstruction module;

[0046] The parsing module is used to receive the manufacturing process model reconstruction requirement text organized according to the sentence structure provided by the manufacturing process reconstruction requirement template; parse the manufacturing process reconstruction requirement according to the manufacturing process reconstruction dictionary, including production units, new manufacturing activities, manufacturing goals, and manufacturing constraints;

[0047] The manufacturing process reconstruction requirement template is constructed based on the terms provided by the manufacturing process reconstruction dictionary, and includes a production unit selection template, a manufacturing activity addition template, a manufacturing target template, and a manufacturing constraint template;

[0048] The preliminary solution generation module is used to match the production units and newly added manufacturing activities analyzed by the analysis module in the manufacturing process model library, extract the manufacturing activities to be performed by the production units, as well as the production unit information, all manufacturing activity information and execution order, and generate a preliminary manufacturing process model solution;

[0049] The optimization module is used to perform multi-objective optimization on the manufacturing activities in the preliminary manufacturing process model scheme according to the manufacturing objectives and manufacturing constraints analyzed by the analysis module;

[0050] The reconstruction module is used to reconstruct the final manufacturing process plan according to the multi-objective optimization results of the optimization module.

[0051] Beneficial effects:

[0052] (1) Standardized manufacturing process reconstructs the way of expressing requirements

[0053] The existing manufacturing process construction and reconstruction technology usually relies on human experience and manual configuration. This method is time-consuming and error-prone, and cannot quickly respond to the changing needs of complex battlefield environments. The present invention provides a standardized expression method by defining a manufacturing process reconstruction dictionary and a manufacturing process reconstruction requirement template. Through word segmentation, part-of-speech tagging and sentence processing technology, the user's textual requirements can be accurately parsed, significantly improving the efficiency and accuracy of manufacturing process reconstruction.

[0054] (2) Manufacturing process optimization based on multi-objective optimization algorithm

[0055] Existing manufacturing process optimization usually only arranges manufacturing activities based on the matching relationship and priority weight relationship between data and manufacturing activities in the manufacturing process, and recombines the location and execution sequence of manufacturing activities in the manufacturing process model. It lacks systematic algorithm support and is difficult to meet the needs of manufacturing process model reconstruction in complex scenarios. After obtaining a preliminary manufacturing process model solution, the present invention further performs multi-objective optimization on the manufacturing activities in the manufacturing process model based on the manufacturing goals and manufacturing constraints obtained from the manufacturing process reconstruction requirement template, generates an optimal set of manufacturing activities, and significantly improves the optimization efficiency and effect.

[0056] (3) Model Reconstruction of Heterogeneous Manufacturing Process Model

[0057] In model-based system engineering, the manufacturing process model can be constructed by BPMN, SysML, UML or a specific domain modeling language. The semantic syntax of different modeling languages ​​is heterogeneous, and the existing technology can only support the reconstruction of the manufacturing process model under a specific modeling language. The present invention constructs a manufacturing process model based on the multi-architecture modeling language KARMA. Since the KARMA language can support the unified expression of multiple heterogeneous manufacturing processes, the solution of the present invention can also support the model reconstruction of heterogeneous manufacturing process models.

[0058] (4) In a preferred embodiment, the information of the production unit objects and manufacturing activity objects in the manufacturing process model library also includes the position and size of the object nodes on the layout, so that the newly generated demand model will not be unable to be viewed due to problems such as position confusion, thereby improving the readability of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A schematic flow chart of a method for supporting complex product manufacturing process model reconstruction provided by the present invention;

[0060] Figure 2 A schematic diagram of the framework of the method for supporting complex product manufacturing process model reconstruction provided by the present invention;

[0061] Figure 3 A schematic diagram of the structure of a device for supporting reconstruction of a complex product manufacturing process model provided by the present invention. DETAILED DESCRIPTION

[0062] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0063] like Figure 1 and Figure 2 As shown, the present invention proposes a method for supporting the reconstruction of a complex product manufacturing process model, comprising the following steps:

[0064] Step S1: Construct a manufacturing process metamodel library.

[0065] First, a domain-specific metamodel library is constructed based on the multi-architecture modeling language KARMA. The domain-specific metamodel library includes six metamodels (graph, object, relationship, node, attribute, and role). The relationship between these six metamodels is as follows: the graph metamodel includes the object metamodel and the relationship metamodel; the object metamodel includes the attribute metamodel and the node metamodel; the relationship metamodel includes the attribute metamodel, and the two ends of the relationship metamodel are bound to the role metamodel; the node metamodel exists in dependence on the object metamodel; the attribute metamodel is the attribute of the object metamodel and the relationship metamodel; the role metamodel is bound to the two ends of the relationship metamodel.

[0066] The specific domain metamodel library constructed for the reconstruction of the manufacturing process model in the present invention is a manufacturing process metamodel library, and its six metamodels are:

[0067] (1) The graphic element model includes:

[0068] The manufacturing process graph metamodel is used to construct a manufacturing process model and describe the manufacturing process. The manufacturing process graph metamodel is configured as follows:

[0069] The configuration object metamodel includes: production unit object metamodel, manufacturing activity object metamodel, start object metamodel, and end object metamodel;

[0070] The configuration relation metamodel includes: control flow relation metamodel;

[0071] Configuration property metamodel includes: None,

[0072] The beginning of the configuration control flow relationship metamodel can be connected to the production unit object metamodel, the manufacturing activity object metamodel, and the start object metamodel; the terminal of the configuration control flow relationship metamodel can be connected to the production unit object metamodel, the manufacturing activity object metamodel, and the end object metamodel.

[0073] (2) The object metamodel includes:

[0074] ·Production unit object metamodel: used to describe the production unit and its related characteristics in the manufacturing process; the configuration style is "rounded rectangle"; the configuration attribute metamodel includes: name attribute metamodel

[0075] Type; configure the color to "default";

[0076] Manufacturing activity object metamodel: used to describe manufacturing activities and their related characteristics in the manufacturing process; the configuration style is "rounded rectangle"; the configuration attribute metamodel includes: name attribute metamodel

[0077] type, time attribute metamodel, cost attribute metamodel, resource consumption rate attribute metamodel; configuration color is "default";

[0078] ·Start object metamodel: used to describe the start node of the manufacturing process; the configuration style is "rounded rectangle"; the configuration attribute metamodel includes: none; the configuration color is "default";

[0079] End object metamodel: used to describe the end node of manufacturing; the configuration style is "rounded rectangle

[0080] "Shape"; configuration attribute metamodel includes: None; configuration color is "default".

[0081] (3) The relational metamodel includes:

[0082] ·Control flow relationship metamodel: used to describe the order relationship between two manufacturing activities; the start of the control flow relationship is bound to the previous manufacturing activity, and the end of the control flow relationship is bound to the next manufacturing activity; the configuration attribute metamodel includes: none; configure the start role as the control flow start role metamodel; configure the end role as the control flow end role metamodel, and the configuration style is "bold straight line";

[0083] (4) The attribute metamodel includes:

[0084] Name attribute metamodel: used to describe the name of the object production unit object metamodel and the manufacturing activity object metamodel. The attribute value type is configured as String type and the unit is configured as "None".

[0085] Time attribute metamodel: used to describe the time consumed by the manufacturing activity object metamodel. The attribute value type is configured as Int and the unit of the attribute value is configured as "hour";

[0086] Cost attribute metamodel: used to describe the cost consumed by the manufacturing activity object metamodel. The attribute value type is configured as Int and the unit of the attribute value is configured as "Yuan";

[0087] ● Resource consumption rate attribute metamodel: used to describe the percentage of resources consumed by the manufacturing activity object metamodel. The attribute value type is configured as Int and the unit of the attribute value is configured as "%";

[0088] (5) The point meta-model is none;

[0089] (6) The role meta-model includes:

[0090] Control flow start role metamodel: used to bind the start of the control flow and configure the control flow start role

[0091] The direction of the color element model is "starting point" and the style is "arrow" style;

[0092] Control flow terminal role metamodel: used to bind control flow terminals and configure control flow terminal roles.

[0093] The direction of the color element model is "terminal" and the style is "diamond" style.

[0094] Step S2: Define a manufacturing process reconstruction dictionary.

[0095] The purpose of the manufacturing process reconstruction dictionary definition is to provide the ability to recognize professional terms in the manufacturing process reconstruction requirements, which helps to accurately match and understand the information in the manufacturing process reconstruction requirements text. The definition of each word in the dictionary includes two parts: word content and part of speech. The definition format of the dictionary is {word: part of speech}. The manufacturing process reconstruction dictionary defined in the present invention includes: {new manufacturing process model: new model}, {should: shall}, {include: contain}, {add: add}, {increase: add}, {name: name}, {is: is}, {production activity: object}, {maximum: max}, {greater than or equal to: morethan}, {greater than: more than}, {less than or equal to: rather than}, {less than: rather than}, {manufacturing process model, model}.

[0096] Step S3: Define a manufacturing process reconstruction requirement template.

[0097] The purpose of defining the manufacturing process reconstruction requirement template is to provide a specific statement template, allowing users to input the reconstruction requirements of the manufacturing process model according to the corresponding template. The manufacturing process reconstruction requirement template mainly includes: production unit selection template, manufacturing activity addition template, manufacturing target template and manufacturing constraint template.

[0098] Production unit selection template: This template allows users to enter structured sentences. The template is made based on the parts of speech of the words provided by the dictionary:<new model> <shall> <contain>…, …, …, according to this sentence, the user can input an example of a structured sentence: the new manufacturing process model should include manufacturing workers, technicians, quality inspectors, manufacturing managers, and maintenance personnel. The manufacturing process model reconstruction device of the present invention can process the sentence using NLP natural language processing technology, and obtain the corresponding production unit object and the manufacturing activities it performs from the manufacturing process model library constructed in step S4 as part of the new manufacturing process model.

[0099] Manufacturing activity addition template: The user defines the manufacturing activities that the production unit needs to perform in the manufacturing process model library, but in the new manufacturing process model, some production units may need to perform additional manufacturing activities, which do not exist in the manufacturing process model library. The purpose of this template is to allow users to enter structured sentences. The template is made according to the part of speech:… <shall> <add> <name> <is> … <object>, the user can input a structured sentence such as: technicians should add a production activity named equipment maintenance. The manufacturing process model reconstruction device can process the sentence using NLP natural language processing technology, and add corresponding manufacturing activities to the production unit specified in the sentence, and obtain a preliminary manufacturing process model solution to meet the needs of the new manufacturing process.

[0100] Manufacturing target template and manufacturing constraint template: Based on the preliminary manufacturing process model, it is necessary to define optimization goals and constraints to assist users in generating a manufacturing process model after algorithm processing.

[0101] The purpose of creating a target template is to allow users to enter structured sentences. The template is made based on parts of speech: <model> … <max>, for example: the resource consumption rate of the manufacturing process model is the smallest, and the cost of the manufacturing process model is the smallest. The manufacturing process model reconstruction device can process the sentence using NLP natural language processing technology and convert it into the optimization target of the mathematical model in the algorithm.

[0102] The purpose of making constraint templates is to allow users to enter structured sentences. Templates are made based on parts of speech: <model>…<rather than / more than> …, for example: if the time of the manufacturing process model is less than or equal to 100 hours, and the cost of the manufacturing process model is less than or equal to 300,000 yuan. The manufacturing process model reconstruction device can process the statement using NLP natural language processing technology and convert it into the constraints of the mathematical model in the algorithm. After algorithm processing, the optimal manufacturing process solution is obtained to help users quickly build a new manufacturing process model.

[0103] Step S4: Construct a manufacturing process model library, which includes production unit construction (constructing the production units and their attribute values ​​involved in the manufacturing process model), manufacturing activity construction (constructing the manufacturing activities and their attribute values ​​performed by each production unit in the manufacturing process model) and the topological relationship between manufacturing activities (sequential relationship, parallel relationship, etc.). This manufacturing process model library provides a basic process model. If the actual demand is greater than this model, it can be added through the manufacturing activity addition template.

[0104] The production unit object construction includes: purchasing department, quality inspection department, production planning department, warehousing department, production workers, assembly workers, surface treatment workers, equipment maintenance workers, packaging workers, logistics department, and data analysts.

[0105] Construct a manufacturing activity object in the purchasing department object: Purchase raw materials (the name attribute value is Purchase raw materials, the time attribute value is 10 hours, the cost attribute value is 10,000 yuan, and the resource consumption rate attribute value is 3.5%).

[0106] Construct manufacturing activity objects in the quality inspection department object: acceptance of materials (the name attribute value is acceptance of materials, the time attribute value is 1 hour, the cost attribute value is 1,000 yuan, and the resource consumption rate attribute value is 0.5%), quality inspection during production (the name attribute value is quality inspection during production, the time attribute value is 10 hours, the cost attribute value is 10,000 yuan, and the resource consumption rate attribute value is 2.5%), product testing (the name attribute value is product testing, the time attribute value is 5 hours, the cost attribute value is 2,000 yuan, and the resource consumption rate attribute value is 0.5%), final product inspection (the name attribute value is final product inspection, the time attribute value is 6 hours, the cost attribute value is 10,000 yuan, and the resource consumption rate attribute value is 1.5%),

[0107] Construct manufacturing activity objects in the production planning department object: formulate production plan (name attribute value is formulate production plan, time attribute value is 8 hours, cost attribute value is 20,000 yuan, resource consumption rate attribute value is 2.5%), arrange production tasks (name attribute value is arrange production tasks, time attribute value is 5 hours, cost attribute value is 10,000 yuan, resource consumption rate attribute value is 3.5%),

[0108] Construct manufacturing activity objects in the warehouse department object: manage raw material inventory (name attribute value is manage raw material inventory, time attribute value is 10 hours, cost attribute value is 20,000 yuan, resource consumption rate attribute value is 1.5%), distribute materials to production lines (name attribute value is distribute materials to production lines, time attribute value is 8 hours, cost attribute value is 10,000 yuan, resource consumption rate attribute value is 0.5%), store finished products (name attribute value is store finished products, time attribute value is 12 hours, cost attribute value is 20,000 yuan, resource consumption rate attribute value is 1.5%), store finished products (name attribute value is prepare product shipment, time attribute value is 8 hours, cost attribute value is 15,000 yuan, resource consumption rate attribute value is 0.5%),

[0109] Construct manufacturing activity objects in the production worker object: pre-processing materials (name attribute value is pre-processing materials, time attribute value is 8 hours, cost attribute value is 13,000 yuan, resource consumption rate attribute value is 2.5%), parts manufacturing (name attribute value is parts manufacturing, time attribute value is 35 hours, cost attribute value is 1,200,000 yuan, resource consumption rate attribute value is 35%), and construct manufacturing activity objects in the assembly worker object: assembling products (name attribute value is assembling products, time attribute value is 20 hours, cost attribute value is 1,000,000 yuan, resource consumption rate attribute value is 28%),

[0110] Construct manufacturing activity objects in the surface treatment worker object: surface treatment and painting (the name attribute value is surface treatment and painting, the time attribute value is 15 hours, the cost attribute value is 900,000 yuan, and the resource consumption rate attribute value is 18%), and construct manufacturing activity objects in the equipment maintenance worker object: daily equipment inspection (the name attribute value is daily equipment inspection, the time attribute value is 6 hours, the cost attribute value is 90,000 yuan, and the resource consumption rate attribute value is 7%), preventive maintenance (the name attribute value is equipment maintenance worker, the time attribute value is 6 hours, the cost attribute value is 90,000 yuan, and the resource consumption rate attribute value is 7%), equipment failure maintenance (the name attribute value is equipment failure maintenance, the time attribute value is 8 hours, the cost attribute value is 60,000 yuan, and the resource consumption rate attribute value is 3%),

[0111] Construct a manufacturing activity object in the packing worker object: pack finished products (the name attribute value is pack finished products, the time attribute value is 18 hours, the cost attribute value is 1,000,000 yuan, and the resource consumption rate attribute value is 15%), and construct a manufacturing activity object in the logistics department object: arrange transportation (the name attribute value is arrange transportation, the time attribute value is 10 hours, the cost attribute value is 20,000 yuan, and the resource consumption rate attribute value is 0.6%),

[0112] Construct a manufacturing activity object in the data analyst object: Analyze production data (the name attribute value is Analyze production data, the time attribute value is 2 hours, the cost attribute value is 3,000 yuan, and the resource consumption rate attribute value is 0.2%);

[0113] Manufacturing activity objects also include: start object and end object.

[0114] After constructing the manufacturing activities, the control flow relational metamodel is used to connect the manufacturing activities in a certain order to form a manufacturing process model. For example, the manufacturing activities are connected in the following order: start, purchase raw materials, accept materials, formulate production plans, arrange production tasks, manage raw material inventory, distribute materials to the production line, pre-process materials, manufacture parts, assemble products, surface treatment and painting, quality inspection during production, conduct product testing, daily equipment inspection, preventive maintenance, equipment failure repair, final product inspection, pack finished products, store finished products, prepare product shipment, arrange transportation, record production data, analyze production data, and end.

[0115] The constructed manufacturing process model will be used as an optimization space to participate in subsequent algorithms.

[0116] Step S5: The user inputs the textual expression of the manufacturing process model reconstruction requirement according to the four statement templates specified in the manufacturing process reconstruction requirement template, and performs manufacturing process reconstruction requirement statement processing on the textual expression.

[0117] The purpose of manufacturing process reconstruction requirement statement processing is to process the textual expression of the manufacturing process reconstruction requirement input by the user, obtain recognizable requirement information as the input of manufacturing process reconstruction. Statement processing mainly includes word segmentation, part-of-speech tagging, chunking and information extraction.

[0118] For example, the textual expression of the manufacturing process refactoring requirements input in this case includes:

[0119] The production unit selection requirements are: the new manufacturing process model should include the purchasing department, quality inspection department, production planning department, warehousing department, production workers, assembly workers, surface treatment workers, equipment maintenance workers, packaging workers, logistics department, and data analysts;

[0120] The requirements for manufacturing activities to be increased are: Production workers should add a manufacturing activity named pre-processing materials;

[0121] The manufacturing target requirements are: the resource utilization of the manufacturing process model is the lowest, and the cost of the manufacturing process model is the lowest;

[0122] The manufacturing constraint requirement is: the time of the manufacturing process model is less than or equal to 350 hours.

[0123] Word segmentation and part-of-speech tagging: This step first removes punctuation marks and stop words in the manufacturing process reconstruction requirement statements to facilitate subsequent word segmentation, part-of-speech tagging, and sentence processing; then, combined with the manufacturing process reconstruction dictionary, a rule-based word segmentation operation is used and part-of-speech tagging is performed on the word segmentation results to facilitate a better understanding of the text content.

[0124] The result of part-of-speech tagging in this case is:

[0125] <new model> <shall> <contain>Purchasing department, quality inspection department, production planning department, warehousing department, production workers, assembly workers, surface treatment workers, equipment maintenance workers, packaging workers, logistics department, data analysts,

[0126] Production Worker <shall> <add> <name> <is>Pre-treatment of materials <object>，

[0127] <model>Resource Utilization <min>cost <min> ，

[0128] <model>time<rather than / more than> 350

[0129] Block segmentation: The purpose of this step is to block the segmentation results according to the corresponding four types of templates in the manufacturing process reconstruction requirement template library, and divide the manufacturing process reconstruction requirements into independent information blocks.

[0130] In this case, the information blocks can be identified based on the sentence structure composed of parts of speech in the manufacturing process reconstruction requirement template. <model>…<rather than / more than> …, can be obtained from <model>time<rather than / more than> 350”, and determine that these two pieces of information belong to manufacturing constraint information.

[0131] Information extraction: Extract the information required for model reconstruction based on the block results. The types of information include four categories: production unit information, manufacturing activity addition information, manufacturing target information, and manufacturing constraint information. The purpose of this step is to extract the information required for model reconstruction, which will be used as input to the manufacturing process model reconstruction device.

[0132] The production unit information in this case is: {Purchasing Department, Quality Inspection Department, Production Planning Department, Warehousing Department, Production Workers, Assembly Workers, Surface Treatment Workers, Equipment Maintenance Workers, Packaging Workers, Logistics Department, Data Analysts}, the manufacturing activity addition information is: {Production Workers, Pre-processing Materials}, the manufacturing target information is {Resource Utilization, Minimum; Cost, Minimum}, and the manufacturing constraint information is: {Time, Less than or Equal to, 350}.

[0133] Step S6: Match the manufacturing process model data and generate a preliminary manufacturing process plan.

[0134] (1) The purpose of matching manufacturing process data is to match the corresponding production units and the manufacturing activities they perform from the manufacturing process model library based on the extracted production unit information and the added manufacturing activity information. Model data matching mainly includes production unit data matching and added manufacturing activity data matching.

[0135] Production unit data matching: The purpose of this step is to obtain relevant production unit data, such as name, parameter value, location, size, etc., from the manufacturing process model library based on the production unit information.

[0136] Manufacturing activity data matching: The purpose of this step is to obtain the manufacturing activity data performed by the relevant production units and the newly added manufacturing activities from the manufacturing process model library based on the addition of manufacturing activity information, including name, parameter type, parameter name, parameter value, position, size, etc.

[0137] (2) The purpose of generating a preliminary manufacturing process is to convert the matched production unit and manufacturing activity data and display the manufacturing process model data in an editable structured form, such as the name of the production unit and manufacturing activity, the attribute name and attribute value, etc., so as to facilitate users to modify or delete the manufacturing activities according to actual manufacturing needs, where the modification includes the change of the attribute value of the manufacturing activity. The generation of a preliminary manufacturing process plan includes the conversion of manufacturing process model data and the generation and display of a preliminary manufacturing process plan.

[0138] Manufacturing process model data conversion: The purpose of this step is to convert the matched manufacturing process model data into form data. The conversion only converts the model data that needs to be displayed, such as the position of the object model and the size of the object model. Only the name, parameter name, parameter value and other information of the object model are converted.

[0139] Generation and display of preliminary manufacturing process plan: The purpose of this step is to display the converted model data, so as to provide users with an interface or carrier for modifying the plan. The preliminary manufacturing process plan mainly displays the relevant data of each manufacturing activity attribute in a tabular form.

[0140] Step S7: Algorithm processing. The manufacturing process model provides optimization space for algorithm optimization. The purpose of the algorithm processing is to optimize and solve the preliminary manufacturing process plan based on the modified preliminary manufacturing process plan, manufacturing target information and manufacturing constraint information through a multi-objective optimization algorithm, so as to obtain a model plan that meets the needs of the user and provide a basis for the subsequent generation of the manufacturing process model. The algorithm used in the algorithm processing is NSGA-II: Non-dominated Sorting Genetic Algorithm II, which is a genetic algorithm for solving multi-objective optimization problems. The algorithm process regards each manufacturing activity in the manufacturing process model as a node, and all manufacturing activities in the model constitute a node set. Based on the manufacturing goals and manufacturing constraints, a mathematical model is constructed, and each node in the node set is adaptively selected to finally obtain the optimal solution set.

[0141] The goal in this case is to minimize the sum of resource utilization values ​​of all manufacturing activities in the manufacturing process model and minimize the sum of cost values ​​of all manufacturing activities in the manufacturing process model. The constraint is that the sum of time values ​​of all manufacturing activities in the manufacturing process model is less than or equal to 350 hours.

[0142] The main steps of the NSGA-II algorithm are as follows:

[0143] a) Setting objective functions and constraints: Convert manufacturing target information and manufacturing constraint information into objective functions and constraints in the mathematical model, providing a basis for subsequent optimization processes.

[0144] b) Initialize the population: A set of node subsets is randomly generated based on the node set as the initial population. Set the number of iterations. Here, the manufacturing activity set is used as the basic set, and some manufacturing activities are randomly selected from the basic set to form an individual. Through different selections, a group of individuals is obtained to form the initial population. The original node set also needs to be used as a population. In this way, it can be finally known which manufacturing activities can be excluded from the manufacturing process.

[0145] c) Evaluation, non-dominated sorting and crowding calculation: First, the objective function of each node subset in the initial population is calculated to measure the performance of each node subset. Then, the node subsets in the population are sorted according to the non-dominated relationship. Finally, crowding calculation is performed to calculate the crowding of each node subset in the non-dominated layer.

[0146] d) Selection, crossover and mutation: First, select individuals (a set of node subsets) for the next generation population, and select node subsets with larger crowding to maintain the diversity of the population. Then, crossover and mutation are performed on the selected individuals (node ​​subsets) to generate new offspring individuals (node ​​subsets).

[0147] e) Merging and generating new populations: combining subsets of nodes that have undergone selection, crossover, and mutation into new populations.

[0148] f) Iterate and generate the optimal solution: Iterate the process from c) to e) until the number of iterations is met, output the final non-dominated sorting result, and obtain the optimal node set.

[0149] g) Determine the manufacturing activity nodes: Based on the optimal node set, the optimal manufacturing activity set is mapped.

[0150] Step S8: Generate the final manufacturing process model.

[0151] The purpose of generating the final manufacturing process model is to match the production units and manufacturing activity data in the manufacturing process model library based on the optimal node set obtained by algorithm processing, and finally obtain a manufacturing activity process model that meets the requirements. Model data generation mainly includes demand model data acquisition and new demand model generation.

[0152] Manufacturing process model data processing: The purpose of this step is to match the manufacturing activities and production units in the existing manufacturing process model library based on the optimal node set processed by the algorithm, and obtain all the data of these manufacturing activities and production units, including the location and size of the model, so that the newly generated demand model will not be unable to be viewed due to problems such as position confusion, thereby improving the readability of the model.

[0153] Manufacturing process model generation: The purpose of this step is to process the acquired manufacturing process model data, such as updating the IDs of existing production units and manufacturing activity data, adding new manufacturing activity data, etc., so as to generate the final manufacturing process model data, so that the modeling tool can directly view or modify the manufacturing process model on the modeling canvas by parsing the manufacturing process model data.

[0154] In the final manufacturing process model, the production unit objects include: purchasing department, quality inspection department, production planning department, warehousing department, production workers, assembly workers, surface treatment workers, equipment maintenance workers, packaging workers, logistics department, and data analysts. The manufacturing activity objects include: purchasing raw materials, pre-processing materials, accepting materials, formulating production plans, arranging production tasks, managing raw material inventory, material distribution to production lines, pre-processing materials, parts manufacturing, assembling products, surface treatment and painting, quality inspection during production, product testing, preventive maintenance, equipment failure repair, final product inspection, packaging finished products, storing finished products, preparing product shipments, and arranging transportation. Among them, the manufacturing activities of analyzing production data, recording production data, and daily equipment inspection were removed because they did not meet the optimal solution.

[0155] Based on the above method, the present invention also provides a device for supporting the reconstruction of complex product manufacturing process models, such as Figure 3 As shown, the application layer of the device includes: a manufacturing process meta-model library construction unit, a manufacturing process model library construction unit, and a manufacturing process model reconstruction unit.

[0156] The manufacturing process metamodel library construction unit is used to build a manufacturing process metamodel library based on the multi-architecture modeling language KARMA.

[0157] A manufacturing process model library construction unit is used to construct a manufacturing process model library based on the manufacturing process meta-model library; the manufacturing process model library includes the production units involved in the basic manufacturing process and the manufacturing activities performed by them, as well as the execution sequence between the manufacturing activities;

[0158] The manufacturing process model reconstruction unit includes an analysis module, a preliminary solution generation module, an optimization module and a reconstruction module;

[0159] A parsing module is used to receive a manufacturing process model reconstruction requirement text organized according to a sentence structure provided by a manufacturing process reconstruction requirement template; and parse the manufacturing process reconstruction requirement according to a manufacturing process reconstruction dictionary, including production units, new manufacturing activities, manufacturing goals, and manufacturing constraints;

[0160] The manufacturing process reconstruction requirement template is constructed based on the terms provided by the manufacturing process reconstruction dictionary, and includes a production unit selection template, a manufacturing activity addition template, a manufacturing target template, and a manufacturing constraint template;

[0161] A preliminary solution generation module is used to match the production units and newly added manufacturing activities analyzed by the analysis module in the manufacturing process model library, extract the manufacturing activities to be performed by the production units, as well as the production unit information, all manufacturing activity information and execution order, and generate a preliminary manufacturing process model solution;

[0162] An optimization module, used for performing multi-objective optimization on the manufacturing activities in the preliminary manufacturing process model scheme according to the manufacturing objectives and manufacturing constraints analyzed by the analysis module;

[0163] The reconstruction module is used to reconstruct the final manufacturing process plan according to the multi-objective optimization results of the optimization module.

[0164] In summary, the present invention supports users to input manufacturing process reconstruction requirements, such as manufacturing process requirements, based on the constructed manufacturing process model library according to structured templates. Secondly, the system extracts information and matches elements for the input manufacturing process reconstruction requirements to obtain a preliminary manufacturing process model solution, and then weighs and optimizes the early demand demonstration process through a multi-objective optimization algorithm to reconstruct the final manufacturing process solution, thereby realizing the model-driven adaptive reconfigurable manufacturing process function.

[0165] The above specific embodiments only describe the design principle of the present invention. The shapes and names of the components in the description may be different and are not limited. Therefore, those skilled in the art in the field of the present invention may modify or replace the technical solutions recorded in the above embodiments; and these modifications and replacements do not deviate from the creative purpose and technical solutions of the present invention and should all fall within the protection scope of the present invention.< / model> < / model> < / model> < / min> < / min> < / model> < / object> < / is> < / name> < / add> < / shall> < / contain> < / shall> < / model> < / max> < / model> < / object> < / is> < / name> < / add> < / shall> < / contain> < / shall> < / object> < / is> < / name> < / add> < / shall> < / max> < / model> < / object> < / is> < / name> < / add> < / shall> < / contain> < / shall>

Claims

1. A method for supporting complex product manufacturing process model reconstruction, characterized in that: include: Step S1: construct a manufacturing process model library, define a manufacturing process reconstruction dictionary and a manufacturing process reconstruction requirement template; The manufacturing process model library includes the production units involved in the basic manufacturing process and the manufacturing activities they perform, as well as the execution sequence between the manufacturing activities; The manufacturing process reconstruction requirement template is constructed based on the manufacturing process reconstruction dictionary, and specifies the sentence structure of the manufacturing process model reconstruction requirement text, including a production unit selection template, a manufacturing activity addition template, a manufacturing target template, and a manufacturing constraint template; Step S2: Reconstruct the demand template according to the manufacturing process, input the manufacturing process model to reconstruct the demand text; parse the manufacturing process reconstruction demand according to the manufacturing process reconstruction dictionary, including production units, new manufacturing activities, manufacturing goals, and manufacturing constraints; Step S3: According to the production units and newly added manufacturing activities analyzed in step S2, matching is performed in the manufacturing process model library, the manufacturing activities to be performed by the production units, as well as the production unit information, all manufacturing activity information and execution sequence are extracted, and a preliminary manufacturing process model solution is generated; Step S4: Based on the manufacturing objectives and manufacturing constraints analyzed in step S2, multi-objective optimization is performed on the manufacturing activities in the preliminary manufacturing process model scheme to reconstruct the final manufacturing process scheme.

2. The method according to claim 1, characterized in that In step S4, non-dominated sorting genetic algorithm II (NSGA-II) is used for multi-objective optimization.

3. The method according to claim 1 or 2, characterized in that When performing multi-objective optimization in step S4, each manufacturing activity in the preliminary manufacturing process model scheme is taken as a node to construct a node set; a group of node subsets are generated as the initial population by randomly selecting from the node set, and each node in the node set is adaptively selected based on the parsed manufacturing goals and manufacturing constraints to obtain the optimal solution set.

4. The method according to claim 1, characterized in that In step 1, the manufacturing process model library is constructed based on the manufacturing process meta-model library built on the basis of the multi-architecture modeling language KARMA; The metamodels contained in the manufacturing process metamodel library include: The graphic element models include: a manufacturing process graphic element model, which is used to describe the manufacturing process; The object metamodel includes: a production unit object metamodel, a manufacturing activity object metamodel, a start object metamodel and an end object metamodel; the production unit object metamodel is used to describe the production unit in the manufacturing process and its related characteristics, including a name attribute; the manufacturing activity object metamodel is used to describe the manufacturing activity in the manufacturing process and its related characteristics, including a name attribute, a time attribute, a cost attribute, and a resource consumption rate attribute; The relational metamodel includes: control flow relational metamodel, which is used to describe the sequence relationship between two manufacturing activities.

5. The method according to claim 4, characterized in that The manufacturing process model library constructed in step S1 includes: The production unit objects include: purchasing department, quality inspection department, production planning department, warehousing department, production workers, assembly workers, surface treatment workers, equipment maintenance workers, packaging workers, logistics department, data analysts; Manufacturing activity objects include: purchasing raw materials constructed in the purchasing department object; accepting materials, quality inspection during production, product testing, and final product inspection constructed in the quality inspection department object; formulating production plans and arranging production tasks constructed in the production planning department object; managing raw material inventory, material distribution to production lines, and storing finished products constructed in the warehousing department object; pre-processing materials and parts manufacturing constructed in the production worker object; assembling products constructed in the assembly worker object; surface treatment and painting constructed in the surface treatment worker object; daily equipment inspection, preventive maintenance, and equipment failure repair constructed in the equipment maintenance worker object; packaging finished products constructed in the packaging worker object; arranging transportation constructed in the logistics department object; analyzing production data constructed in the data analyst object; Manufacturing activities also include: start objects and end objects; After constructing the manufacturing activities, the control flow relational metamodel is used to connect the manufacturing activities in the required sequence to form a basic manufacturing process.

6. The method according to claim 1, characterized in that The manufacturing process reconstruction dictionary includes {new manufacturing process model: new model}, {should: shall}, {include: contain}, {add: add}, {increase: add}, {name: name}, {is: is}, {production activity: object}, {maximum: max}, {greater than or equal to: more than}, {greater than: morethan}, {less than or equal to: rather than}, {less than: rather than}, {manufacturing process model, model}.

7. The method according to claim 1, characterized in that The manufacturing process reconstruction requirement template is specifically: The production unit selection template is:<new model> <shall> <contain> …，…,…，；< / contain> < / shall> The template for adding manufacturing activities is:… <shall> <add> <name> <is> … <object>； The manufacturing target template is: <model> … <max min> ；< / max> < / model> The manufacturing constraint template is:… <shall> <add> <name> <is> … <object>。 8. The method according to claim 1, characterized in that In step S3, the generation of a preliminary manufacturing process model solution includes: manufacturing process model data conversion and preliminary manufacturing process solution generation and display: The step of converting the manufacturing process model data is to convert the data to be displayed in the extracted data into form data; The steps of generating and displaying the preliminary manufacturing process plan display the converted form data in a table format.

9. The method according to claim 1, characterized in that The production unit objects and manufacturing activity objects in the manufacturing process model library also include the position and size of the object nodes on the layout; When the final manufacturing process plan is reconstructed in step S4, the manufacturing activities and production units are matched in the manufacturing process model library according to the optimization plan, and the positions and sizes of these manufacturing activities and production units are obtained, and provided to the modeling tool to display icons of each object on the modeling canvas for users to view and modify.

10. A device for supporting reconstruction of complex product manufacturing process model, characterized in that: include: Manufacturing process metamodel library construction unit, manufacturing process model library construction unit, manufacturing process model reconstruction unit; The manufacturing process metamodel library construction unit is used to build a manufacturing process metamodel library based on the multi-architecture modeling language KARMA; The manufacturing process model library construction unit is used to construct a manufacturing process model library based on the manufacturing process meta-model library; the manufacturing process model library includes the production units involved in the basic manufacturing process and the manufacturing activities performed by them, as well as the execution order between the manufacturing activities; The manufacturing process model reconstruction unit includes an analysis module, a preliminary solution generation module, an optimization module and a reconstruction module; The parsing module is used to receive the manufacturing process model reconstruction requirement text organized according to the sentence structure provided by the manufacturing process reconstruction requirement template; parse the manufacturing process reconstruction requirement according to the manufacturing process reconstruction dictionary, including production units, new manufacturing activities, manufacturing goals, and manufacturing constraints; The manufacturing process reconstruction requirement template is constructed based on the terms provided by the manufacturing process reconstruction dictionary, and includes a production unit selection template, a manufacturing activity addition template, a manufacturing target template, and a manufacturing constraint template; The preliminary solution generation module is used to match the production units and newly added manufacturing activities analyzed by the analysis module in the manufacturing process model library, extract the manufacturing activities to be performed by the production units, as well as the production unit information, all manufacturing activity information and execution order, and generate a preliminary manufacturing process model solution; The optimization module is used to perform multi-objective optimization on the manufacturing activities in the preliminary manufacturing process model scheme according to the manufacturing objectives and manufacturing constraints analyzed by the analysis module; The reconstruction module is used to reconstruct the final manufacturing process plan according to the multi-objective optimization results of the optimization module. < / object> < / is> < / name> < / add> < / shall> < / object> < / is> < / name> < / add> < / shall>

Images