An intelligent production deployment method and system based on industrial knowledge graph

Abstract

The present invention relates to the technical field of production deployment, and discloses an intelligent production deployment method and system based on an industrial knowledge graph; the method comprises screening out effective features from features to be identified, constructing an industrial knowledge graph, extracting demand features from production demand data, obtaining comprehensive deployment parameters of the industrial knowledge graph, calculating a production deployment index, formulating a deployment level, and controlling the intelligent production deployment of nodes and edges according to the deployment level; relative to the prior art, the present invention can provide an optimal graph state for the import of production demand data on the user side, and at the same time provide accurate data basis for the intelligent production deployment of small batches of personalized products, and then realize the accurate production deployment purpose between physical equipment and process procedures within the enterprise according to the production deployment relationship between nodes and edges of the industrial knowledge graph, which can not only improve the accuracy of intelligent production deployment, but also meet the demand for rapid response to small batches of personalized products.

Description

Technical Field

[0001] The present invention relates to the field of production deployment technology, and more specifically, to an intelligent production deployment method and system based on industrial knowledge graph. Background Art

[0002] Intelligent production dispatching systems have a wide range of application scenarios, especially in the fields of automated production lines and intelligent manufacturing. Through intelligent dispatching systems, complex production processes can be effectively managed, equipment utilization and production efficiency can be improved, and the production costs of enterprises can be reduced. The industrial knowledge graph not only includes the unified management of various industrial resources, but also includes the recording of the dynamic relationships between these entities. Therefore, by combining the industrial knowledge graph with the intelligent production dispatching system, enterprises can achieve efficient and low-cost production and manufacturing effects.

[0003] The patent application with reference publication number CN114137921A discloses a real-time deployment system and deployment method for an intelligent production workshop of sanitary equipment. In the cloud and workshop twin data center environment, production resources are scheduled and configured, production tasks are split, and available resources are scheduled. The optimal production resource configuration is performed for each processing node, so that spare resources are minimized, production efficiency is improved, order tasks are tracked in real time, and real-time data in the production process is fed back to the cloud and workshop twin data center; RFID antennas are set in the loading buffer area, contact buffer area, equipment processing area and discharge buffer area, so that the monitoring accuracy of the production process can be greatly improved; through the real-time interaction and scheduling of the manufacturing task management center and the workshop twin data center, as well as high-precision RFID recognition, the required scheduling data can be obtained quickly and accurately, ensuring the accuracy of scheduling, and the operation status of the workshop can be analyzed through the workshop twin data and the assembly time can be predicted. Through the information interaction and iterative optimization of the physical workshop and the virtual workshop, the scheduling plan is continuously adjusted and optimized, and production disturbances are responded to in a timely manner;

[0004] The existing intelligent production deployment method obtains the past production process data of the enterprise side, and constructs a constant-state industrial knowledge graph based on the production process data to carry out production deployment processing for the customized products on the user side. Since the constructed industrial knowledge graph can only represent the historical production situation on the enterprise side, it cannot be updated in real time according to the actual production situation on the enterprise side and the actual production needs of the user side. As a result, the data of the industrial knowledge graph is lagging and cannot be kept in the latest version. When the user side needs to produce small batches of personalized products, the enterprise side will be unable to respond quickly and accurately to the production deployment needs of small batches of personalized products, thereby reducing the quality of intelligent production deployment.

[0005] In view of this, the present invention proposes an intelligent production deployment method and system based on industrial knowledge graph to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: an intelligent production deployment method based on industrial knowledge graph, applied to a production deployment center, comprising:

[0007] S1: Collect comprehensive production data from the industrial end, which includes equipment data, personnel data, material data, process data, and product data, and extract features to be identified from the comprehensive production data;

[0008] S2: Filter out effective features from the features to be identified, convert the effective features into graph points, which include nodes and edges, and construct an industrial knowledge graph;

[0009] S3: receiving production demand data from the user end, and extracting demand features from the production demand data;

[0010] S4: Import the demand characteristics into the industrial knowledge graph to update the graph, obtain the comprehensive allocation parameters of the updated industrial knowledge graph, the comprehensive allocation parameters include node fluctuation frequency, associated edge update degree and update delay value, and calculate the production allocation index;

[0011] S5: Compare the production allocation index with the preset production allocation threshold to formulate an allocation level, which includes the linkage comprehensive allocation level and the independent autonomous allocation level, and control the intelligent production allocation of nodes and edges according to the allocation level.

[0012] Furthermore, the method for extracting the features to be identified includes:

[0013] Through natural language processing technology, semantic recognition is performed on i comprehensive production data one by one to obtain i text semantics and i digital semantics;

[0014] The workshop location, equipment number, manufacturing process, production materials and testing instruments are recorded as standard characters, and the standard characters are combined into standard features;

[0015] According to the order of query time, the i-character semantics are compared with the standard features in turn, and the character semantics that are repeated with any standard character in the standard features are recorded as the semantics to be recognized, and the semantics to be identified;

[0016] Build A blank basic group, and draw the first characteristic bit and the second characteristic bit in the basic group respectively, to obtain The first characteristic bit and The second characteristic bit;

[0017] Will The semantics to be recognized and the corresponding The digital semantics are imported into The first characteristic bit and In the second characteristic bit, we get feature groups, and After the feature groups are numbered in sequence, we get Features to be identified.

[0018] Further, the screening methods of effective features include:

[0019] Through the attribute management system, we can query The attribute status box of the feature to be identified is marked with the remark value in the attribute status box. The first remark value and The second remark value;

[0020] Will The first remark value and the corresponding Multiply the second remark value by Status index;

[0021] The expression of the state index is:

[0022] ;

[0023] In the formula, For the The state index of the feature to be identified, For the The first annotation value of the feature to be identified, For the The second annotation value of the feature to be identified;

[0024] The state index that is less than the preset index threshold is recorded as a valid index, and the feature to be identified corresponding to the valid index is recorded as a valid feature, and s valid features are obtained.

[0025] Furthermore, the method for constructing the industrial knowledge graph includes:

[0026] Use NER technology to identify the key entities of s effective features one by one, obtain s key entities, and identify the attribute information of s key entities respectively;

[0027] The relationship between two adjacent key entities among the s key entities is extracted through the relationship extraction model to obtain w relationships;

[0028] The s key entities are converted into s basic points through virtual reality technology, and information boxes are drawn on the s basic points respectively;

[0029] Import the attribute information of the s key entities into the information boxes of the s basic points, generate s nodes, and number the s nodes one by one;

[0030] Identify the numbers of the nodes where the key entities at the front and back ends of the w relationships are located one by one, and obtain w front numbers and w back numbers;

[0031] Connect the front ends of w relations to the nodes with the same number as the w front numbers, and connect the back ends of w relations to the nodes with the same number as the w back numbers, to obtain w edges;

[0032] The s nodes and w edges are stored as a knowledge graph through the graph database, and a regular update mechanism is set in combination with heterogeneous graph technology to construct an industrial knowledge graph.

[0033] Furthermore, the method for extracting demand features includes:

[0034] The input data of the user end is queried through the enterprise production database, and the information status of the input data is identified one by one, and the input data with the information status of the user is recorded as production demand data, and u production demand data are obtained;

[0035] Use natural language processing technology to identify the demand semantics of u production demand data one by one, and mark the keywords in the demand semantics;

[0036] The demand semantics with the keywords progress, allocation, quality and time are recorded as demand features, and y demand features are obtained.

[0037] Furthermore, the method for obtaining the node fluctuation frequency includes:

[0038] A1: At time T1, record the node states of s nodes in the industrial knowledge graph, identify the state values ​​in the node states, and obtain s initial state values;

[0039] A2: At time T2, record the node status of s nodes in the industrial knowledge graph, identify the state value in the node status, and obtain s updated state values;

[0040] A3: The node state whose updated state value is inconsistent with the initial state value is recorded as an abnormal state, and the node corresponding to the abnormal state is recorded as a fluctuating node, and the number of fluctuating nodes is counted;

[0041] A4: Repeat Repeat steps A1-A3 to obtain The number of fluctuation nodes, and The number of fluctuation nodes is compared with the duration from time T1 to time T2 to obtain sub-frequency;

[0042] The expression of sub-frequency is:

[0043] ;

[0044] In the formula, For the The sub-frequency, For the The number of fluctuation nodes, is the duration between time T1 and time T2;

[0045] A5: The sub-frequencies are accumulated and averaged to obtain the node fluctuation frequency;

[0046] The expression of node fluctuation frequency is:

[0047] ;

[0048] In the formula, is the node fluctuation frequency, For the Sub-frequency.

[0049] Furthermore, the method for obtaining the update degree of the associated edge includes:

[0050] Take T3 as the starting time and the preset monitoring time as the interval standard to mark Monitoring time points;

[0051] exist At each monitoring time point, the association values ​​of the w edges are queried one by one through the association management system to obtain w association values;

[0052] Compare any two association values ​​one by one, record the edges with the same association value as associated edges, and count the number of associated edges to obtain associated value;

[0053] Will The association values ​​are compared with the total number of edges in turn, and the comparison values ​​are accumulated and averaged to obtain the update degree of the association edges;

[0054] The expression of the update degree of the associated edge is:

[0055] ;

[0056] In the formula, is the update degree of the associated edge, For the The associated value, is the total number of edges.

[0057] Furthermore, the method for obtaining the update delay value includes:

[0058] Query one by one by timestamp The time when the node state of the fluctuation node first fluctuates and the time when the node state fluctuates for the last time are obtained. The first moment and a second moment;

[0059] Will The first moment to The duration between the second moments is recorded as the first update duration, and the first update duration greater than the preset first duration threshold is recorded as the first delay duration, to obtain The first delay duration;

[0060] Will Subtract the first delay time lengths from the preset first time length threshold in turn and add up the differences to obtain a first delay value;

[0061] The expression for the first delay value is:

[0062] ;

[0063] In the formula, is the first delay value, For the The first delay time, is a preset first duration threshold;

[0064] The second update durations of all associated edges are queried one by one through the timestamps, and the second update durations greater than the preset second duration threshold are recorded as the second delay durations. The second delay duration;

[0065] Will Subtract the second delay time lengths from the preset second time length threshold in turn and add up the differences to obtain a second delay value;

[0066] The expression for the second delay value is:

[0067] ;

[0068] In the formula, is the second delay value, For the The second delay time, is a preset second duration threshold;

[0069] The first delay value and the second delay value are added and averaged to obtain an updated delay value;

[0070] The expression for updating the delay value is:

[0071] ;

[0072] In the formula, To update the delay value;

[0073] The expression of production allocation index is:

[0074] ;

[0075] In the formula, is the production allocation index, , , is the weight factor, and , , Both are greater than 0.

[0076] Furthermore, the method for formulating the linkage comprehensive allocation level and the independent autonomous allocation level includes:

[0077] Production Allocation Index With the preset production allocation threshold Compare;

[0078] when Greater than or equal to When the government is in a hurry to find a coordinated and comprehensive deployment level,

[0079] when Less than When the government has a plan to allocate resources independently, it should formulate independent deployment levels;

[0080] The control methods for intelligent production deployment of nodes and edges include:

[0081] In the industrial knowledge graph, nodes and edges associated with the keywords of y demand features are marked one by one to obtain n target nodes and k target edges;

[0082] According to the order of the production process in the industrial knowledge graph, the n target nodes and k target edges are numbered in ascending order;

[0083] Establishing a first dispatching line between n target nodes and connecting the n target nodes to the first dispatching line, establishing a second dispatching line between k target edges and connecting the k target edges to the second dispatching line;

[0084] When the linkage comprehensive deployment level is formulated, the production deployment center issues a synchronous deployment instruction to control the n target nodes on the first deployment line and the k target edges on the second deployment line to be deployed in an overall linkage manner at the same time;

[0085] When an independent and autonomous deployment level is formulated, the production deployment center issues an asynchronous deployment instruction to control the n target nodes on the first deployment line and the k target edges on the second deployment line to be independently and autonomously deployed in an ascending order of numbers.

[0086] An intelligent production deployment system based on industrial knowledge graph, applied to a production deployment center, for implementing the intelligent production deployment method based on industrial knowledge graph, comprising a first extraction module, a graph construction module, a second extraction module, an index calculation module and a production deployment module, wherein each module is connected via a wired or wireless network;

[0087] The first extraction module is used to collect comprehensive production data from the industrial end, which includes equipment data, personnel data, material data, process data and product data, and extract features to be identified from the comprehensive production data;

[0088] A graph construction module is used to select effective features from the features to be identified, convert the effective features into graph points, which include nodes and edges, and construct an industrial knowledge graph;

[0089] The second extraction module is used to receive the production demand data from the user end and extract the demand features from the production demand data;

[0090] The index calculation module is used to import demand characteristics into the industrial knowledge graph to update the graph, obtain the comprehensive allocation parameters of the updated industrial knowledge graph, which include node fluctuation frequency, associated edge update degree and update delay value, and calculate the production allocation index;

[0091] The production allocation module is used to compare the production allocation index with the preset production allocation threshold and formulate an allocation level. The allocation level includes the linkage comprehensive allocation level and the independent autonomous allocation level, and controls the intelligent production allocation of nodes and edges according to the allocation level.

[0092] The technical effects and advantages of the intelligent production deployment method and system based on industrial knowledge graph of the present invention are as follows:

[0093] The present invention constructs an industrial knowledge graph based on the effective features of the industrial end, and integrates heterogeneous graph technology to achieve the automatic update effect of the industrial knowledge graph, so that the industrial knowledge graph can always be kept in the latest version, thereby providing the best graph state for the import of production demand data on the user side, and providing a solid foundation for building user, product and resource portraits. At the same time, by obtaining the comprehensive deployment parameters of the industrial knowledge graph to calculate the production deployment index, it can provide accurate data basis for the subsequent small batch personalized product intelligent production deployment, and then according to the production deployment relationship of the nodes and edges of the industrial knowledge graph, the accurate production deployment purpose between the physical equipment and the process procedures within the enterprise can be achieved, which can not only improve the accuracy of intelligent production deployment, but also meet the needs of rapid response to small batch personalized products, greatly improving the effect of the enterprise's intelligent production deployment. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 A schematic diagram of a process of an intelligent production deployment method based on an industrial knowledge graph provided in the first embodiment of the present invention;

[0095] Figure 2 A schematic diagram of an intelligent production deployment system based on an industrial knowledge graph provided in Embodiment 2 of the present invention;

[0096] Figure 3 A schematic diagram of a structural diagram of an electronic device provided in Embodiment 3 of the present invention;

[0097] Figure 4 A schematic diagram of the structure of a computer-readable storage medium provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0098] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0099] Example 1: Please refer to Figure 1 As shown, the intelligent production deployment method based on the industrial knowledge graph described in this embodiment is applied to the production deployment center, including:

[0100] S1: Collect comprehensive production data from the industrial end and extract features to be identified from the comprehensive production data;

[0101] The industrial end refers to the port that can plan production deployment operations for enterprises according to the actual production situation of enterprises. Comprehensive production data refers to the comprehensive information generated in the industrial end that represents the production and manufacturing of enterprises. It can comprehensively represent the actual production information of the industrial end and serve as the data basis for the subsequent construction of the industrial knowledge graph.

[0102] Comprehensive production data includes equipment data, personnel data, material data, process data and product data; equipment data is used to represent the status and information of mechanical equipment used for production and manufacturing within the enterprise, personnel data refers to the number and positions of production workers used for production and manufacturing within the enterprise, material data refers to the type and quantity of raw materials used for production and manufacturing within the enterprise, process data refers to the information representation of the production process used for production and manufacturing within the enterprise, and product data refers to the information representation of finished products after production and manufacturing within the enterprise; equipment data, personnel data, material data, process data and product data are all obtained by querying the enterprise production database;

[0103] After the comprehensive production data is collected, it is necessary to perform preliminary extraction of the comprehensive production data, and to concisely and accurately represent the comprehensive production data through the data to be identified, so as to facilitate the subsequent construction and use of the industrial knowledge graph.

[0104] Methods for extracting features to be identified include:

[0105] Through natural language processing technology, semantic recognition is performed on i comprehensive production data one by one to obtain i text semantics and i digital semantics; text semantics is used to accurately represent the text part of the comprehensive production data, and digital semantics is used to accurately represent the digital part of the comprehensive production data, which can ensure that the comprehensive production data can be fully and accurately parsed;

[0106] The workshop location, equipment number, manufacturing process, production materials and testing instruments are recorded as standard characters, and the standard characters are combined into standard features;

[0107] According to the order of query time, the i-character semantics are compared with the standard features in turn, and the character semantics that are repeated with any standard character in the standard features are recorded as the semantics to be recognized, and the semantics to be identified;

[0108] Build A blank basic group, and draw the first characteristic bit and the second characteristic bit in the basic group respectively, to obtain The first characteristic bit and The basic group is the basic framework for constructing the features to be identified, while the first and second feature bits are the smallest units that make up the basic group and provide the accurate location for data import for the subsequent construction of the features to be identified.

[0109] Will The semantics to be recognized and the corresponding The digital semantics are imported into The first characteristic bit and In the second characteristic bit, we get feature groups, and After the feature groups are numbered in sequence, we get Features to be identified.

[0110] It should be noted that the features to be identified are used to concisely and preliminarily represent the true meaning of the comprehensive production data, and provide a basis for the subsequent construction of the industrial knowledge graph. Therefore, the features to be identified cannot be used directly, but require further optimization and screening.

[0111] S2: Filter out effective features from the features to be identified, convert the effective features into graph points, which include nodes and edges, and construct an industrial knowledge graph;

[0112] Effective features refer to features to be identified that meet the requirements of feature screening and can be directly used in the subsequent construction of industrial knowledge graphs. They can screen out a small number of accurate features from a large number of features to be identified, and provide a solid foundation for the subsequent construction of industrial knowledge graphs.

[0113] The screening methods for effective features include:

[0114] Through the attribute management system, we can query The attribute status box of the feature to be identified is marked with the remark value in the attribute status box. The first remark value and The second remark value; the attribute status box is a framework for storing the remark information of the feature to be identified, and the remark value is the numerical value corresponding to the actual remark information in the attribute status box. The larger the remark value, the more reliable the data support for the industrial knowledge graph can be provided by the feature to be identified;

[0115] Will The first remark value and the corresponding Multiply the second remark value by Status index;

[0116] The expression of the state index is:

[0117] ;

[0118] In the formula, For the The state index of the feature to be identified, For the The first annotation value of the feature to be identified, For the The second annotation value of the feature to be identified;

[0119] The state index that is less than the preset index threshold is recorded as a valid index, and the feature to be identified corresponding to the valid index is recorded as a valid feature, and s valid features are obtained; the preset index threshold refers to the minimum value of the state index of the feature to be identified that can be marked as a valid feature, which can accurately distinguish whether the feature to be identified is a valid feature.

[0120] After obtaining the effective features, the effective features can be integrated, so as to construct an industrial knowledge graph that meets the industrial end based on s effective features, so that the industrial knowledge graph can comprehensively represent the production and manufacturing capabilities and conditions of the industrial end, and the graph points can represent the key points in the industrial knowledge graph and serve as the target and object for subsequent production deployment;

[0121] Graph points include nodes and edges. Nodes refer to the point representation of entities in the industrial knowledge graph, which can refer to each specific entity on the industrial side. Edges refer to the relationship representation between entities in the industrial knowledge graph, which can refer to the dependencies between related entities.

[0122] The construction methods of industrial knowledge graph include:

[0123] The NER technology is used to identify the key entities of s effective features one by one, obtain s key entities, and identify the attribute information of s key entities respectively; the key entity refers to the actual mechanical equipment corresponding to the effective features when the industrial end is producing and deploying, and the attribute information is used to accurately represent the type attributes of the key entity;

[0124] The relationship between two adjacent key entities among the s key entities is extracted through the relationship extraction model to obtain w relationships; the relationship extraction model specifically adopts the CasRel model, which is an entity relationship extraction model based on joint decoding and is used to accurately extract the relationship between two adjacent key entities;

[0125] Through virtual reality technology, s key entities are converted into s basic points, and information boxes are drawn on the s basic points respectively; the information box is a framework used to provide auxiliary information notes for the basic points and facilitate the accurate conversion of subsequent nodes;

[0126] Import the attribute information of the s key entities into the information boxes of the s basic points, generate s nodes, and number the s nodes one by one;

[0127] Identify the numbers of the nodes where the key entities at the front end and back end of w relationships are located one by one, and obtain w front numbers and w back numbers; the front end and the back end represent the two ends of the relationship respectively, the front end specifically refers to the end of the relationship that is located at the front end of the process industry process, and the back end refers to the end of the relationship that is located at the back end of the process industry process;

[0128] Connect the front ends of w relations to the nodes with the same number as the w front numbers, and connect the back ends of w relations to the nodes with the same number as the w back numbers, to obtain w edges;

[0129] The s nodes and w edges are stored as a knowledge graph through the graph database, and a regular update mechanism is set in combination with heterogeneous graph technology to construct an industrial knowledge graph.

[0130] It should be noted that by integrating heterogeneous graph technology into the knowledge graph, the automatic update effect of the process industry knowledge graph is achieved, so that the industrial knowledge graph can always be kept in the latest version, avoiding delays and lags in information within the industrial knowledge graph.

[0131] S3: receiving production demand data from the user end, and extracting demand features from the production demand data;

[0132] The user end refers to the port that can provide personalized production demand information to the industrial end and make production arrangements based on the industrial knowledge graph. The production demand data refers to the comprehensive data contained in the production of small batches of personalized products by the user end, and serves as the original data of the user end. Since the types of production demand data are more complex and the quantity is larger, it is necessary to extract the production demand data and obtain demand features that can directly and accurately represent the true meaning of the production demand data;

[0133] The extraction methods of demand features include:

[0134] The input data of the user end is queried through the enterprise production database, and the information status of the input data is identified one by one, and the input data with the information status of the user is recorded as production demand data, and u production demand data are obtained;

[0135] Use natural language processing technology to identify the demand semantics of u production demand data one by one, and mark the keywords in the demand semantics;

[0136] The demand semantics with the keywords progress, allocation, quality and time are recorded as demand features, and y demand features are obtained.

[0137] It should be noted that the extracted demand features are used as the objects for subsequent integration and interaction with the industrial knowledge graph, which can ensure that the industrial knowledge graph on the industrial side can correspond one-to-one with the demand data on the user side, so as to achieve efficient intelligent production deployment effects.

[0138] S4: Import the demand characteristics into the industrial knowledge graph to update the graph, obtain the comprehensive allocation parameters of the updated industrial knowledge graph, and calculate the production allocation index;

[0139] After obtaining the demand features, the demand features and the industrial knowledge graph can be imported and integrated with each other, so that the demand features can be associated with the corresponding nodes and edges in the industrial knowledge graph, so that the industrial knowledge graph on the industrial side can be updated and upgraded in real time with the demand data on the user side, ensuring that the industrial knowledge graph can be kept in the latest version;

[0140] When the graph is updated, the demand characteristics are integrated with the industrial knowledge graph through heterogeneous graph technology, so that the industrial knowledge graph can build corresponding portraits of users, products and resources, and provide the user end with an industrial knowledge graph with automatic update function to ensure that the user end can perform intelligent production deployment operations based on the industrial knowledge graph;

[0141] When the industrial knowledge graph is updated, it is necessary to identify and monitor the updated situation of the industrial knowledge graph and reflect the changes of the industrial knowledge graph through relevant data, that is, to use comprehensive allocation parameters to numerically represent the relationship between the nodes and edges of the industrial knowledge graph, so as to provide a data basis for the subsequent production allocation index;

[0142] The comprehensive deployment parameters include node fluctuation frequency, associated edge update degree and update delay value;

[0143] The node fluctuation frequency refers to the number of nodes that fluctuate per unit time in the updated industrial knowledge graph, which can be used to indicate the update rate of the nodes in the industrial knowledge graph. The greater the node fluctuation, the more times the nodes fluctuate per unit time in the industrial knowledge graph, the greater the demand for intelligent production deployment, and the greater the production deployment index.

[0144] Methods for obtaining node fluctuation frequency include:

[0145] A1: At time T1, record the node states of s nodes in the industrial knowledge graph, identify the state values ​​in the node states, and obtain s initial state values; the state value is used to truly represent the node state and serves as a basis for judging whether the node fluctuates;

[0146] A2: At time T2, record the node status of s nodes in the industrial knowledge graph, identify the state value in the node status, and obtain s updated state values;

[0147] A3: The node state whose updated state value is inconsistent with the initial state value is recorded as an abnormal state, and the node corresponding to the abnormal state is recorded as a fluctuating node, and the number of fluctuating nodes is counted;

[0148] A4: Repeat Repeat steps A1-A3 to obtain The number of fluctuation nodes, and The number of fluctuation nodes is compared with the duration from time T1 to time T2 to obtain sub-frequency;

[0149] The expression of sub-frequency is:

[0150] ;

[0151] In the formula, For the The sub-frequency, For the The number of fluctuation nodes, is the duration between time T1 and time T2;

[0152] A5: The sub-frequencies are accumulated and averaged to obtain the node fluctuation frequency;

[0153] The expression of node fluctuation frequency is:

[0154] ;

[0155] In the formula, is the node fluctuation frequency, For the Sub-frequency.

[0156] The update degree of associated edges refers to the ratio of the number of edges with associated relationships to the total number of edges in the updated industrial knowledge graph, which can be used to represent the update degree of edges in the industrial knowledge graph. The greater the update degree of associated edges, the greater the number of edges with associated relationships in the industrial knowledge graph, the greater the demand for intelligent production deployment, and the greater the production deployment index.

[0157] The methods for obtaining the update degree of the associated edge include:

[0158] Take T3 as the starting time and the preset monitoring time as the interval standard to mark monitoring time points; the preset monitoring time refers to the time span between two adjacent monitoring time points, which can ensure that the two adjacent monitoring time points will not be too close or too far away from each other, thereby ensuring the independence of the data at the two adjacent monitoring time points;

[0159] exist At each monitoring time point, the association values ​​of the w edges are queried one by one through the association management system to obtain w association values; the association value is a numerical representation of the degree of association of the edges, thereby ensuring that each edge has a corresponding association value;

[0160] Compare any two association values ​​one by one, record the edges with the same association value as associated edges, and count the number of associated edges to obtain associated value;

[0161] Will The association values ​​are compared with the total number of edges in turn, and the comparison values ​​are accumulated and averaged to obtain the update degree of the association edges;

[0162] The expression of the update degree of the associated edge is:

[0163] ;

[0164] In the formula, is the update degree of the associated edge, For the The associated value, is the total number of edges.

[0165] The update delay value refers to the value of the time taken by the nodes and edges in the updated industrial knowledge graph to update the graph exceeding the standard total update time, which can represent the update performance of the industrial knowledge graph. The larger the update delay value, the larger the value of the time taken by the nodes and edges in the industrial knowledge graph to update the graph exceeding the standard total update time, the smaller the demand for intelligent production deployment, and the smaller the production deployment index;

[0166] The methods for obtaining the update delay value include:

[0167] Query one by one by timestamp The time when the node state of the fluctuation node first fluctuates and the time when the node state fluctuates for the last time are obtained. The first moment and a second moment;

[0168] Will The first moment to The duration between the second moments is recorded as the first update duration, and the first update duration greater than the preset first duration threshold is recorded as the first delay duration, to obtain a first delay duration; the preset first duration threshold is used for the maximum value of the first update duration under normal circumstances where no delay occurs, which can provide a numerical basis for comparison of the identification of the first delay duration;

[0169] Will Subtract the first delay time lengths from the preset first time length threshold in turn and add up the differences to obtain a first delay value;

[0170] The expression for the first delay value is:

[0171] ;

[0172] In the formula, is the first delay value, For the The first delay time, is a preset first duration threshold;

[0173] The second update durations of all associated edges are queried one by one through the timestamps, and the second update durations greater than the preset second duration threshold are recorded as the second delay durations. The preset second delay time threshold is used for the maximum value of the second update time under normal circumstances where no delay occurs, which can provide a numerical basis for comparison of the identification of the second delay time;

[0174] Will Subtract the second delay time lengths from the preset second time length threshold in turn and add up the differences to obtain a second delay value;

[0175] The expression for the second delay value is:

[0176] ;

[0177] In the formula, is the second delay value, For the The second delay time, is a preset second duration threshold;

[0178] The first delay value and the second delay value are added and averaged to obtain an updated delay value;

[0179] The expression for updating the delay value is:

[0180] ;

[0181] In the formula, To update the delay value.

[0182] After obtaining the node fluctuation frequency, associated edge update degree and update delay value, the production allocation index representing the intelligent production allocation level of the industrial knowledge graph can be calculated based on the node fluctuation frequency, associated edge update degree and update delay value, and used as the basis for subsequent intelligent production allocation;

[0183] The expression of production allocation index is:

[0184] ;

[0185] In the formula, is the production allocation index, , , is the weight factor, and , , All are greater than 0;

[0186] in, , exemplary, is 0.42, is 0.30, It is 0.28.

[0187] S5: Compare the production allocation index with the preset production allocation threshold, formulate an allocation level, and control the intelligent production allocation of nodes and edges according to the allocation level;

[0188] After the production allocation index is calculated, the production allocation index can be compared with the preset production allocation threshold, and the level of intelligent production allocation at the industrial end can be formulated according to the difference after comparison, so as to obtain the allocation level, so that the allocation level can achieve efficient linkage and allocation effect on the production mode and progress of various equipment and processes at the industrial end;

[0189] The deployment level includes the linkage comprehensive deployment level and the independent autonomous deployment level; the linkage comprehensive deployment level means that the equipment and processes in the industrial end need to be coordinated in a linkage manner, and at this time, the nodes and edges in the industrial knowledge graph need to be coordinated as a whole; the independent autonomous deployment level means that the equipment and processes in the industrial end need to be independently coordinated in production, and at this time, the nodes and edges in the industrial knowledge graph need to be independently coordinated;

[0190] The methods for formulating the coordinated comprehensive allocation level and the independent and autonomous allocation level include:

[0191] Production Allocation Index With the preset production allocation threshold Comparison; The preset production allocation threshold refers to the minimum value of the production allocation index when the linkage comprehensive allocation level is formulated, which can be used to numerically limit the minimum requirements for meeting the linkage comprehensive allocation level, so as to achieve an accurate distinction between the linkage comprehensive allocation level and the independent autonomous allocation level;

[0192] when Greater than or equal to When , it means that the number of nodes and edges that have changed in the industrial knowledge graph is large, and the allocation requirements of equipment and processes on the industrial side have reached the minimum requirements of overall linkage allocation. In this case, it is necessary to carry out linkage allocation of nodes and edges in the industrial knowledge graph and formulate a linkage comprehensive allocation level;

[0193] when Less than When , it means that the number of nodes and edges that have changed in the industrial knowledge graph is small, and the allocation needs of equipment and processes on the industrial side have not met the minimum requirements of overall linkage allocation. In this case, it is necessary to allocate the nodes and edges in the industrial knowledge graph separately and formulate an independent and autonomous allocation level.

[0194] After the deployment level is determined, the entities and processes in the industrial end need to be adjusted according to the deployment level, and the specific deployment methods and operations are represented through the nodes and edges in the industrial knowledge graph, so as to quickly respond to small batch personalized demand issues and improve the emergency production deployment capabilities of enterprises for personalized products during production and manufacturing;

[0195] The control methods for intelligent production deployment of nodes and edges include:

[0196] In the industrial knowledge graph, nodes and edges associated with the keywords of y demand features are marked one by one to obtain n target nodes and k target edges;

[0197] According to the order of the production process in the industrial knowledge graph, the n target nodes and k target edges are numbered in ascending order;

[0198] A first dispatching line is established between n target nodes, and the n target nodes are connected to the first dispatching line; a second dispatching line is established between k target edges, and the k target edges are connected to the second dispatching line; the first dispatching line refers to a connection line that can associate multiple different target nodes as a whole, so as to ensure that all target nodes can be in a mutually associated state; the second dispatching line refers to a connection line that can associate multiple different target edges as a whole, so as to ensure that all target edges can be in a mutually associated state;

[0199] When the linkage comprehensive deployment level is formulated, the production deployment center issues a synchronous deployment instruction to control the n target nodes on the first deployment line and the k target edges on the second deployment line to be deployed in an overall linkage manner at the same time;

[0200] When an independent and autonomous deployment level is formulated, the production deployment center issues an asynchronous deployment instruction to control the n target nodes on the first deployment line and the k target edges on the second deployment line to be independently and autonomously deployed in an ascending order of numbers.

[0201] In this embodiment, by collecting comprehensive production data from the industrial end, extracting features to be identified from the comprehensive production data, screening out effective features from the features to be identified, converting effective features into graph points, the graph points include nodes and edges, and constructing an industrial knowledge graph, receiving production demand data from the user end, extracting demand features from the production demand data, importing the demand features into the industrial knowledge graph to update the graph, obtaining the comprehensive allocation parameters of the updated industrial knowledge graph, and calculating the production allocation index, comparing the production allocation index with a preset production allocation threshold, formulating an allocation level, and controlling the intelligent production allocation of nodes and edges according to the allocation level; Compared with the prior art, by constructing an industrial knowledge graph based on effective features of the industrial end, and fusing heterogeneous graphs Technology is used to achieve the automatic update effect of the industrial knowledge graph, so that the industrial knowledge graph can always be kept in the latest version, thereby providing the best graph state for the import of production demand data on the user side, and providing a solid foundation for building user, product and resource portraits. At the same time, by obtaining the comprehensive deployment parameters of the industrial knowledge graph to calculate the production deployment index, it can provide accurate data basis for the subsequent small batch personalized product intelligent production deployment, and then realize the accurate production deployment purpose between the physical equipment and process procedures within the enterprise according to the production deployment relationship of the nodes and edges of the industrial knowledge graph, which can not only improve the accuracy of intelligent production deployment, but also meet the needs of rapid response to small batch personalized products, greatly improving the effect of the enterprise's intelligent production deployment.

[0202] Example 2: Please refer to Figure 2 As shown, the part not described in detail in this embodiment is described in the first embodiment, and an intelligent production deployment system based on an industrial knowledge graph is provided, which is applied to a production deployment center and is used to implement an intelligent production deployment method based on an industrial knowledge graph, including a first extraction module, a graph construction module, a second extraction module, an index calculation module and a production deployment module, wherein each module is connected via a wired or wireless network;

[0203] The first extraction module is used to collect comprehensive production data from the industrial end, which includes equipment data, personnel data, material data, process data and product data, and extract features to be identified from the comprehensive production data;

[0204] A graph construction module is used to select effective features from the features to be identified, convert the effective features into graph points, which include nodes and edges, and construct an industrial knowledge graph;

[0205] The second extraction module is used to receive the production demand data from the user end and extract the demand features from the production demand data;

[0206] The index calculation module is used to import demand characteristics into the industrial knowledge graph to update the graph, obtain the comprehensive allocation parameters of the updated industrial knowledge graph, which include node fluctuation frequency, associated edge update degree and update delay value, and calculate the production allocation index;

[0207] The production allocation module is used to compare the production allocation index with the preset production allocation threshold and formulate an allocation level. The allocation level includes the linkage comprehensive allocation level and the independent autonomous allocation level, and controls the intelligent production allocation of nodes and edges according to the allocation level.

[0208] Example 3: Please refer to Figure 3 As shown, this embodiment discloses an electronic device, including a processor and a memory;

[0209] Wherein, the memory stores a computer program that can be called by the processor;

[0210] The processor executes the intelligent production deployment method based on the industrial knowledge graph by calling the computer program stored in the memory.

[0211] Since the electronic device introduced in this embodiment is an electronic device used to implement an intelligent production deployment method based on an industrial knowledge graph in Example 1 of this application, based on the intelligent production deployment method based on an industrial knowledge graph introduced in the embodiment of this application, the technical personnel of this field can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application is not described in detail here. As long as the technical personnel of this field implement the electronic device used in the intelligent production deployment method based on an industrial knowledge graph in the embodiment of this application, it belongs to the scope of protection of this application.

[0212] Example 4: Please refer to Figure 4 As shown, this embodiment discloses a computer-readable storage medium on which a rewritable computer program is stored;

[0213] When the computer program is run, it implements the intelligent production deployment method based on industrial knowledge graph.

[0214] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An intelligent production deployment method based on industrial knowledge graph, applied to the production deployment center, characterized in that: include: S1: Collect comprehensive production data from the industrial end, which includes equipment data, personnel data, material data, process data, and product data, and extract features to be identified from the comprehensive production data; S2: Filter out effective features from the features to be identified, convert the effective features into graph points, which include nodes and edges, and construct an industrial knowledge graph; S3: receiving production demand data from the user end, and extracting demand features from the production demand data; S4: Import the demand characteristics into the industrial knowledge graph to update the graph, obtain the comprehensive allocation parameters of the updated industrial knowledge graph, the comprehensive allocation parameters include node fluctuation frequency, associated edge update degree and update delay value, and calculate the production allocation index; Methods for obtaining node fluctuation frequency include: A1: At time T1, record the node states of s nodes in the industrial knowledge graph, identify the state values ​​in the node states, and obtain s initial state values; A2: At time T2, record the node status of s nodes in the industrial knowledge graph, identify the state value in the node status, and obtain s updated state values; A3: The node state whose updated state value is inconsistent with the initial state value is recorded as an abnormal state, and the node corresponding to the abnormal state is recorded as a fluctuating node, and the number of fluctuating nodes is counted; A4: Repeat Repeat steps A1-A3 to obtain The number of fluctuation nodes, and The number of fluctuation nodes is compared with the duration from time T1 to time T2 to obtain sub-frequency; The expression of sub-frequency is: ； In the formula, For the The sub-frequency, For the The number of fluctuation nodes, is the duration between time T1 and time T2; A5: The sub-frequencies are accumulated and averaged to obtain the node fluctuation frequency; The expression of node fluctuation frequency is: ； In the formula, is the node fluctuation frequency, For the sub-frequency; The methods for obtaining the update degree of the associated edge include: Take T3 as the starting time and the preset monitoring time as the interval standard to mark Monitoring time points; exist At each monitoring time point, the association values ​​of the w edges are queried one by one through the association management system to obtain w association values; Compare any two association values ​​one by one, record the edges with the same association value as associated edges, and count the number of associated edges to obtain associated value; Will The association values ​​are compared with the total number of edges in turn, and the comparison values ​​are accumulated and averaged to obtain the update degree of the association edges; The expression of the update degree of the associated edge is: ； In the formula, is the update degree of the associated edge, For the The associated value, is the total number of edges; The methods for obtaining the update delay value include: Query one by one by timestamp The time when the node state of the fluctuation node first fluctuates and the time when the node state fluctuates for the last time are obtained. The first moment and a second moment; Will The first moment to The duration between the second moments is recorded as the first update duration, and the first update duration greater than the preset first duration threshold is recorded as the first delay duration, to obtain The first delay duration; Will Subtract the first delay time lengths from the preset first time length threshold in turn and add up the differences to obtain a first delay value; The expression for the first delay value is: ； In the formula, is the first delay value, For the The first delay time, is a preset first duration threshold; The second update durations of all associated edges are queried one by one through the timestamps, and the second update durations greater than the preset second duration threshold are recorded as the second delay durations. The second delay duration; Will Subtract the second delay time lengths from the preset second time length threshold in turn and add up the differences to obtain a second delay value; The expression for the second delay value is: ； In the formula, is the second delay value, For the The second delay time, is a preset second duration threshold; The first delay value and the second delay value are added and averaged to obtain an updated delay value; The expression for updating the delay value is: ； In the formula, To update the delay value; The expression of production allocation index is: ； In the formula, is the production allocation index, , , is the weight factor, and , , All are greater than 0; S5: Compare the production allocation index with the preset production allocation threshold to formulate an allocation level, which includes the linkage comprehensive allocation level and the independent autonomous allocation level, and control the intelligent production allocation of nodes and edges according to the allocation level.

2. According to claim 1, the intelligent production deployment method based on industrial knowledge graph is characterized in that: The method for extracting the features to be identified includes: Through natural language processing technology, semantic recognition is performed on i comprehensive production data one by one to obtain i text semantics and i digital semantics; The workshop location, equipment number, manufacturing process, production materials and testing instruments are recorded as standard characters, and the standard characters are combined into standard features; According to the order of query time, the i-character semantics are compared with the standard features in turn, and the character semantics that are repeated with any standard character in the standard features are recorded as the semantics to be recognized, and the semantics to be identified; Build A blank basic group, and draw the first characteristic bit and the second characteristic bit in the basic group respectively, to obtain The first characteristic bit and The second characteristic bit; Will The semantics to be recognized and the corresponding The digital semantics are imported into The first characteristic bit and In the second characteristic bit, we get feature groups, and After the feature groups are numbered in sequence, we get Features to be identified.

3. According to claim 2, the intelligent production deployment method based on industrial knowledge graph is characterized in that: The screening method of the effective features includes: Through the attribute management system, we can query The attribute status box of the feature to be identified is marked with the remark value in the attribute status box. The first remark value and The second remark value; Will The first remark value and the corresponding Multiply the second remark value by Status index; The expression of the state index is: ； In the formula, For the The state index of the feature to be identified, For the The first annotation value of the feature to be identified, For the The second annotation value of the feature to be identified; The state index that is less than the preset index threshold is recorded as a valid index, and the feature to be identified corresponding to the valid index is recorded as a valid feature, and s valid features are obtained.

4. According to claim 3, the intelligent production deployment method based on industrial knowledge graph is characterized in that: The method for constructing the industrial knowledge graph includes: Use NER technology to identify the key entities of s effective features one by one, obtain s key entities, and identify the attribute information of s key entities respectively; The relationship between two adjacent key entities among the s key entities is extracted through the relationship extraction model to obtain w relationships; The s key entities are converted into s basic points through virtual reality technology, and information boxes are drawn on the s basic points respectively; Import the attribute information of the s key entities into the information boxes of the s basic points, generate s nodes, and number the s nodes one by one; Identify the numbers of the nodes where the key entities at the front and back ends of the w relationships are located one by one, and obtain w front numbers and w back numbers; Connect the front ends of w relations to the nodes with the same number as the w front numbers, and connect the back ends of w relations to the nodes with the same number as the w back numbers, to obtain w edges; The s nodes and w edges are stored as a knowledge graph through the graph database, and a regular update mechanism is set in combination with heterogeneous graph technology to construct an industrial knowledge graph.

5. According to claim 4, the intelligent production deployment method based on industrial knowledge graph is characterized in that: The method for extracting the demand characteristics includes: The input data of the user end is queried through the enterprise production database, and the information status of the input data is identified one by one, and the input data with the information status of the user is recorded as production demand data, and u production demand data are obtained; Use natural language processing technology to identify the demand semantics of u production demand data one by one, and mark the keywords in the demand semantics; The demand semantics with the keywords progress, allocation, quality and time are recorded as demand features, and y demand features are obtained.

6. According to claim 5, the intelligent production deployment method based on industrial knowledge graph is characterized in that: The method for formulating the linkage comprehensive allocation level and the independent and autonomous allocation level includes: Production Allocation Index With the preset production allocation threshold Compare; when Greater than or equal to When the government is in a hurry to find a coordinated and comprehensive deployment level, when Less than When the government is in a position to provide independent and autonomous deployment levels, The control methods for intelligent production deployment of nodes and edges include: In the industrial knowledge graph, nodes and edges associated with the keywords of y demand features are marked one by one to obtain n target nodes and k target edges; According to the order of the production process in the industrial knowledge graph, the n target nodes and k target edges are numbered in ascending order; Establishing a first dispatching line between n target nodes and connecting the n target nodes to the first dispatching line, establishing a second dispatching line between k target edges and connecting the k target edges to the second dispatching line; When the linkage comprehensive deployment level is formulated, the production deployment center issues a synchronous deployment instruction to control the n target nodes on the first deployment line and the k target edges on the second deployment line to be deployed in an overall linkage manner at the same time; When an independent and autonomous deployment level is formulated, the production deployment center issues an asynchronous deployment instruction to control the n target nodes on the first deployment line and the k target edges on the second deployment line to be independently and autonomously deployed in an ascending order of numbers.

7. An intelligent production deployment system based on industrial knowledge graph, applied to a production deployment center, used to implement an intelligent production deployment method based on industrial knowledge graph according to any one of claims 1 to 6, characterized in that: It includes a first extraction module, a graph construction module, a second extraction module, an index calculation module and a production deployment module, wherein each module is connected via a wired or wireless network; The first extraction module is used to collect comprehensive production data from the industrial end, which includes equipment data, personnel data, material data, process data and product data, and extract features to be identified from the comprehensive production data; A graph construction module is used to select effective features from the features to be identified, convert the effective features into graph points, which include nodes and edges, and construct an industrial knowledge graph; The second extraction module is used to receive the production demand data from the user end and extract the demand features from the production demand data; The index calculation module is used to import demand characteristics into the industrial knowledge graph to update the graph, obtain the comprehensive allocation parameters of the updated industrial knowledge graph, which include node fluctuation frequency, associated edge update degree and update delay value, and calculate the production allocation index; The production allocation module is used to compare the production allocation index with the preset production allocation threshold and formulate an allocation level. The allocation level includes the linkage comprehensive allocation level and the independent autonomous allocation level, and controls the intelligent production allocation of nodes and edges according to the allocation level.

Images