Artificial intelligence-based industrial software design and application platform

Abstract

The application discloses an industrial software design and application platform based on artificial intelligence and relates to the technical field of industrial software.The platform comprises a demand analysis and modeling module, which is used for acquiring industrial business demand information and domain knowledge, performing natural language processing and semantic analysis on the industrial business demand information, and performing structured representation on the domain knowledge to form a preliminary industrial knowledge model; and a knowledge graph construction module, which is used for being based on the preliminary industrial knowledge model.The demand analysis and modeling module is used for acquiring industrial business demand and domain knowledge through multiple channels, forming a preliminary knowledge model through natural language processing and structured modeling, and combining the knowledge graph construction module to construct a unified industrial knowledge graph containing industrial concepts, business processes and the like, so that the problem of scattered knowledge of a traditional platform and reliance on artificial experience to understand demand is solved.

Description

Technical Field

[0001] This invention relates to the field of industrial software technology, specifically to an artificial intelligence-based industrial software design and application platform. Background Technology

[0002] As a core support for intelligent manufacturing and industrial digital transformation, the level of industrial software research and application directly affects the competitiveness and innovation efficiency of the modern industrial system. Industrial software covers multiple key aspects, from computer-aided design and engineering simulation to production execution systems, aiming to improve the intelligence level of industrial product design, manufacturing, and management through digital means.

[0003] The design, development, and deployment of industrial software is a key research area in the field of industrial technology. This area focuses on encapsulating complex industrial knowledge, business processes, and algorithm models into reusable and scalable software components to support industrial enterprises in rapidly building customized digital solutions.

[0004] Existing industrial software platforms generally adopt traditional architectures and development models, resulting in excessive coupling between software modules, poor functional reusability, and difficulty in adapting to rapidly changing industrial scenarios. Although low-code or no-code development technologies have lowered the barrier to software development to some extent, existing platforms still have significant shortcomings in the depth of encapsulation of industrial knowledge, automated integration of algorithm models, and real-time processing capabilities for multi-source heterogeneous data. The design process of industrial software heavily relies on the experience of professional developers, lacking intelligent understanding and adaptive optimization capabilities for industrial business processes, resulting in long software customization cycles and low iteration efficiency. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an artificial intelligence-based industrial software design and application platform, which solves the problems of high coupling, poor functional reusability, insufficient multi-source heterogeneous data processing capabilities, and long customization cycles due to reliance on professional personnel experience in existing industrial software platforms.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an artificial intelligence-based industrial software design and application platform, the platform comprising: The requirement analysis and modeling module is used to acquire industrial business requirement information and domain knowledge, perform natural language processing and semantic analysis on the industrial business requirement information, and perform structured representation on the domain knowledge to form a preliminary industrial knowledge model. The knowledge graph construction module is used to automatically construct an industrial knowledge graph containing industrial concepts, business process definitions, algorithm models, data specifications, and interrelationships based on the preliminary industrial knowledge model and combined with predefined industrial ontology and relationship rules. The industrial knowledge graph serves as the basic knowledge base for intelligent design and application of industrial software. The intelligent component management module is used to select pre-set reusable software components from the component library or generate customized software components based on the business processes and algorithm models defined in the industrial knowledge graph and through artificial intelligence reasoning models. It also performs parameterized configuration and interface adaptation on the selected or generated software components to form software functional units that meet specific industrial business needs. The multi-source data integration module is used to collect, convert, clean and semantically align heterogeneous industrial data from different industrial equipment, sensors, enterprise information systems and external data sources in real time, and integrate the processed data into a unified data view to provide data input and operation support for the software functional units generated by the intelligent component management module. The platform-based deployment and operation module is used to provide a containerized deployment environment and operation support services for software functional units, so as to support the application deployment of the software functional units in the target industrial scenario, and monitor the operation status, resource consumption and business execution of the software functional units in real time.

[0007] Furthermore, the demand analysis and modeling module is used to obtain industrial business demand information, including: Obtain industrial business requirements in the form of text descriptions input by users; Obtain domain knowledge contained in industry standards, regulations, and historical project documents; Access industrial operating procedures and decision-making logic stored in the expert experience database.

[0008] Furthermore, the knowledge graph construction module automatically constructs an industrial knowledge graph, including: Identify key entities in the preliminary industrial knowledge model, including equipment type, production process, failure mode, material properties, and process parameters; Identify the relationships between the key entities, including composition relationships, dependency relationships, driving relationships, and constraint relationships; The identified key entities and relationships are mapped to the predefined industrial ontology, and the industrial knowledge graph is stored using a graph database.

[0009] Furthermore, the business processes defined in the industrial knowledge graph include: Process nodes, process sequence, input and output data types, and decision branches.

[0010] Furthermore, the intelligent component management module, through an artificial intelligence inference model, selects pre-set reusable software components from the component library, or generates customized software components, including: Based on the business process definitions in the industrial knowledge graph, the types and functional boundaries of the required software functional units are determined. Based on the type and functional boundaries of the software functional units, a machine learning model is invoked to evaluate the component matching degree, and candidate software components are selected from the pre-built component library; For business requirements that have no matching or low matching degree, a code generation model is used to generate customized code snippets that meet the functional specifications and encapsulate them into new software components.

[0011] Furthermore, the parameterized configuration of the selected or generated software components includes: Based on the algorithm model parameter specifications defined in the industrial knowledge graph, the initial running parameters are automatically set for the algorithm model in the software component. Based on the data specifications in the industrial knowledge graph, configure the data interface and data processing logic of the software components to adapt to the unified data view provided by the multi-source data integration module.

[0012] Furthermore, the multi-source data integration module is used to collect heterogeneous industrial data from different industrial equipment, sensors, enterprise information systems, and external data sources in real time, including: It communicates with industrial equipment through multiple protocol adapters to collect the operating parameters and status data of the equipment; By subscribing to sensor data streams through message queues, environmental and production process data can be obtained in real time. Business data and historical production records can be obtained from enterprise resource planning systems and manufacturing execution systems through application programming interfaces (APIs) or data service interfaces (DATA APIs).

[0013] Furthermore, the multi-source data integration module integrates the processed data into a unified data view, including: Timestamp alignment and spatial location mapping are performed on data from different sources; Data fusion algorithms are used to merge data with the same semantics but different representations to eliminate data redundancy; Construct a unified data model based on semantic grids to provide the unified data view.

[0014] Furthermore, the platform-based deployment and operation module is also used to adaptively adjust the software configuration and operation strategy based on operational feedback using artificial intelligence optimization algorithms, including: Obtain the performance and business metrics of the software functional units during actual operation; Based on the aforementioned performance and business metrics, a reinforcement learning model is used to evaluate the effectiveness of the current software configuration and operation strategy. Based on the evaluation results, the allocation of computing resources, parallelism settings, and internal algorithm parameters of the software functional units are dynamically adjusted.

[0015] Furthermore, the platform also includes: The human-computer interaction interface module provides a graphical user interface that supports users in inputting industrial business requirements, visually browsing and editing industrial knowledge graphs, drag-and-drop combination and deployment configuration of software components, and monitoring and feedback of the operational status of industrial software after deployment.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention acquires industrial business requirements and domain knowledge from multiple channels through a requirements analysis and modeling module. A preliminary knowledge model is formed through natural language processing and structured modeling. Combined with a knowledge graph construction module, a unified industrial knowledge graph encompassing industrial concepts and business processes is built, addressing the problems of fragmented knowledge and reliance on human experience to understand requirements in traditional platforms. An intelligent component management module, relying on the knowledge graph, uses AI reasoning to select reusable components or generate customized components and complete adaptation, reducing software module coupling, improving functional reusability, and addressing the poor reusability of traditional platforms. A multi-source data integration module collects heterogeneous industrial data in real time and integrates it into a unified data view, solving the problem of insufficient multi-source heterogeneous data processing capabilities in traditional platforms. A platform-based deployment and operation module provides a containerized environment and monitors and adjusts it in real time, further optimizing operational performance. Overall, this significantly shortens the industrial software customization cycle, improves iteration efficiency, reduces reliance on the experience of professional developers, meets the needs of rapidly changing industrial scenarios, and provides core support for intelligent manufacturing and industrial digital transformation. Attached Figure Description

[0017] Figure 1 This is a diagram of the overall platform architecture of the present invention; Figure 2 This is a flowchart illustrating the knowledge graph construction process of the present invention. Figure 3 This is a schematic diagram illustrating the generation of the intelligent component of the present invention; Figure 4 This is a flowchart of the multi-source data processing of the present invention; Figure 5 This is the adaptive operation optimization diagram of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1-5 This invention provides an artificial intelligence-based industrial software design and application platform, the platform comprising: The requirements analysis and modeling module is used to acquire industrial business requirements information and domain knowledge, perform natural language processing and semantic analysis on the industrial business requirements information, and represent the domain knowledge in a structured way to form a preliminary industrial knowledge model. The knowledge graph construction module is used to automatically construct an industrial knowledge graph based on a preliminary industrial knowledge model, combined with predefined industrial ontology and relationship rules. This graph includes industrial concepts, business process definitions, algorithm models, data specifications, and interrelationships. The industrial knowledge graph serves as the basic knowledge base for intelligent design and application of industrial software. The intelligent component management module is used to select pre-built reusable software components from the component library or generate customized software components based on the business processes and algorithm models defined in the industrial knowledge graph and through artificial intelligence reasoning models. It also performs parameterized configuration and interface adaptation of the selected or generated software components to form software functional units that meet specific industrial business needs. The multi-source data integration module is used to collect, convert, clean and semantically align heterogeneous industrial data from different industrial equipment, sensors, enterprise information systems and external data sources in real time, and integrate the processed data into a unified data view to provide data input and operation support for the software functional units generated by the intelligent component management module. The platform-based deployment and operation module provides a containerized deployment environment and operation support services for software functional units, enabling them to be deployed in target industrial scenarios and monitor their operation status, resource usage, and business execution in real time.

[0020] Specifically, in practical applications of this platform, the requirements analysis and modeling module first acquires industrial business requirements information and domain knowledge through multiple channels. It then uses the BERT semantic understanding model to perform natural language processing on the textual business requirements input by users, extracting key requirement elements. Simultaneously, it retrieves relevant regulatory documents and historical project documents from industry standard databases, combines them with structured industrial operation specifications stored in the expert experience base, and transforms domain knowledge into structured data containing entities, attributes, and relationships through ontology methods, forming a preliminary industrial knowledge model and solving the problem of fragmented requirements understanding in traditional platforms.

[0021] The knowledge graph construction module is based on a preliminary industrial knowledge model. It predefines an industrial ontology and uses the OWL language to describe core concepts and relationship rules such as equipment and processes. It identifies key entities such as equipment types and production links through the BiLSTM-CRF model, and uses the Transformer model to mine the composition and dependency relationships between entities. After mapping entities and relationships to the predefined ontology, it uses the Neo4j graph database for storage to build a complete industrial knowledge graph, providing unified knowledge support for subsequent processes and breaking the limitations of traditional platform knowledge dispersion.

[0022] Based on the business processes and algorithm models in the knowledge graph, the intelligent component management module adopts a fusion reasoning model of "rule reasoning plus deep learning". It first selects reusable components from the pre-built component library. For requirements without matching components, it calls the GPT-4 code generation model to generate customized code snippets and encapsulate them into components. Then, based on the parameter specifications and data specifications in the graph, it completes the component parameter configuration and interface adaptation to form software functional units, which greatly reduces the module coupling and improves the component reusability.

[0023] Preferably, the pre-built component library can initially be encapsulated from industry-standard software function modules, components accumulated from historical projects, and standardized algorithm models, and can be continuously enriched and updated during platform use through newly generated customized components.

[0024] Preferably, candidate code snippets are generated through a code generation model, and after formal verification or rule-based code review, they are confirmed by a qualified security engineer before being packaged.

[0025] The multi-source data integration module communicates with industrial equipment through adapters supporting protocols such as Modbus and OPCUA to collect equipment operating parameters; it subscribes to sensor data streams using a Kafka message queue to obtain environmental and production data in real time; and it retrieves business data and historical records from enterprise ERP and MES systems via RESTful APIs. For the collected heterogeneous data, the PCA dimensionality reduction algorithm is used to clean noisy data. After format conversion and semantic alignment, the data is integrated into a unified data view, providing stable data input for software functional units and solving the problem of insufficient data processing capabilities in traditional platforms.

[0026] The platform-based deployment and operation module uses Docker container technology to build deployment environments for software functional units, and Kubernetes to achieve container orchestration. Simultaneously, the Prometheus monitoring tool collects real-time data on the functional units' operating status, CPU utilization, and business execution progress. This implementation significantly reduces software customization cycles, improves the efficiency of industrial software iteration, and solves the problems of high coupling, poor reusability, and reliance on human experience inherent in traditional platforms.

[0027] In this embodiment, the demand analysis and modeling module is used to obtain industrial business demand information, including: Obtain industrial business requirements in the form of text descriptions input by users; Obtain domain knowledge contained in industry standards, regulations, and historical project documents; Access industrial operating procedures and decision-making logic stored in the expert experience database.

[0028] Specifically, when the requirements analysis and modeling module acquires industrial business requirements, it first builds a web-based requirements input interface, allowing users to enter business requirements in text format. The interface includes built-in requirements guidance prompts to help users fully describe their needs. Then, it connects to an industry standard database via a data interface to automatically retrieve and extract domain knowledge from industry standards, regulations, and historical project documents relevant to the current business, forming a knowledge document set. Simultaneously, it accesses an expert experience base, which stores industrial operating procedures and decision-making logic in structured tables, extracting relevant experience data through keyword matching. By acquiring requirements and knowledge through multiple channels, it ensures the comprehensiveness of requirements information, reduces the problem of missing requirements, and provides sufficient data support for subsequent modeling.

[0029] In this embodiment, the knowledge graph construction module automatically constructs an industrial knowledge graph, including: Identify key entities in the preliminary industrial knowledge model. Key entities include equipment type, production process, failure mode, material properties, and process parameters. Identify the relationships between key entities, including composition relationships, dependency relationships, driving relationships, and constraint relationships; The identified key entities and relationships are mapped to predefined industrial ontology, and the industrial knowledge graph is stored using a graph database.

[0030] Specifically, when the knowledge graph construction module automatically builds the industrial knowledge graph, it first uses a BiLSTM-CRF model to identify key entities in the preliminary industrial knowledge model, accurately extracting entities such as equipment type, production process, failure mode, material attributes, and process parameters from text data. Then, it uses a Transformer model to analyze the contextual relationships between entities, identifying relationships such as composition, dependency, driving force, and constraint. Subsequently, it uses the SPARQL query language to map the identified entities and relationships to a predefined industrial ontology, ensuring the consistency of the graph concepts. Finally, it imports the mapped entity and relationship data into the Neo4j graph database to complete the storage of the industrial knowledge graph. This process improves the accuracy and efficiency of graph construction, providing a reliable knowledge foundation for intelligent software design.

[0031] In this embodiment, the business processes defined in the industrial knowledge graph include: Process nodes, process sequence, input and output data types, and decision branches.

[0032] Specifically, when defining business processes in an industrial knowledge graph, process nodes are presented in the form of a visual flowchart, with each node labeled with its corresponding industrial operation task. Directed edges are used to represent the process sequence, clearly defining the execution order between nodes. The input and output data types for each node are defined using JSON format, specifying data fields and format requirements. For processes with branches, decision-making branch conditions are set based on business rules, such as triggering a quality inspection branch when process parameters exceed limits. Clear business process definitions improve the accuracy of intelligent component matching, ensuring that software functional units conform to business logic.

[0033] In this embodiment, the intelligent component management module selects pre-set reusable software components from the component library or generates customized software components through an artificial intelligence inference model, including: Based on the business process definition in the industrial knowledge graph, determine the type and functional boundaries of the required software functional units; Based on the type and functional boundaries of the software functional units, a machine learning model is invoked to evaluate the component matching degree and candidate software components are selected from the pre-built component library; For business requirements that have no matching or low matching degree, a code generation model is used to generate customized code snippets that meet the functional specifications and encapsulate them into new software components.

[0034] Specifically, when the intelligent component management module selects or generates software components through an artificial intelligence inference model, it first breaks down the process into multiple sub-tasks based on the business process definition in the knowledge graph, clarifying the software functional unit type and functional boundaries corresponding to each sub-task; then, it calls the machine learning model to calculate the matching degree between components and functional units in the pre-set component library. The matching degree evaluation uses the cosine similarity algorithm, and the formula is: ; in The feature vector of the functional unit, For the feature vector of the pre-set component, For vector dot product, , Based on the vector magnitude, candidate components with a cosine similarity greater than 0.8 are selected. For requirements with no candidate components or a similarity lower than 0.5, a code generation model is invoked, input functional specification parameters, and customized code snippets are generated. These snippets are then encapsulated to form new software components. This approach improves component reusability and shortens the development cycle of customized components.

[0035] Preferably, the feature vector A is formed by concatenating the process type code (1-10), the number of input data dimensions (integer), and the algorithm complexity level (1-5) after One-Hot encoding and normalization.

[0036] The threshold of 0.8 is determined by training an ROC curve on historical project data, or it can be dynamically adjusted by the user according to the industrial safety level.

[0037] In this embodiment, the parameterized configuration of the selected or generated software components includes: Based on the algorithm model parameter specifications defined in the industrial knowledge graph, the initial running parameters are automatically set for the algorithm models in the software components. Based on the data specifications in the industrial knowledge graph, configure the data interfaces and data processing logic of the software components to adapt to the unified data view provided by the multi-source data integration module.

[0038] Specifically, when configuring the selected or generated software components, the parameter specifications of the corresponding algorithm models are first retrieved from the industrial knowledge graph. For example, the proportional coefficient and integral time ranges for PID control algorithms are automatically set for the initial operating parameters of the algorithm models within the components. Then, based on the data specifications in the graph, the data interface type of the components is configured, using RESTful API or WebSocket protocols. Simultaneously, data processing logic, such as data format conversion and outlier filtering rules, is defined to ensure that the components can adapt to the unified data view provided by the multi-source data integration module. This configuration process requires no manual intervention, improving component adaptation efficiency and reducing configuration errors.

[0039] In this embodiment, the multi-source data integration module is used to collect heterogeneous industrial data from different industrial equipment, sensors, enterprise information systems, and external data sources in real time, including: It communicates with industrial equipment through multiple protocol adapters to collect the equipment's operating parameters and status data; By subscribing to sensor data streams through message queues, environmental and production process data can be obtained in real time. Business data and historical production records can be obtained from enterprise resource planning systems and manufacturing execution systems through application programming interfaces (APIs) or data service interfaces (DATA APIs).

[0040] Specifically, when the multi-source data integration module collects heterogeneous industrial data, it first deploys multiple protocol adapters supporting industrial protocols such as Modbus, OPCUA, and Profinet. These adapters establish communication connections with industrial equipment, periodically collecting equipment operating parameters and status data. Next, a Kafka message queue cluster is built. Sensor devices push data streams to the message queue via TCP, and the module subscribes to queue topics to obtain real-time data such as ambient temperature and production progress. Finally, by calling the application programming interfaces (APIs) or data service interfaces provided by the ERP and MES systems, business order data, historical production records, and other information are retrieved at a set frequency. This multi-protocol and multi-interface support ensures the real-time and comprehensiveness of data collection, meeting the data requirements of the software functional units.

[0041] In this embodiment, the multi-source data integration module integrates the processed data into a unified data view, including: Timestamp alignment and spatial location mapping are performed on data from different sources; Data fusion algorithms are used to merge data with the same semantics but different representations to eliminate data redundancy; Build a unified data model based on semantic grids to provide a unified data view.

[0042] Specifically, when the multi-source data integration module integrates data into a unified data view, it first uses the NTP network time protocol to align the timestamps of data from different sources, and combines this with equipment installation location information to complete spatial location mapping, ensuring spatiotemporal consistency of the data. Then, for data with the same semantics but different representations, such as temperature data collected by different devices, a weighted average algorithm is used for data fusion, with the formula as follows: ; in For data from various data sources, To assign weights to corresponding data sources, each data source's weight is determined based on its device accuracy level, data acquisition frequency, and data integrity. Higher device accuracy, more stable acquisition frequency, and better data integrity result in a larger weight value. These weight values ​​are then normalized to meet certain conditions. At this point, the weighted average formula can be simplified to: The merged data retains the highest precision of the original data from each data source. The number of significant digits is set according to the actual application requirements of the industrial scenario, and is not less than the lowest significant digits of the original data from each data source. If there are differences in the number of significant digits among the data sources, the merged data is rounded to eliminate data redundancy. Finally, a unified data model based on RDF semantic mesh is constructed, and the processed data is stored according to semantic classification to form a unified data view. This process improves data quality and provides unified and reliable data support for software functional units.

[0043] In this embodiment, the platform deployment and operation module is further configured to adaptively adjust the software configuration and operation strategy based on operational feedback using an artificial intelligence optimization algorithm, including: Obtain the performance and business metrics of software functional units during actual operation; Based on performance and business metrics, reinforcement learning models are used to evaluate the effectiveness of current software configuration and operation strategies. Based on the evaluation results, the allocation of computing resources, parallelism settings, and internal algorithm parameters of the software functional units are dynamically adjusted.

[0044] Specifically, when the platform-based deployment and operation modules are adaptively adjusted, the Prometheus monitoring tool is first used to collect performance and business metrics of the software functional units. Performance metrics include response time, memory usage, CPU utilization, and disk read / write speed, while business metrics include task completion rate, error rate, abnormal data identification rate, and business process compliance rate. The metric data is then input into a DQN-based reinforcement learning model. The model aims to maximize the operational stability and business efficiency of the functional units, evaluating the effectiveness of the current software configuration and operation strategy. Based on the evaluation results, the allocation of computing resources for the functional units is dynamically adjusted, such as increasing or decreasing the number of CPU cores and memory capacity; parallelism settings are adjusted, such as adjusting the number of threads; and internal algorithm parameters are optimized, such as optimizing the learning rate of the prediction model. Adaptive adjustment ensures that the software functional units maintain stable operation in different industrial scenarios, improving platform adaptability.

[0045] In this embodiment, the platform further includes: The human-computer interaction interface module provides a graphical user interface that supports users in inputting industrial business requirements, visually browsing and editing industrial knowledge graphs, drag-and-drop combination and deployment configuration of software components, and monitoring and feedback of the operational status of industrial software after deployment.

[0046] Specifically, the human-computer interaction interface module uses the Vue front-end framework to develop a graphical user interface, which is divided into a requirement input area, a knowledge graph browsing area, a component operation area, and an operation monitoring area. The requirement input area provides text input boxes and template selection functions, allowing users to easily enter industrial business requirements. The knowledge graph browsing area uses visual charts to display the industrial knowledge graph; users can zoom and drag graph nodes, click on nodes to view detailed information, and it also supports editing functions, allowing users to add or modify entities and relationships. The component operation area provides drag-and-drop component combination functionality; users can drag components from the component library to the canvas, configure relationships between components through connections, and complete deployment configuration. The operation monitoring area displays the operating status data of software functional units in a dashboard format, allowing users to view real-time feedback information. This interface lowers the user's operating threshold, enabling even non-professional developers to quickly use the platform to complete industrial software design and deployment.

[0047] In summary, this invention acquires industrial business requirements and domain knowledge through multiple channels via a requirements analysis and modeling module. A preliminary knowledge model is formed through natural language processing and structured modeling. This is combined with a knowledge graph construction module to build a unified industrial knowledge graph encompassing industrial concepts and business processes, addressing the problems of fragmented knowledge and reliance on human experience to understand requirements in traditional platforms. The intelligent component management module, relying on the knowledge graph, uses AI reasoning to select reusable components or generate customized components and complete adaptation, reducing software module coupling, improving functional reusability, and addressing the poor reusability of traditional platforms. The multi-source data integration module collects heterogeneous industrial data in real time and integrates it into a unified data view, solving the problem of insufficient multi-source heterogeneous data processing capabilities in traditional platforms. The platform deployment and operation module provides a containerized environment and monitors and adjusts it in real time, further optimizing operational performance. Overall, this significantly shortens the industrial software customization cycle, improves iteration efficiency, reduces reliance on the experience of professional developers, meets the needs of rapidly changing industrial scenarios, and provides core support for intelligent manufacturing and industrial digital transformation.

[0048] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An industrial software design and application platform based on artificial intelligence, characterized in that, The platform includes: The requirement analysis and modeling module is used to acquire industrial business requirement information and domain knowledge, perform natural language processing and semantic analysis on the industrial business requirement information, and perform structured representation on the domain knowledge to form a preliminary industrial knowledge model. The knowledge graph construction module is used to automatically construct an industrial knowledge graph containing industrial concepts, business process definitions, algorithm models, data specifications, and interrelationships based on the preliminary industrial knowledge model and combined with predefined industrial ontology and relationship rules. The industrial knowledge graph serves as the basic knowledge base for intelligent design and application of industrial software. The intelligent component management module is used to select pre-set reusable software components from the component library or generate customized software components based on the business processes and algorithm models defined in the industrial knowledge graph and through artificial intelligence reasoning models. It also performs parameterized configuration and interface adaptation on the selected or generated software components to form software functional units that meet specific industrial business needs. The multi-source data integration module is used to collect, convert, clean and semantically align heterogeneous industrial data from different industrial equipment, sensors, enterprise information systems and external data sources in real time, and integrate the processed data into a unified data view to provide data input and operation support for the software functional units generated by the intelligent component management module. The platform-based deployment and operation module is used to provide a containerized deployment environment and operation support services for software functional units, so as to support the application deployment of the software functional units in the target industrial scenario, and monitor the operation status, resource consumption and business execution of the software functional units in real time.

2. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The requirement analysis and modeling module is used to obtain industrial business requirement information, including: Obtain industrial business requirements in the form of text descriptions input by users; Obtain domain knowledge contained in industry standards, regulations, and historical project documents; Access industrial operating procedures and decision-making logic stored in the expert experience database.

3. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The knowledge graph construction module automatically constructs an industrial knowledge graph, including: Identify key entities in the preliminary industrial knowledge model, including equipment type, production process, failure mode, material properties, and process parameters; Identify the relationships between the key entities, including composition relationships, dependency relationships, driving relationships, and constraint relationships; The identified key entities and relationships are mapped to the predefined industrial ontology, and the industrial knowledge graph is stored using a graph database.

4. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The business processes defined in the industrial knowledge graph include: Process nodes, process sequence, input and output data types, and decision branches.

5. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The intelligent component management module, through an artificial intelligence inference model, selects pre-set reusable software components from the component library or generates customized software components, including: Based on the business process definitions in the industrial knowledge graph, the types and functional boundaries of the required software functional units are determined. Based on the type and functional boundaries of the software functional units, a machine learning model is invoked to evaluate the component matching degree, and candidate software components are selected from the pre-built component library; For business requirements that have no matching or low matching degree, a code generation model is used to generate customized code snippets that meet the functional specifications and encapsulate them into new software components.

6. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The parameterized configuration of the selected or generated software components includes: Based on the algorithm model parameter specifications defined in the industrial knowledge graph, the initial running parameters are automatically set for the algorithm model in the software component. Based on the data specifications in the industrial knowledge graph, configure the data interface and data processing logic of the software components to adapt to the unified data view provided by the multi-source data integration module.

7. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The multi-source data integration module is used to collect heterogeneous industrial data from different industrial equipment, sensors, enterprise information systems, and external data sources in real time, including: It communicates with industrial equipment through multiple protocol adapters to collect the operating parameters and status data of the equipment; By subscribing to sensor data streams through message queues, environmental and production process data can be obtained in real time. Business data and historical production records can be obtained from enterprise resource planning systems and manufacturing execution systems through application programming interfaces (APIs) or data service interfaces (DATA APIs).

8. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The multi-source data integration module integrates the processed data into a unified data view, including: Timestamp alignment and spatial location mapping are performed on data from different sources; Data fusion algorithms are used to merge data with the same semantics but different representations to eliminate data redundancy; Construct a unified data model based on semantic grids to provide the unified data view.

9. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The platform-based deployment and operation module is also used to adaptively adjust the software configuration and operation strategy based on operational feedback using artificial intelligence optimization algorithms, including: Obtain the performance and business metrics of the software functional units during actual operation; Based on the aforementioned performance and business metrics, a reinforcement learning model is used to evaluate the effectiveness of the current software configuration and operation strategy. Based on the evaluation results, the allocation of computing resources, parallelism settings, and internal algorithm parameters of the software functional units are dynamically adjusted.

10. The artificial intelligence-based industrial software design and application platform according to claim 1, characterized in that, The platform also includes: The human-computer interaction interface module provides a graphical user interface that supports users in inputting industrial business requirements, visually browsing and editing industrial knowledge graphs, drag-and-drop combination and deployment configuration of software components, and monitoring and feedback of the operational status of industrial software after deployment.

Images