Method and system for obtaining optimal parameters of machine tool based on industrial mechanism AI data model

Abstract

The invention provides a method and system for obtaining optimal parameters of a machine tool based on an industrial mechanism AI data model, belongs to the technical field of industrial machine tool application, and solves the problems that an existing machine tool cannot obtain the optimal parameters in real time, autonomous intelligent control of the machine tool cannot be achieved and the like. Comprising the following steps: multi-dimensional high-frequency data fusion acquisition and monitoring; performing data synchronization and storage management; analyzing and mining data; mixing intelligent decision engines; constructing an industrial knowledge base and a model base; training an industrial knowledge base and a model base; intelligent decision making and optimization; equipment is automatically controlled; user visual interaction is carried out; carrying out on-line adaptive parameter optimization; continuously optimizing and iterating; and the production business system works cooperatively. According to the method, online self-adaptive parameter optimization can be realized, a closed-loop control system with rapid sensing, intelligent decision-making and accurate execution capabilities is constructed, the method develops in the direction of sensing-decision-execution full-link autonomous optimization, and finally autonomous intelligent control of the machine tool is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial machine tool applications, and relates to a method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model, in particular to a system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model. Background Art

[0002] A machine tool refers to a machine that manufactures machines, and is generally divided into metal cutting machine tools, forging machine tools, woodworking machine tools, etc. There are many methods for machining mechanical parts in modern mechanical manufacturing: in addition to cutting, there are also casting, forging, welding, stamping, extrusion, etc. However, for parts with relatively high precision requirements and relatively fine surface roughness requirements, they generally need to be finally processed by cutting on a machine tool. Machine tools play a major role in the construction of national economic modernization.

[0003] Reasonably optimizing the operating parameters is of great significance for reducing processing energy consumption, reducing processing costs, improving processing efficiency, and ensuring processing quality. However, there are often contradictions between these goals. With the change of cutting parameters, while improving one goal, another goal may be weakened.

[0004] During the use of a machine tool, due to changes in the external environment and wear of the machine tool, etc., unqualified processing may occur. The performance of the machine tool is a dynamic change process, and the obtained optimal parameter solution can only be applicable to a certain dimension in the past, rather than the optimal choice in the current processing environment.

[0005] Therefore, it is an urgent problem to be solved to obtain real-time and reliable data, perform online adaptive parameter optimization, and finally achieve the autonomous intelligent control of the machine tool.

[0006] Based on this, we propose a method and system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model. Summary of the Invention

[0007] The purpose of the present invention is to address the above problems existing in the prior art, and propose a method and system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model. The technical problem to be solved by this invention is: how to achieve online adaptive parameter optimization, construct a closed-loop control system with the capabilities of rapid perception, intelligent decision-making, and precise execution, develop towards the direction of full-link autonomous optimization of perception-decision-execution, and finally achieve the autonomous intelligent control of the machine tool.

[0008] The purpose of the present invention can be achieved by the following technical solutions:

[0009] A method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model, comprising the following steps:

[0010] Step 1, Multi-dimensional High-frequency Data Fusion Acquisition and Monitoring: Deploy a multi-source heterogeneous data synchronization acquisition system and edge computing nodes. Through the multi-source heterogeneous data synchronization acquisition system combined with sensors, real-time monitor and collect mechanism parameters, status data, quality data, equipment data, environmental parameters, and historical experience data;

[0011] Step 2, Data Synchronization and Storage Management: Provide efficient and scalable big data storage, supporting the storage and access of massive data; Regularly back up the data to ensure data security and recoverability, and guarantee the stable operation of the system; Set strict data access permissions, and different users are assigned different permissions according to their roles and responsibilities to ensure data security and privacy;

[0012] Step 3, Data Analysis and Mining: Dynamic feature extraction, deeply analyze production data to discover patterns, trends, and anomalies in the data, providing a basis for decision-making; Deeply mine production data to discover hidden information, patterns, and knowledge in the data, providing support for prediction and optimization; Based on historical data and current data, build prediction models to predict and evaluate future production situations, equipment status, product quality, etc., and make preparations and decisions in advance;

[0013] Step 4, Intelligent Decision Engine Hybrid: Mechanism-AI Data Dual-Driven Model Collaboration Mechanism, mechanism model and AI model, mechanism model: quickly respond to basic parameter constraints; AI model predicts the optimal parameter offset under dynamic working conditions;

[0014] Step 5, Construction of Industrial Knowledge Base and Model Base: Build an industry knowledge base containing but not limited to industry mechanisms, experience knowledge, standards and specifications, and process flows, encapsulate various industrial models, and industrial models include but not limited to equipment fault prediction models, production optimization models, and quality control models for users to call and modify; Provide a platform-based development environment to support users in quickly building and deploying industrial APPs to meet personalized application requirements;

[0015] Step 6, Training of Industrial Knowledge Base and Model Base

[0016] Supervised Learning Model: For labeled data, supervised learning algorithms can be selected, and supervised learning algorithms include but not limited to support vector machines, decision trees, random forests, and artificial neural networks;

[0017] Unsupervised Learning Model: When the data has no clear labels, unsupervised learning algorithms are used, and unsupervised learning algorithms include but not limited to clustering algorithms and principal component analysis;

[0018] Reinforcement Learning Model: Suitable for problems that need to optimize decisions through continuous trial and error;

[0019] Model training and optimization, data set division: Divide the preprocessed data into a training set, a validation set, and a test set; Select an optimization algorithm: According to the selected AI model, select a suitable optimization algorithm to adjust the model's parameters; Model evaluation and improvement: Use evaluation metrics including but not limited to mean squared error, mean absolute error, accuracy, and recall to evaluate the trained model, and adjust and improve the model according to the evaluation results;

[0020] Step 7, intelligent decision-making and optimization: Based on the results of data analysis and mining, combined with industry knowledge and experience, provide intelligent decision-making suggestions to assist in making decisions not limited to production plans, resource allocation, and marketing strategies; Through optimization algorithms, optimize all aspects of the production process, including production processes, process parameters, and equipment operation; Fault warning and diagnosis; Real-time monitor the equipment status and production process, use models and algorithms to predict and warn of potential faults, and at the same time provide fault diagnosis and repair suggestions to reduce equipment downtime and repair costs;

[0021] Step 8, equipment automatic control: Complete the equipment action control according to the decision in Step 7 to achieve automatic operation;

[0022] Step 9, user visual interaction: Provide an intuitive, simple, and easy-to-operate user interface, lower the user's usage threshold, enable personnel with different professional backgrounds to easily use the system, and through various visual forms not limited to charts, dashboards, and maps, display in real time information not limited to production data, equipment status, and operation indicators, facilitating users to monitor and manage, and promptly discover problems and anomalies; Provide intelligent prompts and detailed help documents to help users quickly get started and solve problems encountered during use, improving user experience and work efficiency;

[0023] Step 10, online adaptive parameter optimization:

[0024] Objective function definition: According to actual requirements, determine the optimization objective function, and express the objective function as a functional relationship between machine tool parameters and machining quality indicators;

[0025] Application of optimization algorithms: Use optimization algorithms including but not limited to genetic algorithms, particle swarm algorithms, and simulated annealing algorithms to perform optimization searches on machine tool parameters based on the model; Search for the optimal solution in a complex solution space, and gradually find the combination of machine tool parameters that optimizes the objective function through continuous iteration and evolution;

[0026] Verification of optimized parameters: Experimentally verify the optimal parameters obtained through model optimization in actual machine tool machining, set the machine tool according to the optimized parameters, conduct machining experiments, and detect machining quality and other related indicators;

[0027] Parameter Fine-tuning: According to the experimental results, fine-tune the optimized parameters. Due to various uncertain factors in the actual machining process, further adjustment of the parameters is required to ensure the best machining effect in actual production;

[0028] Step Eleven, Continuous Optimization and Iteration:

[0029] Machine Tool Machining Process Simulation: Use professional machine tool machining simulation software to establish three-dimensional models of the machine tool, tool, and workpiece, and simulate the machining process under different parameters; through simulation, various physical quantity changes during the machining process and the quality indicators of the machined parts can be obtained;

[0030] Optimization Based on Simulation Data: Combine the industrial mechanism AI data model, use a large amount of data generated by simulation to train and optimize the model. By simulating and evaluating different parameter combinations in the simulation environment, quickly find a better parameter range, and then further verify and fine-tune through actual experiments, thereby reducing the number of actual experiments and improving the optimization efficiency;

[0031] Then feedback various types of data into the model for the continuous optimization and iteration of the model. With the continuous progress of production technology and the accumulation of data, regularly update and optimize the model to ensure that the model always remains in the best state, providing continuous support and optimization for production;

[0032] Step Twelve, Collaborative Work of Production Business Systems: The manufacturing execution system, warehouse management system, and enterprise resource planning system work in collaboration with the mechanism-AI data dual-driven model system, sharing updated data in real-time, optimizing the production process, and at the same time, knowledge transfer and model reuse between different machine tools can be carried out to reduce data requirements.

[0033] The mechanism parameters in Step One include but are not limited to spindle speed, feed rate, and cutting depth; the state data includes but is not limited to vibration signals, temperature field distribution, power curve, and acoustic emission characteristics; the quality data includes but is not limited to surface roughness, geometric tolerance, and residual stress distribution; the equipment data includes but is not limited to tool wear and guide rail wear status; the environmental parameters include but are not limited to coolant viscosity, environmental temperature and humidity, and workpiece material batch characteristics; the historical experience data includes but is not limited to the experience data and professional knowledge accumulated by the enterprise in long-term practice.

[0034] The fusion rule for the intelligent decision engine mixture in Step Four is θ 融合 = α·θ 机理模型 +(1 - α)·θ AI ; where α ∈ [0, 1], and it is dynamically adjusted according to the model confidence.

[0035] Issuing commands for device automatic control in Step 8: Update parameters with a cycle of 1 ms through the EtherCAT bus for smooth transition control and avoid step changes: The actual speed change is the smaller value of the command speed change and the constraint condition.

[0036] Δv 实际 = min{Δv 指令 , 0.05v 当前};

[0037] where Δv 实际 is the actual speed change applied to the system, which is the result of imposing constraints on the command speed change; Δv 指令 is the command speed change, that is, the speed change required by the control system or the user; 0.05v 当前 represents the constraint condition of the speed change rate, which limits the speed change to 5% of the current speed.

[0038] The architecture design of online adaptive optimization in Step 10 is [Edge Sensing Layer] - [GTSN] → [Digital Twin] - [OPCUA] → [Cloud Optimization Engine]. The key technologies adopted are real-time mapping of digital twins. The refresh speed at the 1 ms level enables the virtual model to quickly reflect the state changes of physical entities; with the help of the Federated Learning framework, dynamic update and optimization of the model can be achieved without revealing data privacy; dynamic constraint processing is based on the Lyapunov stability theory; the control closed-loop is real-time monitoring → anomaly detection → model recalibration → parameter re-optimization.

[0039] The data flow period of continuous optimization and iteration in Step 11 is less than 5 ms.

[0040] A system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model, including a data fusion acquisition and monitoring module, a data synchronization and storage management module, a data analysis and mining module, an intelligent decision-making engine hybrid subsystem, an industrial knowledge base and model library construction module, an industrial knowledge base and model library training module, an intelligent decision-making and optimization module, a device automatic control module, a user visual interaction module, an online adaptive parameter optimization module, a continuous optimization and iteration module, and a business collaboration working subsystem.

[0041] The data fusion acquisition and monitoring module includes a data acquisition sub-unit, a data preprocessing sub-unit, a data fusion sub-unit, and a monitoring sub-unit.

[0042] With the above structure, the data fusion acquisition and monitoring module is a key part of many complex systems. Its main function is to collect data from different data sources, perform fusion processing, and monitor the entire process and related indicators.

[0043] The data synchronization and storage management module includes a data extraction subunit, a data conversion subunit, a data transmission subunit, a synchronization scheduling subunit, a storage device management subunit, a data storage policy formulation subunit, a data backup and recovery subunit, and a data access control subunit.

[0044] With the above structure, the data synchronization and storage management module plays a key role in the data processing system. It is responsible for ensuring the consistency of data among different storage locations or systems and effectively storing and managing the data.

[0045] The data analysis and mining module includes a data exploration subunit, a data preprocessing subunit, a model selection and training subunit, a model evaluation subunit, and a result interpretation and presentation subunit.

[0046] With the above structure, the data analysis and mining module is the core part of processing and understanding data.

[0047] The industrial knowledge base and model base construction module includes a knowledge acquisition subunit, a knowledge representation subunit, a knowledge storage subunit, a knowledge update and maintenance subunit, a model design subunit, and a model storage and management subunit.

[0048] With the above structure, the industrial knowledge base and model base construction module plays a key role in industrial intelligence. It provides knowledge support and model basis for industrial production, management, decision-making, etc.

[0049] The industrial knowledge base and model base training module includes a feature engineering subunit, a model training subunit, a model evaluation and optimization subunit, and a knowledge fusion and storage subunit.

[0050] With the above structure, it deeply analyzes and models production data, optimizes the production process, ensures product quality to improve production efficiency; can predict faults based on equipment operation data, optimize maintenance strategies, and enhance equipment management; provides a data-driven basis for enterprise decision-making to help cope with market changes; promotes knowledge accumulation, inheritance and innovation, stimulates the innovation vitality of enterprises; can also integrate data resources, promote the development of industrial intelligence, improve the overall informatization level of enterprises, and enhance the competitiveness and sustainable development ability of enterprises.

[0051] The intelligent decision-making and optimization module includes a decision-making reasoning and generation subunit, an optimization strategy formulation subunit, and a decision-making evaluation and feedback subunit.

[0052] With the above structure, by integrating and preprocessing multi-source data, constructing and training appropriate models, using rule engines and intelligent algorithms to infer and generate decisions, formulating optimization strategies with objective functions and constraint conditions, and simultaneously evaluating the decision-making effect in real time and providing feedback for adjustment, it provides a scientific, accurate and adaptable optimal decision-making solution for complex problems, helping enterprises and organizations improve efficiency, reduce costs and enhance competitiveness.

[0053] The user visual interaction module includes a data visualization presentation subunit, a user interaction operation subunit, a visualization configuration and customization subunit, and a feedback and prompt subunit.

[0054] With the above structure, through diversified data visualization presentation, it provides flexible interaction operations and customization functions, and is paired with a feedback and prompt mechanism to help users intuitively understand data, conveniently operate and analyze, quickly obtain valuable information and respond to problems in a timely manner, improving decision-making efficiency and accuracy.

[0055] The online adaptive parameter optimization module includes a parameter evaluation subunit, an optimization strategy generation subunit, a parameter adjustment and execution subunit, and a knowledge learning and evolution subunit.

[0056] With the above structure, by collecting and preprocessing multi-source data in real time, using index calculation and model prediction to evaluate parameter performance, selecting an appropriate optimization algorithm to generate strategies, updating parameters and monitoring the adjustment effect, and at the same time accumulating experience knowledge for continuous learning and evolution, dynamically adjusting parameters to make the system performance reach the optimal, improving production efficiency, reducing costs and ensuring product quality.

[0057] The continuous optimization and iteration module includes a data integration subunit, a data insight subunit, an optimization strategy formulation subunit, an optimization implementation and monitoring subunit, and a knowledge precipitation and inheritance subunit.

[0058] With the above structure, by collecting and integrating data from multiple channels, deeply analyzing to understand the laws, trends and causal relationships, setting goals, generating evaluations and selecting optimization strategies accordingly, promoting the implementation and monitoring of the plan, and at the same time precipitating and inheriting knowledge and experience, enabling the system, product or business to continuously adapt to changes, improve performance and achieve continuous evolution.

[0059] The business collaboration work subsystem includes a data integration and sharing subunit, a process collaboration and scheduling subunit, a task assignment and execution subunit, a communication and collaboration subunit, and a performance evaluation and continuous improvement subunit.

[0060] With the above structure, by integrating and sharing multi-system data, coordinating and scheduling business processes, reasonably allocating and monitoring task execution, promoting internal communication and collaboration, and evaluating performance to drive continuous improvement, the efficient integration of all production links and business systems, information interconnection, and smooth operation of processes are achieved, thereby enhancing production efficiency, product quality, and the overall competitiveness of the enterprise.

[0061] Compared with the prior art, the method and system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model have the following advantages:

[0062] 1. Construct a closed-loop control system with the capabilities of rapid perception, intelligent decision-making, and precise execution, and develop towards the direction of full-link autonomous optimization of "perception - decision - execution", ultimately achieving the autonomous intelligent control of the machine tool.

[0063] 2. Help users quickly get started and solve problems encountered during use, improving user experience and work efficiency.

[0064] 3. Establish a perfect fault warning mechanism and emergency response system, timely detect possible faults and problems in the system, and take corresponding measures for handling to ensure the normal operation of the system and online adaptive parameter optimization.

[0065] 4. The manufacturing execution system, warehouse management system, and enterprise resource planning system work in coordination with the mechanism - AI data dual - drive model system, sharing and updating data in real time, optimizing the production process, and at the same time, knowledge transfer and model reuse between different machine tools can be carried out, reducing data requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is the flowchart of the method of the present invention.

[0067] Figure 2 is the block diagram of data acquisition in Step 1 of the present invention.

[0068] Figure 3 is the block diagram of Step 4 of the present invention.

[0069] Figure 4 is the flowchart of Step 6 of the present invention.

[0070] Figure 5 is the flowchart of Step 10 of the present invention.

[0071] Figure 6 is the block diagram of the system composition of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0072] The following are specific embodiments of the present invention in combination with the accompanying drawings to further describe the technical solutions of the present invention, but the present invention is not limited to these embodiments.

[0073] As Figures 1 - 5 shown, the method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model includes the following steps:

[0074] Step 1, Multi-dimensional high-frequency data fusion acquisition and monitoring: Deploy a multi-source heterogeneous data synchronization acquisition system and edge computing nodes. Through the multi-source heterogeneous data synchronization acquisition system cooperating with sensors, real-time monitor and collect mechanism parameters, status data, quality data, equipment data, environmental parameters, and historical experience data;

[0075] Step 2, Data synchronization and storage management: Provide efficient and scalable big data storage, support the storage and access of massive data; Regularly back up the data to ensure data security and recoverability, and guarantee the stable operation of the system; Set strict data access permissions, and different users are assigned different permissions according to their roles and responsibilities to ensure data security and privacy;

[0076] Step 3, Data analysis and mining: Dynamic feature extraction, deeply analyze production data to discover the laws, trends, and anomalies in the data, providing a basis for decision-making; Deeply mine production data to discover the hidden information, patterns, and knowledge in the data, providing support for prediction and optimization; Based on historical data and current data, build a prediction model to predict and evaluate future production situations, equipment status, product quality, etc., and make preparations and decisions in advance;

[0077] Step 4, Intelligent decision engine hybrid: Mechanism-AI data dual-drive model collaborative mechanism, mechanism model and AI model, mechanism model: Quickly respond to basic parameter constraints; AI model predicts the optimal parameter offset under dynamic working conditions;

[0078] Step 5, Construction of industrial knowledge base and model library: Build an industry knowledge base containing but not limited to industry mechanisms, experience knowledge, standard specifications, and process flows, encapsulate various industrial models, and industrial models include but not limited to equipment fault prediction models, production optimization models, and quality control models for users to call and modify; Provide a platform-based development environment to support users in quickly building and deploying industrial APPs to meet personalized application requirements;

[0079] Step 6, Training of industrial knowledge base and model library:

[0080] Supervised learning model: For labeled data, supervised learning algorithms can be selected, and supervised learning algorithms include but not limited to support vector machines, decision trees, random forests, and artificial neural networks;

[0081] Unsupervised learning model: When the data has no clear labels, unsupervised learning algorithms are used, and unsupervised learning algorithms include but not limited to clustering algorithms and principal component analysis;

[0082] Reinforcement learning model: suitable for problems that require optimizing decisions through continuous trial and error;

[0083] Model training and optimization, dividing the dataset: Divide the preprocessed data into a training set, a validation set, and a test set; Select an optimization algorithm: According to the selected AI model, select a suitable optimization algorithm to adjust the model's parameters; Model evaluation and improvement: Use evaluation metrics including but not limited to mean squared error, mean absolute error, accuracy, and recall to evaluate the trained model, and adjust and improve the model according to the evaluation results;

[0084] Step 7, intelligent decision-making and optimization: Based on the results of data analysis and mining, combined with industry knowledge and experience, provide intelligent decision-making suggestions to assist in making decisions not limited to production plans, resource allocation, and marketing strategies; Through optimization algorithms, optimize each link in the production process, including production processes, process parameters, and equipment operation; Fault warning and diagnosis; Real-time monitor the equipment status and production process, use models and algorithms to predict and warn of potential faults, and at the same time provide fault diagnosis and repair suggestions to reduce equipment downtime and repair costs;

[0085] Step 8, automated equipment control: Complete equipment motion control according to the decision in Step 7 to achieve automation;

[0086] Step 9, user visual interaction: Provide an intuitive, simple, and easy-to-operate user interface to lower the user's usage threshold, enabling personnel with different professional backgrounds to easily use the system. Through various visualization forms not limited to charts, dashboards, and maps, display in real-time information not limited to production data, equipment status, and operation indicators, facilitating users to monitor and manage, and promptly discover problems and anomalies; Provide intelligent tips and detailed help documents to help users quickly get started and solve problems encountered during use, improving user experience and work efficiency;

[0087] Step 10, online adaptive parameter optimization:

[0088] Objective function definition: According to actual requirements, determine the optimization objective function, and express the objective function as a functional relationship between machine tool parameters and machining quality indicators;

[0089] Application of optimization algorithms: Utilize optimization algorithms including but not limited to genetic algorithms, particle swarm algorithms, and simulated annealing algorithms to perform optimization searches on machine tool parameters based on the model; Search for the optimal solution in a complex solution space, and gradually find the combination of machine tool parameters that optimizes the objective function through continuous iteration and evolution;

[0090] Optimize parameter verification: Experimentally verify the optimal parameters obtained through model optimization in actual machine tool processing. Set the machine tool according to the optimized parameters, conduct machining experiments, and detect machining quality and other relevant indicators;

[0091] Parameter fine-tuning: According to the experimental results, fine-tune the optimized parameters. Due to various uncertain factors in the actual machining process, further adjustment of the parameters is required to ensure the best machining effect in actual production;

[0092] Step Eleven, continuous optimization and iteration:

[0093] Machine tool machining process simulation: Use professional machine tool machining simulation software to establish three-dimensional models of the machine tool, cutting tool, and workpiece, and simulate the machining process under different parameters; Through simulation, various physical quantity changes during the machining process and the quality indicators of the machined parts can be obtained;

[0094] Optimization based on simulation data: Combine the industrial mechanism AI data model, use a large amount of data generated by simulation to train and optimize the model. By simulating and evaluating different parameter combinations in the simulation environment, quickly find a better parameter range, and then further verify and fine-tune through actual experiments, so as to reduce the number of actual experiments and improve the optimization efficiency;

[0095] Then feed various types of data back into the model for continuous optimization and iteration of the model. With the continuous progress of production technology and the accumulation of data, regularly update and optimize the model to ensure that the model always remains in the best state and provides continuous support and optimization for production;

[0096] Step Twelve, collaborative work of production business systems: The manufacturing execution system, warehouse management system, and enterprise resource planning system work in collaboration with the mechanism-AI data dual-driven model system, sharing and updating data in real time, optimizing the production process, and at the same time enabling knowledge transfer and model reuse between different machine tools, reducing data requirements.

[0097] The mechanism parameters in Step One include but are not limited to spindle speed, feed rate, and cutting depth; The state data includes but is not limited to vibration signals, temperature field distribution, power curve, and acoustic emission characteristics; The quality data includes but is not limited to surface roughness, geometric tolerance, and residual stress distribution; The equipment data includes but is not limited to tool wear and guide rail wear status; The environmental parameters include but are not limited to coolant viscosity, environmental temperature and humidity, and workpiece material batch characteristics; The historical experience data includes but is not limited to the experience data and professional knowledge accumulated by the enterprise in long-term practice.

[0098] The fusion rule for the intelligent decision engine mixture in Step Four is θ 融合 = α·θ 机理模型 +(1 - α)·θAI where α ∈ [0, 1], which is dynamically adjusted according to the model confidence.

[0099] Instruction issuance for device automatic control in Step Eight: Update parameters with a 1ms cycle through the EtherCAT bus, perform smooth transition control, and avoid step changes: The actual speed change is the smaller value of the instruction speed change and the constraint condition.

[0100] Δv 实际 = min{Δv 指令 , 0.05v 当前};

[0101] where Δv 实际 is the actual speed change applied to the system, which is the result of imposing constraints on the instruction speed change; Δv 指令 is the instruction speed change, that is, the speed change required by the control system or the user; 0.05v 当前 represents the constraint condition of the speed change rate, which limits the speed change to 5% of the current speed;

[0102] The architecture design of online adaptive optimization in Step Ten is [Edge Sensing Layer]—[GTSN] → [Digital Twin]—[OPC UA] → [Cloud Optimization Engine]. The key technologies adopted are real-time mapping of digital twins. The 1ms-level refresh speed enables the virtual model to quickly reflect the state changes of physical entities; with the help of the Federated Learning framework, dynamic update and optimization of the model can be achieved without revealing data privacy; dynamic constraint processing is based on the Lyapunov stability theory; the control closed-loop is real-time monitoring → anomaly detection → model re-calibration → parameter re-optimization.

[0103] The code for anomaly detection based on dynamic threshold control of the 3σ criterion is as follows:

[0104]

[0105]

[0106] where the EMA function is a custom exponential moving average function used to calculate the exponential moving average of data. It accepts a data array data and a smoothing coefficient β as inputs and returns an array containing the exponential moving average;

[0107] dynami c_threshold function: This function takes a data array data_stream as input. First, it checks if the data array is empty. If it is empty, an exception is thrown. Then, it calculates the mean and standard deviation using the EMA function. Finally, it calculates the lower and upper threshold values respectively and returns them. Dynamic thresholds are used to detect outliers in the data. If a data point is below the lower threshold or above the upper threshold, it can be considered an outlier.

[0108] In step eleven, the data stream period for continuous optimization and iteration is less than 5 ms.

[0109] As Figure 6 shown, the system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model includes a data fusion acquisition and monitoring module, a data synchronization and storage management module, a data analysis and mining module, an intelligent decision-making engine hybrid subsystem, an industrial knowledge base and model library construction module, an industrial knowledge base and model library training module, an intelligent decision-making and optimization module, a device automation control module, a user visualization interaction module, an online adaptive parameter optimization module, a continuous optimization and iteration module, and a business collaboration work subsystem.

[0110] The data fusion acquisition and monitoring module includes a data acquisition subunit, a data preprocessing subunit, a data fusion subunit, and a monitoring subunit.

[0111] The data fusion acquisition and monitoring module is a key part of many complex systems. Its main function is to collect data from different data sources, perform fusion processing, and monitor the entire process and related metrics.

[0112] The data synchronization and storage management module includes a data extraction subunit, a data conversion subunit, a data transmission subunit, a synchronization scheduling subunit, a storage device management subunit, a data storage strategy formulation subunit, a data backup and recovery subunit, and a data access control subunit.

[0113] The data synchronization and storage management module plays a key role in the data processing system. It is responsible for ensuring that data remains consistent between different storage locations or systems and effectively storing and managing the data.

[0114] The data analysis and mining module includes a data exploration subunit, a data preprocessing subunit, a model selection and training subunit, a model evaluation subunit, and a result interpretation and presentation subunit.

[0115] The data analysis and mining module is the core part of processing and understanding data.

[0116] The industrial knowledge base and model base construction module includes a knowledge acquisition subunit, a knowledge representation subunit, a knowledge storage subunit, a knowledge update and maintenance subunit, a model design subunit, and a model storage and management subunit.

[0117] The industrial knowledge base and model base construction module plays a key role in industrial intelligence, providing knowledge support and model basis for industrial production, management, decision-making, etc.

[0118] The industrial knowledge base and model base training module includes a feature engineering subunit, a model training subunit, a model evaluation and optimization subunit, and a knowledge fusion and storage subunit.

[0119] Deeply analyze and model production data, optimize production processes, ensure product quality to improve production efficiency; be able to predict faults based on equipment operation data, optimize maintenance strategies, and enhance equipment management; provide data-driven basis for enterprise decision-making, help cope with market changes; promote knowledge accumulation, inheritance and innovation, stimulate the innovation vitality of enterprises; can also integrate data resources, promote the development of industrial intelligence, improve the overall informatization level of enterprises, and enhance the competitiveness and sustainable development ability of enterprises

[0120] The intelligent decision-making and optimization module includes a decision-making reasoning and generation subunit, an optimization strategy formulation subunit, and a decision-making evaluation and feedback subunit.

[0121] By integrating and preprocessing multi-source data, constructing and training appropriate models, using rule engines and intelligent algorithms to reason and generate decisions, formulating optimization strategies with objective functions and constraint conditions, and at the same time evaluating the decision-making effects in real time and providing feedback for adjustment, provide scientific, accurate and adaptable optimal decision-making solutions for complex problems, and help enterprises and organizations improve efficiency, reduce costs and enhance competitiveness.

[0122] The user visual interaction module includes a data visualization presentation subunit, a user interaction operation subunit, a visualization configuration and customization subunit, and a feedback and prompt subunit.

[0123] Through diverse data visualizations (such as charts, maps, 3D models), provide flexible interaction operations (such as filtering, sorting, querying) and customization functions (interface layout, styles, etc.), and match with a feedback and prompt mechanism (operation prompts, error feedback, data warnings), helping users intuitively understand data, operate and analyze conveniently, quickly obtain valuable information and respond to problems in a timely manner, and improve decision-making efficiency and accuracy.

[0124] The online adaptive parameter optimization module includes a parameter evaluation subunit, an optimization strategy generation subunit, a parameter adjustment and execution subunit, and a knowledge learning and evolution subunit.

[0125] By collecting and preprocessing multi-source data in real time, calculating indicators, predicting and evaluating parameter performance using models, selecting and adapting optimization algorithms to generate strategies, updating parameters and monitoring and adjusting effects, while accumulating experience and knowledge for continuous learning and evolution, dynamically adjusting parameters to optimize system performance, improving production efficiency, reducing costs and ensuring product quality.

[0126] The continuous optimization and iteration module includes a data integration subunit, a data insight subunit, an optimization strategy formulation subunit, an optimization implementation and monitoring subunit, and a knowledge precipitation and inheritance subunit.

[0127] By collecting and integrating data from multiple channels, analyzing deeply to understand the laws, trends and causal relationships, setting goals, generating evaluations and selecting optimization strategies accordingly, promoting the implementation and monitoring of the plan, while precipitating and inheriting knowledge and experience, enabling the system, product or business to continuously adapt to changes, improve performance and achieve continuous evolution.

[0128] The business collaboration work subsystem includes a data integration and sharing subunit, a process collaboration and scheduling subunit, a task assignment and execution subunit, a communication and collaboration subunit, and a performance evaluation and continuous improvement subunit.

[0129] By integrating and sharing data from multiple systems, collaborating and scheduling business processes, reasonably allocating and monitoring task execution, promoting internal communication and collaboration, while evaluating performance to drive continuous improvement, achieving the efficient integration of all production links and business systems, information interconnection and smooth process operation, improving production efficiency, product quality and the comprehensive competitiveness of the enterprise.

[0130] The specific embodiments described in this article are only illustrative of the spirit of the present invention. Those skilled in the art of the present invention can make various modifications or supplements to the described specific embodiments or use similar methods to replace them, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A method for obtaining the optimal parameters of a machine tool based on an AI data model of industrial mechanism, characterized in that, It includes the following steps: Step 1, Multi-dimensional high-frequency data fusion acquisition and monitoring: Deploy a multi-source heterogeneous data synchronization acquisition system and edge computing nodes. Through the multi-source heterogeneous data synchronization acquisition system cooperating with sensors, mechanism parameters, status data, quality data, equipment data, environmental parameters, and historical experience data are monitored and acquired in real time; Step 2, Data synchronization and storage management: Provide efficient and scalable big data storage, supporting the storage and access of massive data; Regularly back up the data to ensure data security and recoverability, and guarantee the stable operation of the system; Set strict data access permissions, and different users are granted different permissions according to their roles and responsibilities to ensure data security and privacy; Step 3, Data analysis and mining: Dynamic feature extraction, deeply analyze production data to discover the laws, trends, and anomalies in the data, providing a basis for decision-making; Deeply mine production data to discover the hidden information, patterns, and knowledge in the data, providing support for prediction and optimization; Based on historical data and current data, build a prediction model to predict and evaluate future production situations, equipment status, product quality, etc., and make preparations and decisions in advance; Step 4, Intelligent decision engine hybrid: Mechanism-AI data dual-drive model collaborative mechanism, mechanism model and AI model, mechanism model: Quickly respond to basic parameter constraints; AI model predicts the optimal parameter offset under dynamic working conditions; Step 5, Construction of industrial knowledge base and model library: Build an industrial knowledge base containing but not limited to industry mechanisms, experience knowledge, standard specifications, and process flows, encapsulate various industrial models, and industrial models include but not limited to equipment fault prediction models, production optimization models, and quality control models for users to call and modify; Provide a platform-based development environment to support users in quickly building and deploying industrial APPs to meet personalized application requirements; Step 6, Training of industrial knowledge base and model library: Supervised learning model: For labeled data, supervised learning algorithms can be selected, and supervised learning algorithms include but not limited to support vector machines, decision trees, random forests, and artificial neural networks; Unsupervised learning model: When the data is not clearly labeled, unsupervised learning algorithms are used, and unsupervised learning algorithms include but not limited to clustering algorithms and principal component analysis; Reinforcement learning model: Applicable to problems that need to optimize decisions through continuous trial and error; Model training and optimization, dividing the dataset: Divide the preprocessed data into training set, validation set, and test set; Select an optimization algorithm: According to the selected AI model, select a suitable optimization algorithm to adjust the model parameters; Model evaluation and improvement: Use evaluation metrics including but not limited to mean square error, mean absolute error, accuracy, and recall to evaluate the trained model, and adjust and improve the model according to the evaluation results; Step 7, Intelligent decision-making and optimization: Based on the results of data analysis and mining, combined with industry knowledge and experience, provide intelligent decision-making suggestions to assist in making decisions not limited to production plans, resource allocation, and marketing strategies; Optimize each link in the production process through an optimization algorithm, including production processes, process parameters, and equipment operation; Fault early warning and diagnosis; Monitor the equipment status and production process in real time, use models and algorithms to predict and warn of potential faults, and at the same time provide fault diagnosis and maintenance suggestions to reduce equipment downtime and maintenance costs; Step eight, equipment automatic control: Complete the equipment action control according to the decision in step seven to achieve automatic operation; Step nine, user visual interaction: Provide an intuitive, concise, and easy-to-operate user interface, lower the user's usage threshold, enable personnel with different professional backgrounds to use the system conveniently, and display in real time, through various visualization forms such as but not limited to charts, dashboards, and maps, information such as production data, equipment status, and operation indicators, to facilitate users' monitoring and management and promptly discover problems and anomalies; Provide intelligent prompts and detailed help documents to help users quickly get started and solve problems encountered during use, improving user experience and work efficiency; Step ten, online adaptive parameter optimization: Objective function definition: According to actual requirements, determine the optimization objective function and express the objective function as a functional relationship between machine tool parameters and machining quality indicators; Optimization algorithm application: Use optimization algorithms including but not limited to genetic algorithms, particle swarm algorithms, and simulated annealing algorithms to perform an optimization search on machine tool parameters based on the model; search for the optimal solution in a complex solution space, and gradually find the combination of machine tool parameters that optimizes the objective function through continuous iteration and evolution; Optimized parameter verification: Experimentally verify the optimal parameters obtained through model optimization in actual machine tool machining, set the machine tool according to the optimized parameters, conduct machining experiments, and detect machining quality and other related indicators; Parameter fine-tuning: Fine-tune the optimized parameters according to the experimental results. Due to various uncertain factors in the actual machining process, further adjustment of the parameters is required to ensure the best machining effect in actual production; Step eleven, continuous optimization and iteration: Machine tool machining process simulation: Use professional machine tool machining simulation software to establish three-dimensional models of the machine tool, tool, and workpiece, and simulate the machining process under different parameters; various physical quantity changes during the machining process and the quality indicators of the machined parts can be obtained through simulation; Optimization based on simulation data: Combine the industrial mechanism AI data model, use a large amount of data generated by simulation to train and optimize the model, quickly find a better parameter range by simulating and evaluating different parameter combinations in the simulation environment, and then further verify and fine-tune through actual experiments, thereby reducing the number of actual experiments and improving the optimization efficiency; Then feedback various types of data into the model for the continuous optimization and iteration of the model. With the continuous progress of production technology and the accumulation of data, regularly update and optimize the model to ensure that the model always maintains the best state and provides continuous support and optimization for production; Step Twelve, Collaborative Work of Production Business Systems: The Manufacturing Execution System, Warehouse Management System, and Enterprise Resource Planning System work in collaboration with the Mechanism-AI Data Dual-Driven Model System, sharing and updating data in real time, optimizing the production process, and enabling knowledge transfer and model reuse between different machine tools, reducing data requirements.

2. The method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 1, wherein The mechanism parameters in Step One include, but are not limited to, spindle speed, feed rate, and cutting depth; the state data includes, but is not limited to, vibration signals, temperature field distribution, power curve, and acoustic emission characteristics; the quality data includes, but is not limited to, surface roughness, geometric tolerance, and residual stress distribution; the equipment data includes, but is not limited to, tool wear and guide rail wear status; the environmental parameters include, but are not limited to, coolant viscosity, ambient temperature and humidity, and workpiece material batch characteristics; the historical experience data includes, but is not limited to, the experience data and professional knowledge accumulated by the enterprise in long-term practice.

3. A method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 2, characterized in that, The fusion rule mixed by the intelligent decision-making engine in the fourth step is θ 融合 = α·θ 机理模型 +(1 - α)·θ AI ; where α ∈ [0, 1], and it is dynamically adjusted according to the model confidence level.

4. A method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 3, characterized in that, Issuance of Instructions for Equipment Automation Control in Step Eight: Update parameters at a cycle of 1ms through the EtherCAT bus, perform smooth transition control, and avoid step changes: The actual speed change is the smaller value between the commanded speed change and the constraint conditions. Δv 实际 = min{Δv 指令 , 0.05v 当前}; where, Δv 实际 is the actual speed change applied to the system, which is the result of imposing constraints on the commanded speed change; Δv 指令 is the commanded speed change, i.e., the speed change required by the control system or the user; 0.05v 当前 represents the constraint condition of the speed change rate, which limits the speed change to 5% of the current speed.

5. The method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 4, wherein The architecture design of online adaptive optimization in Step Ten is [Edge Sensing Layer]—[GTSN]→[Digital Twin]—[OPCUA]→[Cloud Optimization Engine]. The key technologies adopted are real-time mapping of digital twins, and the refresh speed at the 1ms level enables the virtual model to quickly reflect the state changes of the physical entity; with the help of the Federated Learning framework, dynamic update and optimization of the model can be achieved without revealing data privacy; dynamic constraint processing is based on the Lyapunov stability theory; the control closed-loop is real-time monitoring → anomaly detection → model recalibration → parameter re-optimization.

6. The method for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 5, characterized in that, The data flow period for continuous optimization and iteration in Step Eleven is less than 5ms.

7. A system for obtaining optimal parameters of a machine tool based on an industrial mechanism AI data model according to any one of claims 1-6, characterized in that, It includes a data fusion acquisition and monitoring module, a data synchronization and storage management module, a data analysis and mining module, a hybrid subsystem of intelligent decision-making engines, an industrial knowledge base and model base construction module, an industrial knowledge base and model base training module, an intelligent decision-making and optimization module, an equipment automation control module, a user visualization interaction module, an online adaptive parameter optimization module, a continuous optimization and iteration module, and a business collaboration work subsystem.

8. A system for obtaining optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 7, characterized in that, The data fusion acquisition and monitoring module includes a data acquisition subunit, a data preprocessing subunit, a data fusion subunit, and a monitoring subunit. The data synchronization and storage management module includes a data extraction subunit, a data transformation subunit, a data transmission subunit, a synchronization scheduling subunit, a storage device management subunit, a data storage strategy formulation subunit, a data backup and recovery subunit, and a data access control subunit. The data analysis and mining module includes a data exploration subunit, a data preprocessing subunit, a model selection and training subunit, a model evaluation subunit, and a result interpretation and presentation subunit.

9. The system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 8, characterized in that, The industrial knowledge base and model base construction module includes a knowledge acquisition subunit, a knowledge representation subunit, a knowledge storage subunit, a knowledge update and maintenance subunit, a model design subunit, and a model storage and management subunit. The industrial knowledge base and model library training module includes a feature engineering subunit, a model training subunit, a model evaluation and optimization subunit, and a knowledge fusion and storage subunit; The intelligent decision-making and optimization module includes a decision reasoning and generation subunit, an optimization strategy formulation subunit, and a decision evaluation and feedback subunit.

10. A system for obtaining the optimal parameters of a machine tool based on an industrial mechanism AI data model according to claim 9, characterized in that, The user visualization interaction module includes a data visualization presentation subunit, a user interaction operation subunit, a visualization configuration and customization subunit, and a feedback and prompt subunit; The online adaptive parameter optimization module includes a parameter evaluation subunit, an optimization strategy generation subunit, a parameter adjustment and execution subunit, and a knowledge learning and evolution subunit; The continuous optimization and iteration module includes a data integration subunit, a data insight subunit, an optimization strategy formulation subunit, an optimization implementation and monitoring subunit, and a knowledge precipitation and inheritance subunit; The business collaborative work subsystem includes a data integration and sharing subunit, a process collaboration and scheduling subunit, a task assignment and execution subunit, a communication and collaboration subunit, and a performance evaluation and continuous improvement subunit.

Images