Research And Application Of Key Technology Of Digital Twin For High-end CNC Machine Tools

The flourishing development of emerging information technology and its deep integration with the advanced manufacturing industry are accelerating the transformation of the global manufacturing industry towards intelligence,informatization,and servitization.The development level of intelligent manufacturing represents the global status of the national manufacturing industry.As the cornerstone of the development of intelligent manufacturing,high-end CNC machine tools with the ability of high-speed,high-precision,intelligence,combined machining,multi-axis linkage,and network communication,account for a critical role in the high-end equipment manufacturing industry in China and represent the core competitiveness of national advanced manufacturing industry.In this context,high-end machine tools will face the new requirement of intelligent transformation and upgrading,from passive execution of instructions to the autonomous agent with self-perception and external cooperation.The improvement of the intelligence level of high-end CNC machine tools is a significant driving force in promoting the implementation of the intelligent manufacturing practice.With the ability of description,diagnosis,prediction,and decision of physical entities,digital twin technology acts as a significant means to promote the intelligent transformation of manufacturing industry,which can realize the interactive integration of physical space and information space and provide a new idea for the direction of intelligent transformation of high-end CNC machine tools.Based on the current research work both at home and abroad,this thesis makes a thorough analysis of the new requirements for the intelligent transformation of high-end CNC machine tools and carries out research work around the key technologies of digital twin for high-end CNC machine tools,such as data-aware method,knowledge-driven fault diagnosis method and production scheduling decision method.The main research content of this thesis is as follows.(1)In view of the intelligent demand of high-end CNC machine tools,the typical characteristics of digital twin are deeply analyzed from the system perspective and the implementation perspective,and the architecture of digital twin for high-end CNC machine tools is proposed,summarizing and analyzing the data-aware method,knowledge-driven fault diagnosis method and production scheduling decision method.Finally,the construction process of geometry model,behavior model,and knowledge model of digital twin for high-end CNC machine tools is studied.(2)A data-aware architecture based on OPC UA is proposed,including the device layer,the network layer,and the application layer,to meet the demand of digital twin of high-end CNC machine tools for sensing data.The standardization of real-time data acquisition of high-end CNC machine tools is realized by constructing OPC UA server at the edge gateway.For the problem of information island caused by data structure and semantic heterogeneity among different CNC systems,sensors,and other data-aware devices,a unified object dictionary of CNC machine tools is established and mapped to the address space of OPC UA server to realize interconnection and interoperability of different CNC machine tools at the semantic level,which provides data support for the construction of digital twin and intelligent services.Finally,the feasibility and effectiveness of the proposed method are verified by the application.(3)A fault diagnosis method for high-end CNC machine tool based on convolutional neural network algorithm and CNC machine tool fault knowledge graph is proposed to solve the difficulties in fault diagnosis of high-end machine tools and effective integration and utilization of fault maintenance experience and knowledge.Firstly,the convolutional neural network algorithm is trained with the historical fault data of high-end CNC machine tools to obtain the fault classification model used to fit the real-time fault data of machine tools to realize the intelligent diagnosis of faults.Then,the scattered fault maintenance knowledge in the enterprise is integrated to build a fault knowledge graph of machine tools,which is used to retrieve and reason about the diagnosed machine tools fault information to obtain the maintenance knowledge of corresponding faults,improving the fault maintenance efficiency.Experiments demonstrate the effectiveness of the proposed method.(4)For the problem of digital twin production rescheduling decision of high-end CNC machine tools,a production rescheduling decision method based on Monte Carlo Tree Search algorithm(MCTS)is proposed for intelligent fault maintenance and production rescheduling service.In response to the invalidation of initial scheduling plan due to the CNC machine tools failures in the dynamic production environment,a production rescheduling strategy based on MCTS algorithm is proposed,which adaptively adjusts the production plan based on the current fault information to reduce the loss caused by failures.Based on the characteristics of rescheduling problem,the algorithm establishes a production rescheduling model with the makespan and stability as the optimization objectives,and proposes three rescheduling rules for constructing the legal behavior space and the scheduling state space of the MCTS algorithm in the rescheduling process,and designs an evaluation function to train the reinforcement model based on the gradient strategy.The experimental results show that the proposed method improves the effectiveness and efficiency of the rescheduling strategy.(5)A multi-objective optimal scheduling decision method based on an improved second generation non-dominant sorting genetic algorithm(NSGA-II)is proposed for the problem of digital twin multi-objective optimal scheduling decision of high-end CNC machine tools,which integrates the predictive maintenance and production scheduling problems of machine tools,enabling the CNC machine tools to dynamically adjust the scheduling plan according to the prediction results and production objectives.An improved NSGA-II algorithm is proposed to combine tool prediction maintenance and production scheduling decision,establishing a mathematical model of the production scheduling decision problem based on the remaining useful life constraint of tool with the optimization objectives of minimizing makespan and total tardiness and the improvement of chromosome encoding and decoding method,the population initialization strategy and the evolutionary operator adopted in the NSGA-II algorithm.Finally,it is verified that the proposed method improves the quality and efficiency of digital twin optimal scheduling decision of high-end CNC machine tools through comparative experiments.