Research And Application Of Response Mechanism Of Digital Twin Machining System

Improving the stability of machining process is an important way to ensure the quality of machining.The digital twin technology of constructing virtual model of machining process can realize the real-time monitoring.However,the perception and response mechanism of the existing digital twin technology has the problems of poor perception robustness and low response efficiency.In this regard,combining the advantages of wide range,strong scalability and high response efficiency of biological perception,a responsive digital twin system integrating biological perception network is proposed,and improves the performance of digital twin system by means of biological cognition and innate and acquired response mechanisms.The main work of this paper is as follows:Firstly,aiming at the problems of weak ability of multi-source data perception and insufficient scalability of the existing digital twin machining system,a machining system perception module integrating biological perception and cognitive ability is proposed.The proposed perception module includes two parts: perception and cognition.Firstly,the perception module completes the primary evolution of the model based on dimensional topology modeling and graph information interaction and outputs the primary perception.Then it uses different types of primary perception to construct the goal-based topology and completes the advanced evolution of the model by information interaction and outputs the advanced perception.On the basis of advanced perception,the cognitive generation mechanism based on hidden Markov network is used to complete the cognitive coding of advanced perception sequence.The perceptual module drives the innate and acquired response mechanisms of the system through the perceptual and cognitive information.The perceptual comparison experiment results show that the perceptual module based on model evolution and information recognition can significantly improve the speed and accuracy of new task expansion compared with other models such as convolutional neural network.Secondly,aiming at the problems of chaotic response mechanism and low efficiency of the existing digital twin machining system,a machining system response module integrating biological sequential collaborative response mechanism was proposed.Based on the characteristics of biological innate response mechanism such as clear and fast,the digital twin innate response mechanism based on rule base is proposed.The innate response mechanism is based on the expert knowledge base,supplemented by the coding results of the biological habit-forgetting mechanism,and takes advantage of the low dimension and strong interpretability of the rule base to fully integrate the artificial professional knowledge and experience,so as to provide simple and accurate feedback.Based on the characteristics of high precision and wide range of biological cognitive response,an acquired response mechanism of digital twin based on state transition was proposed.The acquired response mechanism,driven by the cognitive coding results,searches for similar information in the policy library based on the neighbor neural network and modifiers it to complete the recommendation task.The experiments of system response speed under various models show that the innate and acquired response mechanisms are more efficient and the hierarchical response is used to enhance the response ability.At last,based on B/S architecture,the responsive digital twin system is constructed,and the milling experiment and system simulation are used to verify the above capabilities.The biological perception-response mechanism can effectively improve the perception ability and response efficiency of the digital twin system,which has certain value.