Research On Autonomous Sensing Method Of Intelligent Machining Machine Based On Multi-Mode Fusion And Deep Learning Of Multi-Physics Information

With the development of computer science and advanced manufacturing technology and the pursuit of people’s quality of life.People put forward higher requirements on the appearance and quality of the products.Traditional manufacturing began to be challenged.A new manufacturing method--intelligent manufacturing began to develop.Intelligent manufacturing technology makes the system highly integrated,it also makes it difficult for managers to predict the extent to which their decisions will affect future system performance.Therefore,the intelligent processing machine with the basic processing ability of the processing machine,also need to have the ability to self perception.Not only to their own parameters,the state of perception,but also perceive the external environment.Therefore,the intelligent machine has the ability to collect,process and fuse the multi-physical domain information of itself and the outside world,and the ability of using this information to realize autonomous awareness(state identification,fault diagnosis and health assessment).Intelligent processing machine is also a microcosm of intelligent manufacturing system(IMS),and the realization of intelligent manufacturing system,will be reflected in the intelligent processing machine as the core of the intelligent unit.Therefore,the level of intelligence of an intelligent manufacturing system is determined by an intelligent processing machine.The degree of intelligence of intelligent machining machines is mainly determined by their autonomous perception ability.So,in order to effectively solve the problem of autonomous perception of intelligent machines,a method for autonomous perception of intelligent machining machines based on multi physical domain information,multi-mode fusion and depth learning is proposed in this paper.The information systems of intelligent processing machines and manufacturing environments are coupled by multiple physical fields.Therefore,the intelligent machining machine and its manufacturing environment for multi physical domain information and multi-mode fusion information are used in this paper,refining reliable fusion features is provided for autonomous sensing,and in this paper,it is applied to the automatic perception of spindle motors in intelligent machining machines.This paper mainly studies the content: First of all,by comparing the advantages and disadvantages of various information fusion technologies at home and abroad,a modified GMM-HMM and with stochastic time-series modeling capability and decision fusion capability is proposed.The multi physical domain information fusion of intelligent machining machine and its manufacturing environment is realized.Then,compare the traditional machine learning method with the current deep learning algorithm,a deep learning network model to support the autonomic perception of intelligent processing machines based LSTM is proposed.Combining deep situational awareness with the advantages of this advanced approach to perception,an intelligent processing machine autonomic perception model based on depth learning and depth situation awareness is proposed.Finally,the intelligent processing machine self sensing method is applied to machine spindle motor simulation test system.Data of multiple physical domains such as vibration and speed of motor are collected,to validate the effectiveness of multimodal fusion based on GMM-HMM and D-S for multi physical domain information.The fused feature information is used as the input of the autonomous perception model based on deep learning and deep situational awareness,to verify the accuracy and reliability of the model for autonomous sensing of spindle motors.