Research On The Key Technologies Of Intelligent Generation Of CNC Process Knowledge

Faced with the trend that market demands individualized,small batch and short cycle manufacturing,manufacturing enterprises adopt advanced and intelligent manufacturing modes to be responsive to the dynamic market demands.As the basic component of the intelligent manufacturing modes,Computerized Numerical Control(CNC)machine tools with the intelligent process planning function is one of the key points of intelligent CNC equipment research.CNC machining process knowledge is the basis for CNC machine tools realizing intelligent process planning function.Due to its diversity and complexity,the collection and management of process knowledge is difficult,which limits the implementation of various intelligent process planning methods.In recent years,due to the improvement of computing power,machine-learning technology has developed rapidly.One of the key models,generative model,can learn the distribution and generate similar samples by observing existing samples,which provides a new way to solve the difficulty of acquiring process knowledge.Therefore,based on processing data,using generative models,knowledge engineering and other related technologies,this paper studies the key technologies of intelligent generation of CNC machining processing knowledge.With the intelligent generation system,process of knowledge accumulation can be accelerated to support the knowledge base,which will be used for intelligent process decision.In this paper,facing the requirements of process knowledge generation,the overall schema of the intelligent generation system of process knowledge is designed.The architecture of intelligent generation system with knowledge representation,knowledge generation,knowledge evaluation and knowledge management functions is established.Then,the representation methods of CNC process knowledge is studied.Aiming at the application scenarios and the characteristics of instantial process knowledge,generative process knowledge and evaluative process knowledge,the representation models of CNC process knowledge based on STEP paradigm,production rules and neural network is constructed.After that the intelligent generation method of CNC process knowledge is studied.The generation of processing route is based on decision tree and production rules,and the intelligent generation of processing parameters is realized based on Generative Adversarial Networks(GAN).Moreover,a multi-channel and multi-level comprehensive fuzzy evaluation method is proposed to evaluate generated knowledge,which is consisted of neural network pre-evaluation,qualitative expert evaluation and quantitative use evaluation.In addition,the application strategy in the intelligent CNC system with cloud process knowledge base is given to solve the problem of process knowledge application.Finally,the implementation of the generation system is expounded.The generation results show that the generation system captures the distribution of CNC process knowledge well and increases the diversity of process knowledge,which confirms the feasibility and effectiveness of the method.