Study For Intelligent Prediction System Of Cylindrical Longitudinal Grinding

Intelligent manufacture (IM) is a result for combination of the artificial intelligence (AI) technique and manufacturing technique, and its purpose is to realize automatization and intelligence during the whole manufacturing process. Its target is to replace of parts of person's intelligence mental work, and the ultimate target is to realize the overall intelligence of the organizational skill and manufacturing environment of the manufacturing process. The customers request the performance of the product perfect and dependable, and making manufacture qualified once possibly. Therefore, it requests not only to make the system having the higher production efficiency, much lower original material and energies to consume, but also requests the components to have the higher accuracy, roughness of surface, integrity of surface and strict manufacturing consistency. Grinding technique usually is the last process of the mechanical processing, and the working quality affects the service result of the workpiece directly. The main parameters which characterize the quality of the workpiece includes roughness of surface, surface burn, roundness and the value of deforming of the workpiece etc. . But the grinding process is very complicated, there is a lot of influencing factors, and the relation of the factors are also complicated, so it's very hard to establish mathematics model exactly, the selection of grinding parameters influence the finish quantity of workpiece directly. In the reality grinding process, it depends on a great deal of the operators' experience and experience knowledge mostly, Therefore the existing state of grinding technique has become one of the key techniques which hindrance the development of some advanced manufacturing techniques. The inexorable trend of the development of the grinding technique is to make it intelligent.Currently intelligent prediction system has already played an important role in many fields of modern science technique. As precondition and foundation of decision, the intelligent prediction system is very important to the ultimate blue print. This paper established Intelligent Prediction System of Cylindrical Longitudinal Grinding based on the research of the precision finishing method -cylindrical longitudinal grinding which widely applied by the mechanical manufacturer, and it was the first step for further studying for the Intelligent Prediction Control System of Cylindrical Longitudinal Grinding.Intelligent prediction system of cylindrical longitudinal grinding basing on artificial intelligence built the predictive model of surface roughness and workpiece size to predict these parameters, which are difficulty to detect. Intelligent prediction system of cylindrical longitudinal grinding is mainly composed by the man-machine interface,knowledge base,surface roughness prediction module and workpiece size prediction module.1. Man-machine interfaceThe program of the man-machine interface is based on Visual C++ 6.0--the visual development environment and the object-oriented programming method. The system has friendly man-machine interface, and it made operation of customer convenient and flexible.2. Knowledge baseThe knowledge base of this system gathers the rule database and the model database as a whole, making expression of knowledge more generalization. The rule database stores the knowledge based on the rule expression and the rule that to hunt the knowledge information; The model database stores surface roughness model and workpiece dimension model, all these models above were verified through experiment.3. Surface roughness prediction moduleThe surface roughness prediction module predicts surface roughness by building intelligent predictive model based on FBFN intelligently, and solves the difficulty of on-line detection. The predictive model is based on FBFN using ALS algorithm to optimize.4. Workpiece size prediction moduleThe workpiece size prediction module predicts workpiece size by building intelligent predictive model based on the dynamic Elman neural network intelligently. It solves the problem which the size detection position lags the grinding position.Our study for intelligent prediction system of cylindrical longitudinal grinding is based on a renovated MMB1320 precise semiautomatic cylindrical grinding machine. Visual C++,MATLAB and SQL Server is used to realize the software, we use Visual C++ to program the interface, MATLAB to finish building,training and simulating the predictive model behind the scenes,SQL Server to maintain and manage the database, they connect with program interface. This system is practical, convenient and has good interaction and expansion, so it has some applied foreground.