Basic Study On Quality Controlling Of ECABG

With the development of modern science and technology, it becomes more and more difficult to simply rely on mechanical technology to meet the performance requirements. So the composite processing technology which sets mechanical, physical, chemical technology role in the integration has been developing rapidly. Electrochemical abrasive belt grinding (ECABG) technology is one of the Electrochemical-Mechanical composite processing technology. It makes a good match to choose reasonable abrasive belt grinding manner and technologic parameters, and processes the workpiece surface alternately, then meets the quality requirements.ECABG is a nonlinear process which many factors affect the manufacturing process. It is necessary to choose and optimize the reasonable technologic parameters so as to obtain good quality, that According to the different parts (shape, material, processing technology, etc.) to determine the best technologic parameters to control the machining process and the required processing quality.On the basis of relevant ECABG theory, taking bearing raceway crown processing as an object to study the Quality control of processing is put forward as following:1. The basic principles and characteristics of ECABG theory have been deeply studied. The appropriate experiment equipment was set up, and the manufacturing process and the ways to improve processing efficiency were analyzed.2. The influence of technologic parameters, such as electric current density, processing time, granularity of abrasive belt, gap between cathode and anode, processing lineal speed, has been researched primarily by orthogonal experiments, which provide the foundation for optimization of technologic parameters.3. A similar mathematical model between the machining precision and influencing factors is set up with the BP neural network theory, thus avoids the complex interrelationship between technologic parameters and processing quality. On the basis of BP neural network, the technologic parameters and the dimension of the middle and the hidden layer are determined. Taking predicting and controlling the surface roughness as an example, the model trained by the experimental datum turns into the final BP neural network model. The experimental results indicate that this BP neural network model forecast has a high quality accuracy. In addition, The BP neural network model can be used to optimize technologic parameters.