Research On Surface Quality Of Micro Turn-milling

With the increasing demand for micro parts in fields such as civil market and national defense, much attention has been given to the issues involving in the design and processing. Micro machining has become one of the most important methods in microminiature parts processing because it has many advantages such as the low cost of processing, high machining precision and high machining efficiency, strong three-dimensional processing capacity. The micro Turn-milling which is as a kind of micro cutting and compared with micro-turning, shows that it has higher machining efficiency and better surface quality. At present,it is very difficult for micro-part to measure it’s roughness and residual stresses. Therefore, study on the roughness and residual stress of micro turn-milled surface is significant for production and scientific test.Firstly.by analyzing the literature of micro milling and combining the micro milling characteristics, the main factors influencing of roughness is determined. The orthogonal experiment was implemented by selecting the orthogonal test table and the level of each factor according to the actual situation. The formation mechanism of the surface morphology are analyzed through the observation on surface contourgraph after the test. The factors that affect the surface roughness are analyzed by range analysis, on this basis, the analysis of variance is used to determine significance of each factor.Secondly, in view of the difficulty of measuring the roughness of workpiece surface which processed by micro turn-milling, Prediction method based on BP neural network is proposed to solve this problem. The BP neural network was improved because of its shortcomings. Number of nodes in the input layer and output layer are determined according to the number of factors and the index. The number of hidden layer nodes is selected through empirical formula preliminarily, finally generalization ability test for the selected number of nodes is used to determine the number of nodes in the hidden layer. MATLAB software programming is used to implement the BP algorithm and the corresponding improvement measures. The created BP network is trained and simulated, and the results shows that roughness model of micro turn-milling established by BP neural network has high prediction accuracy.Finally, Aiming at the difficulty of measuring residual stress of microminiature parts, micro milling cutter and workpiece are set up by Solid Works, The established models are analyzed by the finite element software ABAQUS in order to estimate the influence rule of residual stress in different levels.