Modeling Of Cutting Forces And Surface Roughness, And Parameter Optimization During High-Speed Machining Titanium Alloy

During high-speed machinig titanium alloy, high temperature near the cutting edge usually leads to the rapid wear of cutting tool, which result in the lower machining efficiency, higher machining costs and poor machining quality. Therefore, investigation on cutting mechanism of titanium alloy and optimizing the cutting parameters are essential to increase machining efficiency and to control the cutting process. Titanium alloy Ti-6A1-4V was selected as the workpiece material. Based on the combination of experimental research, neural network and optimization technique, the optimal cutting parameters were searched to satisfy the demands on machinng efficiency, quality and costs when high-speed machining titanium alloy.1) Experimental research and empirical models for cutting forces and surface roughnees when high-speed machining titanium alloy. Based on orthogonal experiments, Ti-6A1-4V was high-speed milled with the carbide inserts. The effects of cutting parameters (axial depth of cut, radial depth of cut, cutting speed, and feed per tooth) on cutting forces and surface roughness were investigated by means of range analysis. The empirical models for cutting forces and surface roughnees were established with the multiple linear regression and polynomial regression, which were experimentally verified by the validation tests.2) Predictive mnodes for cutting forces and surface roughness based on neutral network. BP (Back Propagation) neutral network, which consists of a number of processing unit, has a power capacity of parallel computing, studying, fault-tolerant and anti-interference. BP neutral network trains the predictive model by means of the inputting variables and ouput results. In this research, error back-propagation algorithm was employed to train the predictive model with the aid of adaptive the learning rate method and the additional momentum, which could overcome the slower convergence rate and easily trapping in the partial small inadequate of the traditional BP network. The validation tests indicated that the predictive models based on BP neutral network can accurately predict cutting forces and surface roughness during high-speed cutting Ti-6Al-4V.3) Cutting parameters optimization based on given material removal rate. When establishing the optimization model, the material removal rate was set as one constrain. That is to say, the optimal combination of axial depth of cut, radial depth of cut, cutting speed, and feed per tooth for the minF(x) or minRα(x) was searched with the given constrain of metrail removal rate. Finally, the optimal cutting parameters and corresponding results (cutting forces and sucfaee roughnes) were verified by validation tests.This research was supported by The Specialized Research Fund for the Doctoral Program of Higher Education (Grant No.20070422033), The Scientific Research Foundation for the Outstanding Young Scientist of Shandong Province (Grant No.2007BS05001). and The Important National Science & Technology Specific Projects (Grant No.2009ZX0400-032).