Improved GEP Based Physical Modeling And Process Parameters Optimization Methods For CNC Milling Process

With the development of CNC machining technology, how to model CNC machining process and optimize process parameters becomes the key factor affecting the CNC machining capacity. In traditional CNC machining, because of lacking modeling of machining process, the understanding and analysis of machining mechanism are inadequate; the process parameters are obtained mainly through experience, thus impeding the improvement of machining efficiency and quality. Therefore, this paper focuses on the physical modeling of CNC milling process and optimization of process parameters.Physical modeling of machining process is very important for understanding and analyzing machining mechanism, and is the basis for optimization of machining parameters. However, highly nonlinear, time-varying and uncertainty of the machining process makes it very difficult to construct explicit mathematical expressions of the machining process accurately. Among conventional methods, Artificial Neural Networks can only achieve implicit model while empirical formula and response surface methodology can obtain an explicit model whose accuracy is not very high. To construct an accurate, explicit mathematical model of the machining process of CNC milling, this paper proposes Greedy Randomized Adaptive Search Procedure (GRASP) based Gene Expression Programming (GGEP). The basic GEP evolves without a guiding direction and often converges prematurely. To overcome these disadvantages, the idea of greedy and iterative restart of GRASP is introduced into GEP. Experimental results show that GGEP method has a good performance in convergence efficiency, success rate of calculation, and overcoming premature convergence.Cutting force is the key factor that affects the machining process of CNC milling. This paper studies cutting force modeling techniques. A GGEP based cutting force modeling and forecasting method has been proposed, which can efficiently construct high-accuracy, explicit cutting force model. At the mean time, single factor analysis, analysis of variance (ANOVA) and Taguchi method have been used to analyze the influencing factors of cutting force. Experimental results show that the maturity and matching degree of model proposed in this paper is very high. Results of analysis of the influencing factors show that the feed rate has the biggest contribution on the cutting force and they are positively correlated. To minimize the cutting force, the optimal machining parameters are increased feed rate, reduced inclination and cutting speed.Surface roughness is an important indicator to measure the performance of CNC milling process and quality of the workpiece, and it plays a key role in wear resistance and fatigue resistance of the workpiece. This paper studies surface roughness modeling techniques. A GGEP based surface roughness modeling and forecasting method have been proposed, which can efficiently grub high-accuracy, explicit surface roughness model. At the mean time, single factor analysis, ANOVA and Taguchi method have been used to analyze the influencing factors of surface roughness. Experimental results show that the maturity and matching degree of surface roughness model constructed in this paper is very high. Results of analysis of the influencing factors show that the feed rate has the biggest contribution on the surface roughness and they are positively correlated. To minimize the surface roughness, the optimal machining parameters are increased spindle speed, reduced feed rate and proper depth of cut.With the rise of low carbon manufacturing, reducing energy consumption becomes one of hot issues for the CNC machining process. This paper studies energy consumption modeling techniques. A GGEP based energy consumption modeling and forecasting method have been proposed, which can efficiently grub high-accuracy, explicit energy consumption model. At the mean time, single factor analysis, ANOVA and Taguchi method have been used to analyze the influencing factors of energy consumption. Experimental results show that the maturity and matching degree of energy consumption model constructed in this paper is very high. Results of analysis of the influencing factors show that cutting speed has the biggest contribution on the energy consumption and they are positively correlated. To minimize the energy consumption, the optimal machining parameters are reduced cutting speed, feed rate and depth of cut.Optimization of machining parameters has a significant impact on machining quality of the workpiece. Improper selection of machining parameters will impede the CNC machine to exert its processing capacity. This paper proposes models of machining parameters optimizing problem, based on the models of cutting force, surface roughness and energy consumption. Combined with chaotic sequence and constraint handling mechanism, Chaotic Imperialist Competitive Algorithm (CICA) is proposed to solve the optimization of machining parameters. Through cutting strategy formulation, single-pass parameters optimization and multi-pass parameters optimization, the optimization of multi-pass face milling has been solved. The results show that the proposed algorithm can effectively solve the problem of optimization of process parameters.Based on the actual processing of a CNC milling machine, the above theories are applied to the physical modeling and optimization of process parameters of practical CNC milling process. Feasible and effectiveness of the proposed method are further validated.Finally, a summary of the full text is given, and the future work is expected.