Application of CMAC Neural Network in Noncircular Turning Process
Based on the characteristic analysis of mid-convex and varying ellipse pistons, a scheme of a fast tool servo (FTS) system is brought forward, which is driven by a piezoelectric actuator. Both the dynamic hybrid model for hysteresis compensation of piezoelectric actuator and the noncircular turning process model for cutting force control are expatiated in use for the real-time and high-precision tracking control of FTS. Built on these models, two control loops are devised using cerebellar model articulation controller (CMAC) neural networks. Integrated with the Preisach concept, the inner CMAC loop is designed to learn on-line and compensate for the nonlinear hysteresis and frequency dependent effect of a piezoelectric actuator. Combined a CMAC identifier with a CMAC controller, the outer loop is constructed to perform the adaptive inverse control of cutting force in the noncircular turning process. Experimental results show that the proposed control policy is effective.
Shuyan Yang Haifeng Wang
School of Mechanical Engineering Qingdao Technological University Qingdao, 266033 School of Mechanical and Power Engineering Shanghai Jiao Tong University Shanghai, 200030
国际会议
青岛
英文
2006-07-21（万方平台首次上网日期，不代表论文的发表时间）