Support Vector Fuzzy Adaptive Network in the Modeling of Material Removal Rate in Rotary Ultrasonic Machining

Abstract

Rotary ultrasonic machining (RUM) is one of the cost-effective machining methods for machining difficult to process material. It is a hybrid machining process that combines the material removal mechanisms of diamond grinding with ultrasonic machining. However, due to the lack of understanding of the mechanisms of these operations, models for these machining processes are difficult to establish. In this paper, the support vector fuzzy adaptive network (SVFAN), a parameter free nonlinear regression technique, is used to model the material removal rate in RUM. The SVFAN retains the advantages of both the fuzzy adaptive networks and the support vector machines. The former possesses the linguistic representation ability and the latter is a very effective learning machine. The results are compared with that obtained by the use of fuzzy adaptive network and it is shown that the combined approach is a more effective algorithm for the modeling of complex manufacturing processes.