Abstract
This study analyzed the workpiece surface quality (Ra) and the material removal rate (MRR) on process parameters during machining SKD11 by medium-speed wire electrical discharge machining (MS-WEDM). An experimental plan for composite design (CCD) has been conducted according to methods response surface methodology (RSM) and subsequently to seek the optimal parameters. The experimental data were utilized to model MRR and Ra under optimal parameter condition by a backpropagation neural network combined with genetic algorithm (BPNN-GA) method. Eventually, the comparisons between the results from BPNN-GA and those from the RSM demonstrate that BPNN-GA method is a more effective way for optimizing MS-WEDM process parameters.
ACKNOWLEDGEMENT
This research was supported by the National Natural Science Foundation of China (NSFC) under Grant No. 51175207 and Grant No. 51121002, the National Key Technology R&D Program No. 2012BAF13B07, and the Science and Technology Planning Project of Guangdong Province No. 2012B011300015. Authors Guojun Zhang and Zhen Zhang contributed equally to this work.
Notes
Ra: Surface roughness (Ra), MRR: Material removal rate.
DOF: Degrees of freedom, Seq SS: sequential sum of squares, Adj SS: adjusted sum of squares, Adj MS: adjusted mean squares.
SHL network: single-hidden-layer network, DHL network: double-hidden-layer network, No.of 1st: No.of neurons in 1st hidden layer, No.of 2nd: No.of neurons in 2nd hidden layer.