Radial Basis Function Neural Network for modelling an Electrical Discharge Machining drilling process

Experimental research for prediction of Material Removal Rate in an Electrical Discharge Machining drilling process is presented in this study. The process was performed in High Strength Steel which is extensively used in automotive industry for truck chassis manufacturing. The prediction model was computed by a Radial Basis Function Neural Network. With the aim to increase the prediction capability of the Neural Network, a Genetic Algorithm was used to tune its weights. An experimental series were carried out to get raw data and explore the influence of the machining parameters, namely: gap voltage, machining current and pulse frequency, as well as electrode configuration (single or multi-channel tubular). Experimental results reveal that the modified Neural Network, satisfactorily predicts the Material Removal Rate for each experimental condition.