Evaluation of Surface Roughness and Material Removal Rate in CWEDM Using DOE and ANN

In this work, an experimental study in cylindrical wire EDM machining parameters followed by artificial neural network is presented. Experiments have been done by means of the technique of design of experiments (DOE) with the three levels fractional factorial method. An Artificial neural network (ANN) has been used for predicting machining parameters, using the experimental parameters as input and results as target. It is shown that DOE combined with ANN can be very useful in experimental applications, especially in the field of manufacturing which input parameters are too many to design full-factorial experiments. The superiority of this approach is highly noticeable when there is only experimental data which demonstrate the process behavior, and little or no explicit mathematical relationships, based on the physics of the process, are available to correlate the input and output parameters.

Keywords:

Cylindrical Wire Electro Discharge Machining , Design of Experiments , Artificial Neural Network