Smart Machining of Titanium Alloy Using ANN Encompassed Prediction Model and GA Optimization

Titanium alloys are substantially used in the aerospace industry due to their high strength to weight ratio. During turning of titanium alloys, obtaining smooth surface with dimensional accuracy is a requirement in machining of parts. The quality of machining is a function of machining parameters used in turning. Twenty-seven experiments were conducted with different combination of cutting speed(Vc), Feed rate (f) and Depth of cut (ap) and the corresponding surface roughness (Ra) was measured. Response surface methodology (RSM) was firstly developed with the experimental data and Genetic algorithm(GA) was then applied to minimize the surface roughness. Validation experiments showed that the error rate was 4.27%. In order to minimize the error rate, a systematic approach was implemented to develop optimal Artificial neural network (ANN) model with the contemplation of effect of ANN architecture and activation functions on the prediction. The predicted ANN data were further used to develop revised RSM model. The later prediction model and the respective optimization resulted the best cutting parameters for achieving the minimum Ra. The predicted surface roughness from GA is \( R_{a} = 1.42 \mu {\text{m}} \) for the optimum turning condition of \( V_{c} = 148\,{\text{m/minute}},\, f = 0.1\,{\text{mm/rev}}, ap = 0.5 \,{\text{mm}} \). The error rate between the predicted results and the validation results is only 2.27%. The proposed model can be used practically in the turning of titanium alloys.