An approach to intelligent machining in face milling

Quality control calls for building up an Intelligent Machining (IM) for parallel acquisition of performance parameters and deducing the process/tool status by aptly fusing the acquired data. This paper presents a methodology to implement IM in Face Milling operation, by acquiring Acoustic Emission (AE) and cutting force and fusing the same through Artificial Neural Network (ANN) model such as Multiple Layer Perceptron (MLP) using BackPropagation Algorithm (BPA), which is the most popular algorithm. It highlights the importance of selecting an apt network configuration for better prediction of the tool status. Also the paper attenuates the significance of stopping (of the training process) criterion and critically discussed some of the already fomulated criteria. Further, the paper highlights the importance of evaluating the performance of the network after every epoch and stresses that the training should not be stopped merely based on rate of error convergence or small error threshold or small gradient threshold but by adopting hybrid criterion of any one of the above mentioned criteria along with minimum validation error.