Chip form classification and main cutting force prediction of cast nylon in turning operation using artificial neural network

Abstract

In this paper classification of chip form and main cutting force prediction of cast nylon in turning operation by using artificial neural network (ANN) are described. The multi-layer perceptron of back-propagation neural network (BPNN) was employed as a tool to classify a chip form following ISO 3685-1977(E) and predicted the tangential cutting force. The turning operation was performed by a conventional form of high speed steel cutting tool with various cutting speeds, feed rates and depths of cutting. The BPNN structure had two models consisting of classification and prediction model. Each model composes of an input layer, two hidden layers and one output layer. Input layer composes of three input parameters, including cutting speed, feed rate and cutting depth. Hidden layer contains twenty nodes on each layer. A node of output layer was determined for obtaining the results. The sixty data from the experiments were used for neural network training with optimum parameters equal 0.6 of training rate and 0.6 of momentum. A set of data from the fifteen turning operation experiments were employed for prediction. The results revealed that the classification accuracy for classification chip form was 86.67%; and the main cutting force prediction was 91.130% of accuracy. Therefore, the chip form and main cutting force in cast nylon turning operation can be classified and predicted with reasonable accuracy for a given set of machining conditions using ANN model.