OPTIMISATION OF PARAMETERS FOR IMPROVING DIMENSIONAL ACCURACY IN CNC MACHINING

摘要

The objective of this work is to demonstrate how the single hidden layer Multilayer Perceptron (MLP) neural network could be used to model a typical NC turning process. The Network configuration was decided after performing many trials, and then a generalized MLP neural network with a single hidden layer was used to establish the process model with the available experimental data. The neural network was then used to predict the diameter error and the cutting force for different operating conditions and the testing process was conducted. From the results obtained it was found that the predicted values were within the allowable error tolerance. Therefore, it was found that the implemented single hidden layer back propagated Neural Network approach yields a relatively more accurate process model for the turning process.