Prediction Of Drilling Force And Temperature With High Manganese Steel Based On Artificial Neural Networks

In recent years with the rapid development of modern manufacturing industry and the railway construction in china the machining market of turnout rail will keep stable growth in the next few years. But the turnout rail's material mainly is high manganese steel ZGMn13, which belongs to difficult-to-cut materials and its research is still a key project. A new method with the artificial neural network which would be used to study drilling force and temperature during the high manganese steel process was introduced in this paper.In this paper, data input and results interfaces were designed by using the artificial neural networks technology and the MATLAB software. The drill diameter, feed rate and drilling speed were selected as the input values of the three-layer BP neural networks and drilling force, torque and temperature as the output values. The high manganese steel drilling force, torque and temperature prediction models were established by selecting the various parameters of BP neural networks.In order to collect experimental data for the prediction model training and validation, the experimental system and program for the drilling force and torque data acquisition were designed. The experiments were completed and the original data were obtained by using new carbide drill. According to some problems during the drilling railway turnout process, drill material was reselected and the geometric parameters of the drill were determined. The new carbide drill with the whole structure was designed.The training and validation data samples were obtained from the experimental raw data. The drilling force, torque and temperature prediction models were trained by using the training data sample to achieve the requisite approximation accuracy and the model generalization errors were verified by using the validation data sample. The forecast function could be used if the model accuracy meets the requirement. The drilling force, torque and temperature of high manganese steel can be forecasted with the model which had been established in this paper.