Artificial Neural Network Modeling of Injection Upsetting Load Prediction

Lateral extrusion, which requires less forming load comparing the closed-die forging, has been presenting a gradually increasing field of application as a type of metal forming wherein advantageous of forging and extrusion are combined. In this study production of a work piece with a single tapered tooth manufactured by the lateral extrusion method has been made and experimental load and die fullness rates have been measured. After that the data obtained from experiments was submitted to the developed artificial neural networks (ANN) model. The ANN model was trained by taking diameter, height and stroke as the input variables and the injection upsetting load as the output parameter. The ANN predicted results were found to be in terms of prediction accuracy with experimental results.