Development and Modelling of Hydro-formed circular sheet Using Neural Networks

Abstract

Hydro forming process is one of the new technologies in the manufacturing process in which hollow pieces and sheet works are formed under fluid pressure. This paper deals with the development of a sheet hydro forming (SHF) set up and model to predict the deformation caused by the hydro forming using Artificial Neural Networks(ANN). During the sheet hydro forming (SHF) process, the hydraulic pressures were measured and simultaneously the required hemispherical products were obtained with dome height and specific thickness. To validate result of this set up work, a hemispherical part was formed in a fabricated hydro forming set up using single sheet. The data used to train the Artificial Neural Network (ANN) model was collected from sheet hydro forming (SHF) set up that was designed and built in our laboratory. Single-and two-hidden-layer feed forward neural network models were used to capture the nonlinear correlations between the input(hydraulic pressure, thickness and diameter of work piece) and output (dome height of deformed product). Neural network model was able to predict the deformed dome height, thickness of circular sheet with good accuracy.