Artificial Neural Network Application In Cylindrical Bending Process

Abstract

The cylindrical section are back born of fabrication industry, as it is used to prepare the skeleton of pressure vessel, reactors etc. Spring back is the serious problem in air V bending and three roller bending processes. To overcome this problem there exist number of different analytical methods. Among many such analytical method FEM has proven its application for the manufacturing process, however it cannot be used for the online controlling the spring back, as it takes very long time for prediction. In order to get optimum solution of spring back ANN can be used. In present work an attempt has been made to use Neuro solution 4.3 for the ANN applying to cylindrical bending process. Effect of network parameters on Mean Square Error for prediction of spring back was studied. The result of ANN model was compared with Experimental and analytical results. The result was found in good agreement.