Expert system to predict forging load and axial stress

Finite element (FE) analysis of forging process generally takes a long time to carry out. Sometimes, it might be required to predict the results of FE analysis as accurately as possible in a less processing time. A pre-analysis prediction of the results could also be helpful in some cases. Soft computing-based expert system has been developed in the present work, to predict forging load and axial stress developed. Forging load and axial stress have been calculated for an axi-symmetric part by varying the values of maximum contact friction stress and normal contact stiffness factor in ANSYS FE package. Data generated in the process have been utilized for developing a fuzzy logic-based expert system, on the basis of the authors’ knowledge, which has been optimized subsequently using a genetic algorithm (GA). Two other expert systems (ESs) have been developed automatically using the GA, without taking any aid from the manually-designed fuzzy logic system. It has been found that the expert systems are able to make predictions of forging load and axial stress as accurately as the FE package can do. Results of the ESs have also been seen to be comparable with the experimental results reported in the available literature. Instantaneous prediction capability of the developed expert systems proves their suitability for on-line implementations.