Soft computing approaches for comparative prediction of ram tensile and shear strength in aluminium-stainless steel explosive cladding

This study deals with the application of soft computing techniques viz., response surface methodology, artificial neural network, radial basis function network and support vector regression in analyzing and predicting the ram tensile and shear strengths of aluminium 5052-stainless steel 316 explosive clads, having different interlayer. 60 explosive cladding experiments were conducted, based on central composite design of experiments, by varying the process parameters viz., loading ratio (mass of the explosive/mass of the flyer plate: 0.6-1.0), distance of separation (6-10 mm), preset angle (6 degrees-8 degrees) and interlayer (aluminium 1100/pure copper/stainless steel 304). The responses viz., ram tensile and shear strengths obtained from 90% of the experiments and trial experiments are used for training artificial neural network, radial basis function network and support vector regression in a Matlab environment, altering training algorithms and number of neurons in the hidden layer. The remaining 10% of the experimental outcome is used for testing the developed models. Likewise in RSM, regression equations are generated for the responses, based on analysis of variance. All the four models are capable of predicting the ram tensile and shear strength effectively, as the average percentage deviation with the experimental outcome are less than 10%. Of the three models, artificial neural network model predicts the ram tensile strength and shear strength in a better manner.

Automated summary

This study applies various soft computing techniques to analyze and predict the strength of aluminum-stainless steel explosive clads. Experimental data is used to train models, resulting in accurate predictions.