Inverse analysis and multi-objective optimization of single-point incremental forming of AA5083 aluminum alloy sheet

This paper presents soft computing-based modeling and multi-objective optimization of process parameters in single-point incremental forming (SPIF) of aluminum alloy sheet in order to obtain desired deformed shape with optimal formability satisfying multiple objectives. Response surface methodology and adaptive neuro-fuzzy inference system (ANFIS)-based models were developed to predict the responses based on the experimental data collected according to central composite design of experiments considering tool diameter, feed rate and step height as inputs, and outputs, namely forming wall angle, deformed sheet thickness and surface roughness. Inverse analyses were also performed to determine the set of input parameters to achieve desired outputs. Two different algorithms, namely back-propagation and hybrid, were employed to train the ANFIS in batch mode with the help of experimental data. The performances of the developed models were tested through real experimental data and also cross-validation methods. ANFIS trained by hybrid algorithm was found to be slightly better than that trained by the back-propagation algorithm in terms of prediction accuracy. Desirability function and a non-dominated sorting genetic algorithm were utilized for performing multi-objective optimization in SPIF, and the obtained optimal results were found satisfactory compared to the experimental data. The proposed approach could provide a reliable guidance for selection of suitable parameters in SPIF to achieve desired formed parts.