Parametric study and multi-objective optimization of process design variables in single-point incremental forming of low-carbon steel/CP-titanium bimetallic sheets via a full factorial DoE, GRA and RSM

Single-point incremental forming process (SPIF) of monolithic sheet was comprehensively investigated over the last decades; contrarily, few efforts were conducted on the SPIF processing of composite sheets. Unlike the conventional manufacturing processes, this emerging technique is not yet strongly controlled. In fact, the SPIF process exhibits its capability to shape the dissimilar laminated metal composite sheets into 3D geometries. However, this process is not really industrialized, yet, it still suffers from geometrical inaccuracy, non-homogeneous thickness distribution and significant forming forces that can damage tools. In the present study, the incremental forming of CP-titanium/low-carbon steel bimetallic sheets has been investigated both experimentally and numerically. In the first attempt, the authors demonstrated that the order of the layers' stacking of the bilayer sheet strongly influences the process. Secondly, a full factorial design of experiment (DoE) approach was first conducted to examine the main and interaction effects of the SPIF process parameters, namely the total initial sheet thickness of all the considered layers, the wall-drawing angle, the vertical increment step size and the process strategies on the forming forces evolution and the final sheet thickness distribution. Then, a multi-optimization technique by combining the grey relational analysis (GRA) and the response surface methodology (RSM), associated to a global approach, is used to determine an optimal combination of input process parameters, required to obtain favorable responses. Furthermore, the analysis of variance (ANOVA) for the grey relational grade is conducted to identify the best levels and significant contribution of parameters. Subsequently, promising results were obtained from the optimization, and the optimal responses have been validated through confirmatory experiments demonstrating that the process could be effectively improved and that the optimization technique is robust. Finally, it could be observed that the optimization results demonstrated the compatibility and reliability of these processing techniques.