Multi-objective optimization of a multi-field actuated, multilayered, segmented flexible composite beam

Multi-layered, self-actuated devices have been the focus of recent studies due to their ability to exhibit large displacements and achieve complex shapes. Such devices have been constructed using active materials responsive to varying stimuli including electro-active and magneto-active materials to perform useful functions and achieve a wider variety of target shapes compared to single-field actuated unimorph/bimorph structures. However, fabrication of these devices for experimentation is time-consuming and expensive, which warrants the use of simulations as a means of designing high-performing structures. This work seeks to optimize structures employing materials response to magnetic and electric fields for multiple objective functions selected based on the needs of soft robotics applications such as grippers. A multi-objective optimization problem is constructed, utilizing a model developed for any arbitrary number of segments, layers, and material types, accommodating for large displacements and simultaneously applied fields. Three objective functions are chosen: (1) target shape approximation, based on the errors between the coordinates of the computed and desired shapes, (2) cost based on volume of magnetic material, and (3) work performed on a tip-force. The arbitrary optimization problem is reduced to a specific case study containing eight segments to alleviate the computational cost of an unwieldy number of parameters. The parameters are narrowed to: (1) segment lengths, (2) magnetic material in each magneto-active layer. The structure is pre-set to three material types: electro-active polymer, magneto-active elastomer, and a passive substrate. The case study's optimization problem is performed by a genetic algorithm developed by MATLAB for multiple objective functions. The results of the optimization on the case study are analyzed by studying the feasible designs on the Pareto front of the objective functions. Different trade-offs between objective functions are identified, and various feasible designs are found more suitable than others, based on the needs and priorities of an application.