Lightweight design of lattice-embedded brake pedals using artificial intelligence -based optimization

Abstract

The application of lattice structures has become increasingly important in designing complex components due to additive manufacturing (AM) advancements. Various types and design parameters of lattice structures allow weight reduction while maintaining the required strength and improving mechanical properties, with the strength varying based on these parameters. One common approach to calculating this strength is by using software solvers like SimSolid, which employs the meshless analysis solution (MAS). Considering the variety of parameters, the complexity of lattice structures, and the computational difficulties in analysis methods, identifying the optimal lattice structure for a design is highly challenging. To overcome this challenge, artificial neural networks (ANNs) are integrated into the optimization algorithm used in this study. The training data for the ANN are obtained from the analysis results of the designs generated using the design parameters selected by the Latin hypercube sampling (LHS) method. The ANNs integrated non-dominated sorting genetic algorithm II (NSGA-II) optimization algorithm is used to minimize the mass while ensuring the strength of the material by keeping the maximum stress within the permissible limits. The method is applied to the weight reduction of the brake pedal, approximately 26.96 % is achieved while maintaining the required strength under existing conditions.