Multi-objective design optimization of an engine accessory drive system with a robustness analysis

Designing a good engine accessory drive system becomes a hard work with its increasingly complicated configuration and high demands on its dynamic characteristics. In this work, a hybrid mutation particle swarm optimization (HMPSO) algorithm is presented to optimize the key structure parameters of an engine accessory drive system for its vibration control. The superiority of the HMPSO algorithm against several other concerned metaheuristic algorithms in terms of solution quality and stability are verified by non-parametric statistical tests on ten benchmark functions. The design problem of the engine accessory drive system is a multi-objective optimization problem; the weighted sum method and main target method are applied to convert it to a single-objective one. Optimization on an example engine accessory drive system using the HMPSO algorithm demonstrates obvious improvement in system vibration after optimization. A robustness analysis is conducted to identify the robustness of dynamic responses of the engine accessory drive system with respect to small variations of the design variables relative to the optimal design in the design space, and suggestions on design of an engine accessory drive system are given according to it.