Reliability-Based Robust Design Optimization of Mechanical Systems in the Presence of Uncertain Parameters Based on Bayesian Inference

This competitive commercial space forces designers and manufactures to produce and supply products with high quality and low prices at a desirable level of reliability. On the other hand, during the design and production process, engineers are always faced with uncertainty. In recent years, to encounter these uncertainties and guarantee the quality and reliability of a system subsequently, reliability-based robust design optimization (RBRDO) algorithms have been developed based on robust design optimization (RDO) and reliability-based optimization (RBDO). In practical engineering, uncertainties of some design parameters or variables are epistemic and only a few samples are available for designer. Generally, some of the RBRDO methods ignore the information in the design process. This approach can lead to an enormous error. Other RBRDO methods ignore this valuable information in the design process. This study, a comprehensive RBRDO framework is developed by combining Bayesian reliability analysis and dimensionality reduction method (DRM) using NSGA2-II multi-objective optimization algorithm. For verification of the proposed algorithm, an engineering example is selected and the effects of epistemic uncertainty on objectives are studied. Moreover, the results of the proposed approach are compared with other existing approaches at a specific case of available data about epistemic uncertainty.