Multilevel nested reliability-based design optimization with hybrid intelligent regression for operating assembly relationship

Designing assembly relationship during operation always involves the analyses of many components and multi-discipline interaction, and seriously influences the reliability and work efficiency of complex machinery. To improve the assembly relationship design, distributed collaborative improved support-vector regression (DCISR) method and multilevel nested model are developed to effectively perform the reliability-based design optimization (REDO) of the assembly relationship. In the DCISR method, the improved support-vector regression (ISR) is developed as the basis function of the DCISR model for reliability analysis, by adopting multi-population genetic algorithm (MPGA) to find the optimal model parameters. The proposed multilevel nested model is considered as optimization model for optimizing the assembly relationship. The developed approach and model were applied to the REDO of turbine blade-tip running clearance in respect of nonlinear material parameters and transient loads. As revealed in this study, all optimal solutions satisfy the design requirements of both the blade-tip clearance and the corresponding assembly components. The optimized clearance is reduced by 10% approximately under the reliability premise, by optimally balancing the working efficiency and reliability of the blade-tip. In term of the comparisons of methods and models, it is illustrated that the presented DCISR method holds higher computational efficiency and precision, and the multilevel nested model has higher precision in the REDO of operating assembly relationship. The efforts of this study provide the efficient method and model to optimally design the complex operating assembly relationship, and thereby enrich mechanical reliability method. (C) 2020 Elsevier Masson SAS. All rights reserved.