Modeling of creep-fatigue features of hydrogen storage bed and its parameter optimizing based on finite element method and orthogonal experimental design with artificial neural networks

Abstract

Regarding the creep-fatigue behavior of hydrogen storage bed (HSB) structures under high-temperature loading and internal pressure loading under repeated hydrogen absorption and desorption process, the heat transfer characteristics and creep-fatigue life during a complete service process in HSB were calculated using Ansys Workbench and Ncode software, and the reliability of the life results calculated by finite element method were validated using ASME code. Based on orthogonal experimental design (OED) theory, an orthogonal experimental table with 4 factors and 5 levels was compiled for the four process parameters: diameter of cooling tube, thickness of cylinder wall, spacing between inner and outer layers, and heat transfer coefficient of condensate. Moreover, an optimization method based on OED coupled with BP ANNs was proposed to find the best process parameters. And the importance of each parameter was analyzed based on range analysis and analysis of variance (ANOVA) to divide the process parameters into rough optimized parameters and precise optimized parameters in this method. In order to identify the optimal process parameter combination of HSB, the maximum life and corresponding process parameter combinations were identified based on OED and the combined method, respectively. The maximum life identified by the combined method proposed in this work is increased by 10.23 % than that identified by OED, which shows the superiority of this combined method.