Reliability-based thermal-fluid-structural topology optimization for PEMFCs’ turbulent coolant channels by Q-SORA strategy

Abstract

In order to attain a desirable trade-off between the flow dissipation and thermal reliability level of proton exchange membrane fuel cells (PEMFCs) when considering the random positions of porous multi-heat sources on bipolar plates (BPs), this work establishes a reliability-based topology optimization (RBTO) framework for coolant channels. In the RBTO, the flow dissipation of the channel is minimized subject to the target structural compliance and thermal failure probability, and the RBTO framework is decoupled by quantile-based sequential optimization and reliability assessment (Q-SORA) strategy efficiently, which involves a sequential loop of deterministic topology optimization and quantile estimation under the current channel topology. When the target failure probability ($P_f^T$) is set to 3.8 % (0.2 %), the minimized flow dissipation corresponding to uniform and gradient heat sources is 0.000729 (0.001844) kg·m²/s³and 0.000674 (0.001476) kg·m²/s³, respectively. In addition, the flow dissipation increases significantly with the decrease of the $P_f^T$, and the same flow directions of coolant channels and the reaction porous gas channels can effectively decrease the flow dissipation cost in RBTO of coolant channels. Moreover, the impacts of random multi-heat source positions on the maximum temperature (T_max) of the BPs can be reduced with the increase of inlet velocity. When the inlet velocity is higher than 2m/s and the $P_f^T$ is 3.8 %, T_max of the BPs at the mean heat-source positions and the flow dissipation of the topological coolant channel are both lower than those of the traditional straight multi-channel configuration.