Fuel Cell Stack Module Modal Analysis

Unlike IC engines, fuel cells have no moving parts, thus are not a source of vibration. However, they are subject to road induced vibration in automotive applications. Therefore fuel cell Stack Modules (STM) have to be designed to withstand severe road vibration besides other dynamic loads such as mechanical shock and crash impulse. Modal analysis and testing of an automotive fuel cell STM is discussed in this paper. A 3D FEA model of the STM was built to calculate its natural frequencies (eigenvalues) and mode shapes (eigenvectors) using ANSYS. Modal analysis of the STM was essential for the mounting bracket design since, as for every automotive component design, the space is very limited and the part must be strong enough to meet the requirements, while keeping its weight as low as possible. Based on FEA calculations the mounting strategy for the STM changed by switching from a 3-point mounting system to a 4-point mounting system. The calculations clearly showed that based on space and weight requirements in place, a 3-point mounting system could not be used to support the STM. By switching to a 4-point mounting system, the strength problem was resolved but there was a new challenge; the first natural frequency of the STM was too low causing the entire fuel cell stack to resonate, thus causing other parts of the system to have serious problems. The 4-point mounting bracket then was designed to have the maximum stiffness within the allowed space by optimizing its geometry. Based on the FEA results, the bracket footprint was increased and four M8 bolts per bracket were used instead of two. Bracket bolt-joint modeling was crucial for accurate calculation of the natural frequencies. The analysis results were compared against test results run on an electro-dynamic shaker and the correlation was superb. The difference between the calculated and measured first natural frequency was less than 1%. This level of correlation between the FEA and test results allowed the STM mounting system resonance problem to be resolved a lot faster, with minimum design iterations.

View Now