Numerical investigation of flow field configuration for accelerated helium mixing in battery pack leak detection

Helium leak detection is one of the essential methods for current battery pack enclosure sealing performance testing. However, owing to the absence of a quantitative criterion for assessing helium mixing uniformity, the helium leakage detection method is compelled to adopt a conservative testing cycle of 375 seconds (300-second pre-mixing followed by 75-second sampling), which constitutes a major limitation of this approach. This study comparatively evaluates fan-array flow configurations within a seal-testing bench using three-dimensional transient computional fluid dynamics (CFD) simulations. Four initial flow topologies were compared, among which the cyclone layout provided the strongest transport but generated a vortex stagnation core at the geometric center. Three refined cyclone-derived configurations were then proposed to destabilize this core. The quadrantal antagonistic circulating layout achieved a better performance by breaking the geometric symmetry of the fan array. For the quadrantal antagonistic circulating layout, the predicted average t ₉₀ over four leakage locations was 122.6 seconds, and the corresponding simulation-based testing cycle was reduced by 59.2% compared with the conservative 375-second procedure. The average leakage rate was 10.05 sccm, with an error of only 0.5%. The simulation-based testing cycle evaluation method established in this article provides a numerical basis for shortening the cycle of the helium detection process.