Adaptive hybrid cooling strategy to mitigate battery thermal runaway considering natural convection in phase change material

The use of a hybrid cooling system, which combines phase change material (PCM) and liquid channels, has been extensively researched to mitigate the risk of battery thermal runaway. However, the existing studies on hybrid cooling systems have overlooked the impact of buoyancy-driven natural convection in PCM. This results in a uniform flow rate allocation among liquid channels, leading to inadequate heat dissipation performance. To this end, this paper proposed a prediction-informed adaptive flow rate allocation strategy for the hybrid cooling system considering natural convection in PCM. First, a novel thermal runaway model was developed, followed by a PCM model that considers natural convection. Then, the proposed strategy utilized Bi-directional Long Short-Term Memory (Bi-LSTM) to predict the outlet temperature, which represents the cooling demand. The flow rate was regulated based on the predicted temperature to avoid an initial flow rate of 0 in the liquid channel and achieve a timely response. The results showed that the proposed strategy extended the thermal runaway propagation interval by 60.6% (85.9 s) and reduced the total flow rate by 31.7%. This strategy significantly enhanced the hybrid cooling system's performance, without increasing the parasitic energy or reducing the battery pack's energy density.