A surrogate model method for aircraft cabin thermal analysis based on equivalent heat sink

Thermal analysis of an aircraft cabin is challenging to simulate accurately because of the complexity of the involved heat transfer processes. The natural convection heat transfer coefficient, a critical parameter for thermal analysis, is difficult to determine reliably through direct numerical simulation or experimental measurement, largely because of the presence of numerous instruments and unknown operational conditions. To overcome these challenges, a surrogate model with physical significance was proposed, in which the instruments are represented as equivalent heat sinks. The parameters required for the surrogate model were determined by analyzing the heat transfer processes within the cabin. Using a set of real temperature field data, genetic algorithms were employed to identify the relevant surrogate model parameters, including the natural convection heat transfer coefficient, heat transfer surface area, volume, and heat generation power. To validate the proposed surrogate model, the temperature field predicted by the surrogate model was compared with experimental results from the real aircraft cabin. It was found that the surrogate model can accurately predict the temperature data of the real aircraft cabin, and the more data utilized in the parameter identification process, the more accurate the temperature prediction. Additionally, as long as the cabin structure remains unchanged, the surrogate model is applicable across different scenarios, significantly improving the efficiency of thermal analysis. Finally, temperature predictions under various boundary conditions were performed and compared with directly simulated temperature fields. The maximum error of the convective heat transfer coefficient obtained through parameter identification was approximately 6 %, while the maximum error of the temperature field predicted by the surrogate model was 2.2 K.