In order to obtain better air quality and improve the energy saving level of cabin air conditioning, computational fluid dynamics (CFD) method was used to establish a simulation model of the economy class of Boeing 737 passenger aircraft, and particle image velocimetry (PIV) technology was used to verify the accuracy of the simulation model. Based on this model, the effects of different return air ratios on the CO₂ concentration field in cabin air conditioning were studied, and the ventilation efficiency index of the passenger breathing area was used to evaluate the effects of different return air ratios on cabin air quality. The fuel compensation loss under various return air proportion conditions with the same air supply volume was also calculated using the total takeoff mass method, and a functional relationship between the fuel compensation loss, the ventilation efficiency of the passenger breathing area, and the return air proportion was fitted. The efficiency coefficient approach was used to design the evaluation function, and the ideal return air proportion for the cabin air conditioning system was 64.864%. This method can provide a basis for the proportional control of return air in cabin air conditioning.

An improved sine-cosine optimization (ISCA) deep belief network (DBN) prediction model for ground air conditioning energy consumption is suggested in order to increase the prediction accuracy of ground air conditioning energy consumption when the aircraft cabin is cooled by ground air conditioning. In contrast to the standard sine-cosine optimization algorithm, the improved sine-cosine algorithm introduces a cosine adjustment factor to create a new non-linear oscillation adjustment factor to balance the algorithm's overall performance. It also suggests an improved logistic chaotic map, which increases population diversity. In order to prevent the algorithm from reaching a local optimum, a learning technique based on the concept of mutation evolution is finally suggested.Search and local optimization capabilities; finally, a learning strategy is proposed based on the idea of mutation evolution to avoid the algorithm from falling into local optimum. The ISCA-DBN model is applied to the prediction of ground air-conditioning energy consumption of Boeing 737-800 aircraft, and the performance is compared with back propagation (BP)、support vector machine (SVM)、DBN algorithms. There is a certain improvement in both prediction accuracy and real-time performance.