Research on online monitoring Technology of Refrigerant Charge in airborne liquid-cooled - evaporative refrigeration system

Liquid-cooled evaporation refrigeration systems are key efficient thermal management solutions for aircraft, and refrigerant charge amount exerts a critical impact on their performance and operational stability, making online monitoring of refrigerant charge an essential research focus. This study proposes an online refrigerant charge monitoring model based on integrated machine learning to ensure the system's efficient and stable operation under variable working conditions. A one-dimensional numerical simulation model of the aircraft liquid-cooled evaporation refrigeration system was established on the AMESim platform, based on which the effects of different refrigerant charge levels on key performance indicators were analyzed to determine the optimal charge amount. Simulation experiments were conducted to build a relevant dataset, which was processed through preprocessing and feature selection to screen out the optimal feature subset. Four individual regression models for refrigerant charge estimation were constructed with this subset, and further integrated via the Stacking ensemble method to reduce prediction errors significantly. The results verify that the proposed Stacking-based online monitoring model can effectively realize accurate online monitoring of refrigerant charge in aircraft liquid-cooled evaporation refrigeration systems, thus guaranteeing the system's long-term operational stability.