关键词: 大数据；数据预处理；静态压力；回归分析；权重系数

Abstract: Monitoring refrigerant leakage in automotive air conditioning systems has long been a persistent challenge for the industry. With the widespread adoption of pressure and temperature sensors in automotive thermal management systems, establishing a robust pressure-temperature (P-T) model to meet leakage monitoring requirements has become a pressing issue. This paper focuses on the R134a refrigerant static pressure model. Based on big data, a robust P-T regression model correlating R134a static pressure with ambient temperature is established through data collection, labeling, regression analysis, and the introduction of weighting coefficients. The refrigerant static pressure value (P) fitted by the model is compared with the actual value collected via the controller area network (CAN) bus to serve as the basis for leakage judgment. The optimized model demonstrates excellent robustness, satisfying refrigerant monitoring needs throughout the entire vehicle lifecycle.

Key words: big data; data pre-processing; static pressure; regression analysis; weight coefficients