Abstract: In the detection of refrigerant leakage of household air conditioner, the selection of characteristic parameters has an important influence on its judgment accuracy, so the selection of characteristic parameters and the influence of leakage amount on characteristic parameters are studied. Principal component analysis was used to select 21 characteristic parameters such as suction temperature and exhaust temperature, and six different characteristic parameters were obtained, namely, suction temperature, exhaust temperature, evaporation temperature, suction pressure, exhaust pressure and operating current, so as to determine whether the refrigerant leaks. In addition, through the experiment and simulation is obtained under different leakage six characteristic parameters change trend, on the basis of experiment for six characteristic parameters, the importance of from high to low is: the evaporation temperature, suction temperature, suction pressure, exhaust pressure, current operation, exhaust temperature, for the writing of the refrigerant leakage detection software provides the basis.

Key words: Principal component analysis, Important degree, Characteristic parameters, Refrigerant leakage judgment