SVM-based FDD method for refrigerant charge in variable refrigerant flow system

Based on the support vector machine (SVM) algorithm, establishes a fault detection and diagnosis (FDD) model of refrigerant charge. Optimizes the model using grid search method and ten-fold cross validation method, and verifies model performance by testing data. The results show that the accuracy rate of fault detection and diagnosis is very high when the refrigerant charge is insufficient. But the accuracy rate is obviously low when the refrigerant charge is over charged. After the optimization, the total accuracy rate increases from 82.2% to 94.6%.