Central air conditioning system cooling load prediction method based on SVR algorithm

The invention discloses a central air conditioning system cooling load prediction method based on an SVR algorithm. The central air conditioning system cooling load prediction method comprises the steps that a cooling load prediction cycle tau is determined; a dataset with historical data of outdoor meteorological parameters and indoor air conditioning system relevant parameters as a cooling loadprediction calculation model is built; a cooling load prediction training dataset is built; an input dataset and a training set are built; and a prediction model is generated and saved. Indoor coolingload needs in the future time are predicted according to real-time data of the outdoor meteorological parameters and indoor central air conditioning operation parameters, air conditioning cooling load prediction is conducted through the SVR regression algorithm based on an RBF kernel, the training dataset is selected optimally through a PCA (principal component analysis) method, and the needs forthe air conditioning cooli