Space cooling load prediction for building using online LS-SVM

Space cooling load affected by lots of factors is transient, multi-dimensional and highly interactive. A model of Online Least Squares Support Vector Machine (LS-SVM) is established to forecast the space cooling load of building, and the correlation analysis is used to find out the main parameters of influencing factors, such as, dry-bulb temperature, solar irradiance, and so on. As an example, the hourly space cooling load of an office building in shanghai is investigated. The hourly space cooling load is firstly calculated by the simulation software of DEST, and then, an online LS-SVM model is presented to forecast the load. The simulation results show that the online LS-SVM model is effective for space cooling load prediction.