Gaussian Process Model for Control of an Existing Building

AbstractThis study reports the development of a Gaussian Process (GP) model based on BEMS data. The GP model is a data driven model and requires a few dominant inputs. It provides a quick prediction with a far less computation than the whole building simulation tools (e.g. EnergyPlus). The GP model developed in this study is capable of predicting the behavior of a building system (fan energy consumption). This paper reports how the authors developed the GP model. In particular, this paper addresses how to deal with outliers existing in BEMS data set. In this study, RANdom Sample Consensus (RANSAC) was selected for detecting the outliers from the data set. The RANSAC method can be beneficially applied to improve the accuracy of the GP model