Modelling and predicting energy consumption in laboratory buildings using multiple linear regression

Abstract

This study was carried to improve the energy saving by investigating the influence factors that contribute to high energy consumption in a building particularly related to the building in Technology Campus, UTeM. Correlation analysis was performed to measure the strength of relationship between the influence factors whereby all the factors proven to have a strong linear correlation with the energy consumption. The Stepwise Selection of Multiple Linear Regression (MLR) were used to determine and modelling the most influence factors that affects the energy consumption. The final linear regression models was developed based on the amount of lighting in a building and surrounding temperature in the building which is considered as major influence factors that affect the energy consumption. Comparing the actual and predicted energy consumption in Technology Campus, UTeM showed that the MLR model obtained can be used to predict energy consumption and accounted for around 81% of the variance.