Markovian Models for Electrical Load Prediction in Smart Buildings

Developing energy consumption models for smart buildings is important for studying demand response, home energy management, and distribution network simulation. In this work, we develop parsimonious Markovian models of smart buildings for different periods in a day for predicting electricity consumption. To develop these models, we collect two data sets with widely different load profiles over a period of seven months and one year, respectively. We validate the accuracy of our models for load prediction and compare our results with neural networks.