No abstract

Energy modeling and energy forecasting play an important role in Energy management systems. Since residential consumptions have a large share in load consumption, an accurate load prediction can avoid most of failures in power grid. Furthermore, new players (Electric Vehicle, Renewable Energy) have emerged in the electrical systems that have caused demand forecasting to gain special interest. So, apart from demand forecasting, power generation and power saving forecasting models have also received increasing attention, especially. In this paper we propose a comprehensive model for electricity load prediction in smart home. We cover all sources of electricity consumption, generation and storage in this model to achieve a prediction with high accuracy (e.g. user power consumption behaviors, electric vehicles, renewable energies and etc.). In the other words, our model consists of three sub models: user electricity consumption sub model, renewable energies sub model (solar cells and wind turbines) and EVs power consumption and storage sub model. All these sub models together build an accurate model that provides low error predictions for EMSs. We use statistical models for power consumption/generation forecasting; also we considered preprocessing algorithms for preparing raw historical data. These preprocessing algorithms are helping us to smooth and omit unusable data that are collected during some transient events.