GreenNet: Agent based Energy Load Prediction Techniques for Smart Grid

Abstract

To maintain the reliable delivery of energy with increasing demand is becoming a challenging task with such a dumb electricity distribution system. The ongoing reformation of ancient distribution infrastructure is an effort to enhance its performance so that energy can be consumed with greater proficiency. Smart grid is an advanced concept, which adds intelligence, networking and bi-directional communication features to the existing energy infrastructure. To efficiently utilize the system capabilities, prediction of upcoming energy load on the network is an important task. With a more accurate load forecasting, the smart grid can enhance the management of its resources and expand the economics of energy commerce with electricity markets. A new agent based energy load prediction technique is proposed in this paper, which will predict the load of smart home, one hour prior use. Agents are divided into a group of experts, which will use weighted average prediction methodology to predict the upcoming demand of energy. Simulation results show that by implementing the proposed methodology; we can get 80 percent accurate results of load prediction that will make the electricity grid more reliable and efficient.