Energy Optimization in Smart Grid using Tensor Decomposition

Abstract

In India energy crisis has become one of the primary concern for its development and economic growth. The gap between power demand of users and its supply is incrementing day by day. Moreover, a large portion of the power plants depend on petroleum derivative and have the danger of being phased out in future. In this paper, we consider these issues by adjusting the power supply and demand, concentrating fundamentally on Disseminated Energy Resources (DER). During off peak hour’s residual energy from DER will be stored in the proposed storage arrangement. The proactive consumers (pro-summer) in the demand side will have the scope to sell this stored energy to the national grid during peak hours in a proposed smart bidirectional network. Finally, grid monitoring and metering interface with an advanced control mechanism has been developed, which is expected to increase the extendibility of the pro-summer to handle their energy usage and costs. We have proposed a mechanism known as Tensor Decomposition (TD), in which we combined the smart meter rating produced from different sources into single dimension using MATLAB and finally applied Principle Component Analysis (PCA) to determine the error rate as compared to that of actual rate. Result obtained after applied the TD mechanism over Smart Grid (SG) data set are encouraging as compared to other preexisting approaches.