Load control for supply-demand balancing under Renewable Energy forecasting

Abstract

This paper integrates the conception of forecasting Renewable Energy (RE) sources and the user's load demands with intelligent Demand Side Management (DSM) under smart DC micro-grid (SDMG) architecture. The RE are mainly consisting of intermittent solar and wind generators, while the load demands are classified as base (uncontrollable) loads and flexible (controllable) loads. The base loads are priority loads and are served in real time, while flexible loads could be operated intelligently according to the availability of the supply. We integrate a day-ahead prediction mechanism for RE, so that we can schedule a day-ahead consumption accordingly. Practically, these predictions are attained with certain level of forecasting errors, causing imbalance in supply and demands at real-time. This imbalance also known as RE uncertainty, will make the power system unstable. To address the dynamic behavior of SDMG and to balance supply and demands, we propose a novel robust control strategy for controllable flexible demands. To simplify our system we make the generation and demands deterministic, by employing intelligence of Support Vector Machine (SVM) learning algorithm. We then incorporate SVM with novel Sliding Mode Control (SMC) for scheduling consumer's flexible loads to make DSM more efficient and accurate. The energy allocation mechanism to consumer demands is made analogous to non-linear fluid flow model. The simulations have established an effective forecasted data using SVM and efficient balancing results of supply and demand using SMC.