Efficient Energy Optimization Day-Ahead Energy Forecasting in Smart Grid Considering Demand Response and Microgrids

Abstract

In smart grid, energy management is an indispensable for reducing energy cost of consumers while maximizing user comfort and alleviating the peak to average ratio and carbon emission under real time pricing approach. In contrast, the emergence of bidirectional communication and power transfer technology enables electric vehicles (EVs) charging/discharging scheduling, load shifting/scheduling, and optimal energy sharing, making the power grid smart. With this motivation, efficient energy management model for a microgrid with ant colony optimization algorithm to systematically schedule load and EVs charging/discharging of is introduced. The smart microgrid is equipped with controllable appliances, photovoltaic panels, wind turbines, electrolyzer, hydrogen tank, and energy storage system. Peak load, peak to average ratio, cost, energy cost, and carbon emission operation of appliances are reduced by the charging/discharging of electric vehicles, and energy storage systems are scheduled using real time pricing tariffs. This work also predicts wind speed and solar irradiation to ensure efficient energy optimization. Simulations are carried out to validate our developed ant colony optimization algorithm-based energy management scheme. The obtained results demonstrate that the developed efficient energy management model can reduce energy cost, alleviate peak to average ratio, and carbon emission.

Optimal Demand-Side Management Using Flat Pricing Scheme in Smart Grid

“…Smart metering system and MPC collect energy consumption data of users and pricing signals utility. In this work, which is the continuation of the previous work in [84], a joint scheduling of load (power flexible, time flexible, and base appliances) and energy storage (ESS and EVs) using FPS is exploited for a smart home to achieve objectives such as energy cost, carbon emission, and PAR minimization. A brief comparison of the existing work compared to the proposed model is summarized in Table 1.…”

“…The forecaster module uses ANN for prediction of wind speed and solar irradiance for accurate microgrid energy generation estimation based on history data received as input. The ANN layers like input, hidden, and output having artificial neurons are activated via sigmoidal activation function [84] to generate output. These layers are connected in feed-forward fashion, where the preceding layer gets input from earlier layers.…”

“…These layers are connected in feed-forward fashion, where the preceding layer gets input from earlier layers. The sigmoidal activation function is defined as follows in Equation (1) [84]:…”

“…where S represents input signal with selected attributes, parameter β determines the steepness of sigmoidal function, and b shows bias value. The ANN forecaster employed multivariate autoregressive rule to learn from history data to predict wind speed and solar irradiance for accurate energy estimation [84]. The prediction model is shown as first block in Figure 2.…”

“…The prediction model is shown as first block in Figure 2. The history data is adopted from [84]. The history data is divided into training and testing sets of 80% and 20%, respectively.…”