Optimal energy management in a microgrid under uncertainties using novel hybrid metaheuristic algorithm

Abstract

Smart home users play a significant role to implement a demand response strategy in managing the total power network by curtailing or shifting their electricity usage during peak periods in response to Time-of-Use rates. In Time-of-use metering, utility companies’ electricity charge is higher during peak periods. This paper presents the development of an optimal energy management system by incorporating a demand response strategy to operate and sustain the efficiency and reliability of microgrids at a satisfactory level. The long short-term memory recurrent neural network model has been developed based on a novel hybrid metaheuristic grasshopper and flower pollination algorithm to forecast load demand and renewable energy generation for planning day-ahead schedules. Price-based demand response strategy is applied for a microgrid with the effect of penetration of plug-in hybrid electric vehicles to shave off the peak load of the main grid. The simulation study has been done to show the efficacy of the energy management system in terms of reduced operational cost, unit commitment, and load forecasting under uncertain conditions. The accuracy of the forecasting model is satisfied with an RMSE value of 0.17861 for PV power and 0.47844 in the case of wind power. The load demand is reduced by 26.14% using the time-of-use price-based demand response strategy.