Optimal allocation of customer energy storage based on power big data and improved LSTM load forecasting

Abstract

This research explores the potential of energy storage investment with a focus on regional power users. An incentive-based demand response framework is constructed, emphasizing the aggregation of demand-side resources for active participation in the power market. Utilizing a data-driven approach, the improved Long Short Term Memory (LSTM) model is employed to predict customer behavior in response to incentives. The primary objective is to maximize the life cycle benefit while minimizing the payback period for users investing in energy storage. By harnessing big data analytics, suitable users for energy storage investment are identified and optimal capacity allocation is determined. Given the current energy storage parameters and dynamics of electricity pricing, boundary values for profitable user energy storage capacities are derived. This methodology filters users with capacities exceeding these thresholds, marking them as potential candidates for energy storage deployment. Furthermore, a measurement model for iron phosphate batteries is introduced. Analysis of adjustable resource capacity, duration, and benefits for potential users provides insights into optimal energy storage investment strategies. Integrating configured energy storage batteries with time-of-use tariffs substantially reduces energy storage costs, leading to enhanced economic efficiency.