Capacity Planning of the Hybrid Energy Storage System Based on Load Prediction with Confidence Interval

Microgrid is an effective way to realize the expansion of power system. With the high penetration of renewable energy, microgrid systems are faced with challenges in reliability and security. Hybrid energy storage systems (HESSs) are widely used in microgrid to mitigate the threat of renewable energy fluctuations. However, the high cost of HESSs restricts its development. Reasonable capacity planning will improve the economic competitiveness and promote the feasibility of HESSs. The traditional capacity planning method of HESSs usually relies on historical data or certain forecasting results but ignores the uncertainty caused by new energy fluctuations which may have certain security risks. This paper proposes the capacity allocation method of HESSs considering load prediction margin. First, CEEMDAN-GRU-Attention model is used to predict the net load power of the microgrid. Then reasonable confidence interval (CI) is established for the load prediction results based on Copula theory. Last, particle swarm optimization (PSO) algorithm is used to optimize the capacity of HESSs, aiming to achieve the lowest life cycle cost (LCC) of the HESSs under the premise of system safety. The effectiveness of the proposed method is verified by a practical example, which provides technical support for the safe and economical operation of microgrid.