Economic Optimization of Energy Storage Systems in Active Distribution Networks Using Salp Swarm Algorithm

Integration and Operation of Energy Storage Systems in Active Distribution Networks: Economic Optimization via Salp Swarm Optimization

This paper addresses the integration and operation of lithium-ion battery energy storage systems (ESS) within active distribution networks that have a significant penetration of renewable distributed generation. The primary objective is to minimize total system costs, which encompass energy purchases at the substation node as well as integration, maintenance, and replacement costs of ESS over a 20-year planning horizon.

The proposed methodology employs a master-slave framework utilizing the Salp Swarm Optimization Algorithm to identify optimal locations for ESS, select appropriate technologies, and devise daily operational schemes. A successive approximation power flow technique is used to evaluate the objective function and ensure constraint compliance. This approach incorporates a discrete-continuous encoding strategy, which helps to minimize processing times and enhance the possibility of achieving a global optimum.

The methodology was validated using a 33-node test system tailored to Medellín, Colombia, demonstrating superior performance compared to five other metaheuristic algorithms. The Salp Swarm Optimization Algorithm achieved the highest annual savings of USD 16,605.77, the lowest average cost of USD 2,964,139.99, and the most efficient processing time of 345.71 seconds. These results indicate that the proposed methodology effectively enables network operators to reduce distribution network costs while ensuring high repeatability, solution quality, and computational efficiency.

Introduction

The energy sector is currently facing significant challenges due to climate change and the need for energy security and economic stability. The global demand for energy continues to rise, primarily driven by technological advancements and population growth, creating substantial pressures on conventional power systems. Key challenges include maintaining stability and reliability, minimizing power losses in transmission and distribution networks, and reducing the environmental impact associated with fossil fuel-based generation.

Renewable energy sources, such as solar and wind, have become essential alternatives to traditional energy sources, offering a way to reshape the relationship between environmental, economic, and social values. However, the transition to these sources presents challenges due to their variability and intermittency, which can threaten the stability and efficiency of power grids. The integration of energy storage technologies, especially battery energy storage systems (ESS), has emerged as a strategic solution to enhance the reliability and flexibility of power systems with high renewable penetration. ESS not only facilitate a sustainable energy transition but also comprehensively address technical and environmental challenges.

ESS are particularly beneficial as they can store energy during periods of low demand and high renewable generation, releasing it during peak consumption times. This capability enhances operational efficiency and reduces costs by decreasing reliance on fast-response fossil fuel generation units. Furthermore, ESS provide various ancillary services, including frequency regulation and voltage control, which are crucial for maintaining electricity supply quality, especially in systems with high levels of distributed energy resources.

Researchers have explored various aspects of battery energy storage systems, focusing on optimization, control, and forecasting techniques to improve performance, economic feasibility, and integration into energy systems. For instance, reinforcement learning models have been employed to optimize charging and discharging decisions under uncertain market conditions, while risk-preference optimization models have been utilized to evaluate ESS investments by balancing risk and return.

The integration of ESS into smart grids has been exemplified through various frameworks and strategies that address peak shaving and grid stability, outperforming conventional methods. Innovations include adaptive multi-functional strategies for managing power and voltage in microgrids and predictive control frameworks that improve economic dispatch in isolated ESS networks. Additionally, advanced forecasting techniques have been leveraged to enhance the reliability of decision-making in ESS management.

Despite the advancements, a critical gap remains in the simultaneous optimization of ESS location, technology selection, and operation, particularly within the radial structure of distribution networks. This study proposes an integrated mathematical formulation that addresses these dimensions within AC microgrids, aiming to optimize economic performance and reduce operational costs amid high renewable penetration.

Mathematical Formulation

The integration of ESS in active distribution networks necessitates optimal selection of locations, technology, and power dispatch for daily operations. This is modeled using mixed-integer nonlinear programming, where binary variables denote ESS location and technology, while continuous variables represent ESS operation. The constraints include power flow, ESS state of charge, and time-coupled behavior, resulting in a nonlinear and nonconvex solution space.

The goal is to minimize total annual operational costs, which include energy purchase costs, ESS and distributed generation operation and maintenance costs, initial investments, and replacement costs. The mathematical model incorporates various constraints that represent operational conditions, such as complex power balance, voltage regulation, and thermal limits of conductors.

Proposed Methodology

The proposed master-slave methodology utilizes a discrete-continuous coding vector to integrate strategic and operational decisions. The master stage focuses on determining ESS locations, technology selections, and daily operational schemes, while the slave stage evaluates the objective function and checks for compliance with technical and operational constraints.

The Salp Swarm Optimization Algorithm serves as the optimization technique in the master stage, employing leader-follower dynamics to explore the solution space effectively. The slave stage uses a matrix-based successive approximation approach to assess the impact of ESS configurations on annual operational costs, ensuring compliance with operational constraints through a penalized fitness function.

Numerical Results and Discussions

The proposed methodology was validated using a widely recognized 33-node test system, with results indicating a significant reduction in annual costs when ESS are integrated. Comparisons with other methodologies demonstrated that the Salp Swarm Optimization Algorithm achieved the lowest annual operating costs while maintaining efficient computational performance.

Key findings include the identification of strategic nodes for ESS installation, the preference for higher capacity ESS types, and the effectiveness of the proposed methodology in managing operational strategies for optimal performance.

Conclusions, Limitations, and Future Work

This study highlights the effectiveness of the proposed methodology in optimizing ESS integration and operation within active distribution networks. The Salp Swarm Optimization Algorithm outperformed all compared approaches, demonstrating its capability to minimize costs and enhance computational efficiency. Future work will focus on testing the methodology in larger networks and exploring enhancements to address additional objectives, such as minimizing emissions and improving energy efficiency.