Optimal Battery Energy Storage System Scheduling Based on Mutation-Improved Grey Wolf Optimizer Using GPU-Accelerated Load Flow in Active Distribution Networks

In this paper, a novel Mutation-Improved Grey Wolf Optimizer (MIGWO) model is introduced in order to solve the optimal scheduling problem for battery energy storage systems (BESS), considering the mass integration of renewable energy sources (RES), such as solar and wind generation, in active distribution networks. In this regard, four improvements are applied to the conventional GWO algorithm to modify the exploration-exploitation balance for an enhanced convergence rate. The validity and performance of the proposed model are tested on 23 classical benchmark functions and compared to the original algorithm. The new technologies present in active distribution networks lead to increased complexity in the efficient coordination of existing resources, making it necessary to resort to advanced optimization and calculation methods. As operational planning and control functions in power systems are computationally demanding and require multiple power flow calculations, the necessity of simultaneous (parallel) computing techniques emerged. In order to reduce the computing time, an accelerated GPU parallel computing technique is also applied in the proposed model. The MIGWO algorithm is further applied on the modified IEEE-33 bus system aiming to minimize the total power losses, based on the optimal coordination of BESS operation scheduling and RES generation for multiple load demand and local generation scenarios, as well as for various initial state-of-charge values of BESS.

Automated summary

A novel Mutation-Improved Grey Wolf Optimizer (MIGWO) model is introduced in this paper to solve the optimal scheduling problem for battery energy storage systems, with improvements made for enhanced convergence rate. The complexity of modern power systems requires advanced optimization methods, such as parallel computing techniques, to improve efficiency. The proposed model is tested on benchmark functions and applied to minimize power losses in active distribution networks, showcasing its effectiveness in solving complex coordination problems.