Distributed Control of Behind-the-Meter Energy Resources for Multiple Services

Behind-the-meter battery energy storage is used to mitigate non-dispatchability of customer-owned renewable sources. While serving their owners' objectives, their underutilized capacities can he used to provide grid services to justify their cost. However, the integration of these resources extends the size and complexity of the network operation problem, imposing a processing burden on the central control system. In this paper, an agent-based distributed control framework is proposed that enables utilizing of the underutilized behind-the-meter capacities. In this framework, while the prosumer agents initially schedule their resources for daily cost management, a local customer aggregator agent performs day-ahead active and reactive power control on the underutilized behind-the-meter batteries to provide multiple grid services. A novel grouping algorithm is developed based on voltage sensitivity and a modified k-means clustering method to assign customer groups to the local agents and establish the network subsystems. Mixed-integer linear programming is used to model and solve multiple optimization problems. The proposed framework is applied to the 123-bus distribution network. The simulation results show its effectiveness in terms of complexity and optimality, compared to the central control approach in allocating the unused behind-the-meter resources for two common industry needs: network bottleneck congestion relief and grid-edge voltage support.

Automated summary

This paper proposes an agent-based distributed control framework that utilizes the underutilized behind-the-meter battery energy storage to provide grid services. With the use of a novel grouping algorithm and mixed-integer linear programming, the framework shows good performance in terms of complexity and optimization.