BESS Scheduling for Two Communities of an Active Distribution Network

Effective integration of battery energy storage systems (BESS) in active distribution network's (ADN) communities entails optimal charge/discharge rate scheduling in the day ahead market. Performing quasi-static time-series (QSTS) co-simulations becomes computationally intensive in search of an optimal heuristic solution at each simulation time-step. This necessitates the development of intelligent multi-community energy resource scheduling solutions with reduced execution time and by leveraging the parallel processing feature of modern CPUs. This paper implements a two-stage intelligent approach in a MATLAB-OpenDSS parallel processing environment. In the first stage, a rule-based decision tree provides the community BESS initial charge and discharge power set-points, input to the second stage to minimize community net load power deviations while considering ADN and BESS equality and non-equality constraints. Each community in the ADN has been assigned to a dedicated workstation CPU core to execute the two-stage approach independently, in a controlled parallel-processing environment. The results indicate a 96.1 % reduction in the QSTS co-simulation execution time, also decreasing utility peak demand, branch power losses and nodal voltages deviations in an IEEE 13-Bus DN with two incorporated PV-BESS-based communities.