Multi-Objective Dispatch for Batteries and Renewable Generation Sources in Distribution Grids Using a Weighting-Based Convex Approach While Considering Uncertainty

Abstract

This research presents a multi-objective dispatch (MOD) for energy storage systems (ESS) utilizing batteries and renewable energy resources (RES) in distribution network applications. This proposal employs a convex weighting-based approach. The exact MOD nonlinear programming model, which is non-convex due to the power balance constraint, is approximated using a second-order cone equivalent in order to ensure a global optimum, leveraging the convex properties of the objective functions and the conic equivalent in the complex-variable domain. The MOD, based on the linear combination of the objective functions via the weighting-based method, enables the construction of the optimal Pareto front, as each combination of the weighting factors generates a convex optimization sub-problem. The MOD analysis simultaneously considers the minimization of the expected grid operating costs with regard to energy purchasing at the substation terminals, the reduction of the operating costs associated with the RES and the ESS, and the minimization of the expected daily energy losses. This research makes two contributions: 1) the incorporation of the active and reactive power capabilities of the power electronic converters that interface with the ESS and the RES and 2) a robust analysis via convex optimization to address the uncertainties related to the expected daily generation and demand profiles. Numerical results obtained in the IEEE 33- and 85-bus test system confirm the effectiveness and robustness of the proposed MOD in comparison with the continuous genetic algorithm, the particle swarm optimizer, and the vortex search algorithm. In addition, the best metah-euristic technique was employed for constructing the Pareto front and comparing its performance against the proposed convex approach for the 33-bus grid. In the case of the 85-bus grid, battery power losses were included to demonstrate the effectiveness of this solution methodology in medium-scale distribution grids.