Optimal Operation of Renewable Energy Communities under Demand Response Programs

TL;DR

This paper addresses optimizing the operation of Renewable Energy Communities (RECs) with battery storage under price-volume Demand Response (DR) programs. It introduces a three-step approach: (i) compute standalone profits for each REC member, (ii) solve a low-complexity MILP to schedule community storage while maximizing a chosen performance index and guaranteeing a baseline profit for all members, and (iii) redistribute the DR rewards fairly among members based on a proportional fairness rule. Two objective functions, $H^M$ and $H^E$, are compared to reflect REC manager versus entity-centric goals, and a linear-programming per-member step ensures computational tractability even for large memberships. Numerical results show that the method yields increased profits for REC members compared to operating alone, while maintaining fairness in DR reward distribution and maintaining low computation times (e.g., ~0.15 seconds per day for 30 members). The approach supports practical deployment of DR-enabled RECs and offers a foundation for extensions to stochasticity and alternative reward schemes.

Abstract

Within the context of renewable energy communities, this paper focuses on optimal operation of producers equipped with energy storage systems in the presence of demand response. A novel strategy for optimal scheduling of the storage systems of the community members under price-volume demand response programs, is devised. The underlying optimization problem is designed as a low-complexity mixed-integer linear program that scales well with the community size. Once the optimal solution is found, an algorithm for distributing the demand response rewards is introduced in order to guarantee fairness among participants. The proposed approach ensures increased benefits for producers joining a community compared to standalone operation.

Figures (7)

• Figure 1: Overall reward related to the $j-$th DR request as a function of the net energy injected in the grid.
• Figure 2: Example 1. Results for Entity 1 on day 24 under Problem \ref{['pb:alone']}. Energy production forecast $\hat{E}_u(t)$ (blue), sold energy $E^g_u(t)$ (dashed red), energy selling price $\pi^g_u(t)$ (green), storage energy level $S_u(t)$ (purple), and time periods of DR requests (yellow).
• Figure 3: Example 1. Results for Entity 1 on day 24 under Problem 2 with $H=H^{E}$. Energy production forecast $\hat{E}_u(t)$ (blue), sold energy $E^g_u(t)$ (dashed red), energy selling price $\pi^g_u(t)$ (green), storage energy level $S_u(t)$ (purple), and time periods of DR requests (yellow).
• Figure 4: Example 1. Daily additional profits $\delta_1$ and $\delta_2$ under the objective functions $H^E$ (top) and $H^M$ (bottom).
• Figure 5: Example 1. Daily DR reward profiles for requests 1 and 2, under the objective functions $H^E$ (top) and $H^M$ (bottom). Maximum achievable DR reward $\overline{\gamma}_1+\overline{\gamma}_2$ (orange), actual DR reward at community level $\gamma_1^*+\gamma_2^*$ (blue), DR reward of Entity 1 $\xi_1^*$ (red) and 2 $\xi_2^*$ (yellow), and total reward of entities $\xi^*=\xi_1^*+\xi_2^*$ (green).

Theorems & Definitions (5)

• Remark 1
• Remark 2
• Remark 3
• Example 1
• Example 2