Preprint of "Multi-Objective Optimization of Priority-Based Energy Sharing in Renewable Energy Communities Using NSGA-II"

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Provides a multi-objective optimization method (NSGA-II) for REC operation, with a detailed decision variable analysis showing priority ordering as the key lever.

🏢実務担当者:Offers a framework for designing priority rules in community energy management, reducing grid imports and switching activity, with regime-dependent performance.

🏛政策担当者:Demonstrates the potential of optimized REC operation to support energy transition goals, relevant for REC regulation and incentive design.

This paper presents a two-layer simulation and optimization framework for the operation of renewable energy communities (RECs) with multiple metering points. The lower layer is the KOMEN deterministic simulation model, which evaluates power flows among individual connection points comprising local photovoltaic (PV) generation, fixed and flexible loads, battery energy storage systems (BESS), electric vehicle (EV) charging, and grid import/export within 15 min settlement intervals. The upper layer applies the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to jointly optimize the priority ordering and allocation weights that govern intra-community energy dispatch. The continuous decision vector contains 32 variables (of which 26 are active in the default configuration) encoding priority ranks and fractional weights for three energy sources across three connection points. The framework targets six competing operational objectives—grid import, grid export, shared energy, PV curtailment, battery cycling throughput, and flexible-load switching—of which the active subset depends on the community's operating regime in the studied period. The framework is demonstrated on a three-node REC featuring a 15kW PV system, a 10kW/20kWh local BESS, a community-scale 20kW/50kWh BESS, 11kW EV charging, and two shiftable controllable loads. Profiles are derived from five-minute measured data of a real PV installation in Pohořelice, Czech Republic; October 2025 was selected as the representative month via a full-year baseline simulation. Five operational scenarios are compared. For the selected month, the optimization reduces grid import by 7.1% and flexible-load switching activity by 83% relative to the baseline. Isolating the decision variables reveals that the ordering of dispatch priorities is the dominant control lever—priority-only optimization attains the full improvement, while allocation-weight tuning alone cannot reach the low-import region—with the combined optimization matching the priority-only result and extending the sharing trade-off only marginally. Repeating the analysis for an export-dominated summer month and a deepwinter month confirms that the benefit is strongly regime-dependent—largest in the transition-deficit month—while the effect on switching activity is also regime-dependent, with substantial reductions in the deficit-dominated months but an increase in the selected summer solution. The measured generation and household-consumption profiles are linearly scaled to the community size; the multi-node topology, storage, EV charging, and controllable loads are modeled, making this a hybrid measured–synthetic study.

• Zenodo https://zenodo.org/records/21105602first seen 2026-07-02 04:14:10 · last seen 2026-07-02 04:16:59