CityLearn v3 is a configurable simulation platform for Renewable Energy Communities (RECs) that unifies modeling of member turnover, asset availability, flexible‑load deadlines, demand‑response requests, local energy sharing, and data or equipment failures within a single environment. The framework enforces building and phase power limits on controllable requests and uses a declared timestep to keep power‑to‑energy accounting consistent.
During each simulation step, CityLearn v3 logs controller inputs and distinguishes between actions requested by the controller and those actually applied to the simulated equipment, exposing service gaps caused by constraints or failures. Built‑in reference controllers, service‑ and constraint‑aware performance metrics, and trajectory export utilities enable comparative studies both within a community and across different communities.
Software sanity checks and several application examples illustrate end‑to‑end workflows for service delivery, electrical constraints, settlement, and scenario transitions. A synthetic high‑frequency trace replay demonstrates how aggregation can mask short‑duration peaks without altering annual energy consumption. These records allow researchers to interpret aggregate performance alongside service failures, action reductions, and participant‑level outcomes.
Review