Neurosymbolic inference and fuzzy control enable risk-aware reconfiguration of active distribution networks
Year: 2026
Type: article
Abstract: Distributed generation, electric vehicles, and variable demand complicate distribution-network reconfiguration. We propose a hybrid architecture combining graph convolutional inference, symbolic topology verification, and an interval type-II fuzzy controller tuned offline by quasi-opposition arithmetic optimization. The symbolic layer rejects non-radial or disconnected configurations, while the fuzzy controller adjusts reconfiguration priorities under uncertain operating conditions. MATLAB simulations on a modified IEEE 33-bus system distinguish the effects of topology and distributed generation: reconfiguration alone reduces losses by 31.15% on the benchmark without distributed generation, whereas their combined effect reduces losses by 53.72% relative to the unreconfigured benchmark. In the reported fault-recovery comparison, energy not supplied is 77.2% lower than with particle swarm optimization. The online policy minimizes lost energy by 25.87% with 16 switch operations and no dispatched radiality violations during four seven-day seasonal profiles. The tests also include missing measurements, unseen configurations, and transfer to a 69-bus feeder. The results justify decoupling learnt recommendations from deterministic topology tests, whereas reconstruction errors and electrical-limit verification urge further validation before implementation.
Source: Scientific Reports
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7221031153