Answers are grounded in this paper and cite their sources.
As the penetration ratio of distributed photovoltaic (PV) systems continues to rise, the uncertainty in power generation and load during grid operation has increased. Moreover, the coupling of multiple operational constraints makes traditional capacity assessment methods inadequate for accurately characterizing PV integration capacity and associated operational risks. To address these challenges, this paper proposes a method for assessing the openable capacity of PV systems and predicting overload risks by integrating multi-source measurement data. First, a two-layer capacity assessment optimization model is constructed, and the variation particle swarm optimization algorithm is employed to solve for the PV integration capacity boundary. Subsequently, a deep time-series prediction model is developed, incorporating a convolutional neural network, a convolutional block attention module, and a bidirectional Gated Recurrent Unit (GRU), assessing operational risks. Experimental results demonstrate that the proposed method achieves a capacity gain rate of 6.8% on the test dataset, an F1-score for risk prediction of 0.935, and maintains an average absolute percentage error below 6.1% under noisy interference conditions. When the PV penetration rate reaches 60%, the method can still maintain an openable capacity of 768.5 kW and reduce the maximum voltage deviation at the terminal node from 0.045 to 0.022 p.u. The findings indicate that the proposed method enhances both the accuracy of PV integration capacity assessment and the capability for operational risk prediction, thereby providing effective support for the safe operation of distribution networks under high PV penetration scenarios.
No takes yet. Share an insight, caveat, or question.
Guo et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: