Risk evaluation for low-voltage distribution network based on VMD-LSTM load forecasting model

Addressing the risks of voltage violations and transformer heavy-overloading in low-voltage distribution network caused by impact loads and distributed energy integration, this paper proposes a dynamic comprehensive risk evaluation framework. First, a composite risk index (CRI) is constructed, integrating the system voltage violation risk index (SVRI) and the system active power heavy-overload risk index (SAPHORI). A risk period composite index (RPCI) based on time-decay weighting is also designed. Second, a VMD-LSTM model is built for ultra-short-term load forecasting. Combined with Monte Carlo simulation for future system state sampling, this enables comprehensive future risk evaluation. Case studies using a modified IEEE 13-bus system and actual data from a distribution network in Southwest China demonstrate that the forecasting model achieves over 97 % accuracy. The evaluation results provide warnings several hours in advance regarding risk accumulation trends, offering crucial data support for power grid dispatching.