Real-time risk assessment and regulation strategy of new energy islanded distribution network based on probabilistic prediction and stochastic response surface methodology

This study introduces an innovative framework for real-time risk assessment and adaptive control in islanded microgrids with high renewable energy penetration. Utilizing bi-directional long short-term memory (Bi-LSTM) networks coupled with Gaussian kernel density estimation (KDE), we develop a predictive model for accurately forecasting power dynamics, essential for managing the variability of renewable sources. Simultaneously, we apply the stochastic response surface method (SRS Method) to evaluate operational risks and optimize control strategies efficiently. Our integrative approach combines advanced machine learning and stochastic modeling, providing a robust tool for microgrid operators to enhance system resilience and reliability. Simulations in MATLAB and Simulink validate the efficacy of our model in capturing temporal variations and addressing the inherent uncertainties of renewable integration. The findings advocate for the proposed methods as vital for sustaining microgrid stability and ensuring energy self-sufficiency in isolated setups.