Assessing Extreme Rainfall Impacts on Power Grid Reliability Using Composite Threshold and Chicago Method

Amid rising extreme precipitation driven by global climate change, urban power grid stability encounters significant challenges, threatening infrastructure reliability. This study develops a systematic framework to evaluate operational risks of power grid infrastructure under extreme rainfall through detailed precipitation characteristic analysis. To address methodological gaps in impact assessment and short-duration rainfall pattern analysis, a hybrid threshold method integrates the 99th percentile dynamic threshold with fixed 50 mm and 100 mm thresholds, achieving balanced climate adaptation and disaster response capabilities. The Chicago method simulates short-duration storm patterns across 2 ~ 100 year return periods. A CNN-LSTM model analyzes interannual trends and projects future extreme rainfall events, bolstering long-term risk warning systems. Validation using Shanghai meteorological data (1961 ~ 2020) confirms a clear increase in extreme precipitation events. The methodology overcomes static limitations of conventional approaches, enhancing urban flood prevention and climate resilience of critical power infrastructure. This framework provides valuable support for infrastructure planning in vulnerable coastal megacities facing intensifying climate change impacts.