Probabilistic analysis of extreme rainfall using statistical distributions for nuclear power plant site safety studies.

Extreme rainfall events pose potential risks to critical infrastructure, including nuclear facilities. This study presents a probabilistic analysis of extreme rainfall at the Kalpakkam Nuclear Power Plant (NPP) site in India using extreme value analysis (EVA) with the Gumbel distribution. Two datasets were considered: a historical dataset (1968–1999), used during the original site selection, and an updated dataset (1968–2023) that includes recent observations. Gumbel distribution parameters – location and scale – were estimated using four methods: Least squares method (LSM), method of moments (MOM), maximum likelihood estimation (MLE), and L-moments method (LMM). Among these, the MLE method provided the best fit based on statistical performance. The results indicate that while the location parameter remained relatively stable between datasets, the scale parameter showed a decreasing trend, suggesting reduced variability in extreme rainfall events over time. Bootstrapping was employed to estimate 95% confidence intervals for the parameters and to quantify uncertainty in return level estimates. These findings support the use of updated rainfall data and statistically validated methods for improving hydrological inputs in site safety studies for nuclear facilities. Research highlights: A modest increase in annual rainfall and a narrowing range of extreme rainfall values over decades suggest evolving monsoon dynamics and a potential trend toward more stable, less intense extreme rainfall events. The Gumbel distribution, combined with Maximum Likelihood Estimation (MLE), successfully models extreme rainfall events, providing accurate predictions for rare and high-intensity events in Kalpakkam. Uncertainty analysis reveals an upper limit of 637 mm for extreme rainfall at a 95% confidence level, offering crucial input for flood risk assessments and the design of resilient infrastructure, especially for nuclear facility safety. Observed changes in rainfall patterns highlight the need for updating regulatory frameworks to account for these shifts and incorporate conservative design parameters for future risks. The study emphasizes the importance of incorporating additional climate factors, such as temperature and atmospheric dynamics, along with regional rainfall data, to refine predictions and improve climate adaptation strategies. [ABSTRACT FROM AUTHOR]