A Study on Prediction of Water Leakage Damage Locations Using Logistic Regression Analysis

Key Points

• The aim is to predict locations of water leakage damage based on pre-disaster information using logistic regression analysis.
• Analyzed historical data from the 2012 Yabegawa leakage incident in Northern Kyushu.
• Used logistic regression analysis to model leakage risk.
• Considered factors like layer thickness, permeability ratio, old riverbed, construction age, curvature, average hydraulic gradient, and high water level duration.
• Logistic regression identified potential leakage damage locations effectively.
• Significant factors included layer thickness, curvature, and high water level duration.