Author(s): Yusuke Hida; Hiroshi Chiba; Yoshihiro Asaoka; Hisao Nagabayashi
Linked Author(s):
Keywords: Real-time prediction; Machine learning; Inundation simulation; Water level monitoring system
Abstract: This paper presents a real-time machine learning method for forecasting water levels in sewers. A physically-based numerical model is usually used to calculate water levels. However, calculations are time-consuming. This is a serious issue for real-time sewer management in emergencies. We resolve this issue using machine learning. By comparison, numerical simulations, calculation time and accuracy were improved. Prediction with machine learning took no longer than5 seconds for normally one hour of calculation, resulting in a practical forecasting method.
Year: 2017