Author(s): Wangjiayi Liu; Guanghua Guan

Linked Author(s):

Keywords: Canal; Digital twin; Prediction intervals; Real-time; State assessment

Abstract: Reliable operation of canal systems requires timely and accurate knowledge of their hydraulic state. In practice, however, operators often rely on sparse observations and deterministic models that provide limited information on uncertainty and evolving operational risk. To address this limitation, this study proposes a realtime safety quantification method within a Digital Twin framework for regulated canal systems. A hybrid physics–data-driven hydraulic model is first developed to generate probabilistic prediction intervals for key variables such as water level and discharge. Based on these intervals, a real-time hydraulic state index is constructed to quantify system safety in a continuous manner The index dynamically evaluates system behaviour by considering two aspects. It captures (i) the magnitude, duration, and temporal evolution of . deviations outside the prediction intervals, and (ii) the relative position and trend of observations within the intervals. Multiple hydraulic variables are integrated to improve sensitivity to emerging anomalies. Application to a real-world regulated canal shows that the proposed index effectively captures abnormal events, tracks their evolution over time, and provides an intuitive, risk-aware representation of operational status. The framework supports canal operation and scheduling by complementing conventional monitoring and early-warning approaches .

DOI: https://doi.org/10.64697/iahr.proc.hic2026.167

Year: 2026