A Real-Time Refined Roughness Estimation Framework for the Digital Twin Model Calibration of Irrigation Canal Systems

Abstract

Digital twin (DT) models can mirror irrigation canal systems and monitor the hydrodynamic processes in real-time to help create scheduling schemes. As for the DT model of the open channel, an important parameter that needs to be calibrated is Manning’s roughness coefficient (n). To establish a refined and high-fidelity DT model, the spatial variability of n along the longitudinal direction needs to be considered. Parameter optimization or identification method can estimate the values of n in different longitudinal segments along the canals. However, the existing relevant studies overlook the hydraulic conditions and estimation accuracy in canal segmentation. Therefore, this study proposes a comprehensive segmentation scheme for roughness estimation of irrigation canal systems. Particularly, a practical real-time segmented estimation (SE) framework using the ensemble Kalman filter (EnKF) is proposed and embedded into the DT model calibration. Verified by two canal reaches and two real-world cases, our results show that, compared with the empirical equation, the SE with the EnKF improves the model prediction accuracy by 45%–60%, especially for the canal reach longer than 10 km. This study provides a generic means for DT model calibration of irrigation canals, leading to more refined and precise monitoring and prediction of hydraulic variables.