Data Assimilation of Mobile Sensors in Hydrological Models of Unsteady Flow

Abstract

Abstract In this paper, the estimation of the spatio-temporal variation of water bodies for state variables, velocity (m/s) and water surface elevation (m) for unsteady flows in open channels has been investigated. For data assimilation, average velocity measurements are obtained from mobile sensors such as Lagrangian sensors which have the ability to float passively in water bodies and provide their GPS location. One dimensional Saint-Venant equations are used for a system model linearized by a Taylor series expansion. To obtain a discrete-time state-space model, the coupled PDEs are discretized by Lax diffusive method in time and space. For state estimation of the open channel, a Kalman filter is set up with suitable filtering parameters for the channel’s model. Eulerian (fixed) sensors present at the head and tail of the canal provide the minimally required boundary conditions to run the model. The system is simulated using HEC-RAS simulation software. Water velocity profiles are used to predict the movement of the float, providing measurements for the Kalman Filter which is run in MATLAB. The estimated states are compared with actual values.