Simulation of flood flow in a river system using artificial neural networks

Abstract

Artificial neural networks (ANNs) provide a quick and flexible means of developing flood flow simulation models. An important criterion for the wider applicability of the ANNs is the ability to generalise the events outside the range of training data sets. With respect to flood flow simulation, the ability to extrapolate beyond the range of calibrated data sets is of crucial importance. This study explores methods for improving generalisation of the ANNs using three different flood events data sets from the Neckar River in Germany. An ANN-based model is formulated to simulate flows at certain locations in the river reach, based on the flows at upstream locations. Network training data sets consist of time series of flows from observation stations. Simulated flows from a one-dimensional hydrodynamic numerical model are integrated for network training and validation, at a river section where no measurements are available. Network structures with different activation functions are considered for improving generalisation. The training algorithm involved backpropagation with the Levenberg-Marquardt approximation. The ability of the trained networks to extrapolate is assessed using flow data beyond the range of the training data sets. The results of this study indicate that the ANN in a suitable configuration can extend forecasting capability to a certain extent beyond the range of calibrated data sets.

Jordan Pi-Sigma Neural Network for Temperature Prediction

“…The MLP emulates the biological nervous systems by distributing computations to processing units termed neurons to perform computational tasks (Brath et al, 2002;Kröse & van der Smagt, 1996). The MLP have the ability to learn from input-output pairs (Güldal & Tongal, 2010) and are capable of solving highly nonlinear problems (Shrestha et al, 2005). The MLP with nonlinear transfer functions offer universal function approximation capability based entirely on the data itself, which theoretically mimic any relationship to any degree of precision (Hornik et al, 1989).…”

Flood Management, Flood Forecasting and Warning System

“…general structure of flood warning [12] physical process of flow such as filling and emptying the tanks, water deviation from channels, dams and … are also effective on flood problem. Flood warning technique in these catchments include forecasting precipitation before or collecting that in real time, converting forecasted precipitation or observation to flood, flood routing in river and reservoirs using one of hydraulic or regression methods and estimating flood in target points, following that, necessary information will be given [13].…”

Hydrological model coupling with ANNs

“…There are several publications on simulating hydraulic flow, e.g. Bobovic and Abbott (1997); Dibike (2002); Price et al (1998);Campolo et al (1999); Shrestha et al (2005). Duflow is based on the one-dimensional partial differential equations that describes non-stationary flow in open channels.…”