Nonlinear flood routing based on simulated annealing

Abstract

The Linear Muskingum model has been used most widely in comparison with other flood routing methods in hydrologic engineering. The basic model provides a forecasting form that may be used to predict the downstream flow, a finite time ahead on the basis of the current flow rates. However the shortfall of this method would be the linearity theorem which becomes restricted when the system exhibits a non-linear relationship between weighted flow and storage volume. This study aims to investigate higher forms of non-linear equations using an optimisation technique known as Simulated Annealing. The underlying logic is, the greater the nonlinearity, the higher the degree of freedoms and hence a better estimation of parameters. This then leads to a better outflow predictability. The result of the analysis clearly draws a distinction between usage of the more common non-liner forms and the complex non-linear forms.