Using Nonlinear Programming for Flood Routing and Comparing the Results with those of Dynamic Wave Hydraulic and Muskingum Hydrological Methods

Development of precise and simple methods in flood simulation has greatly reduced financial damage and life loss. Various methods and procedures have been implemented based on Saint-Venant''s one-dimensional equation governing unsteady flows. To simplify the solution for these flows، analytical and numerical methods have been used. In the present study، a new method that provides the optimal outcome is introduced using non-linear programming. Penalty function has also been used to convert nonlinear programming (NLP) constrained problems into unconstrained optimal issues. To verify the accuracy of decision variables، the study covered 60 cross-sections of Gharasu River and 25-year flood hydrographs. After determining the model correctness، the 50 and 100-year flood hydrograph were routed in 18 Kilometers. The results were statistically compared with hydraulic and Muskingum hydrological methods. To sum up; the routed hydrographs introduced by NLP method were very close to the hydrographs produced by dynamic wave method. The R2 of calculated discharge of routed hydrograph by NLP and dynamic wave method were 0. 948، 0. 990، and 0. 989، respectively، with the return period of 25، 50 and 100-year flood being 0. 989. It can be concluded that NLP method is more accurate than Muskingum method، especially when predicting the peak discharge of flood hydrograph.