مقايسه روشهاي رگرسيون آماري و فازي جهت برآورد بار رسوبي رودخانه هاي كارون و دز

Comparison of Statistical Regression and Fuzzy Methods for Estimating Sediment Load for Karoon and Dez Rivers

Estimation of rivers sediment load is one of the most important problems for design of hydraulic structures. Feasibility study of dam constructions, hydropower plants, intakes, bridges and the other hydraulic structures without investigations on river sediment transport is not possible. Existence of numerous empirical methods for estimation of river sediment load, a wide range of calibration coefficients shows that a suitable analytical or empirical method does not yet exist to accurately estimate the sediment load. Therefore, the measured discharges and sediment concentrations in hydrometry stations are statistical analyses for an accurate estimation of sediment loads in rivers. In usual statistical methods a power function is generally fitted on the data sets of flow and sediment discharge and thus the total sediment load could be calculated using this function. These methods are not able to recognize and separate the specific data measuring conditions. Therefore, they are not only able to accurately estimate the sediment load, but also cannot show the temporal variation of sediment loads. In this paper the measured suspended sediment load, at 10 hydrometry stations of Karoon and Dez rivers, which are the most important rivers in Khoozestan province, are analyzed using the usual statistical methods and the results are compared with the corresponding values obtained from the fuzzy method. This research has shown that the temporal variation of sediment loads can be analyzed using a fuzzy method. Also the results obtained using the fuzzy method in comparison with the corresponding values obtained using the usual statistical methods show a better correlation with the observed values. In all stations the fuzzy method estimated the sediment loads more than the ones obtained using statistical methods.