Discharge Estimation for Different Bed Slope Conditions Using Entropy-Based Concept: An Experimental Investigation

Abstract

In hydraulic engineering, river discharge estimation is an important requirement for managing and planning the channel flow. Discharge measurements are of utmost importance for the purposes such as water availability analysis, reservoir operation, flood forecasting, designing of hydraulic structures, etc. In the discharge estimation process, the spatial velocity distribution in the transverse cross-section at the desired location of measurement is required. In traditional approaches such as Prandtl-Von Karman logarithmic law and power law, the velocity is employed deterministically, making its utility easier and providing accurate results for the wide channels only. In contrast, Shannon’s information entropy concept, which evaluates random variables probabilistically, is used in hydrology to determine entropy-based velocity distributions. In this approach, the velocity distributions obtained depends on a parameter called as entropy parameter, which is considered to be a fundamental measure of information about the channel characteristics such as channel bed slope and roughness. It provided better results for both the clear water and sediment-laden flow as compared to the former. In the present study, experiments for discharge estimation were performed on the experimental flume to collect the velocity data at different channel bed slope conditions to demonstrate the accuracy of the entropy-based concept. To prove the truthfulness of the entropy-based concept, the results were compared with the ones obtained from the classical method (velocity area method). Both the approaches have their respective advantages and limitations. Therefore, error analysis was necessary to check the efficiency and accuracy of the entropy-based model, which was performed by comparing the percentage error between the observed and computed discharge values. The final results revealed that the entropy model was a quick and accurate technique for discharge estimation as absolute percentage errors were less than 5% and the 95th percentile was 3%.