Flood Discharge Estimation with Swarm Based Algorithm Models


GÜNAY A., Kumantas M., Kayhan A., ANILAN T.

JOURNAL OF POLYTECHNIC-POLITEKNIK DERGISI, 2024 (ESCI) identifier

Abstract

Accurate flood discharge estimation is critical for the economical, and safe design of hydraulic structures. In this study, various models were developed to estimate the maximum flow rate based on the flood values of the flow observation stations in the Eastern Black Sea Basin, Trabzon province. In the analysis, the area and elevation data of these stations were used. Data from 16 stations in Trabzon, with streamflow records ranging from 9 to 42 years, were utilized in the analyses. To predict future maximum discharge, the study employed not only classical regression (CRA) but also artificial bee colony (ABC) and teaching-learning based optimization (TLBO) algorithms. These algorithms optimized multiple linear regression, hyperbolic, and exponential regression functions. In the modeling for the future maximum flow rate forecasts, the error values of the hyperbolic regression function optimized with TLBO were lower. This reveals that the TLBO performs better than the CRA and ABC methods. Therefore, using the hyperbolic regression model optimized with TLBO for maximum flow rate estimation where there are no measurements for Trabzon province is recommended.