Discharge predicted in compound channels using adaptive neuro-fuzzy inference system (ANFIS)

Abstract

Some hydraulic structures and phenomena, including compound channels, must be studied in relation to open channel flow. Despite the fact that the primary channel and watersheds share a similar degree of roughness, estimating discharge in composite channels with mainstreams and flood plains has proved tricky. The flow discharge for a compound channel with different roughness in the primary and flood plain channels has been studied, and the results computed experimentally using horizontal division level have been compared with those predicted using dimensional analysis and an adaptive neuro-fuzzy inference system. The results show good agreement between experimental and numerical for discharge calculation according to root-mean-square error, MARE, R 2, SI, and Nash–Sutcliffe efficiency, with a percentage error not exceeding ±5%.