AI Summary & Key Takeaways
AI InsightThe results indicated that the analytical method achieved a correlation coefficient of approximately 0.87, whereas the ANFIS model provided superior predictive capability with a correlation coefficient of 0.92.
📌 3 Key Takeaways
Abstract
Accurate estimation of river discharge is a fundamental requirement for flood management, hydraulic design, and river engineering applications. Compound open channels provide a realistic representation of natural rivers due to the interaction between the main channel and adjacent floodplains. However, predicting discharge in compound channels through conventional analytical methods often involves solving complex nonlinear equations, which can be computationally demanding and time-consuming. To overcome these limitations, intelligent data-driven approaches have gained increasing attention in recent years. In this study, an Adaptive Neuro-Fuzzy Inference System (ANFIS) was employed to model and predict flow discharge in compound open channels. The performance of the proposed ANFIS model was evaluated and compared with conventional analytical approaches, including the divided channel method using horizontal interface lines. The results indicated that the analytical method achieved a correlation coefficient of approximately 0.87, whereas the ANFIS model provided superior predictive capability with a correlation coefficient of 0.92. Sensitivity analysis further revealed that the depth ratio, hydraulic radius ratio, and area ratio were the most influential parameters affecting discharge prediction. The findings demonstrate that the ANFIS approach offers a reliable and efficient alternative to traditional methods for estimating discharge in compound open channels and can be effectively applied to complex hydraulic systems.
Özet
Accurate estimation of river discharge is a fundamental requirement for flood management, hydraulic design, and river engineering applications. Compound open channels provide a realistic representation of natural rivers due to the interaction between the main channel and adjacent floodplains. However, predicting discharge in compound channels through conventional analytical methods often involves solving complex nonlinear equations, which can be computationally demanding and time-consuming. To overcome these limitations, intelligent data-driven approaches have gained increasing attention in recent years. In this study, an Adaptive Neuro-Fuzzy Inference System (ANFIS) was employed to model and predict flow discharge in compound open channels. The performance of the proposed ANFIS model was evaluated and compared with conventional analytical approaches, including the divided channel method using horizontal interface lines. The results indicated that the analytical method achieved a correlation coefficient of approximately 0.87, whereas the ANFIS model provided superior predictive capability with a correlation coefficient of 0.92. Sensitivity analysis further revealed that the depth ratio, hydraulic radius ratio, and area ratio were the most influential parameters affecting discharge prediction. The findings demonstrate that the ANFIS approach offers a reliable and efficient alternative to traditional methods for estimating discharge in compound open channels and can be effectively applied to complex hydraulic systems.