Application of soft computing techniques to estimate the scouring depth formed by crossing jets

Area of Science:

• Hydraulic Engineering
• Computational Fluid Dynamics
• Soft Computing Applications

Background:

• Scouring depth from dam water jets is critical for structural design.
• Laboratory studies exist, but soft computing offers a novel approach for modeling this phenomenon.
• Understanding scour from symmetrical crossing jets is essential for dam safety.

Purpose of the Study:

• To explore soft computing techniques for estimating plunge pool scouring depth.
• To model scour caused by symmetrical crossing water jets.
• To identify key parameters influencing scour depth.

Main Methods:

• Investigated various soft computing techniques for scour depth prediction.
• Utilized parameters including densimetric Froude number, tailwater depth, and jet angles.
• Employed the Multi-Layer Perceptron (MLP) model for prediction and analysis.

Main Results:

• The Multi-Layer Perceptron (MLP) model exhibited the best performance in predicting scour depth.
• MLP model achieved a Pearson correlation coefficient of 0.9527 and RMSE of 0.9039.
• Sensitivity analysis indicated tailwater depth has the highest impact, while vertical jet angle has the lowest.

Conclusions:

• Soft computing, particularly MLP, is effective for modeling and predicting scour depth from dam jets.
• Accurate prediction of scour depth is vital for dam design and maintenance.
• Tailwater depth and vertical jet angle are significant factors influencing scour dynamics.