Numerical and sensitivity analysis of hydraulic characteristics of triangular labyrinth side weir

Triangular labyrinth side weirs have significantly more discharge capacity than traditional nonlinear weirs, and the complex hydraulic parameter interaction mechanism has been the research focus. This study used Computational Fluid Dynamics (CFD) to analyze the side weir's flow characteristics. Then, the Bayesian optimization algorithm and Extreme Learning Machine (BELM) developed a prediction model for the side weir's discharge coefficient. Finally, Sobol's method performed a sensitivity analysis for hydraulic parameters. The results show that the main channel's streamline is evenly distributed and begins to shift when it is close to the side weir. The overflow front is increasing and secondary flow also increases. BELM's Mean Absolute Percentage Error and Root Mean Square Error are 8.793 % and 0.455 in the testing stage, respectively, declined by about 56.24 % and 32.29 % compared with ELM; Froude number F_r, weir crest angle θ and the ratio of overflow front length to weir head l/h₁ are the most critical hydraulic parameters affecting the discharge coefficient, the global sensitivity coefficients are 0.4393, 0.4218 and 0.4152, respectively.