Optimization of Vanes in Vortex-Settling Basins Based on Projection Pursuit Regression

To determine the optimal dimensions and attack angles for vanes installed in vortex settling basins (VSBs), the dimensionless length (L/Hc), dimensionless height (H/Hc), attack angle (θ) of the vanes and dimensionless inflow discharge (QI) were selected as influencing factors in this study, and orthogonal tests were conducted. Based on the experimental results, projection pursuit regression (PPR) was used to separately establish predictive models for sediment deposition ratio, water abstraction ratio, and sediment trapping efficiency, and their stability was validated. Subsequently, the nondominated sorting Genetic Algorithm II (NSGA-II) was employed to obtain Pareto fronts under different QI values, aiming for low sediment deposition ratio, low water abstraction ratio, and high sediment trapping efficiency. Finally, the technique for order preference by similarity to ideal solution (TOPSIS) method was used to determine the optimal vane dimensions and attack angles for each QI value: when 0.47<QI<0.9, the optimal parameters are L/Hc=0.6, H/Hc=0.5, and θ=0.17; when 0.9<QI<1.18, the optimal parameters are L/Hc=0.36, H/Hc=0.3, and θ=0.4. To verify the applicability of these optimized designs, validation experiments were conducted, demonstrating that the predictive models and optimized vane results are applicable to different vortex settling basins. These research findings provide guidance and a theoretical basis for the optimized design of vanes in vortex settling basin projects.