Mechanism-guided optimization of a localized expansion elbow for reducing soybean particle-wall impact intensity in pneumatic conveying

Mechanical degradation of soybean particles during pneumatic conveying is a practical challenge, particularly at pipeline elbows where concentrated particle-wall impacts can cause kernel breakage and fines generation. However, how local elbow geometry modifies particle-wall impact behavior remains insufficiently understood. In this study, a localized-expansion elbow was developed and optimized using a coupled CFD-DEM framework combined with response surface methodology. Feed rate, expansion length, and expansion depth were simultaneously optimized by minimizing the time-averaged particle-wall normal impact force. The numerical model was evaluated using pressure-drop measurements and particle-transport visualization through transparent 3D-printed resin elbows. Simulated pressure drops agreed with experimental measurements within 5%, and the observed particle-transport patterns showed qualitative agreement with CFD-DEM predictions. The optimized condition, corresponding to a feed rate of 0.113 kg/s, an expansion length of 38.77 mm, and an expansion depth of 3.06 mm, reduced the average normal impact force by 30.5% compared with the original elbow at 0.1 kg/s. The optimized condition also weakened the concentrated high-velocity region near the outer wall and produced a more dispersed particle stream. More oblique particle-wall interactions were accompanied by a stronger rotational response, with the average particle angular velocity increasing from 13.55 to 16.65 to 22.50 rad/s. The average particle energy reached 0.00192 J. These results provide a practical basis for reducing severe particle-wall impacts during pneumatic conveying of fragile granular materials.