Optimization of hopper design by genetic algorithms

by

Optimization of hopper design by genetic algorithms

Yaowei Yu, Frank Pettersson, Henrik Saxén

Thermal and Flow Engineering Laboratory, Department of Chemical Engineering,Åbo Akademi University, Biskopsg. 8, FI-20500 Åbo, Finland.

DOI:

https://doi.org/10.7494/cmms.2011.1.0308

Abstract:

In the handling of particulate matters, hoppers are frequently used as intermediate storage, but during their filling and emptying, particle size segregation may occur. The hopper geometry is known to affect the outflow pattern (mass flow or funnel flow), and possible inserts in the hopper can also affect the patterns and particle segregation. The present work studies the size segregation in hopper by discrete element modeling (DEM). Due to the considerable computational effort required by the numerical technique, a factorial plan was applied to design a set of DEM experiments, where the insert geometry and position were varied. The results form the basis for a black-box modeling, where the outflow patterns were described by a neural network. Using the arising neural model, the geometry was optimized using genetic algorithms with respect to particle segregation of the outflow. The most promising solution was finally verified by DEM modeling. Thus, by the proposed method complex design problems can be tackled avoiding excessive computational burden.

Cite as:

Yu, Y., Pettersson, F., & Saxén, H. (2011). Optimization of hopper design by genetic algorithms. Computer Methods in Materials Science, 11(1), 28 – 33. https://doi.org/10.7494/cmms.2011.1.0308

Article (PDF):

Keywords:

Hybrid modeling, Particulate flow, Size segregation, Optimal design, Discrete element method, Neural networks, Genetic algorithms

References: