Optimization of Multistage Depressed Collectors

Abstract

Random walk and genetic algorithm techniques have been implemented in a 3-D collector simulation code to automate the design optimization of multistage depressed collectors. An axisymmetric four-stage collector and an asymmetric two-stage collector have been optimized following both approaches. Procedures for the implementation of these methods in any suitable code and simulated performance of the collectors are demonstrated. The results show significant improvement in the collector performances due to optimization. A comparison between the optimum collector performances that were obtained using these methods has also been carried out.