| In industry, mixing tanks are used to homogenize two or more different products that have been combined. This work investigates the use of graph-based evolutionary algorithms (GBEAs) coupled with computational fluid dynamics (CFD) to seek an enhanced design for a hydraulic mixing nozzle. This work also outlines a method for a numerical specification of a nozzle design and generation of a properly formulated CFD model to analyze it. The impact of choice of graph on the optimization process is examined using four different graphs: cycle, hypercube, toroid, and Peterson. Conclusions about the general type of nozzles designed by each graph type are included in order to identify the design strengths associated with each graph. This work provides insight into the considerations required to obtain more desirable performance from a mixing nozzle. Characteristics that influence jet distance and trajectory are reviewed. Some of these characteristics include the shape and size of horn used to guide the jet and the open entrainment area between the horn and the jet. The relative importance of each of these traits and which traits have the most significant impact on the quality of a given design are also explored. Suggestions for nozzle design are summarized. This information allows the most limiting factors of a tank mixing design to be mitigated to the largest extent possible.;This work illustrates that using magnification ratio or velocity alone does not result in an optimized design. These factors must be weighted to obtain a design that balances these factors for the geometry of fluid volume to be mixed. Additionally, this work shows that nozzle placement is perhaps more significant than nozzle design for optimum mixing with minimum power consumption. |
