Geometric Optimization of Jet Pump Used in Vacuum ‎Distillation Applications under Different Operating Conditions ‎using Genetic-algorithm Methods

Abstract

Genetic-algorithm methods are used here for single-objective (SO) and multi-objective (MO) geometrical optimizations of jet pumps used in vacuum distillation of ethanol, an application not deeply studied in scientific literature. These devices are particularly suitable to allow the azeotrope-breaking below the atmospheric pressure at ambient temperature. Based on this, different working pressures (Pp), five non-dimensional geometrical parameters that can influence the jet pump operation, and three performance parameters (drag coefficient, pressure recovery ratio and energy efficiency) are considered in this work. Furthermore, using a central composite, face-centered, enhanced experimental design, 89 simulation experiments are run to obtain Response Surfaces (RS) by genetic aggregation, applying afterwards the SOGA and MOGA optimization methods. Also, Spearman Rank-order correlation matrix is employed as initial screening, finding strongly negative correlation of drag coefficient and efficiency with the working pressure, Pp. Computational Fluid Dynamic (CFD) model is validated with other numerical and experimental works, obtaining satisfactory results. Additionally, the change of the optimized input and output parameters with Pp is studied, along with the behavior of Mach number. It can be concluded that the optimal nozzle parameters evidently influenced by Pp for the SO optimization are: outlet diameter and length of divergent part, conicity of convergent part, and ratio of inlet to throat area. For the MO optimization, changes of optimized geometrical parameters with Pp are negligible. In contrast, performance parameters are importantly influenced by Pp for all optimizations.

Keywords

• computational fluid dynamics
• jet pump performance
• geometrical optimization
• genetic-algorithm methods
• working ‎pressure‎