MODELING AND OPTIMIZATION OF CENTRIFUGAL PUMPS USING ANSYS FLUENT® AND GENETIC ALGORITHM ANALYSIS

In this research, optimization and design of fluid flow in a centrifugal pump is investigated and modeled using numerical analysis of computational fluid dynamics (CFD) and FLUID. First, geometry simulation is done using Gambit program and developing the geometry of the pump considered. After meshing fluid volume for mathematical analysis and then defining the volume of control in accordance with conditions, the simulated pump is ready to be implemented in ANSYS FLUENT (R). Then, the results obtained from FLUENT are investigated to optimize the pump performance in group inference using GMDH neural network model (Group Method of Data Handling). Finally, using polynomials related to efficiency and net positive suction head (NPSH) associated to geometric variables by parametric functions that are introduced as neural networks, multi-objective genetic algorithm is used to optimize centrifugal pump based on two objectives of increasing pump efficiency and decreasing NPSH, which according to Pareto Front, as one of the best approximations, will result in optimal design of the pump. (C) 2019 INT TRANS J ENG MANAG SCI TECH.