Perpustakaan Digital ITB

Centrifugal pumps are critical components in various industrial applications, including air supply systems, power generation, and the oil and gas industry. The performance of a pump largely depends on the design of the impeller, which serve to increase the fluid energy. In this final project, a multi-objective optimization of the impeller and pump casing design of the Ebara EZ-200 originally redesigned by Philip Gunawan is carried out using the Non-dominated Sorting Genetic Algorithm II (NSGA-II). The optimization aims to simultaneously enhance hydraulic efficiency and pump head by varying several geometric parameters of the impeller. The research methodology involves Computational Fluid Dynamics (CFD) simulations using ANSYS CFX to evaluate the baseline performance of Philipe’s design, followed by a geometry optimization process using NSGA-II to produce optimal designs represented by a Pareto front. The design parameters considered include the outer diameter of the impeller, outlet blade angle, and outlet channel width. The optimization results are then compared to the initial design to assess improvements in both efficiency and pressure ratio. Based on the research results, multi-objective optimization of the centrifugal pump using the NSGA-II algorithm successfully improved hydraulic performance compared to the baseline design. The selected optimal design achieved an increase in pump head of 3.44% and an improvement in hydraulic efficiency of 0.76% relative to the baseline configuration. These results demonstrate that genetic algorithm-based optimization can effectively enhance pump performance without requiring significant modifications to the original design configuration.