Geometric optimization of a heart assist pump
2024
0 views
0 downloads
Advisor: Prof. Rafet Yapıcı
Abstract (EN)
Due to the problems encountered in clinical applications of rotodynamic heart support pumps, which are designed with traditional methods to be used for the treatment of heart failure patients, there is a need for improvement studies in their design. In these improvement studies, efforts are made to reduce efficiency and damage to red blood cells. In this study, it was aimed to improve a centrifugal heart assist pump designed with traditional methods in terms of hemolysis index and hydraulic efficiency. For this purpose, a study was conducted on hemolysis index estimation methods on centrifugal blood pumps. The coefficient sets to be used in the Power-Law model with the Euler approach, which is widely used in hemolysis index estimation, and viscosity definition and turbulence models to be used in computational fluid dynamics (CFD) were compared. As a result of these comparison studies, the coefficient set, viscosity definition, and turbulence model that gave the closest match to the experimental results were determined and applied in the following stages of the study. In the optimization study, 7 geometric parameters that were considered to have the most intense effect on hydraulic efficiency and hemolysis index were selected, and the search region was formed by determining the lower and upper limits of these geometric parameters. Within this search area, 183 sample pumps were formed and CFD simulations were performed. The adaptive neuro-fuzzy inference system (ANFIS) model of the pump was obtained by using the hemolysis index and hydraulic efficiency results obtained from CFD simulations. The optimization study was carried out with a genetic algorithm (GA) using these ANFIS models. The pump model with optimum dimensions found by GA was simulated with ANSYS Fluent and the obtained results were compared with the simulation results of the basic pump. As a result of the optimization, it was seen that the basic pump was improved by 3.9% in terms of hemolysis index and 8.7% in terms of hydraulic efficiency. The optimum pump and the basic pump were also compared out of the design rotational speed. At all rotational speeds tested, it was observed that the optimum pump was better than the base pump in terms of both hemolysis index, hydraulic efficiency, and pressure difference, at levels ranging from 1.7% to 8.7%. Likewise, comparisons were made out of the design flow rate, and it was determined that the optimum pump had higher performance at levels ranging from 1.7% to 8.7% at each flow rate tested. The impeller versions of the optimum pump with three different shroud types: curved, conical, and flat were simulated with CFD and compared. In these trials, the version with a conical shroud impeller showed 2.8% higher performance in terms of hemolysis index and 2.7% higher performance in terms of hydraulic efficiency than the version with a curved shroud impeller.
Author
Dr. Ömer İncebay
How to Cite
Ömer İncebay (Doctorate thesis). Geometric optimization of a heart assist pump, 2024, Konya Technical University.
Keywords
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Konya Technical University
- Numerical and experimental in vestigation of optimization of Pelton turbine rotor design parameters in micro turbine size(2018)
- Comparison of some manufacturing costs according to various analysis parameters and other regulations of reinforced concrete structures with different floor systems(2018)
- The use of silica fume in self-compacting concretes affects the concrete compressive strength and adherence(2018)
- Load-bearing carrier system properties in the historical buildings repair and strengthening techniques for damages model analysis of Zenburi masjid(2018)
- Lateral rigidity improvement of deficient reinforced concrete structures with the use of user friendly systems(2018)
- Application of artificial intelligence methods to estimate monthly pan evaporation using meteorological data(2018)
