Investigation of leakage flow in centrifugal blood pump

Abstract

Miniaturization of third generation centrifugal blood pump is important as a smaller pump allows for a less invasive and complicated surgery. Furthermore, it will allow the blood pump to be fitted into people with small chests such as children and women. Therefore, it is important to study the leakage flow characteristics in smaller pumps to prevent haemolysis and thrombus formation. In this study, two smaller pumps of 40mm and 45mm diameter impeller is designed and fabricated to be tested alongside the original 50mm diameter impeller pump. The impeller width of the smaller pumps is increased to compensate for the lower volumetric efficiency. Leakage flow analysis is done by using both the impeller and dummy impeller. The impeller measures the pressure difference in the clearance gap (mean ΔP) of the real pump with regards to the operating flow rate. The dummy impeller is used to measure the mean ΔP with regards to leakage flow rate. This is achieved by operating the pump with the dummy impeller in a reverse flow condition. The reverse flow condition is created by an auxiliary pump unit and the flow rate measured from the outlet of the pump can then be interpreted as the total leakage flow rate. The results obtained from the impeller show that the mean ΔP slightly decreases with increasing operating flow rate. The results obtained from the dummy impeller show that there is a linear relationship between leakage flow and mean ΔP with a positive gradient. From the results of the impeller and dummy impeller, the leakage vs operating flow rate can be obtained by relating them at the similar mean ΔP. The results obtained shows that leakage flow slightly decreases with increasing operating flow rate and is maximum when operating flow rate is zero. The experimental results are then compared with the Computational Fluid Dynamics (CFD) simulation results. From the comparison, it is observed that the CFD simulation is able to predict the experimental leakage flow for the 0.2mm gap size fairly accurately. It is also observed that as the gap size increases, the error between CFD and experimental data increases. Lastly, it is noted that the CFD is unable to predict the leakage flow of the 0.5mm gap due to the large deviations obtained. This is attributed to the negligence of the inertia forces which is more dominant at large gap sizes. In conclusion, the method of measurement conducted in this experiment shows good agreement with the CFD simulations in the region of the normal operating flow rate (5L/min) for the 50mm pump with 0.2mm, 0.3mm and 0.4mm gap sizes. However, the 0.5mm gap size shows a large deviation between CFD and experimental method. Lastly, the large gap sizes of 0.4mm and 0.5mm shows low volumetric efficiency as their leakage flows are higher if not almost as high as the operating flow rate.