Design of Flow Loop to Characterize Nuclei Contetn

Mechanical heart valves are the most common replacements for damaged heart valves as a result of heart valve disease. Cavitation, the rapid formation of vapor filled bubbles due to the local fluid pressure dropping below the liquid vapor pressure, has been detected during mechanical heart valve closure. Cavitation can cause hemolysis and initiate the clotting cascade resulting in thrombus formation. Part of the clot could break off and form an embolism that could be deadly. The objective of this study was to develop a method to characterize the number and size of nuclei, which are small gas bubbles or impurities that are present in fluids and are susceptible to cavitation. A flow loop was built with a cavitation susceptibility meter (CSM), a device that induces cavitation at known conditions. Two different methods to characterize fluids were studied: the root mean square (RMS) method and the wavelet method. The RMS method was initially thought to be a straightforward method of counting cavitation events, but difficulties were encountered when trying to trigger off cavitation events alone. All of the data were post processed using MATLAB^®. Similar problems were encountered with detecting each cavitation event. Other methods of characterizing the nuclei content were tested using the wavelet method. The RMS values and the time widths of each event were used to compare fluids. These methods gave useful insight into the nuclei content without knowing an accurate estimate of the number present. Counting events is not possible with the methods tested because of limitations of the CSM.

Heart valve disease is responsible for about 20,000 deaths each year [1]
275,000 heart valve replacements worldwide each year [2]
Mechanical heart valves are the most common type of replacement and they have been shown to cavitate in vitro [3] as can be seen in Figure 1
Cavitation can damage mechanical heart valves or initiate the clotting cascade, which can result in thrombus formation.
Nuclei are small gas bubbles and impurities present in fluids that are susceptible to cavitation
Knowing the number and size distribution of these nuclei can help to predict the likelihood of cavitation.
Important for evaluating and improving current mechanical heart valve designs

Need to design a flow loop that will induce cavitation in a controlled manner using a CSM so that the number and size distribution of the nuclei content can be determined
The CSM, which can be seen in Figure 2, consists of 3 main regions: the nozzle, the throat, and the diffuser
Cavitation inception occurs in the throat so it is important to know the throat velocity and the throat pressure in order to calculate the critical radius which is the minimum radius needed for cavitation inception
A diagram of the flow loop design is shown in Figure 3
Bubble collapse occurs in the diffuser, which transmits an acoustic pulse
An accelerometer was used to detect the acoustic signal of each event and can be seen in Figure 4

Each cavitation event produces several positive and negative spikes as can be seen in Figure 4, which makes it difficult to count each event. The accelerometer signal was post processed using the wavelet toolbox in MATLAB®. The raw accelerometer signal was decomposed based on frequencies and then reconstructed using wavelet coefficients and is shown in Figure 5. The signal was then band pass filtered (20 – 500 kHz) and denoised as is seen in Figure 6.

Figure 5: Reconstructed Raw Accelerometer Signal using Wavelet Toolbox in MATLAB^® Figure 6: Filtered and Denoised Accelereometer Signal

Difficulties were encountered in counting the number of cavitation events due to limitations of the CSM. One problem was the presence of noise as can be seen by the narrow spikes in Figure 6. Setting a threshold level was somewhat arbitrary and there was no clear way of doing it to include only cavitation events without any noise. Also, at high tension pressures, cavitation events occurred in rapid succession which caused the signal to overlap, making it difficult to separate individual events. Spot cavitation and choking also occurred at high flow rates which limited the working range that could be tested. These discrepancies are apparent in the cavitation event number vs. throat pressure graph as can be seen in Figure 7. The total number should increase with decreasing throat pressure, but this is not seen.

Using the average RMS values and average time width values of each event as seen in Figures 8 & 9 provide insight into nuclei content even though the size and number distribution could not be determined. The time ratio plot also provided some useful insight into the nuclei content of different fluids, although in a less quantitative manner than originally planned as can be seen in Figure 10.

Figure 7: Cavitation vs. Throat Pressure Graph of Different Fluids Figure 8: Average RMS vs. Throat Pressure Graph of Different Fluids

Figure 9: Average Width vs. Throat Pressure Graph of Different Fluids Figure 10: Time Ratio vs. Throat Pressure Graph of Different Fluids

[2] E. Rabkin and F. J. Schoen, “Cardiovascular tissue engineering,” Cardiovascular Pathology, vol. 11(6), pp. 305-317, 2002.

[3] L. A. Garrison, T. C. Lamson, S. Deutsch, D. B. Geselowitz, R. P. Gaumond, and H. M. Tarbell, “An in vitro Investigation of Prosthetic Heart Valve Cavitation in Blood,” Journal of Heart Valve Disease, vol. 3, pp. S8-S24, 1994.

[4] S. D. Chambers, “Determination of the InVivo Cavitation Nuclei Characteristics of Blood,” ASAIO Journal, pp. 541-549, 1999.