A Numerical Analysis of Pressure Pulsation Characteristics Induced by Unsteady Blood Flow in a Bileaflet Mechanical Heart Valve
Round 1
Reviewer 1 Report
1) The authors should use their institutional emails, not a qq.com
2) While the use of Newtonian properties is common when simulating the blood flow, many other groups argue that non-Newtonian properties are more accurate. At least discuss why you chose Newtonian in your discussion section.
3) On row 103, make a space between cases and [17,18]
4) Which opening angle was used for validation in Figure 3?
5) In Figure 3(b) also include the computational curve from the cited experimental and computational study in [22]
6) In Caption of Figure 3, cite [22], too.
Author Response
Dear Reviewer:
Thank you for your comments concerning our manuscript (ID: 477759). Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with your approval. The specific response are as follows:
Point 1: The authors should use their institutional emails, not a qq.com
Response 1: We are very sorry for our negligence of this problem. We've corrected the authors’-emails in the text. The specific changes are as follows:
Corresponding author Shu-xun Li 's email is replaced by lishuxun@lut.cn;
Author Xiao-gang Xu 's email is replaced by xiaogang@lut.edu.cn;
Author Tai-yu Liu's email is replaced by lty0108@126.com.
Thank you for your kind reminding.
Point 2: While the use of Newtonian properties is common when simulating the blood flow, many other groups argue that non-Newtonian properties are more accurate. At least discuss why you chose Newtonian in your discussion section.
Response 2: Thank you for your valuable advice. As you pointed, blood is actually a non-Newtonian fluid. The results are more accurate by using non-Newtonian properties on the study of hemolysis and blood damage. Meanwhile, previous study results showed that Newtonian properties could meet the accuracy requirement, especially in the flow field of velocity and pressure distribution. In some of the experimental studies, a mixture of 37% glycerol, 48% water and 15% NaCl was used, which could be consider as Newtonian fluid.
Considering the main objective of this study is the velocity and pressure distribution near the leaflets, we adopt the Newtonian properties to simplified the calculation.
We have discussed the reason for using Newtonian in the text, thank you for your guidance.
Point 3: On row 103, make a space between cases and [17,18]
Response 3: We are very sorry for this writing error. Thank you for your valuable advice, we have corrected it in the revised manuscript on row 108.
Point 4: Which opening angle was used for validation in Figure 3?
Response 4: The opening angle of 85° was used for validation in Figure 3. We are very sorry for that, we have clarified it in the revised manuscript. Thank you.
Point 5: In Figure 3(b) also include the computational curve from the cited experimental and computational study in [22]
Response 5: We are very sorry for this error, we have cited [15,22] in Figure 3(b). Thank you for your valuable advice.
Point 6: In Caption of Figure 3, cite [22], too.
Response 6: We are very sorry for this error, we have cited [15,22] in Caption of Figure 3. Thank you for your valuable advice.
Thank you very much again for your kind reminding and great help. If there are still any problems in the revised manuscript, please give us the opportunity to communicate with you again.
Kind regards,
All authors.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Summary
In this study by Xu and coworkers, numerical simulations of the unsteady flow through a bileaflet mechanical heart valve (BMHV) under different flow rates and leaflets fully opening angles are presented.
If the reviewer understands well, the aim of the study is to investigate the impact that pressure fluctuations related to turbulence have in the flow field, because pressure fluctuations could be associated with leaflets vibration, that could bring to valve failure in the long perspective.
The Authors use a RANS approach to model turbulence, and the quantity coefficient of pressure pulsation to analyze the impact of different flow regimes (N=5) and leaflets opening angle (N=3).
The Authors report that: (1) the level of pressure pulsation is highly influenced by velocity distribution, with higher coefficient of pressure pulsation associated with lower flow velocity along the main flow direction; (2) pressure pulsation is higher near the trailing edges than near the leading edges of leaflets; (3) based on the obtained results, setting of leaflets fully opening angle at about 80° is recommended.
Overall, there is a commendable intention at the basis of this study, and it would be of interest to the PROCESSES community. However, the manuscript cannot be published as it is, in the opinion of this Reviewer.
Per the points below, I do have some major reservations about some methodological flaws that limit the impact of the presented results.
Major Comments
– the Authors simulate only part of the cardiac cycle, with leaflets in fixed positions. This choice does not allow to properly model the whole valve kinematics. Moreover, it is not clear how, being the leaflets modeled in fixed position, pressure instabilities could give reason of the phenomenon of leaflets fluttering which, in turn, could locally modify the flow field. The Authors should discuss in depth this issue.
– It is unclear to this Reviewer why the Authors call pressure pulsation what is due to turbulent flow instabilities. Please clarify.
– The Authors should give proof of evidence that the k-ω model is appropriate for the fluid dynamics through the valve, in particular at the lower flow regimes investigated. It is well known that the k-ω model is not very effective to model transitional flows (See, e.g., Gallo et al J Biomech 2014). Please clarify.
– In eq. 3, only pressure is presented in the form of variable decomposition. Please add also the velocity.
– As for the geometry of the aortic root, is this last modelled as an axial-symmetric expansion, or are the three sinuses of Valsalva modelled with a 120 degree angle symmetry (see, e.g., Grigioni et al. Asaio 2005)? Please clarify.
– Please add to the text in the methods explicitly all the flow rates investigated and in figure panels which flow rate they refer to.
– The use of the term “validation” in Figure 3 is debatable. The Authors should clarify which flow rate they are comparing with experimental results. Moreover, they should also clarify what they mean whit “the maximum deviation is within the permitted error range”. From the comparison with experimental results in Figure 3, it also clearly emerges that the RANS model does not replicate properly the turbulent mixing in the regions between the three jets. The Authors should comment on this.
– Eq 5 is unclear to this Reviewer. Which is the delta T the Authors are considering for calculations?
– Data related to 25 l/min should be also presented, in figs 8 and 9.
– The limitations of the study should be discussed more in depth, in particular with respect to the fact that leaflets are in fixed position, and with respect to the choice of the RANS k- ω model.
Author Response
Dear Reviewer:
Thank you for your comments concerning our manuscript (ID: 477759). These comments are
all valuable and very helpful for revising and improving our paper, as well as the important
guiding significance to our researches. We have studied comments carefully and have made
correction which we hope meet with your approval. The specific response are as follows:
Point 1: The Authors simulate only part of the cardiac cycle, with leaflets in fixed positions.
This choice does not allow to properly model the whole valve kinematics. Moreover, it is not
clear how, being the leaflets modeled in fixed position, pressure instabilities could give reason
of the phenomenon of leaflets fluttering which, in turn, could locally modify the flow field.
The Authors should discuss in depth this issue.
Response 1: We are very sorry for this point. We should discuss it in depth. In this study, we
simulated only part of the cardiac cycle, with leaflets in fixed positions. The reasons are as
follows:
In the previous experimental study, the phenomenon of leaflets fluttering was only
observed when the leaflets in the fully open positions. The total time of the leaflets in the
fully open positions is from the 200thms to the 500thms in one cardiac cycle (see, Fig.2).
Thus, we simulate only the blood flow of leaflets in fully open positions to simplify the
calculation.
Additionally, as you pointed, only the simulation of the whole cardiac cycle can
properly model the whole valve kinematics. Because the influence of the change of pressure
and velocity boundary conditions to the phenomenon of leaflets fluttering are not clear at the
present time. At the first, we have studied the mechanism and characteristics of pressure
pulsation induced by unsteady blood flow, with leaflets in fully open positions. In the next
step, we will simulate the whole cardiac cycle and model the whole valve kinematics by
moving mesh technique in the future work.
We have discussed in depth this issue in the revised manuscript on row 94, 288. Thank
you for your valuable advice.
Point 2: It is unclear to this Reviewer why the Authors call pressure pulsation what is due to
turbulent flow instabilities. Please clarify.
Response 2: We are very sorry for this unclear point. This study mainly focused on the
mechanism and characteristics of pressure pulsation induced by unsteady blood flow. As you
pointed, the pressure pulsation is actually induced by the turbulent flow instabilities. In our
humble opinion, the meaning of turbulent flow instabilities is broader than pressure pulsation.
So, we adopt the pressure pulsation to decribe the pressure instabilities of turbulent flow near
the leaflets. We have added explanation in the revised manuscript on row 14, 40. Thank you
for your valuable advice.
Point 3:
The Authors should give proof of evidence that the k-ω model is appropriate for the
fluid dynamics through the valve, in particular at the lower flow regimes investigated. It is well
known that the k-ω model is not very effective to model transitional flows (See, e.g., Gallo et
al J Biomech 2014). Please clarify.
Response 3: We are very sorry for this point. we didn’t give enough proof of evidence that
the k-ω model is appropriate for this study. We have added the comparisons of the velocity
profiles between the simulation results by three different turbulent models (standard k-ε,
RNG k-ε, k-ω) and the previous experiment result (see, Fig.3). Under the low Re conditions,
the comparisons showed that the results calculated by the k-ω model, in particular the
velocity distribution near the leaflets, were in good agreement with the experimental data.
Considering the main objective of this study is the flow field near the leaflets and calculation
time, we chose the k-ω model for simulation. We have clarified this point in the text on row
124-134. Thank you for your valuable advice.
Point 4:
In eq. 3, only pressure is presented in the form of variable decomposition. Please add
also the velocity.
Response 4: Thank you for your valuable advice. We have added the velocity, as eq. 4
shown in the revised manuscript.
Point 5:
As for the geometry of the aortic root, is this last modelled as an axial-symmetric
expansion, or are the three sinuses of Valsalva modelled with a 120 degree angle symmetry
(see, e.g., Grigioni et al. Asaio 2005)? Please clarify.
Response 5: We are very sorry for this unclear point. A three-dimensional geometry of St.
Jude bileaflet mechanical heart valve investigated in this research, which was chosen to be
similar to previous studies (see, e.g., Shahriari et al. J. Biomech 2012). The geometry of the
aortic root is modelled as an axial-symmetric expansion, we have clarified it in the revised
manuscript on row 96. Thank you for your valuable advice.
Point 6:
Please add to the text in the methods explicitly all the flow rates investigated and in
figure panels which flow rate they refer to.
Response 6: Thank you for your valuable advice. We have added the flow rate conditions in
figure panels and manuscript on row 141.
Point 7:
(1) The use of the term “validation” in Figure 3 is debatable. (2) The Authors should
clarify which flow rate they are comparing with experimental results. (3) Moreover, they
should also clarify what they mean with “the maximum deviation is within the permitted error
range”. (4) From the comparison with experimental results in Figure 3, it also clearly emerges
that the RANS model does not replicate properly the turbulent mixing in the regions between
the three jets. The Authors should comment on this.
Response 7: We are very sorry for these problems. The correction are as follows:
(1) We have changed the caption of Figure 3 with “Comparisons of the velocity profiles
between the different turbulent models and the experiment result”.
(2) The flow rate condition of 25L/min was chosen for comparing with experimental results.
We have added details in the revised manuscript on row 126.
(3) The meaning of “the maximum deviation is within the permitted error range” is not clear
enough, we have deleted it. Additionally, more discussions have been added in 2.5 section.
(4) As the Response 3 above, we have added the comparisons of the velocity profiles
between the simulation results by three different turbulent models (standard k-ε, RNG k-ε, k-
ω) and the previous experiment result (see, Fig.3). Under the low-Reynolds conditions, the
comparisons showed that the results calculated by the k-ω model, in particular the velocity
distribution near the leaflets, were in good agreement with the experimental data.
Considering the main objective of this study is the flow field near the leaflets, we chose the
k-ω model for simulation. We have added discussion in the revised manuscript in 2.5 section,
and also, we have discussed the limitation of this choice in the conclusion section.
Thank you for your valuable advice.
Point 8:
Eq. 5 is unclear to this Reviewer. Which is the delta T the Authors are considering
for calculations?
Response 8: We are very sorry for our unclear presentation. The delta T in Eq. 5 is the time
of one calculation cycle of 300ms, from the 200thms to the 500thms in one cardiac cycle. We
have added more details in the revised manuscript on row 168.
Point 9:
Data related to 25 l/min should be also presented, in figs 8 and 9.
Response 9: Thank you for your valuable advice. We have added the flow rate of 25 L/min
in figure 8 and 9.
Point 10:
The limitations of the study should be discussed more in depth, in particular with
respect to the fact that leaflets are in fixed position, and with respect to the choice of the RANS
k- ω model.
Response 10: Thank you for your valuable advice. We have added the discussion on the
limitations of this study in the revised manuscript on row 288-293, including the leaflets in
fixed position and the choice of the k- ω model. We have discussed more in depth of the
reasons and potential influence of our choice, as well as the next plan in future work.
Thank you very much again for your kind reminding and great help. If there are still any
problems in the revised manuscript, please give us the opportunity to communicate with you
again.
Kind regards,
All authors.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The Authors have satisfactorily answered to almost all my comments, clearly reporting the limitations of the proposed approach and the impact they have on the presented results. I have not major concerns against the publication of this study.
