[ashtonJ]

Dear all,
Does anybody know how to calculate the oscillatory shear index (OSI) by ANSYS CFX, the OSI equation is attached.

Thank you.
Regards,
AshtonJ

[ghorrocks]

What are all the terms in the equation?

[ashtonJ]

OSI is equal to absolute of wall shear stress integral over time divided by integral of absolute wall shear stress over time.

[ghorrocks]

Is this evaluated at each wall node, or averaged somehow?

You will probably need to activate the transient statistics on the results tab in CFX-Pre to do this.

[Dave442]

Hi Ashkan,

Glenn - OSI is a non-dimensional scalar variable that is often used to characterise the oscillatory nature of vascular flows. It is calculated at each node and describes the transient deflection of the instantaneous WSS vector from the direction of the time-averaged WSS vector.

Ashkan - You can approximate both integrals using the trapezoidal rule and the instantaneous values of the WSS vector/components. From there it is straightforward to calculate the OSI. I did this in the past using a small python script but I'm sure you can do it in CFD post and/or any data manipulation package (Matlab, SciLab etc.)

Good luck,
Dave

[ghorrocks]

OK, thanks.

You might be able to do this using the transient statistics options in CFX. Then you will be able to have this as a variable accessible in CFD-Post like any other variable. So you should be able to do this inside the solver without any external post-processing.

[emily.imdieke]

Were you able to calculate it via Ansys?

[ashtonJ]

Yes.
In CFX-Pre, what you need to do is to calculate the transient statistics of wall shear stress (WSSX, WSSY, WSSZ) and absolute wall shear stress components. Then OSI can be calculated via the following formula:
OSI=0.5*(1-(trans_statistic(abs(WSSX))+trans_statistic(abs(WS SY))+trans_statistic(abs(WSSZ)))/(abs(trans_statistic(WSSX)+trans_statistic(WSSY)+t rans_statistic(WSSZ))))

[emily.imdieke]

Thanks for your response.
I have never seen the trans_static function used like that before. You may not know, but is this equation then not possible to implement just in CFD-Post using the WSS.Trnavg variables that you can set to solve for Pre?

[ashtonJ]

Yes, it needs to be done in CFD-Post. In CFX Pre you just save the arithmetic average of WSS and abs(WSS) components and then in CFD-Post you define the above-mentioned formula to find the OSI.

[emily.imdieke]

I have tried this and it has given me negative numbers for some reason. I am going to keep playing around with it, but do you know why this might be?

[ashtonJ]

OK. There is a mistake in the OSI equation, swap the numerator and the denominator of the OSI equation, then it should work ...

[emily.imdieke]

I finally got it to work in CFD-Post. I was trying to manipulate the data in Matlab as well to see if I could get it to run that way too and it didn't work, but doing it straight in CFD-Post is much easier.
Thanks for the help! I really appreciate it!

[Alfamaniac]

Dear all,

I've tried all the instructions above but it seems that i have a problem. When i'm changing the time step in CFD-Post the value of the OSI is changing too, which is not right. It suppose to be a time average WSS ratio, which means that has to has a fixed value. Any thoughts??Have i done anything wrong or I've just missing something?

Best regards,

Dimitris

[ghorrocks]

You don't set time step in CFD-Post so I do not understand what you mean. Also, time step does not appear in the equations shown above so I and now confused x2.

[Lance]

By changing time step in post you mean that you are looking at different .trn files I presume? The OSI calculation will change when looking at different results (in time), because your time-averaging is changing too. I guess OSI is defined over one single heart beat and not over a fraction of a heartbeat?

[Alfamaniac]

Originally Posted by ghorrocks

You don't set time step in CFD-Post so I do not understand what you mean. Also, time step does not appear in the equations shown above so I and now confused x2.

Originally Posted by Lance

By changing time step in post you mean that you are looking at different .trn files I presume? The OSI calculation will change when looking at different results (in time), because your time-averaging is changing too. I guess OSI is defined over one single heart beat and not over a fraction of a heartbeat?

Exactly!When i'm changing the time step in CFD-Post the value of OSI is changing too. But now i can understand that it's happening because we are looking on a different .trn file as you mentioned above.How it would be possible to calculate OSI for one single heart beat?Should I do anything differently in CFX-Pre or should I define there any expressions?The expression that i'm using for the OSI (in CFD-Post) is

0.5*(1-((abs((Wall Shear X.Trnavg)+(Wall Shear Y.Trnavg)+(Wall Shear Z.Trnavg)))/((abs(Wall Shear X.Trnavg))+(abs(Wall Shear Y.Trnavg))+(abs( Wall Shear Z.Trnavg))))).

[Lance]

Load the .trn file that is the last time step in the cardiac cycle. It will contain data that is time-averaged over one heart beat. But I guess you will need to run several cycles to get rid of initial value effects? Then delay the start of the averaging.

[r_abdollahi70]

Originally Posted by ashtonJ

Yes.
In CFX-Pre, what you need to do is to calculate the transient statistics of wall shear stress (WSSX, WSSY, WSSZ) and absolute wall shear stress components. Then OSI can be calculated via the following formula:
OSI=0.5*(1-(trans_statistic(abs(WSSX))+trans_statistic(abs(WS SY))+trans_statistic(abs(WSSZ)))/(abs(trans_statistic(WSSX)+trans_statistic(WSSY)+t rans_statistic(WSSZ))))

hi
I working with fluent. how can I use this function?
I can use cfd-post.

[ghorrocks]

No idea. Try the fluent forum.