[amirhossein khoddami]

Hi,

I'm amirhossein khoddami,master's science student of mechanical engineering at Tehran university in Iran, and I'm working on my master thesis in Fluid mechanics and turbomachinary field.My project is an axial water turbine called Kaplan turbine that besides having main runner has a sub-runner in series with itself which identifies upstream flow characters from upstream main runner and uses them to adjust main runner blade angles through a control mechanism. I obtained blade angles in 3 radial spans for both runners through a design code in MATLAB software and designed the turbine blades geometry in ANSYS BLADEGEN software and meshed it in ANSYS TURBOGRID. Finally, I simulated my meshed geometry in ANSYS CFX solver. The run proceeded well and the solution converged to the prescribed residuals for the flow equations. But my problem is that obtained Head from cfd isn't equal to the pre-determined Head at the beginning of the design but close with it, although the net power and ratio of the two runners Head is approximately equal to its pre-determined value. I'll be glad to help me solve this issue.

Best regards

[ghorrocks]

Before you try to interpret the results of any CFD simulation you need to do some basic checks to ensure your result is accurate. Have you done any verification and validation? See https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

In particular mesh size, convergence tolerance and time step size are key parameters you need to show are adequate.

[amirhossein khoddami]

hello sir,

firstly thank you for your reply.no,i have not done any verification or validation in my work.because there is no or at least i did not find any similar work of two-runner kaplan turbine in the literature.

thank you very much if you have any and send it to me.

and the last point is that what should i do if i simulated that work and get similar result to this??

[ghorrocks]

You cannot look at the detailed results until you have shown that your results are accurate enough to be worth looking at.

I do not have any data on turbines, it is not my field. But you can do internal validations by comparing progressively finer meshes until the results asymptote out - this does not require external data at all. But comparing to external, trusted data is by far the best, so if you can get data on any turbine (rather than just your specific case) that would be good enough.

[amirhossein khoddami]

thank you again for your reply to me.

what you say is exactly what i had not cared about.thank you.but two other questions i has firstly is that i get strange and illogical result(the net head of the two runners is sum of heads of the two runners basically and regularly).now not only are the two heads obtained for both runners too different to what i had considered for both at first and naturally the net head is different either,but also the net head obtained is not equal to sum of the two heads in the simulation itself(not comparing to the design code)and this is a strange event.if i want to explain more clearer we face an incontinuity here because of heads.

and my second question is related to above and is that what if this incontinuity happens again even if i asymptote out by fining my mesh???

[ghorrocks]

I do not understand your question.

If you are getting unexpected results for the head at different locations please post an image of what you are modelling and where and how you are measuring the head.

[amirhossein khoddami]

hello sir,

how can i upload a screenshot in my laptop here??

i have taken some screenshots of cfx software and i want to share with you.

excuse me i ask you this.it is the first time i do this

thanks

[Gert-Jan]

https://www.cfd-online.com/Wiki/Ansy...n_the_forum.3F

[amirhossein khoddami]

hello sir,

these are the images.

thank you for your link

screenshot1.png[ATTACH]screenshot1.png

screenshot2.png

screenshot3.png[/ATTACH]

[Opaque]

You may need to explain the figures in more detail or add enough legends.

[amirhossein khoddami]

hello

these are my screenshots of cfx software screen.we have three domains here.the first is guide vane domain which is stationary.the next two domains are runner 1 and runner 2 domains that are rotary with an specified angular velocity.i monitored mass flow inlet,power and the runners head for convergence investigation in cfx solver.the main problem is that net head obtained from cfd is not equal to sum of the two heads in the simulation itself(not comparing to the design code)and this is a strange event.actually we face an incontinuity from cfx here.
and the net head is not equal to the pre-determined head at the design stage.

thank you for your help to solve the problem.