[aja1345]

Is solution converged?







Thanks

[-Maxim-]

Originally Posted by aja1345

Is solution converged?

You should change the range settings in your torque monitor so that you can see how much your torque value is still changing. In case your user points don't change a lot any more, it is worth checking out the results in Post.

As a next step, you read over the FAQ again:

Originally Posted by ghorrocks

This FAQ covers this topic: http://www.cfd-online.com/Wiki/Ansys...gence_criteria

Maybe change the time steps a bit. What is your current setting?

If you would like to simulate a turbulent flow correctly, you can use that steady-state result as a initial condition for your transient simulation now.

[aja1345]

Originally Posted by -Maxim-

You should change the range settings in your torque monitor so that you can see how much your torque value is still changing. In case your user points don't change a lot any more, it is worth checking out the results in Post.

As a next step, you read over the FAQ again:

Maybe change the time steps a bit. What is your current setting?

If you would like to simulate a turbulent flow correctly, you can use that steady-state result as a initial condition for your transient simulation now.

Thanks.

How do i set this case? I am new in CFX?

Result is good. But I think that it can be better. Yes i will change time step.

presently i want to simulate steady state.

[aja1345]

I studied the following case in FAQ:

Use Local Timescale Factor. A factor of about 5 is a good guess to start with. If this is successful you should run the final few iterations to convergence with a physical timescale (not local timescale all the way to convergence).

What is the exact meaning of bold sentences?

Can't I run with a local timescale factor to end?

[-Maxim-]

right click on your monitor - monitor properties - range settings - set manual scale (linear).
Since you are new, you should try to figure those little things out yourself. Click through all the settings over and over again - that way you will learn more quickly than us telling you every step on the way.

What is your current time step setting? Auto time scale = 1?

I wouldn't touch local and physical timescales as a beginner. For steady-state auto timescale is usually good enough.

[aja1345]

Originally Posted by -Maxim-

right click on your monitor - monitor properties - range settings - set manual scale (linear).
Since you are new, you should try to figure those little things out yourself. Click through all the settings over and over again - that way you will learn more quickly than us telling you every step on the way.

What is your current time step setting? Auto time scale = 1?

I wouldn't touch local and physical timescales as a beginner. For steady-state auto timescale is usually good enough.

Thanks.

I use physical timescale=0.0005.....

[aja1345]

Should I decrease physical timescale or increase?

I think that I should decrease it.

[ghorrocks]

If the convergence is jumpy then decrease it. If the convergence is steady but slow then increase it. If the convergence is steady and rapid then leave it.

Also if the convergence proceeds for a while then flat-lines then try increasing it (but read the FAQ on further details for what to do here).

[aja1345]

I decrease physical timescale to 0.00005....But residuals are as follows:

[aja1345]

Thanks.

on the suction and pressure side is almost 3 and it is ok.

But on the tip is almost 1. I think this value is not suitable. Right?

When i increases number of node on the tip, total number of mesh is much more than 1500,000 cells. and My Laptop is weak!!! number of 1400,000 cells is better for me. Have you any idea?

[ghorrocks]

Your convergence looks like it is converging for a while then flat lining. So then you should try increasing the time step.

If that does not work the FAQ describes the other options available to you.

You need to do a mesh sensitivity study to determine whether your near wall resolution is acceptable. The acceptable resolution changes for different simulation types.

If your simulation is too large for your computer then get a bigger computer, or network computers together. CFD is very computer intensive and always has been.

[aja1345]

With physical timescale=0.000275, Residuals are as follows:








I think that convergence is again related to mesh a bit. particularly tip gap mesh because on the tip is almost 1..what is your opinion?

[aja1345]

[aja1345]

[Opaque]

Whenever residuals remain constant, or oscillate around a level (mean value remains constant), it is a good practice to visualize the region where this value is located.

Post-process the region of interest to determine if the problem is a mesh quality issue, or a unresolved physics. Anything else will consume a lot of time of guessing what is next.

Hope the above helps,