[waynezw0618]

hi everyone:

if i want to calculate a blade passage of turbomachine with a constant rotating speed.which is better?which will with less numerical error?

[cnsidero]

If it's only a rotor, SimpleSRFFoam is the most appropriate. 'SRF' stands for Single Reference Frame, hence the applicability for the rotor (i.e. one rotating reference frame).

However, if you are interested in both the rotor and stator then you will need MRFFoam. 'MRF' stands for Multiple Reference Frame, hence the applicability for the rotor passage (i.e. rotating reference frame) and the stator passage (i.e. stationary reference frame) - more than one reference frame.

Hope that helps.

[waynezw0618]

Hi Chris:
i know the meaning of MRF and SRF. what i want is the explain in the way of numerical error.
here is some thing from ANSYS CFX help

Alternate Rotation Model(Alternate Rotation Model in CFX is the MRF in OF)
The alternate rotation model is a model for the advection term in the momentum equations. Instead of advecting the relative frame velocity, the flow solver advects the absolute frame velocity.

The main advantage of this model comes into play when the absolute frame velocity is a constant, but the relative frame velocity has a high swirl component. In this case, advecting the relative frame velocity has a high component of error, while advecting a constant absolute frame velocity will have much less error. If the relative and absolute frame velocities are both changing equally, for example, inside a blade passage on a rotating machine, then the alternate rotation model has a similar level of numerical error compared to the default advection model.

The alternate rotation model makes a significant reduction in numerical error when the absolute frame flow is essentially a constant flow parallel to the axis of rotation. For example, the approach flow to a fan or propeller is nearly constant in the absolute frame, but highly rotating flow in the relative frame. At very large radii, small errors in the advection model of the relative frame flow become large errors in the computed flow in the absolute frame. When the alternate rotation model is used in this situation, the numerical error is greatly reduced, because the absolute frame velocity is close to constant. In fact, the numerical error in the advection model reduces to zero as the absolute frame flow becomes axially constant, independent of the radial location and rotation rate.

Conversely, the model can increase errors when, for example, the exit stream is highly swirling. Errors are magnified as the length of the inlet or outlet regions is increased. The choice of whether to apply the model therefore depends on the nature of the inlet and outlet flow conditions, and the length of the inlet and outlet regions.

It has been found that in most realistic flow situations, the model reduces numerical error, and/or does not increase numerical error. The one exception is solid body rotation, where the relative frame flow is constant (=0) and the absolute frame flow is non-zero and varying with radius.


but i still can not understand if i want to calculate the impeller passage which will cause the less numerical error