Trouble modeling wind turbine NREL Phase IV

Seervan · November 2, 2020, 11:53

I am having a hard time simulating NREL PHASE IV Wind Turbine. I use K-w turbulence model, tried polyhedral and trimmer messier, different sizes of prism layer thickness and numbers and even different turbulence models. I wan to get pressure coefficient over span-wise foil section and the power. To lower my Y+ I added extra prism layers but that resulted in much lower power calculation. For higher inlet wind speeds this just gets much worse so any help would be greatly appreciated.

fluid23 · November 6, 2020, 09:46

You mesh needs a lot of work. Starting with image 1, your growth rate away from the blade is too aggressive. I am also not sure that your domain is large enough, but its hard to tell. Also there is a prism mesh at your rotating region interface that is not needed.

Image 2 - lots more issues. Your surface mesh isn't refined enough and your leading edge is junk. Likewise, your trailing edge prism mesh totally breaks down. I see no wake refinement at all and a mesh that is generally too coarse to get results that can be validated.

Image 3 - the scale on your y+ graph makes it impossible to see what the distribution is. If you are using k-w and the near wall model then you need y+<=1. However, I think the default is an k-w SST with an all wall model that uses a blending function. With k-w SST/all wall model you can have 1<y+<100 (some say as high as 300), but you won't be capturing the viscous sublayer. if that is important to you. This also breaks down in the blending function, I think between 5<y+<50 it interpolates the near wall and high wall model to approximate a solution.

In general, pay attention to volume change, aspect ratio and prism layer retraction. These can all get you into trouble with your mesh.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
IDDES NREL wind turbine	mehran_ansary	FLUENT	0	September 22, 2017 06:10
CFD analysis on wind turbine rotor	Ken (Wind Turbine CFD Super Rookie)	Main CFD Forum	45	February 9, 2016 15:07
Negative lift in Modeling an airfoil design for wind turbine	Ram209e	Main CFD Forum	3	August 8, 2012 23:39
CFX Wind turbine mesh trouble and solver failure	tallliam	CFX	9	April 29, 2012 23:55
Basics in modeling a wind turbine	CFDbeginner	ANSYS Meshing & Geometry	0	October 21, 2010 15:27

November 6, 2020, 09:46		#2
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 18	You mesh needs a lot of work. Starting with image 1, your growth rate away from the blade is too aggressive. I am also not sure that your domain is large enough, but its hard to tell. Also there is a prism mesh at your rotating region interface that is not needed. Image 2 - lots more issues. Your surface mesh isn't refined enough and your leading edge is junk. Likewise, your trailing edge prism mesh totally breaks down. I see no wake refinement at all and a mesh that is generally too coarse to get results that can be validated. Image 3 - the scale on your y+ graph makes it impossible to see what the distribution is. If you are using k-w and the near wall model then you need y+<=1. However, I think the default is an k-w SST with an all wall model that uses a blending function. With k-w SST/all wall model you can have 1<y+<100 (some say as high as 300), but you won't be capturing the viscous sublayer. if that is important to you. This also breaks down in the blending function, I think between 5<y+<50 it interpolates the near wall and high wall model to approximate a solution. In general, pay attention to volume change, aspect ratio and prism layer retraction. These can all get you into trouble with your mesh.