Verified turbulent case help #1685
-
|
I have been using AMR-Wind for about a year now and got great results for cases with uniform inflow. By that I mean I have been able to reproduce the results from other published papers. However, I'm trying to reproduce the results for a paper with uniform turbulence, so I can be sure everything is set up correctly for some turbulent tests we want to do. Here is the paper I'm trying to reproduce. Specifically, I'm interested in reproducing their case of a single turbine with 5% turbulence intensity in "greedy" mode. This seems feasible to reproduce since it's a NREL 5MW turbine operating in 9m/s turbulent flow at a constant RPM. Here are the best results I've been able to get: 1D Velocity profile comparison: It seems pretty close, but my advisor is concerned about the amount of spread my AMR-Wind wake has at 3 diameters downstream. I think the difference is due to differences between the turbulent flow I created and the turbulent flow used in the paper. Even though they have the same turbulence intensity, in the paper the flow was generated algorithmically whereas I generated the flow with an AMR-Wind spinup/precursor simulation. As far as I can tell, there is no feature in AMR-Wind for creating uniform turbulence. The best option I have found is to make a spinup file and have the velocity perturbations extend the entire height of the domain, run it for a bit, then check the turbulence intensity after. I've attached my spinup and turbine simulation files. If you're familiar with best practices for modeling turbulence in AMR-Wind, please let me know if there is something I could improve. Also, if you know of a better way to generate uniform turbulence, or know of a better case to verify my parameters against, please let me know. p.s. I have checked and double checked that all constants such as viscosity, RPM, velocity, epsilon, and cell size match between my simulations and the paper's. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
Hi, one option if you want to more fully control the inflow is that you could make up your own boundary planes. There's a demo script here: https://github.com/Exawind/amr-wind/blob/main/tools/generate_native_boundary_plane.py which shows how to generate a simple time varying Gaussian. You could hook up something to write out anything you want. |
Beta Was this translation helpful? Give feedback.
-
|
Awesome! I didn't know that script existed but yes I imagine I can use the output from OpenFAST's TurbSim module as an input for that script. That's even better because ultimately, I want to compare the performance of OpenFAST's OLAF to AMR-Wind and it would be great if I can guarantee the inflow matches perfectly, rather than just characteristics such as turbulence intensity. |
Beta Was this translation helpful? Give feedback.
Hi, one option if you want to more fully control the inflow is that you could make up your own boundary planes. There's a demo script here: https://github.com/Exawind/amr-wind/blob/main/tools/generate_native_boundary_plane.py which shows how to generate a simple time varying Gaussian. You could hook up something to write out anything you want.