Tutorial: Generation of a high-quality mesh for CFD simulations of wind turbines (STAR-CCM+)
HTML-код
- Опубликовано: 11 сен 2022
- This video presents a step-by-step tutorial to set up a CFD simulation of a wind turbine using STAR-CCM+.
Here, we are gonna learn how to set up a URANS-based CFD model for the evaluation of aerodynamic behavior of wind turbines. In particular, a systematic methodology is presented for the generation of a CFD grid with a proper resolution, which provides a reasonable compromise between computational cost and accuracy.
Link to geometry:
github.com/MojtabaMaali/NREL-...
The simulation set up is performed in the following steps:
00:03:13 Definition of the Computational Domain
00:16:15 Definition of the Regions
00:17:20 Mesh Generation
00:45:38 Checking the Mesh Quality
00:49:21 Definition of Physics, Boundary Conditions and Solver Settings
01:00:58 Definition of Monitors and Stopping Criteria
01:10:31 Post-Processing
![Sean Rii, Karyon, Sharzkii - Taungule [My Love] (Official Music Video)](http://i.ytimg.com/vi/eJoMuypbSzQ/mqdefault.jpg)
Thank you for the very well explained tutorial
Great video! I love your tutorials!
Great video! Thank you!
Hi, very nice and descriptive tutorial. The request is to create a tutorial of the turbine with a moving reference frame tutorial.
So good, nice video
very good video !!!! but I wanna know why the Y+ that you choose is 200, and if the wind is parallel to the blade ,how the grid encryption area should be set up ?
Excelente vídeo, mt didático. Fiquei com dúvida em relação ao comprimento da blade, o valor segunda a NREL n seria 5.532 m ?
Olá Lucca, obrigado pelo comentário! O tamanho da pá é 5,029 m, você pode verificar aqui: www.nrel.gov/docs/fy04osti/34755.pdf
Att,
Mojtaba Maali.
Great Video! Sir Could you please make a video on Mesh independence study in starccm+ ( more specifically, how can we plot quantities such as Cl or Cd vs number of cells ) thank you !
Dear joseph ntumba,
Thank you for writing! Sure, I will consider making a video about this topic.
Respectfully,
Mojtaba Maali.
Great video! Are you able to share what the values within the Power report definition represent?
Dear Adam Robertson,
Thank you for writing! The power is the product of the torque and rotor angular velocity (rad/sec).
I hope this helps.
Respectfully,
Mojtaba Maali.
@@adamrobertson7472 First you need to convert the rpm to rad/sec:
rad/s = rpm*(2*pi/60)
Then you can calculate the power as follows:
Power = Torque*rad/s
Vídeo extremamente interessante, muito explicativo! Teria como disponibilizar o arquivo desta simulação, seja por google drive ou outro meio?
Prezado João,
Obrigado pelo comentário! Envie um e-mail para o endereço de e-mail que está na descrição do canal para que eu disponibilize o arquivo para você.
Att,
Mojtaba Maali.
Hi
your tutorial is great!!
To find the angle of attack we chould calculate it?
Dear Bassel Tabaja,
Thank you for writing. To calculate the angle of attack with CFD results, you can have a look at this paper: www.sciencedirect.com/science/article/pii/S0960148118303239
I hope it helps!
Respectfully,
Mojtaba Maali.
@@CFD_Mojtaba Thank you for your replay!
its just i want to calculate the drag and lift coefficient ,so i think i need the AoA to input the direction right?
Kind Regards
@@basseltabaja8838 In this case, you need to calculate the AoA along the blade span.
Respectfully,
Mojtaba Maali.
6:28 Can you provide a link to this paper please. I am doing part of this simulation (+ some other stuff) as a part of my master's degree.
You can use this link to access the paper: www.sciencedirect.com/science/article/pii/S0360544220318739
@@CFD_Mojtaba Thank you!
very quality information from your video as always, however I want to take a little bit more help from you if possible; I designed my own 2.2 m dia fixed pitch propeller and I want to calculate the perfomance such as I want to measure the Thrust & Torque values at selective RPMs, so how can I proceed, is it correct way to iteratively guess the rotation rate (RPM) each time repeatively in order to get thrust and torque values ? Waiting to hear from you soon.
Dear shayant nandi,
I recommend that you have a look at papers, where experimental methods are used to evaluate the wind turbine aerodynamics such as this paper: doi.org/10.1016/j.jweia.2018.03.001
Respectfully,
Mojtaba Maali.
@@CFD_Mojtaba I read the paper attached by you, thankyou for sharing that. Can you please tell me a rough guess on base size for my 2.2 m dia propeller for coarser mesh settings, I am trying to run coarse mesh on 32 GB RAM but memory is getting failed. Need your help sir!!
Good afternoon! Such a problem when modeling a wind turbine. In front of the wind turbine, at the beginning of the calculation area, a wall is being built, although there was no wall at the beginning. The input type is the velocity inlet, the speed profile is tabulated. To observe the vortices behind the wind turbine, an isosurface is used, I suspect that it is the case, but I can't figure it out. I change the parts in the isosurface, but this area appears like a wall in front of the wind turbine. Please tell me what I'm doing wrong. I will be grateful for the answer
Dear Marina,
This problem can be due to a coarse mesh at the region close to the ground (as you are using tabulated data) and interface between the two domains. This can also be associated with the quality of mesh. You need to check the mesh quality based on the methodology presented here: ruclips.net/video/ALGnQFLlrkg/видео.html
The size of time step is another important parameter that you need to define carefully.
I hope this helps!
Respectfully,
Mojtaba Maali.
Hi! Do you recommend any literature that investigate this rotating domains dimensions?
Hello, you can find some information in this paper: www.sciencedirect.com/science/article/abs/pii/S0360544217312033
@@CFD_Mojtaba thank you!!
hello, can you help me for XY plot in start CCM+
Of course, Rachid.
Hi, the rotation rate is given under specific income wind speed. And now, if I consider the wind profile with arbitrary magnitude and direct, or I consider the wake interaction of wind turbines. This method doesn't give the correct result.
Dear Thermal,
Why this strategy is used is twofold. First, reducing the computational cost. Further, using the obtained flow behavior to perform the fluid-structure interaction simulations, which are used to assess the structural behavior.
I hope this helps!
Respectfully,
Mojtaba Maali.
Hello, dear Friend! I wonder how to assign the rotation if we have two turbines? Do we need to create two coordinate systems and two origins, two rotations? Or we can use only one "motion" for two turbines?
Dear Mary,
Yes, you need to create two separate rotational motions for this purpose.
Respectfully,
Mojtaba Maali.
@@CFD_Mojtaba Thank you so much, I did it!
@@Maria-73 Great!
Let me know if you had any questions.
Respectfully,
Mojtaba Maali
@@CFD_Mojtaba Hello, Dear Mojtaba Maali! I have some interesting situation, when I simulated one wind turbine, I could observe the near wake behind turbine. But the wakes don't appear when I simulate wind farm, only slowing down the speed behind the turbines. The wind farm is made in scale, the other parameters of the model are configured similarly to the case with a single turbine? What reasons of this, or can it be correct results? Respectfully, Maria
Hi your simulations are very impressive. I need a suggestion on my simulations. I should make a DFBI simulation (with Star CCM+) in which the body turn around it self (arround y axis, obviously arround its center of mass) with a time dependent angual velocity. The problema is that with a DFBI simulation I can impose just the initial velocity of the body. Do you have an idea about how I can impose a time dependent angular velocity of the body?
Thank you for your interest in this video, You may need to impose this rotational motion using the rigid body option.
@@CFD_Mojtaba Can we chat (in some way) about this problem? Because it is not so easy.
1:10:50 It took six days to complete with 64gb of ram??
I need this to complete within a week with significantly less RAM. Here is my plan of action:
1. Change the inlet boundary to one diameter before the rotor instead of two and nine diameters to the pressure outlet versus fifteen as specified in the video.
2. Reduce the scale of the wind turbine blade down even further.
3. Change the time step from every one degree of rotation to every five degrees of rotation.
4. Change the max run time to an equivalent of two total rotations versus four total rotations.
5. I could make the simulation NONSPINNING and this would significantly cut down on the computational time, but then I would not be able to measure useful metrics like power.
What are your thoughts on this plan of action?
Dear Michael Carnahan,
The reason behind such a high computational cost is the refinement of the grid over both the wake and blade surface.
The best way to decrease the computational cost, may be the use of a steady state approach using a rotating refernece frame. This way, what you obtain is the steady-state solution rather than the temporal evolution of the variables.
In which document are the experimental results?
Data is available on the official website of the NREL.
@@CFD_Mojtaba
I tried to find it, but there is no specific data, only the approximate data in the picture, how did your data come from？
@@user-gi4jh2uo8n have a look at this link: forums.nrel.gov/c/wind-water/uae-phase-vi-data/24