|
WINDMILLS
I AM TRYING TO ANALYZE A BLADE OF A WIND TURBINE USING FLUENT . COULD ANYONE GIVE ME SOME HINTS , OR TELL ME SOMETHING ABOUT HOW COULD I MESH CORRECTLY ... ? I AM A STUDENT OF INDUSTRIAL ENGENEERING IN VIGO , SPAIN , AND I AM MAKING MY FINAL PROJECT ABOUT THIS . MY INTENTION IS TO DO AN INVERSE DESING WITH XFOIL , OR EPPLERS CODE , AND THE ANALYSE THE 3D BLADE WITH FLUENT . I HAVENT WORK TOO MUCH WITH IT , BUT OF COURSE I HAVE SEEN ALL THE TUTORIALS , INCLUDING THE 3D FAN EXAMPLE . WHERE CAN I FOUND MORE INFORMATION ?
|
Re: WINDMILLS
hi
we are doing CFD for wind-turbines sine 1997 it's a hard job. first of all go into the literatur the best beeing the publications of the risoe/dtu group leadee by jens soerensen. i hope also that you are familiar with all classics starting from betz, glauert, wilson, tangler and many others. doing cfd the mesh has to have at leat 40 sections with round about 200 points on each section. most important is the turbulence-model in connection with a method for describing the laminar - turbulent transition. the simplest model is michael's all in all y+ has to be 2 with menter's modified ka omega regards peter schaffarczyk |
Re: WINDMILLS
thank you very much for your answer . have you worked with fluent in such analysis ? I havent start to analyze wind-turbines , so i am not sure to make it convergence . Could you give me some hints about the mesh and the parameters to use in Fluent ? My intention was to make an inverse design , and then analyse the 3-D blade whith Fluent . Do you think this could be achievable ?
|
Re: WINDMILLS
yes we worked with fluent. but to have spanwise resolved data for thrust tangential-force etc w ealso used CFX
for the mesh use the common C-O-topology. fluent also has only k-eps and spalart allmeras turbulenc models, may that this is not enough. 2nd order differencing schemas ara also mandatory. the "inverse design" is usually good for a 2d profile but my main expericence is only with NACA 634xx, DUxx FFA-W3 and Sxxx. do have to design a stall or pitch blade ? the last one is much easier. the answer wheather this would be achievable or not depends on the operating point . >>you can't compute what isn't understood<< pw anderson, physics nobel prize winner 1997 regrads peter schaffarczyk |
Re: WINDMILLS
That's a quite challenging CFD-problem !!
What is the Reynoldsnumber based on the blade's chord of your problem ? How large is the difference between calculations with transition and fully turbulent. It which way did you implement the Michel criterium in your turbulence model ? Did you get the best results with Menter SST model ? What are your experiences concerning the convergence behaviour of that model ? Thanks for your answer, Mark |
Re: WINDMILLS
the re-number denpends on T(ip)S(peed)R(atio). usually the power-prodyucing part of the blade is the outer half. her we have REs up to 10 millions.
depends on what is happening at he nose. we have seen for the naca 63212 quiet large differnces because without transtion there is no re-attaching leading-edge sep. bubble. see details in the risoe-reports in fortran i just cut the the k sources if his criterion is not reached. also here as well as everywhere the danish group has made better progress. i have to say that i have only tested the menter model for a 2d FFA-W3-211 at 12 deg aoa. i have the measurement from anders bjoerk. unfortunately i have not seen any improvement to k-eps. so i decides not to use it for 3d WT baldes convergence is so far no problem in the un-stalled region. all in all: anybody in this area knows about the shortcomings of BEM: inbourd stal-dealy and bad (prandtl,goldstein) tip-corrections. so 3d cfd is the only realistic change to get better understanding instead of difficult and expensice measurements ses the NASA?Ames 10 meter trubine measuremnt program just finsihed and presented at the 14th iea meeting. regards peter |
Re: WINDMILLS
Peter,
thanks for your answer. first of all I do not know the exact form of the criterion. So my question might be senseless but ... Do you check the criterion everywhere in the fluid domain or only near the blade to a certain extend or only in the cell layer adjacent to the blade ? Thanks, Mark |
Re: WINDMILLS
FIRST OF ALL , THANK YOU VERY MUCH FOR YOUR ANSWERS . THE QUESTION IS , IF I COULD BE ABLE TO OBTAIN THE POWER CURVE OF A 600KW STALL REGULATED WIND TURBINE , BY USING THE DEFAULT TURBULENT MODELS OF FLUENT , WITH REASONABLY ACCURANCY . WHERE CAN I FIND THE AIRFOIL DATA , DISTRIBUTION OF PITCH ... OF SUCH A WINDMILL , WITH THE ACCURANCY YOU TOLD ME (200 POINTS ) ...
DO YOU KNOW IF THE PROPID CODE , DEVELOPED BY M.DREKA IN THE UNIVERSITY OF ILLINOIS IS GOOD ENOUGH PREDICTING THE POWER CURVE? IF YOU DIDNT KNOW ANYTHING ABOUT IT , YOU CAN FIN IT IN http://amber.aae.uiuc.edu/~m-selig/propid/ REGARDS |
Re: WINDMILLS
look at white "viscous fluid flow" chap 5.5 eng predict. of transition
if you give me yout email adress i can send you a CFX fortran user routine for implementation mine is alois.schaffarczyk@fh-kiel.de |
Re: WINDMILLS
my mail is ruben_carranza_fernandez@hotmail.com i suppose you knew about propid ; if not i insist on visiting the page i told you , because it has very interesting information . your information has been very useful for me
|
Re: WINDMILLS
|
Re: WINDMILLS
that's very difficult. see the iea reports with calculations of mol hansen (TU denmark) and me (dewek 98, and iea 14, petten) if you are lucky, the blade-designer give you the chord(r), twist(r) and profile data . the naca 63 series ara in abbot doenhoff. there are also periodic couse in lyngy leaded by jens soerensen.
drela prop-it is usula Bem theory in fact not better then betz (1927) or glauert (1934) prop is described in wilsons thesi of 1974. regards peter schaffarczyk |
wind turbine desing help
i m a student doing a reseach in wind turbine rotor desing if anyone have related tutorial please help me thanks lot.my mail is sanathja@gmail.com
|
power out put
Quote:
|
| All times are GMT -4. The time now is 01:47. |