HI ALL,

I'm modelling the flow around buildings and I've got a atmospheric boundary level profile specified at the inlet with inform. The problem is when I try to alter inform to skew the flow at different angles the profile becomes distorted. Can anybody help me? Thanks for any help!

John,

I am assuming that you have set up a function with inform that set the velocity of the inlet at a point as a function of height v = F(h).

Assuming that the direction of wind is blowing NORTH.

What technique did you use to try to get the wind velocity to be North East.

When I do similar problems I resolve the wind speed into the X and Z coordinates to use at each point.

I personally do not bother with the continous wind profile. I have found that setting up the wind speed at discreete elements, say every meter or so, is more than accurate for my needs.

Hope this helps,

Leon

Hi Leon,

The following is the inform formula I am using where INL is the inlet. I got this straight from the cham website. If I want to get North-East I just add (SOURCE OF V1 AT INL IS....) and it works ok . But if I try multipying U1 and V1 by the SIN(theta) and COS(90-theta)components to get various other wind incidences, then I start to get a distorted profile. I am relatively inexperienced at this so any help would be great. Your method for creating a profile seems straight forward, I'll give it a shot, Thanks

inform13begin REAL(VREF,REFHEIGHT,POWER,RHIN,VTAU,ROUGHH,AKAP,ZR FDRH) VREF=5.0;ROUGHH=0.022;AKAP=0.41 REFHEIGHT=10.0 POWER=1./7. ZRFDRH=REFHEIGHT/ROUGHH VTAU=VREF*AKAP/LOG(ZRFDRH) VTAU RHIN=1.189 CHAR(FORM1,FORM2,FORM3) FORM1=VTAU/AKAP FORM2=VTAU^2/0.3 FORM3=VTAU^3/AKAP (SOURCE OF P1 AT INL IS RHIN*(:FORM1:*LOGE(ZG/:ROUGHH)WITH IMAT<100) (SOURCE OF U1 AT INL IS RHIN*(:FORM1:*LOGE(ZG/:ROUGHH)^2 WITH IMAT<100) (SOURCE OF KE AT INL IS RHIN*(:FORM1:*LOGE(ZG/:ROUGHH)*:FORM2:WITH IMAT<100) (SOURCE OF EP AT INL IS RHIN*(:FORM1:*LOGE(ZG/:ROUGHH)*:FORM3:/ZG WITH IMAT<100) inform13end

If the inflow direction is changed so that it is at an angle to the west face of the solution domain, then if the north (or south) boundary is open to the wind, you will also need to introduce an log-law inflow boundary condition at this domain boundary. Whether it is at the north or south boundary depends on the direction of the wind.

I am using two inflow and two outflow boundaries at the necessary locations. I think my problem is in the altering of the inform formulae to suit different incidences.

John how is the profile distorted ???

A simple question, are you using radians or degrees for the angle ??? From memory fortran uses radians unless told otherwise, and inform uses fortran. Maybe this is your problem ???

Also it appears that the formulea presented above are not general equations to solve this type of problem.

By this I mean (SOURCE OF U1 AT INL IS RHIN*(:FORM1:*LOGE(ZG/:ROUGHH)^2 WITH IMAT<100) Can U1 in this term reprsented by u1(thetha) = u1 * sin(thetha) and how does the change in direction effected the KE terms.

It appears to me that the formulea presented above are a simplified set of equations that work for wind perpendicular to the boundary. I could be wrong here but without seeing more on the base equations this may be the problem.

Hope this helps

Leon

Leon,

The base equations used are the standard log-law atmospheric boundary layer , the same as those used in library case t306. I think that maybe I should use discrete elements instead.Would you be able to send me a q1 where you have the inflow specified using this method? I would be very grateful if you could. I have tried to follow t306 but have been unsuccessful in applying it to my application.

Many thanks,

John

Try library case I 107.

I do not have a current q1, at the place were i currently work, that has this type of air inflow condition.

Leon

John,

I have sent a Q1 file to your colleague Tom Norton at Teagasc that exemplifies the use of INFORM for a 3d boundary layer inflow at incidence. Leon is correct in saying that the existing formulae in the 2d examples assume 2d inflow, but they can be generalised quite readily by assuming that the velocity and shear stress vectors are aligned at inlet.

Thanks for your help Leon!