pedestrian comfort

[liamh]

.... This means, that in order to achieve the Lawson’s Comfort Assessment Criteria to be positive, the maximum wind velocity on the plaza, must not exceed 4 meters per second?

Well, kind of.

You have to prove that for 95% of the time that 4m/s is not exceeded. Due to the fact that CFD is so computationally expensive, and you can't analyse the wind environment around your site for a whole year, you must do some statistical analysis using local weather data.

You have also ignored the fact that Lawson's criteria addresses wind gusts as well. To be honest, I tend to ignore it as well, due to the difficulty in getting weather data of sufficient frequency to capture wind gust data.

So the way I go about it is....
1) find local weather data. The weather data in the VE is annual hourly data so for ease I would suggest using that. You may have to purchase weather data closer to your site than that bundled with the VE.
2) determine the annual mean wind speed from local weather data set
3) run 8 steady state CFD simulations modelling 8 wind directions at the annual mean wind speed, e.g. N,NE,E,SE,S,SW,W,NW. 16 would be more accurate, 4 less so.
4) determine the locations around your site where you want to calculate Lawson's criteria
5) find the wind speed at each of these locations for each of the wind directions. Hint: in Microflo use monitorings point to dump the CFD data to text file which you can open to find the wind speed at your locations. If you want to analyse 5 locations and have 8 wind directions this will mean you need to determine 5*8=40 wind speeds. This will need 5 monitoring points.
6) divide the 40 wind speeds from step (5) by the wind speed used for the Microflo sim (the annual mean wind speed) to determine a non-dimensional wind speed up factor
7) using the local weather data and the 40 wind speed up factors calculate the wind speed for each hour over the whole year for each of the 5 locations
8) determine the threshold wind speed at each of the 5 locations, i.e. the wind speed that for 95% of the year is below

I have written a program to do 7 and 8 which could save you a lot of effort. I will PM you.

Notes....
a) there is no method to calculate pedestrian comfort that has been set in stone as THE method to use, so you could come up with your own methodology if you can justify it to your clients
b) the use of steady state simulations and the k-epsilon turbulence model in Microflo ignores transient effects such as vortex shedding. To address this would need an LES turbulence model and a transient simulation. This is very expensive computationally and although it is best practice it is hardly ever done in our industry due to the fact (for isothermal flows) it is probably more expensive than basic wind tunnel testing
c) the wind speed up factor method ignore the fact that as Reynold's number (the ratio of inertial to viscous effects) changes then so does the flow regime (the path of the flow)

[st_george]

Thank you for the answer, Liam.

I am so ashamed I did not tell you, I do not use the VE for the CFD analysis. Sorry once again.
I already completed the CFD analysis, but in Solid Works.

Here is the way I did it, and I am asking you if this is the correct way:

1. From the website of the local Hydrometeorogical station, I took data for each year, from 2001-2010. The mentioned "data" is the mean wind velocity during particular year, for 8 directions (N, NE, E, SE, S, SW, W, NW) and the wind frequencies (how many times per year the wind from this direction showed up).
Then summed the wind velocities, and wind frequencies for all 10 years (from 2001-2010) and divided by 10.
The result is this table:



From it, I took only the most dominant winds, that appeared. By most dominant, I meant the most frequent ones, not the most strongest ones. It appeared that the most dominant ones, were:
N (appeared 132.1 times, with mean wind velocity 1.9 m/s),
NE (appeared 128.5 times, with mean wind velocity 1.91m/s),
E (appeared 145.3 times, with mean wind velocity 1.9m/s),
NW (appeared 175.8 times, with mean wind velocity 2.4).

I took all four of these, and made a CFD analysis in Solid Works.

On none of the four cases, the wind velocity did not reach 4 meters per second.

Is this somewhat of a "proof" that Lawson’s Comfort Assessment Criteria for this location is positive, fulfilled?

Is this the right way to do it?

I uploaded the video of the CFD analysis:
http://www.youtube.com/watch?v=QQTyhsr_ ... e=youtu.be

[liamh]

I remember that you don't use Microflo but my explanation plus program works with any CFD software. The only difference is that Microflo automatically takes into account different wind profiles due to terrain type. In other CFD codes, if the terrain type for the weather data is different to that used for the CFD model then you should take that into account.

You haven't proved Lawson's criteria because you haven't calculated the wind speed that will exceeded for 5% of the time.

Your methodology only proves that on typical days you will meet Lawson's suggested 4m/s threshold for comfortable sitting.

You probably could get that passed by your clients but you should word it carefully so you are not saying you are proving Lawson's criteria is met.

[st_george]

Thank you for the reply Liam.
I am a student of architecture, so there is no client.
The purpose of all this was to create an object on the city plaza. The project ended last year, but I still want to learn how to determine if the Lawson's criteria for pedestrian comfort has been met.

So by doing all this, I covered only 95% of all the mean wind influence, right?

Then what data do I need more, to implement that additional 5%?

Or maybe I misunderstood the meaning of the Lawson's Criteria?! It has nothing to do with 95%?
It says that during year (or decade in my case - from 2001-2010) the wind speed will not exceed more than 4 meters per second, more than 5% of total wind influence during the year?

[liamh]

Not sure I can help you much more.

From your explanation it seems that your study has taken some averages, and averages of averages, and then did some CFD sims of those averaged conditions. You haven't analysed the wind speed over your site, or particular locations, for a prolonged period, typically a year but ten years would be good too.

You need to prove that the wind speed is below 4m/s for 95% of the time for sitting to be comfortable. How does running a few CFD simulations at average wind speeds, and prevailing wind directions do that?

The CFD study has to go along side statistical analysis of the weather data to interpolate the wind speed for the whole year, typically at each hourly step. The usual way to do this, by me and others, is to use a speed up factor. It is a simplification in that it ignores some important effects like the change in flow regime with an increasing Reynold's number etc, but as far as I know it is used widely.

If you Google for Lawson's Criteria and CFD you should be able to find a paper or someones report that uses speed up factors. They may call it something different like "normalised wind speed" or something but they will be dividing the CFD wind speed prediction by a reference velocity.

[st_george]

Liam may I dare to ask you something more about the steps 6, 7 and 8. To be precise, I do not understand them.

I went trough the previous 6 steps with no problem (I guess).

I took the annual wind data for the city:



Run 8 steady state CFD analysis, got the speed at let's say three point (there are more, but for the simplicity I will say I only took three points).
There for I got 3x8= 24 wind speeds.
Then what?

You say:

[quote=liam]6) divide the 40 wind speeds from step (5) by the wind speed used for the Microflo sim (the annual mean wind speed) to determine a non-dimensional wind speed up factor[/quote]

I do not understand this part.
I have 24 wind speed points with 8 wind speed directions.
This means that I need to divide each group of 3 points I got from the particular wind speed, with it's wind speed I used for the CFD steady state?

For example I run a CFD steady state analysis for the northern wind, with it's speed 1.3 m/s. What I need to do is to divide the three values of speed I got on three particular points with 1.3?

The same for North-Eastern wind (1.7 m/s) the same for Eastern (1.7 m/s) and so on...
Right?

But what about the steps 7, and 8?
I do not understand them.
I did not received your program you sent my by Private Messages, as I can not use the PMs on forum for some reason. Maybe I do not have enough posts.

Sorry for disturbance I am so ashamed.

[liamh]

Hi St_George. I don't think I'll get chance to answer this till tomorrow. Hope you can wait.

Liam

[st_george]

No problem Liam.

As long as you need it. You were very patient with me, I think except the gratitude, I owe you the patience also.

Take your time.

[liamh]

St_George. I've put together an Excel file to show the calculations. Can you email me your email address and I will send it off to you?

Regards,
Liam