For me it was Jane Birkin (Je T'aime) and Marianne Faithful (in "Girl on a Motorcycle").

Reminds me, I must try and get a DVD copy of that film, I haven't seen it in years, but still have fond memories of that black leather catsuit with the bejewelled zipper fastening.........................

The Mars bar story, is (so I was told from a supposedly reliable source that was there that evening), apparently, a myth. The bust at Redlands made the headlines only because the Stones were seen as the bad boys at the time, and Keith was supposedly badder than any of them. Marianne Faithful was, apparently, in a state of partial undress at the time of the raid, but the rest of the tale, including the apparent placement of the Mars bar and Mick Jaggers alleged antics with it, was pretty much a figment of the imagination of the press, aided and abetted by the local constabulary, I believe.

How do I happen to know this? Long story, but I lived just down the road from Redlands in the late nineties and Keith Richards was a very active supporter of the village. He made a big donation towards a new village hall in West Wittering, something I'd been involved with on the periphery. Always struck me as a bit odd for such a rebel to live in the same house for around 50 years.

PS: The story goes that she was dressed only in a fur rug (plus, one supposes, the alleged Mars bar), so I don't think she was on a sofa.

Posted By: JSHarris
The major problem is that focussing infrasound is difficult, and needs large devices (of the scale of wind turbines), so isn't really a feasible proposition.

Yes, that's one reason why the proposed device used two higher frequency beams and relied on interference in the target area for its effect, IIRC. Another reason was to better confine/focus the target area.

The standard calculations for wind shear certainly don't take any effects of turbulence into account.

I suppose trees and buildings will have an effect on wind shear, but it would require a complex computer fluid dynamics model to calculate it. The net effect would be relatively small I would think. Building turbulence can increase wind speeds at certain points.

But normally wind shear is being calculated based on wind speeds at a wind farm anemometer site where turbulence would be expected to be very small in any case. ETSU-R-97 requires that any data from an anemometer that has been affected, by wake from a turbine for example, to be rejected.

Aren't trees and buildings (eg ground roughness) the very cause of wind shear?

In "our" situation the anemometer mast "pointed" south west so data from the north east was discounted to avoid shadow effects from the mast. However wind from that direction passes over a site with trees and buildings. So to my mind they may have discarded data from a wind direction that features increased wind shear as well as turbulence?

In that wind direction trees would be just 250-300m upwind of the 126m turbine.

Dave, AFAIK anemometer readings are just for the benefit of the developer to decide that there is enough wind to economically justify a turbine or farm. I've not heard of it being a planning issue as such although it might feature as part of an argument over the size of a turbine. Few developers want to put turbines where they won't generate.

Colin, yes ground roughness is the main cause of wind shear but it is affected by other variables such as temperature and topology. Over forests (my particular interest) the height of trees generally is taken as a displacement of ground level. But all simple wind shear calculations (1/7 rule etc) just assume smooth laminar flow and ignore the added complexities caused by turbulence.

It seems many developers are now trying to include wind shear effects in the noise calculation. However from what I've been reading wind shear also appears to be a cause of Excess Amplitude Modulated noise which they claim can't be predicted/calculated.

One of the things I read recently is that wind shear involves a change in wind direction with height - so that the blade pitch is less than optimal/wrong somewhere around the disc.

I recall something similar and, following the line of thought prompted by these last few posts, tried to find it again last night without success.

However, a synapse obviously made a connection during the "I'll sleep on it" phase and I've found the reference in Mike Stigwood's paper, 'The effect of common wind shear adjustment methodology on the assessment of wind farms when applying ETSU-R-97'...

http://www.wind-watch.org/documents/effect-of-a-common-wind-shear-adjustment-methodology-on-the-assessment-of-wind-farms-when-applying-etsu-r-97/

'Introduction to wind shear', p3.

I need a cup of tea after that lot!

Posted By: CWattersOne of the things I read recently is that wind shear involves a change in wind direction with height - so that the blade pitch is less than optimal/wrong somewhere around the disc.

There may be other effects but there's certainly a fundamental geometrical effect caused by the different relative velocities of the blade and the wind at different points, even without any change of direction of the wind. In sailing, the resultant adjusment needed to the sails is called twist. For a turbine or propellor or rotating wing it gets more complicated because of the rotational motion. Apparent wind vs true wind and all that. Helicopters have special mechanisms to deal with the problem.

But yes, in general the wind at ground level doesn't blow in the same direction as wind higher up.

I guess wind turbine designers are still evolving the best method of handling it all most efficiently. Maximising design life whilst minimising cost complicate the problem beyond just making the electrical generation efficient.

Posted By: CWattersI suppose it's too hard to make a turbine that can very the pitch as the blades rotate like a helicopter.

I thought about that too, but interestingly, that's apparently not how most helicopters deal with the geometrical problem (they don't have the Coriolis shear problem). It's not a straightforward issue to solve. I'm sure other people here know a lot more.

http://en.wikipedia.org/wiki/Retreating_blade_stall
http://www.copters.com/aero/lift_dissymetry.html

Posted By: tedOf particular interest to me would be any comments on section 2.9 and the amount of data required to derive valid background noise levels?

What monitoring equipment are you going to use? and can you vary the sample rate.

You will need to know the type of turbine and the cut in speed and rated power (check what that means, I assume maximum).
So a turbine may work in the range of 3 to 12 m/s windspeed.
Sampling is often carried out every 6 minutes, this is often a mean (stick with mean for this) from the previous 6 minutes sampled every minute (depends on the equipment). This is where you are going to have a problem as a minutes loud noise does not look very loud when it is surrounded by 5 minutes of silence.
Filtering for rainfall (it really does not rain that often in the UK, about a 1/3rd of the time) and wind direction is easy enough.
You will need to draw a map of the surrounding area, where other source of noise are (roads, factories etc) and correlate that with wind direction (perfectly acceptable to use a correlation here).

Windspeed is not constant, but varies between 0 and up to 50 m/s (nearly 100 mph), but you are only interested up the 'rated power', which may be 15 m/s, so you are working down at the lower end of the skewed curve.
Precision comes into this as you have to work within 1 m/s bins, so your Standard Error has to be lower that this.

Standard Error = Standard Deviation / Square Root of Sample Count

Stand Deviation has two formula, one if you know the entire 'population' or in English, your entire dataset and the other if you take a sample from the entire population (this is what you have to use as you have had to filter out for rain and wind direction).

Standard Deviation =Square Root (Sum(value - mean of values)^2) / Number of Values - 1).

Now all that sound pretty horrible as you are working from an unknown (unless you already have windspeed data for the site), but luckily windspeed distribution follows, fairly well, a Weibull distribution, though you will have to calculate this for every wind direction that you are interested in that is not affected by rain. This means that you will have to make sure that you sample in every condition to just get what you want, which may be only 15% of all the data (all depends on that site map, wind direction, rainfall and other influences).

So to capture what you are after, with out fudging figures, you will need, probably a years worth of data to get all wind directions and rainfalls and then slice it up into the appropriate bins., A year is 8760 hours and, if sampled every 6 minutes that will be 87600 samples, of which you may only be able to use 15%, so 13140.

If you doubled that sample number (sampled every 3 minutes), you would improve the standard error but make little difference to the overall figures.

Do you have the wind and rain data already, if you do it is easier to work backwards as spreadsheets have all the formula built in and it makes life very simple.

Thanks Steamy, I was hoping you'd reply.

Data gathering has already happened, using industry standard sound meters (class I). I'm not sure if the data is then extracted 'raw' and processed or if the internal function within the meter for calculating a 10 minute average is used.

The LA90 10-min average values are then plotted on a graph against wind speed. They can be found as various Figures (one for each property) in this report:

http://infrastructure.planningportal.gov.uk/wp-content/ipc/uploads/projects/EN010008/2.%20Post-Submission/Application%20Documents/Environmental%20Statement/Chapter%2016%20Appendices.pdf

This also contains a lot more detail on how the measurements were taken. Figures 1 and 2 (p31) shows day and night wind speeds and directions over the measurement period. You can see that only about 50% of wind directions were experienced covering the main body of data. p32 onwards then has the graphs for the individual properties.

Rain affected data (or other such as morning chorus) is ignored and then a 4th order polynomial regression is used to draw an 'average' line through the remaining valid data points.

The IOA are suggesting that 1 or 2 weeks worth of data is sufficient which I feel is laughable from a statistical point of view.

2.9 of the IOA report says: you need a minimum of 10 data points for each 1 m/s wind speed bin. And further suggests that as few as 5 would be enough in some circumstances. I can't see how such a small sample would be of any useful statistical significance.

One concern I have is how noise data is presented. For example they are allowed to show measured and average curves on one chart and the resulting ETSU limit and predicted data on another chart. It's only when you plot it all on one chart do you realise that for a significant percentge of the time the predicted noise would be louder than background and sometimes by a lot.