Re: Wind Study Projections

[Courier]

For all those with active wind turbines - Just wondering how the actual wind speeds compare to those which were projected by NOABL and VMM results ?

In our case we have had 25 months at an actual average of 5.62 m/s @ 57 metres high compared to a NOABL of 6.9 m/s @ 45 metres high and a VMM of 6.5 m/s @ 30 metres high.

Can we expect to have future years somewhat higher to re-create the a forecast average ?

Another 1 m/s at that level would make a massive difference to the surplus financial output at this early stage in the life of the project .

 

[linga]

Around here last year had the lowest average wind speeds for 35 years

 

[linga]

Variations in wind power generation in Ceredigion (1981-2016)

Dr xxxxxxxx

1. Aim
To investigate the variation in wind speeds in West Wales between 1981 and 2016, and to estimate the impact of these variations on the potential wind power generation.

2. Data
Wind data were obtained from the US National Aeronautics and Space Agency (NASA) Modern Era Retrospective Reanalysis (MERRA). Atmospheric reanalyses like MERRA combined physics-based calculations with available observations of atmospheric conditions to create a 'best estimate' of atmospheric conditions at a given time and location.


The data used for this analysis are averages for each day between 1 January 1981 and 21 December 2016. The data are averaged over the 0.5-degree grid box centred on a latitude of xxN and a longitude of xW, corresponding to the Ceredigion area of Wales. Although these data cannot precisely correspond to the conditions on xxxxxxx Farm, daily-averaged wind speeds do not vary sharply over distances of tens of kilometres. Thus, the variation of wind speeds over the period in question will correspond very closely to the actual daily variation of wind speeds at xxxxxxx Farm.


The wind data are available at a height above the surface of 10 metres. Although this is not the hub height of the turbine at xxxxxx Farm, the variations of wind speeds at different heights above the surface are very closely related to each other. Therefore, although the actual values will not correspond to the values at hub height, the variation over time will correspond very closely to the variation at hub height.

3. Analysis
Data from the MERRA reanalysis were downloaded from the NASA MERRA website. The data are plotted below as a 365-day moving average.

3.1 Variation in wind speed
The graph above shows the variation of wind speeds over time. A 365-day rolling mean has been applied to smooth the data. Thus, each data point represents the mean wind speed of the 365 days preceding that date. The graph shows that the 365 days preceding 21 December 2016 were the least windy 365 days in the 35-year period. The figures below

show this record low was 0.68 m/s lower than the average wind speed over the 35-year period. This corresponds to a speed 13% lower than the average.


Mean wind speed 1981-2016

5.18 m/s

Last 365 days wind speed

4.50 m/s

Last 365 days wind anomaly

-0.68 m/s

Last 365 days percentage wind anomaly

-13.12 %

 

[linga]

No idea if that analysis is replicated across the UK though.

 

[Still Farming]

Last 6 months very low here ?

 

[Princess Pooper]

Wouldn't have thought 25 months was long enough to provide a useful amount of data, it is only 2 yrs which is not enough to get a long term annual average.

Having said that our first year was about right in terms of actual generation v projection (don't care what the wind speed was it's the generation that's important) - certainly the E3120 and I presume most have an efficiency curve so eg an extra 5m/s does not create a proportionate increase in generation at the top of the operating range (and of course it will cut out above that!)

 

[linga]

True but if the annual average increased by 5 m/s I would be very happy.
Whilst it's true it's generation that matters that generation is dependent on wind speed and generally the curve is fairly straight from say 6 m/s to about 12 or 14 depending on turbine.

 

[How much]

Interesting about the average , i had a email newsletter last year from opus in which i think recall seeing them state that average wind speed being low for last year in general.
We have some wind turbines that are under performing against what the modelled data says they should be doing so I commisioned a further wind report on the modeled data for the past year vs the actual scada data records altough only 1/2 hourley data . So some pogentual for varying results
The results was that wind was apparently 5% greater in strength than the forcast average but wind direction as recorded on turbine was out compared to the model .
The so called EXPERTS thought this change of direction ( only a few degrees )was accounting for more turbulance and thus less output , and was our problem , personally i am not certain thats the reason but cant prove anything one way or the other .

 

[Courier]

Thanks for the replies, the graph showing 20% variations over time is certainly most enlightening.
Our output at the recorded wind speed is highly satisfactory and on months where the average speed is recorded at the forecast average then the £production was well over target.
Living in a predominately arable area we have neighbours who do lots of spraying and for the past two years they have never complained once about being behind due to wind conditions which seems unusual so maybe we should be expecting higher production in the next couple of years [emoji6]

Our turbine works at rated output from 9.5 m/s and produces 1/3 of that at around 6.3 m/s.

 

[linga]

What turbine have you got

 

[Courier]

Ours is a Vergnet GEV MP-C 32 downrated from 275 kWh to 200 kWh .

The PP and infrastructure was for 2 turbines at up to 250 kWh each but with the manufacturers would only warrant 250 kWh machines with 30 metre blades and the modelled annual output on a 200 kWh machine with 32 metre blades is higher at our site with a moderate amount of turbulence.

http://www.vergnet.com/wp-content/uploads/2016/01/DC-11-00-01-EN_GEV_MP-C_275_kW.pdf

http://www.vergnet.co.uk/our-medium-wind-turbines/

Our dear friend Amber Rudd killed off the second turbine which meant that we spent more on infrastructure than we otherwise would have needed [emoji35]

 

[linga]

Thought you must have something out of the ordinary.
We have done something similar

 

[windy_hollow]

Interesting about the average , i had a email newsletter last year from opus in which i think recall seeing them state that average wind speed being low for last year in general.
We have some wind turbines that are under performing against what the modelled data says they should be doing so I commisioned a further wind report on the modeled data for the past year vs the actual scada data records altough only 1/2 hourley data . So some pogentual for varying results
The results was that wind was apparently 5% greater in strength than the forcast average but wind direction as recorded on turbine was out compared to the model .
The so called EXPERTS thought this change of direction ( only a few degrees )was accounting for more turbulance and thus less output , and was our problem , personally i am not certain thats the reason but cant prove anything one way or the other .

Interesting about the average , i had a email newsletter last year from opus in which i think recall seeing them state that average wind speed being low for last year in general.
We have some wind turbines that are under performing against what the modelled data says they should be doing so I commisioned a further wind report on the modeled data for the past year vs the actual scada data records altough only 1/2 hourley data . So some pogentual for varying results
The results was that wind was apparently 5% greater in strength than the forcast average but wind direction as recorded on turbine was out compared to the model .
The so called EXPERTS thought this change of direction ( only a few degrees )was accounting for more turbulance and thus less output , and was our problem , personally i am not certain thats the reason but cant prove anything one way or the other .

Small changes in direction should not impact too greatly unless your site has a lot of variation. Turbulence as a rule is badly modelled so unlikely to be picked up by the analysis you summarise- caveat of not seeing the report withstanding.
Are they underperforming at p90 or p50? P 90 means they may have underestimated risks. P50 points to an underlying issue with the base wind data.
Should also note that turbulence can help or hinder production. Just depends on the factors.

 

[Bob]

For all those with active wind turbines - Just wondering how the actual wind speeds compare to those which were projected by NOABL and VMM results ?

In our case we have had 25 months at an actual average of 5.62 m/s @ 57 metres high compared to a NOABL of 6.9 m/s @ 45 metres high and a VMM of 6.5 m/s @ 30 metres high.

Can we expect to have future years somewhat higher to re-create the a forecast average ?

Another 1 m/s at that level would make a massive difference to the surplus financial output at this early stage in the life of the project .

Are you taking the wind measurement from the Turbine? If the anemometer is located behind (Downwind) of the blades some of the energy has been taken out so naturally will record a lower windspeed.

A more accurate figure is obtained by a stand alone mast located near the turbine at a similar hub height as seen on some of the larger wind farms.

Around here 2013-14 was best and this year is looking good up to now

 

[windy_hollow]

Modelled data

Variations in wind power generation in Ceredigion (1981-2016)

Dr xxxxxxxx

1. Aim
To investigate the variation in wind speeds in West Wales between 1981 and 2016, and to estimate the impact of these variations on the potential wind power generation.

2. Data
Wind data were obtained from the US National Aeronautics and Space Agency (NASA) Modern Era Retrospective Reanalysis (MERRA). Atmospheric reanalyses like MERRA combined physics-based calculations with available observations of atmospheric conditions to create a 'best estimate' of atmospheric conditions at a given time and location.


The data used for this analysis are averages for each day between 1 January 1981 and 21 December 2016. The data are averaged over the 0.5-degree grid box centred on a latitude of xxN and a longitude of xW, corresponding to the Ceredigion area of Wales. Although these data cannot precisely correspond to the conditions on xxxxxxx Farm, daily-averaged wind speeds do not vary sharply over distances of tens of kilometres. Thus, the variation of wind speeds over the period in question will correspond very closely to the actual daily variation of wind speeds at xxxxxxx Farm.


The wind data are available at a height above the surface of 10 metres. Although this is not the hub height of the turbine at xxxxxx Farm, the variations of wind speeds at different heights above the surface are very closely related to each other. Therefore, although the actual values will not correspond to the values at hub height, the variation over time will correspond very closely to the variation at hub height.

3. Analysis
Data from the MERRA reanalysis were downloaded from the NASA MERRA website. The data are plotted below as a 365-day moving average.

3.1 Variation in wind speed
The graph above shows the variation of wind speeds over time. A 365-day rolling mean has been applied to smooth the data. Thus, each data point represents the mean wind speed of the 365 days preceding that date. The graph shows that the 365 days preceding 21 December 2016 were the least windy 365 days in the 35-year period. The figures below

show this record low was 0.68 m/s lower than the average wind speed over the 35-year period. This corresponds to a speed 13% lower than the average.


Mean wind speed 1981-2016

5.18 m/s

Last 365 days wind speed

4.50 m/s

Last 365 days wind anomaly

-0.68 m/s

Last 365 days percentage wind anomaly

-13.12 %

varies greatly

Are you taking the wind measurement from the Turbine? If the anemometer is located behind (Downwind) of the blades some of the energy has been taken out so naturally will record a lower windspeed.

A more accurate figure is obtained by a stand alone mast located near the turbine at a similar hub height as seen on some of the larger wind farms.

Around here 2013-14 was best and this year is looking good up to now

The anemometer reading in the SCADA file should be corrected for this to an upstream value.

Correct though that it has some issues with accuracy and if you are really worried about performance then a Power performance study might be worthwhile - the mast bit.

So you have two options. A more detailed sacra study to see if the turbine is performing against your warranty power curve, look at direction in more detail and then correlate the output against other typical/similar neighbouring farms. Or do a power performance.

 

[linga]

The analysis I posted was done in about 15 minutes on Christmas day and in no way was it intended to be "accurate".
It was simply a response to my musing about the reasons for our turbines lack of output against prediction.
Nevertheless , bearing in mind the assumptions made it gives an interesting indication of how variable the wind is and also just how low the wind speed was throughout last year in this area.

 

[Courier]

Are you taking the wind measurement from the Turbine? If the anemometer is located behind (Downwind) of the blades some of the energy has been taken out so naturally will record a lower windspeed.

A more accurate figure is obtained by a stand alone mast located near the turbine at a similar hub height as seen on some of the larger wind farms.

Around here 2013-14 was best and this year is looking good up to now

Our turbine has a hub height of 55 metres and a swept area at 32 metres diameter and is a downwind configuration so the anemometer is effectively in front of the blades. I doubt a separate mast would make any difference.

 

[windy_hollow]

Our turbine has a hub height of 55 metres and a swept area at 32 metres diameter and is a downwind configuration so the anemometer is effectively in front of the blades. I doubt a separate mast would make any difference.

at the risk of being an anorak on this...............

Wind approaching a wind turbine slows down as it "sees" the structure upwind. The zone where the wind slows down is the induction zone, and according to standards starts around 2-2.5 rotor diameters upstream of the turbine.

Good turbine manufacturers account for this by applying a correction factor called the nacelle transfer function to the anemometer data. This is meant to project the anemometer reading to the freestream wind speed and is calibrated at the time of turbine certification. Should happen for big and small wind if they use any form of nacelle measurement.

The reason for using a mast or other form of measurement would be to place the mast upstream of the turbine in the main wind direction and measure the wind speed against the power output of the turbine and then compare to your warranty.

This can be expensive though and only a solution if you really feel the turbine is underperforming. 1st steps are to do what you are doing and looking at the output as it happens and maybe looking to see how other similar machines in the area, if you know of them, are doing. after that you can do some work using the turbine data itself and query the manufacturer form there. Only if it gets to a point of falling out would a the mast validation really be a step.