Small wind turbine power curves - Ampair historic practice

I thought it worth writing down how Ampair has been doing its power curve testing over the years. This is not intended to show ourselves as some sort of gold standard for technique but simply to say what we are doing in the spirit of full disclosure. Anyway historically Ampair has measured and reported its power curves:

1. Using real turbines, regulators, batteries, etc.
2. In open air.
3. Either on short poles or on vehicles.
4. With manual data logging of cup type anemometers at approximately hub height, and with manual data logging of voltmeters and ammeters.
5. Noting power into battery at typical battery voltages, and more recently of power into grid for grid connected (G83) inverters.
6. Publishing the results in our manuals and marketing literature.

Our underlying intent has always been to try and report a power curve that a client could reasonably expect to reproduce if they had commonly available instrumentation and patience. Taking the points in turn shows how things can be done differently.

1. Real systems have not been tuned for abnormally high performance, and include all the components that should be in a client's system.
2. An important part of the way Ampair tests turbines is that it is in the open air. The open air has a lot of variation in it and this causes turbines a lot of problems trying to respond. If you test in unreal conditions such as the smooth and stable flow of a wind tunnel it is easy to get much more power out of a wind turbine system than of the same system in real wind. This is because the mechanical and electrical system responses don't have any tracking error in steady flows and doubling of system power can be achieved in wind tunnels like this (assuming the wind tunnel is large enough, if it is smaller then the power out is even higher in a wind tunnel). By the way there is a role for wind tunnels and small wind turbines - but only for R&D work, not for published power curves.
3. Historically we've tested on short poles, or on vehicles. Both ways give similar results if you avoid too ideal a set up, or distortions induced by putting instrumentation in the wrong place. We typically do our vehicle testing in cluttered environments so we get very unclean wind on to the turbine, similarly our poles are ordinarily very low down and again we get foul wind. This is exactly how our clients tend to use our turbines.
4. We have used manual recording of the data from standard instrumentation. We do not use high precision instrumentation (we keep an eye out for calibration errors, but do no more than this) and we manually record what is best thought of as the visual average of the data we see. This visual filtering is important as it means we watch the three meters (volts, amps, knots) and look for the periods when they are steady enough that we can write down all three quickly. We could try and be very optimistic and write down only the peaks but that wouldn't be helpful, and if we were pessimistic and wrote down only the troughs we would be out of business as nobody would ever buy our turbines (this is what Hugh Piggott refers to as "marketing suicide"). We use cup type anemometers so that we are always recording the maximum wind available irrespective of direction and we set the anemometer at a height that gives it representative wind on to the turbine. A feature of using pretty basic instrumentation is that our results can be reproduced by clients (so in the olden days we used Avometers, these days we use Flukes) who aren't able to buy fancy kit. Until recently dataloggers were barely affordable for the smallwind manufacturers (such as Bergey and Proven) and not affordable for microwind manufacturers (such as Ampair).
5. We are interested in 'useful power for the client' so we are logging power into the battery or power into the grid. This is important as some people measure power out of the turbine which is often much higher than power into battery or grid because regulator and inverter inefficiencies can be much higher than people think.
6. It may seem obvious but we publish our results in our manuals and our marketing literature. Not all manufacturers are quite this transparent.

What I have just described above is a low tech equivalent of what is done in large wind turbines. The real issue is the genuineness of the good intent of the people writing the power curve. No amount of fancy instrumentation makes up for a person who is trying to push the data to mis-represent reality. In fact high tech instrumentation can make it easier to artificially over report a power curve. So when we read a test report (such as the yachting magazines produce every few years, or we have a conscientious client write in, or folk like Piggott and Gipe do tests) we pay attention if our turbines are behaving differently than we think they should be. And when people like Piggott say nice things like "It has never (except with Ampair) been my experience that the outputs I measure match manufacturer's predictions" then we think we are getting it about right.

In the future it is likely that we will start using higher tech instrumentation but we will do so in the context of a very tightly defined standard that will lay out exactly how to report power curves, which we are almost at with the BWEA small wind standard (and the related parallel AWEA one). We've been upgrading our long term test site at Misty View Farm in Cornwall with that in mind and we continue to install more and more kit as we search for a better way to do it. Until we are happy with that methodology we'll carry on doing it the way we always have. Amazingly very few people have ever done any serious work on comparative testing of small wind turbines and manufacturers' power curve claims and to this day the best work has been done by Paul Gipe (see Windworks ) who comes up with some alarming reports about almost all manufacturers' power curves. Fortunately Ampair come out well with the quote from Gipe of ".. the Ampair 100 were the only machines to exceed their power curves at any time". The other independent person who has looked seriously at this topic is Hugh Piggot of Scoraig Wind who has been turned down by the UK government when he has proposed to do independent testing on small wind turbines in the past which is a shame and which has limited his ability to do work in this area.

Why does all this matter ? Simple: clients make purchasing decisions on the basis of power curves and governments give grants on the basis of power curves. So if one manufacturer gives a power curve that is over-optimistic they attract a lot more (taxpayer funded) government grant and a lot more eager customers than another manufacturer that is more cautious. I would like to say that honesty is always rewarded in the long term, but unfortunately in the short term history reports that manufacturers can quickly be out of business.

Over the last two years Ampair have put a lot of effort in to move the standards agenda forwards and we've left it for others to do comparative testing of modern grid-connected microwind turbines. Now results are starting to come from public testing in but that's a topic for another post.

Supplier part numbers in MYOB

I like to include supplier part numbers in purchase orders, and MYOB thoughtfully allows for this. I couldn't understand why they weren't coming through on the printed forms (of course they never show on the screen because of the MYOB lay out issues I've described before) and finally I've realised that its because of the supplier selection. For each item one can set up a normal supplier. If one then issues a PO to another supplier MYOB thoughfully conceals the supplier part number. This prevents us all inadvertently upsetting our secondary suppliers by revealing alien part numbers to them. Unfortunately MYOB doesn't have a secondary supplier pick and secondary supplier part number.

We also make use of the supplier part number to link our drawing registry and our accounting system. We use the supplier part number as our drawing number (for parts which are being manufactured to our drawings) and our own part number is our (MYOB) accounting part number. It is a work around but the accounts people are determined not to use drawing numbers as part numbers, and the drawing people simply can't use the acounting part number convention (because of the way the drawing package file structures are arranged).

I guess we're guilty of turning MYOB into a cheapskate ERP but there really aren't many good alternatives out there. Recently a competitor changed from Sage FC 500 to Exchequer and they tell me it is good. They were driven by pretty much the same issues that drive any SME manufacturer. I hadn't heard of this before but it appears they shelled out £20-£40k for the move whereas we've done a pretty good job with MYOB for £4k inc consulting. Equally interestingly our consultant has dropped out of the MYOB consulting scheme because she doesn't want to support MYOB until they take the networking issues seriously. She was a silver and really knew her stuff - much better than the gold tier people we checked out. But equally she says Sage is a poor network user as well.

El Reg carried an interesting article on ERP packages the other day. I'm not sure they ever consider stuff as small and insignificant as us, but given the way Sage are hoovering up mid to low end package vendors I guess we'll be rereading that article in a few years time. Hopefully making more microwind and water turbines by then but it might be something else given the way our R+D team are throwing out ideas.

Oh and this new blogger text editor is a disaster - it doesn't allow ampersands. What a nightmare. Turns out the offending amersand is in the label field !!