BRE report FB17: Ampair comments part 2

(continued from previous post on the BRE urban micro wind turbine assessment)

In calculating electricity generation BRE make the careful statement “the conditions under which the the [manufacturers’] power curves were obtained is, in most cases, uncertain, it is expected that they were obtained in wind tunnels or using free standing mast-mounted turbines”. Herein lies my biggest quibble with the FB17 methodology as BRE should have put this in much bigger letters and written an additional caution, namely:

WIND TUNNEL POWER CURVES ARE NOT COMPARABLE WITH OPEN FIELD POWER CURVES.

At the moment to use all wind turbine manufacturers’ power curves to predict electricity generation as if they were equal in quality is to introduce a large error especially as (to the best of my knowledge) of the three systems analysed only the Ampair 600 is an open field power curve. The reason for the difference is that in a wind tunnel one can hold the wind speed absolutely constant for the measurement period, and the wind direction is necessarily constant, and the turulence can be reduced to a minimum. This means that the turbine system performance is not degraded by a variety of tracking errors irrespective of whether they are mechanical in nature or electrical in nature. (This unfortunate situation will change as very rigorous standards for generating power curves are being introduced, but it is so important a quibble that we are disappointed with BRE for not emphasising their predicament).

Once this quibble is out of the way the actual values for electricity production in an urban environment are of approximately the right order of magnitude (at least for the Ampair). But the net electricity generation is different than the gross electricity production. This is because BRE have assumed (or not even stated the assumption) that the turbines are not importing electricity. Again for some systems this is an unwise assumption as the annual standby power consumption can be of the same order of magnitude as the annual production leading to a net generation of approximately zero, or even negative. For an Ampair system the extra money we put in to a high quality inverter pays off in very low standby powers so for Ampair the BRE assumption is valid.

The energy production figures are not corrected for any wind quality issues. As BRE themselves state this is a known issue and it is too early to try and make such adjustments.

The CO2 payback results appear sound. From an Ampair perspective it is pleasing that the Ampair pays back CO2 more quickly than the other turbines. We believe that this would be even more noticeable if the power curves were put on a level playing field and if the electricity import issue was accurately modelled.

The lifecycle costings contain an important error regarding the actual cost of System 3 (RD Swift). Obtaining pricing for this system has been notoriously difficult for the last two years and so it is unsuprising that BRE have used the only public domain data they could locate, that of £3.5k as provided by bettergeneration.co.uk who are not an RD Swift distributor and who are themselves misled. Unfortunately it is very misleading and a more accurate cost would have been about £7k. Whilst RD Swift are working hard to reduce both price and cost and certainly aspire to much lower than £7k the other two systems are being evaluated on their costs in the market today and once again a level playing field should have been applied.

The other issue with lifecycle costings is the maintenance and longevity one discussed earlier. We think that a quality turbine should aim to last on average 15 years with perhaps a five yearly on demand maintenance visit (i.e. four visits: once to install; two for maintenance; and one to remove), and a lower quality turbine should aim to last on average 10 years with four visits in total (at reduced intervals). We don’t mean lower quality in a pejorative sense: it may be more cost effective to make a cheaper turbine with a shorter life.

These two costing problems contaminate the finncial payback calculations sufficiently that one cannot make any further comments.

Further research is of course required and the proposed list is a sensible one. Overall I think this FB17 report from BRE is a pretty good assessment of the aspects of urban microwind and look forward to more pieces of the puzzle being slotted into the jigsaw in due course. Whether one thinks urban microwind is a good idea is a different thing and one we'll comment on separately.

(concluded)

BRE report FB17: Ampair comments part 1

These notes on the BRE microwind assessment are not intended to be a comprehensive scientific critique, just an informal and necessarily limited set of comments. We are a manufacturer not an academic institution.

This area is short of good quality work so anything as sound as this is to be welcomed. To ask for perfection at this stage would be mind numbingly boring and so any niggles must be put in that context.

Our most important criticisms are:

1. The cost of the System 3 is completely inaccurate. Our sources tell us that the true cost of this system is typically £5k-10k installed not the £3.5k assumed in the report which came from a very dated website which was quoting a 'target' price.
2. The manufacturer’s power curves have been used. The three systems have very different ‘quality’ power curves only one of which is measured in real wind conditions.

System 1 is an Ampair 600-230 EU2. We assisted Bath University in their Life Cycle Analysis (LCA) in the data gathering phase of their work, and put them in touch with BRE when we were approached by BRE to conduct another LCA. We thought another LCA would be repetitious and instead we asked BRE to invest their resources in a complementary piece of work which has become the resource piece of this BRE assessment. So in a sense we were the matchmaker but beyond that both Bath and BRE are autonomous and independent organisations and Ampair has not had any influence over either beyond commenting on an early BRE draft with respect to references to the Ampair product. In due course we look forwards to seing the full analysis that Bath have been conducting.

System 3 is the Renewable Devices Swift. This can be seen at a glance since it is the only 1.5kW turbine in production. More academically rigorous is the fact that ref 5 (Rankine, Chick, Harrison) has been quoted by RD Swift in the Swift marketing literature of 2006/2007.

System 2 must be the Windsave WS 1000 system complete with the Plug’n’Save inverter. This can be deduced because there are only two systems that meet the basic specification on p10 (albeit with some errors) and the price of £1798 on p40 narrows it down to the Windsave.

I’m writing the above paragraphs because it will assist researchers who need an independent reference for identifying the three systems which are being examined as exemplars of the microwind industry. They are also the only three systems currently on the market which pass the basic adequacy tests for grid connected urban microwind products so they are truly typical.

The inventory analysis is pretty sound. The WS 1000 is largely imported from Asian sources and so I think the impact would be higher than is predicted. The RD Swift data is for the Mk1 Swift and they have changed inverter for the Mk2 so it may no longer be representative, also I’m not sure that the impact data for the carbon fibre blades truly represents end of lifecycle costs (but that’s just a suspicion of mine re the HSE costs). The Ampair data is for the EU2 version and so is up to date. The most important issue I have with the inventory section is the assumption regarding maintenance. The authors seem to disregard visual observation as a good way of initiating condition-based maintenance which is a pity. Our experience is that the combination of visual observation and a passable ear is appropriate. Maintenance issues don’t come on neatly in annual cycles – they arise much faster – so an annual inspection is very costly (economically and environmentally) and ultimately futile. Instead it is better to either assume condition-based maintenance or a range of lifetimes. In this respect I would be extremely suprised to see all these turbines lasting very long in a coastal environment as the only one that appears to be marine grade is the Ampair although again I am happy to be corrected. Constructing a marine grade turbine is an expensive business and is directly reflected in the purchase price of the units.

The urban wind resource estimation is as good as it gets at the moment in the public domain from a theoretical perspective. To a certain extent theoretical work can only go so far and practical work can only go so far and the final picture will only become clear when all the pieces of the jigsaw are fitted together (several times). The experimental wind tunnel data re flows over building roofs look fairly similar to results from CFD modelling work carried out by the Loughborough University Centre for Renewable Energy Technologies (CREST) team led by Simon Watson and are not unsuprising, i.e. higher = better, and ends = better, and ridges = better, and best of all is of course to be as far away from buildings and trees and other obstructions as possible.

The BRE comparison of Met Office windspeeds, NOABL windspeeds and predicted actual windspeeds is very interesting. In Appendix B they show the relative location of the Met Office instrumentation and the five sites for which they have predicted wind speeds. The first thing is that the shape of the wind speed distribution is very important in assessing the energy yield and that can be observed in the Fig 5 on p15 and Fig 7 on p16 where the more ‘marine’ a wind is the more energy it contains. This change in a wind’s character is of course not described by the NOABL database. Then the correction factor that BRE calculate using their BREVe tool (based on BS 6399-2) to produce a prediction is as yet not tested against reality (from the perspective of small scale wind) and so for now is a harmless exercise. Soon they will no doubt be cranking the same BREVe tool for comparison with actual winds measured from various trials and then we will get closer to the holy grail of accurate site specific prediction. Ideally such a tool will be predicting mean wind speed, distribution, and turbulence but that is of course an ideal which will have a rather large error bar on it.

(to be continued)