Aperture areas, panel outputs and test reports

[wookey]

I've started a wiki page on areas and test reports : http://www.navitron.org.uk/pmwiki/pmwiki.php?n=SolarThermal.PanelAreaAndEfficiency

That needs more work, but I need to clarify some things first.

I've put up the definitions of gross, aperture and absober area that are used by Fraunhofer ISE and the gross and aperture defintions for SPF. I can't find the definition of absorber area that SPF use.

The definitions are not the same, which may explain why the numbers from the two institutions differ.

Here are the numbers quoted by SPF and ISE for a selection of panels:

Panel
Navitron SFB20	2.403	1.747	2.655
AMK OPC-10	1.145	1.143	1.650
Consol 22-47-1500	2.624	1.356	1.17
Eco-nomical 20x47 (calculated)	2.74	1.26	1.15
Eco-nomical 18x58	2.100	1.706	1.451
Xinwang ANK20x58	2.853	1.869	1.614

I've tried to choose ones that either we know something about or are similar to the nav 20x47 tested. This has only been somewhat successful.

Looking at the Fraunhofer test sheet (link below) for the economical 20x58x18 panel (the only one they got tested) we find some numbers that make sense. The area of the inside of 18 tubes 1.8m long does indeed come to the 1.706m2 calculated. It's easy to do the same sums for the 20x47 panel and get convincing aperture and absorber areas for that too. These agree with the ones quoted on eco-nomical's website:  http://www.eco-nomical.co.uk/SolarWaterProduct.htm no doubt because they've done the same sums.

However the SPF numbers are much harder to make sense of. One big problem when comparing panels is that nowhere in their report do they say how many tubes a panel has or what their diamter is. Without these numbers it's impossible to do the sums. However we can find some panels for which we do know the number of tubes and their sizes (although you have to be careful - eco-nomical got an 18-tube panel tested but now list,sell and quote areas for, a 20-tube panel - such things might apply to other tested panels too). The consol has 22 tubes. My understanding of the 47 and 58mm tubes is that they pretty much all have identical external dimensions so panels with same number and size of tube should have very similar (the same?) aperture and absorber areas.

Navitron quote "The 20 tube double-walled tube collector (47mm tubes) is 1760x1500x130mm (LxWxH)" in their FAQ. That suggests a gross area of 2.64m2. SPF say it is 1640mm*1470mm=2.4m2. SPF aren't including the pipe spigots. I guess Navitron are. You'd think that measuring the simple panel area was simple enough, but we already have a 9% discrepancy here.

Looking at the various aperture and absorber areas in the table above you can see that the numbers are in proportion for the 2 58mm panels (with 18 and 20 tubes respectively). Similarly the aperture areas for the consol and eco-nomical panels match up (although the absorber areas don't). However the Navitron numbers are completely different. The only explanation for this is that the SPF defintion for absorber area changes if the panel has a reflector: 

The aperture area is the surface of the collector, through which non-concentrated sunlight enters. (tubular collectors without reflector: inner tubular diameter x length x number of tubes, with reflector: largest projected area of the collector)

Perhaps if you totted up the area of the reflectors+tubes on a navitron SFB20 you'd get 1.75m2 ? Anyone care to check?

The only other panel in the list where the absorber area is bigger than the aperture area is the AMK OPC panels so I've added that to the above table. The ratio of aperture to absorber is similar in both cases, but not exactly the same. That one has a reflector drawn and listed in the diagram, however the Navitron one doesn't. Did the tested panel have reflectors fitted? I guess it did but they have been missed off the diagram (which is rather slapdash).

Now I haven't even got on to stuff about incident angles and the efficiency co-efficients but this post is long enough already and it's bedtime(!) so I'll stop now.

My conclusion from the above is that Fraunhofer write much better reports than SPF - they make sense, definitions are given, details are recorded. Actual logged data is tagged on the end, and paicture of the actual panel under test is included. However they don't put them up on their site, and I can't see it on economical's either. So I've put a copy here for people who wish to compare reporting styles: http://wookware.org/solar/economical_test_report.pdf

The thing I haven't worked out yet is how to normalise the SPF Nav SFB20 data for what we can all understand as a sensible definition of aperture area (i.e. ignoring the reflectors), and recalculate the efficiency coefficients. I don't think it's very hard, but it may be difficult without the original data. Any stat's gurus here?

Sorry if this is a bit rambling, but there is definately an issue here it would be good to properly understand so we can do meaningful sums.

[ericw]

I'm with CeeBee on this one. - The only thing that really matters is the actual output of the panel.

SPF obviously measure the output characteristics and then calculate the various parameters based on some definition of area. When you use their efficiencies quoted with the appropriate area (regardsless of how obsurely it may be defined), you get the same output. 

Efficiencies taken alone are only of use for point scoring 'marketing' statistics.

Interestingly enough if you look at all the panels with the same diagram for the cross-section as the Navitron one, the vast majority have a very similar outputs.

As regards the accuracy of the results - as there are only a few actual manufacturers, it is quite likely that the same panels are available branded as something else, which may have been tested by another test house.
Does anyone in the industry know of these cases?

[martin]

The vacuum tube system was invented in a Chinese university, and the vast majority of panels flogged worldwide come out of China - there are many companies in Europe particularly who "rebrand" the tubes to try to make them appear to have been made in Europe, and charge premium prices!
What's important is that the importer has a constant trade with known suppliers, who maintain high standards. Some of the real "cheapos" as are found on Fleabay can be of very inferior manufacture/finish

[wookey]

eric, I agree, we only care about actual outputs (except possibly people with tiny roofs and big families :-)

But, (apart from doing our own measurements) the only definitive source of said outputs is these test reports. So understanding what their numbers mean is helpful to avoid cockups like me using completely the wrong number for 'area of panel' when working out how much I needed. The 3 characteristic values they derive for a panel define the output curve against ambient/fluuid temp diff, and that, combined with the angle modifiers and weather allow us to work out how much heat we can potentially get out a given panel, and to allow for the way this varies as the ambient and circulating fluid temps change. Just approximating with one fixed 'output' is a little too crude for my taste, especially if trying to understand system performance as circulating temps change. On the other hand  I note that the solar association agreed in 2004 that you can multiply any sort of panel area (unglazed, glazed, vacuum) by '0.7' to get output for the purpose of quoting installed capacity as in fact they all come out pretty similar for typical (summer) conditions: http://www.estif.org/fileadmin/downloads/Technical_note_solar_thermal_capacity.doc

Thinking about it, I agree that so long as you use the quoted area for the panel in question correspondoing to the efficiency numbers (no matter what it actually is) then how that area is defined does not matter. This does underline the almost completely arbitrary nature of the quoted efficiency numbers.

[jwt]

Does the aperture area matter?

If you are mounting panels all you are interested in is output v price v total size.

To take a stupid example imagine an evacuated tube system using 2 47mm tubes in a frame 2m x 2m. All I really care about is how many of the panels could I fit on the roof, how much each one costs and what the output per panel is.

JWT

[CeeBee]

Does the aperture area matter?

No - I think not. But it's a pity that the published SPF report of the Navitron panel uses a seemingly artificially high area, hence a low 'efficiency', which makes the panel look bad when compared to others on the SPF site. If everyone who went to the SPF site comparing panels looked at all the numbers in the fine detail that we do here on the forum, then they might realise that the 'efficiency' figure doesn't really matter, but on a cursory glance, the SFB-20 looks to be a poor panel.

I don't know whether the Navitron folk have any gripes over the SPF-measured total output of the panel, or just with the meaningless area/efficiency figures.

[ericw]

This seems an appropriate thread to ask the question.

Is the published variation of output with angle (IAM), real or a peculiarity of the measurement method ?

I'm not convinced by the Apricus explanation (http://www.apricus.com/html/solar_collector_efficiency_iam.htm )

I would have thought that once the incident angle of the sun gets to the region where the tubes are not shaded by each other, because of the cylindrical nature of the tubes, a constant absorber area is presented to the sun and so one should expect a constant output until the tubes become shaded again.

One might further argue that once the gaps between the tubes are visible to the sun, it is possible to reflect the radiation going through these gaps to the back of the tubes and so increase the output.

[CeeBee]

Is the published variation of output with angle (IAM), real or a peculiarity of the measurement method ?

Referring (I guess) to the 'IAM' graph on the SPF report for the Navitron panel. The factor increases as one moved away from normal (perpendicular) incidence, before dropping as the angle becomes more oblique.

I've not attempted to read up precisely what the measurement is supposed to mean, but I think it must be that as one moves away from normal incidence, the amount of solar energy hitting the panel decreases (as the cosine of the deviation from normal), but for a tube panel, then the energy produced remains the same until one reaches the point where the tubes shade each other. So: less energy hitting panel, but same output produced leads to a figure greater than 1.

The above is talking about the 'transverse' IAM. The 'longitudinal' one tends to stay at 1 until large angles are reached, and for 'flatties' both remain at 1, which is what makes me think that the above is the way it works. Indeed why does it fall away at all at large angles - must be becuase of things like the energy just being reflected at oblique angles and not entering the collector at all. And yes - the Apricus explanation sounds like 'flannel' to me, but it's not easy to try to explain.

ETs perhaps have many advantages, but I think that it's pushing it to argue this 'passively tracks the sun' thing. That's only because there are gaps between the tubes. If we built a panel without gaps, then I expect it would collect more at its peak, but the curve would no longer go >1 at either side. It's a compromise how big to make the gaps between tubes.

[ericw]

CeeBee,

Does the cosine function not just apply to a horzontal surface where the incident angle does change?
In the case of cylindrical tubes, a same sized fraction of the absorber is always illuminated by the sun.

Eric

[CeeBee]

Does the cosine function not just apply to a horzontal surface where the incident angle does change?
In the case of cylindrical tubes, a same sized fraction of the absorber is always illuminated by the sun.

Hoping we're getting at the same thing here. Yes - I just meant that the cosine function affects the amount of solar energy hitting the whole panel - the total rectangular area bounded by the frame. It's the amount of this area that's occupied by useful absorbing suface that varies with panel type. For flatties, it's always ~100%, for ET's it's (say) 80% at normal incidence because of gaps, rises to 100% as the gaps close up, then stays at 100%.

And I'm ignoring reflectors etc. which try to 'fill in' the gaps.

[dhaslam]

The efficiency graph shows a big performance drop when Tm-Ta is maximized, as much as 50% drop.    Tm seems to be the average water temperature across the collector and Ta ambient temperature.    This suggest that high pump speeds are best to keep the collector temperature down.    I wonder if this is the reason for some of the poor June performance figures.

[ericw]

The efficiency graph shows a big performance drop when Tm-Ta is maximized, as much as 50% drop. 

An alternative way to look at this effect is to assume the energy absorbed by the panel is independent of temperature but there is a temperature dependant loss associated with the panel.

Using the SPF data this loss calculates at around 3w/degree C difference between the mean panel temperature and ambient.

While this does show that higher flow rates giving lower temperature rises are advantageous, remember that the panel temperature is also set by the temperature of the cylinder coil which is likely to be the bigger contributor.

[CeeBee]

While this does show that higher flow rates giving lower temperature rises are advantageous, remember that the panel temperature is also set by the temperature of the cylinder coil which is likely to be the bigger contributor.

Indeed. In the mathematically limiting case of an infinite (well very high!) flow rate, the panel would only heat the water by a smidgin as it passed, the water everywhere in the circuit would be at essentially the same temperature, and this temperature would settle at whatever above the cylinder temp was necessary in order for the energy produced at the panel to be dissipated by the cylinder coil.

At lower flow rates, the panel heats the water significantly - flow and return are at different temps - such that (very approximately here) their average is the same as at the single temp for very high flow rate.

1 litre of water heated by 1 degree requires 4184 joules of energy. So if flow rate was 1 litre/minute, then 4184 joules per minute would be required to raise it by 1 degree. That's about 70 watts (4184/60). Let's say a panel produces 1kW (might be 2kW,  might be 500W, doesn't matter for our purposes here - we know this is in the right ballpark). At a flow of 1 litre/minute, a 1kW panel raises the temp by 1000/70 = 14 degrees. At 2 litres/minute, it would be half that, 7 degrees, etc.

So, with cylinder at 60 degrees, at a very high flow rate you might have water at 70 degrees everywhere in the circuit (I just made that up - I don't know the efficiency of the cylinder coil, but it's about right, and doesn't need to be accurate). At a flow of 2 litres/minute, you might have flow at 73.5 and return at 66.5 (to get the 7 degrees difference mentioned above).

So indeed, unless your flow rate is very low indeed (say 1 litre/minute or less), or you have a large number of panels, then this isn't going to make a vast difference to the heat loss, since we're talking about a panel around 70, and ambient around (say 20).

[ericw]

Is the published variation of output with angle (IAM), real or a peculiarity of the measurement method ?

After a long search of the web this does seem to be a real effect and the output of ET panels is higher when the sun is at an angle 50 degrees rather than at 90 degrees.
For those of us who are long enough in the tooth to remember Fortran90 there is even a program to calculate it http://www.colostate.edu/orgs/SEAL/research/IAM/iam.html

The usual explanation given is that its due to reflection from the surface of the neighbouring tubes increasing the active absorbsion area.

[Simon]

...
For those of us who are long enough in the tooth to remember Fortran90 there is even a program to calculate it http://www.colostate.edu/orgs/SEAL/research/IAM/iam.html
...

Fortran90!!, I wonder how long my teeth must be; I still remember Fortran IV which I used extensively when working at Lloyd's Register of Shipping in the early '70s.

Cheers

Simon

ps I still have all my teeth!