They don't perform as well as expected/documented. More to do with the testing regime though.
You can work out the performance pretty well from first principles. Just treat it as a wall with a known U-Value.
They are cylindrical for 3 main reasons.
Low surface area to volume ratio.
They take pressure better (hoop stresses).
They are cheaper to make.

Thanks both. I'll do some working. I'm finding it strangely difficult to find good stats on DHW cylinders. E.g. what's the average amount and quality of insulation a new one arrives with.

They appear to be obvious weak points in the efficiency of home heating systems. I know that you can retrofit extra insulation easily, but it feels like (I may be wrong) the standard of insulation of new 'off the shelf' cylinders isn't that great. Strange, because adding an extra 50-100mm of PUR shouldn't be expensive, and would make a real difference.

As I say, I may be completely wrong. Struggling to find good statistics on 'average' performance. And I'd love to know what the best performing ones manage, but can't find much on that either.

Posted By: GarethCStrange, because adding an extra 50-100mm of PUR shouldn't be expensive,

I wonder how much that's is down to practicality of transport, fitting through doors, etc, rather than cost of materials and application.

Posted By: ringiJust having pipework pre-installed that goes from the top of the tank, to exit the insulator lower down, could help cut heat lose.

Problem is where it goes after that. Since every installation is different, wherever you choose to bring the outlet is likely to be wrong, plus you somehow have to arrange to remove air from the loop and be able to drain it, So I suspect that putting the outlet in its natural place and then taking care with the external plumbing and insulation design is a better strategy. Plus open-vented tanks need it to go upwards anyway.

What do the loss ratings for cylinders assume for the pipework? Ignore it? Assume copper lagged to the same thickness as the cylinder? … ?

Using stainless for the first 300 mm or so would help a lot, I think. About 20 times less thermally conductive than copper.

[ Cross-posted with gravelld: another advantage of PCMs is they can store more heat at a lower (fixed) temperature. ]

Posted By: ringithere is a LOT of heat lose from the top of a tank where the pipes connect as it is hard to insulation

Yes, what I mean is that I think producing some better technique or products for doing this type of insulation would be more beneficial. It's obviously possible to use expanding foam in place once the plumbing's done but that can easily be or look messy. Something better might make a significant difference.

Posted By: SteamyTealower temperatures

That depends on heat source. With a PV-powered tank, it makes sense to try to get it as hot as possible, regardless of losses, because that maximises the length of dull period you can outlast. Of course, that kind of strategy also means you need to determine a long term average to get any meaningful numbers.

Posted By: GarethCthe standard of insulation of new 'off the shelf' cylinders isn't that great.

I agree - my one seems to be heating the basement cellar in which it is installed...

My electrician says that this is not a problem, "as it is in any event warming the house".
In effect, since our fresh air intake is in this room, I suppose that the water tank is preheating our inside air also.

I have made some attempts at calculating the heat-loss:
The tank is 250 liters and 3000 watts. Generally runs on off-peak hours only (= 8 hrs/day, of which six overnight and two at mid-day).

Townwater temperature = 10.5°C
No thermostat on the tank, just a mixing valve (set for “HOT”): HW temp = 50°+ (guessing at 52°C)
Therefore TD = (52)-(10) = 42K
fully charged tank holds:
250 x 42 = 10,500 kcal
x 1.163 = 12,211 watts

I only fully charge the HWC fully for 3 days at the weekend – in the week, I charge it in mid-day off-peak only = 2 hours. So it generally consumes 6 kWh per day, in the week, as verified by meter reading.

Our (estimated) HW usage is 2 showers per day + various washing/rinsing – total estimated usage = (2 x 50) + (5x5) = total 125 litres. = one-half of the tank.

Estimated electrical consumption for warming one-half of the tank = 6100 watts, which tallies more or less with the meter reading… Q.E.D.

(I did the estimates without any fiddling of the numbers – honest…)

So contrary to expectations and on the face of it, if the numbers are right, we do not appear to be losing much heat out of the tank at all…

However, we must be losing *SOME* heat, because the rest of the basement is between 16.1°C and 17.3°C, whereas the cellar where the tank is, is at 18.6°C.

Some of this heat might be due to lighting = 4 ft fluorescent tube (? 30 W), which is generally left ON.
(not a good idea, I know…).
=====================
I tried running online calculator at http://www.chromalox.com/resource-center/calculators/comfort-heater.aspx to determine the cellar heat-loss: it came up with 680 watts which sounds like A LOT to me (our bedroom gave 560 watts, and it is twice the size) plus has a 1 sq.m window…). (we do not have any heating in the bedroom) (just ourselves ).

The cellar is totally buried, 9 ft square x 7.5 ft high = 600 cft; no windows, bare cement floor; outside wall lined with insulated gypsum board. Backwall (= blockwork) insulated idem; other two walls = 3-inch hollow brick. Ceiling uninsulated.

Given that the air is changed continually due to the fresh-air intake, the numbers -- contrary to expectations again -- might in fact be right: because of the cold air draught, the cellar requires 680 watts of heat: the fluo tube probably delivers about 10 watts of useful heat (?), meaning that 660 watts are being delivered by the hot-water cylinder - seems like a lot: it means a loss-rate of 660 watts, out of an average daily input of 6 kW.

OK, the top half of the HWC will be generally HOT and account for 6 kW, and the bottom half will be "generally cool" so account for ? 500 watts ?

all enlightenment welcome

gg

(contrary to appearances, this is an attempt to contribute to the debate...)

er...Added factor, just realized: in the simulator, I set the design temps as "12°C external, 20°C internal" TD = 8°C.

So the 680 watts heat requirement is to get the cellar to 20°C, and we are NOT at 20°C, only at 18.6.
Can't decide (at such a late hour...) what this means !! eek

Posted By: Ed DaviesAll very confusing.

hmm - feared it might be !

gg

Ideally you want to heat the DWH tank top to bottom up to about 50c, then heat the complete tank as hot as possible. That way you get usable DWH even if you can't heat the complete tank.

Heat lose increases with dT, so it may be better to have a lots of water stored at 60c, then less water stored at 85c.

Are my maths wrong here?

A 200l (0.2 cubic metre) volume rectangular tank of dimensions 0.36mx0.36mx1.5m has surface area of 2.48 square metres.

Assume hot water at 45 degrees, and room temperature of 18 degrees, so 27 degree delta.

Assume 100mm of foil faced PUR insulation giving a u-value of 0.235W/m2K (so ignore thermal resistance of tank walls, and don't account for 'piercings' yet).

So heat loss is 0.235x2.48x27=16W

So 24 hour heat loss is about 24x16=378Wh, 0.4kWh

Too low surely, what am I doing wrong?

45 °C is probably a bit low. The lowest usable is really around 38 °C (shower temperature). With a thermostatic mixer you'd need a few more degrees as they generally let a bit of cold through whatever. Then to store a useful amount of heat you need more degrees still.

Also, “piercings” are, I understand, a significant part of the losses.

[Huh, cross-posted with DJH. Perhaps using the middle of the insulation for the area calculation would be a good approximation.]

Posted By: gravelld'Piercings'? You mean the pipes in/out?

yes - and sensor ports etc.

Posted By: Paul_BI have temperature sensors at four levels

I'm trying to decide where to put my sensors. Are there any particular places on the tank you would recommend? Do you think four is the right number, or doesn't it matter?