Thanks for your interest and your detailed and very interesting explanations. I've re-arranged the quotes of your message a bit to facilitate this reply! ...
The dT50 rating is a hangover from pre-condensing boilers that needed high return temps to prevent condensing and then corrosion at boiler return.
Since condensing boilers came in typical design would be ( (75+55)/2 )-21 = 44. correction factor for a dT50 rating would be 0.846
Current practice would be ((55-35)/2)-21 = 24 correction factor 0.383.
.... The correction factor falls off a little faster than Pro-rata and can be looked up on internet lists or calculated (actual dT/spec sheet dT)^1.3
Fair enough, and thanks for answering what was going to be my next question (how to determine that 'correction factor').
The examples you quote assume a temp loss in the rad (flow-return) of 20 C, for differing flow temps. Why is that,and is it necessarily the case in practice? In fact, one presumably can change dT by altering flow rate (hence altering temp drop in the rad) without changing flow temp, can't one?
* Having calculated a heat loss for room, adjust for dT of system which will be average rad temp ie flow+return/2 minus room temp, this obviously varies subject to the type of heat source and the level of efficiency at which you want the design to operate (lower is ALWAYS better).
* Having corrected for spec sheet to operating dT you would apply a correction factor for the method of water connection - in domestic work connections are invariably Bottom Opposite Ends or BOE which meean rad derated by 0.943
* Any significant cill or shelf over rad would usually attract a derate factor of 0.95 and a rad cabinet 0.7
* The final adjustment would be applied on basis of operation - if the heating is going to be regularly turned off with energy saving controls and then required to bring a room back to temperature in a relatively short period then an output uplift would be required somewhere from 10-20% for resonably well insulated property reheating in half hour or so.
Having estimated heat lost as accurately as one can, and then applied all of those adjustments is all very well (and sounds like an attempt at 'pretty good science') but then seems to be completely overshadowed/undermined by ....
Having calculated a heat loss for room, adjust for dT of system .... <for> the type of heat source and the level of efficiency at which you want the design to operate
AND
The new flow temps are to make rad systems heat pump ready but make rads impractically large in most domestic premises so you will find many ignore the regs!!
It therefore seems that (in domestic settings) having attempted to undertake a series of fairly accurate calculations and adjustments thereto, one then applies an essentially arbitrary (and potentially massive) 'alteration' dependent upon what flow temp/dT one 'wants' and what is 'practical'. You refer to the 'many who ignore the regs' somewhat pejoratively, but if to do otherwise would be impractical (or just unaffordable), then it would be impractical, and therefore would/could not be done.
So this is where a science seems to turn into more of an art - with the bottom line that, regardless of all the calculations and adjustments (and 'regs'), the rads which are installed will often simply be 'the rads that will fit' (i.e. 'practical'). Is that not the case?
.... if the heating is going to be regularly turned off with energy saving controls and then required to bring a room back to temperature in a relatively short period then an output uplift would be required somewhere from 10-20% for resonably well insulated property reheating in half hour or so.
It's not really something which concerns me very much, but the 'warming up' process needs somewhat more sophisticated modelling, and probably happens faster than some would imagine - because,if one starts heating a room from cold, the temp difference across walls/floors/ceilings etc. (and hence heat loss) is initially pretty (or very) small. Our usual heat loss calculations refer to the steady state after 'warming up', such that 'heat loss' of a room is a constant. However,during the warming up phase, heat loss starts low and then gradually increases as the room warms up.
Rads are traditionally sited under a window because glazed areas will usually have significantly greater heat loss than walls and the cold air against the glass will create a downdraught that can then be counteracted by the rad and the resultant turbulent mixing will lessen the convection current effects already discussed in this thread. ....
Indeed, and ...
More uncharitably the space under the window gives a rad size for those that can't manage the above process
Not really 'uncharitable', since it's an undeniable truth that such is what many have done (and probably still do). As I've said,it's certainly what I've done in the (fairly distant) past,and it always seems to have resulted in a system which worked reasonably well, if not necessarily very efficient! Furthermore, as above, it seems that this is probably essentially what very often still happens in domestic settings - since the science/calculations are likely to call for impractically large radiators , such that what gets installed is "what will fit"
Trv sensor are best mounted horizontally where possible for exactly the reasons you have mentioned ...
I can see that would help a bit to address the issue I mentioned, but I have very rarely seen it done, and cannot recall ever having seen any TRVs for sale that would enable that to be done directly - since, whether 'straight', 'angled' or 'corner', the sensor head seems to invariably be at right angles to the spigot that goes into the rad. If one wanted a horizontal sensor, it therefore looks as if one would have to improvise with a 'straight' TRV and a M/F 1/2" elbow - which would be a little messy (and a little'wide') ... or am I missing something?
... but as they are rarely calibrated in actual temperatures any offset caused by local 'microclimate' can just be adjusted for with an appropriate setting.
You speak as if there is just a fixed 'offset' between the temp seen by the sensor and the temp of the room, but is it really that simple? My 'concern' was that, for example, when a radiator is first turned on in a very cold room, the temp 'seen by' the sensor (which is 'up close to the radiator' will presumably initially be a lot higher than the actual temp of the room air in general (as a result of radiation, if not also some convection) but that difference between sensor temp and general room-air-temp will gradually decrease as the room warms up?. However, as I've said, they seem to work fairly well, so my concerns are presumably largely unwarranted!
Kind Regards, John
