FWIW, a modulating boiler is more efficient at its lower outputs, so as long as you have one that can meet your needs at the max, on those days when it doesn't need as much, it can modulate. Some have a larger range of modulation than others. Mine can go from 20% to 100% (a Viessman unit). This gives you a little more leaway when matching to the house's needs (but the unit cost is higher). It works especially well with radiant heat. Efficiency is quite good, but depending on the system, may or may not pay you back depending on the fuel costs (likely to only go up).
Most mod-cons have an efficiency sweet spot somewhat above the lowest fire. At lowest fire the transfer efficiency on the heat exchanger begins to fall due to insulating laminar flow of the gases on the fire side. Turbulence is necessary to scrub away the insulating boundary layer, and that only happens at higher burn rates. This is why most mod cons are limited to ~1/4 or 1/5 turndown ratios. If they go much lower they rapidly fall off an efficiency cliff (if dropping from 98% combustion efficiency to ~80% is considered a disaster, which it might well be from a stack temperature point of view when using a PVC vent.)
But a short-cycling mod-con is still a short-cycling boiler- oversizing them only to have them short cycle at low fire is not a recipe for maximum efficiency. With high mass low temp radiation (like radiant slabs) you can still get excellent performance out of them even 3x oversized, but it takes buffer tanks & design skill to keep an oversized mod-con running a true 90%+ with low-mass radiation in micro-zoned systems. Sizing a mod-con so that the average load during the heating system is well-aligned with the efficiency sweet-spot is ideal- better than oversizing. Most don't go much below 4-5kw (14-17kbtu/hour), which is already over my heating-season average, which means buffering or high-mass radiation would be required to keep the true efficiency up to snuff.
flamefix said:
I don't follow why the envelope of the building isn't relevant when calculating heatloss?
Because the load on the boiler as measured by the fuel use per degree day tells you the heat load at any outdoor temperature far more accurately than any wild-assed guesses as to the true U-values of all of the as-built structures, or the actual vs. presumed rate of air leakage, etc. The boiler is completely agnostic as to the construction details, responding only to the state of the thermostat. Yet it's fuel use & modeled partial load efficiency curve tells the entire heat-load story with significant precision, quite independently of ANY construction detail.
Heating systems designed for the heat load a minimum outdoor temperature that falls in the 95th-99th percentile of historical binned hourly weather data over several heating seasons are deemed "perfectly sized" for delivering adequate comfort without losing too much by way of efficiency. When the full output of the boiler matches the "design day" (~97th percentile temperature) heat load, the duty cycle & efficiency of the boiler are maximised.
flamefix said:
But whats a homeowner to do in your country if they had just moved in and hadn't got all the previous history on the boiler use age?
Manual-J is the standard here, and many software packages are based on it, as are some building-codes. IIRC California Title 24 now disallows oversizing of heating/cooling systems by more than 10%
per Manual-J. Of course, the calculation is easily skewed by the whims & prejudices of the software data-entry person, so the 10% limit isn't perfect. But it probably keeps most systems from being more than 1.5-1.7x oversized, in which case they're still at least somewhat close to optimal. (Among it's many shortcomings, the US seasonal efficiency "AFUE" rating for heating equipment presumes 1.7x oversizing relative to the design-day heat load.)
flamefix said:
Sometimes performing heat calculations seem nonsensical for instance: the heatload is calculated as requiring 15kW heat input but (a majority of homes have combination boilers) to get the flow rates for hot water anything up to a 35kW boiler may be specified to provide that flow rate (for power showers).
That is an issue here as well- the instantaneous load of a sustained hot water draw such as a shower can be several times the design-day heat-only load. The better-efficiency solution is to design in sufficient mass to the hot water storage rather than increase the size of the burner. (This is of course easier said than done.) There is a standby loss associated with the high mass storage, but it's a small loss relative to the efficiency loss of oversizing a boiler by a factor of two or three.
eg: The peak heating load at my home is ~8-9kw, with an average mid-winter daily load of 3-6kw, but the instantaneous load of a winter shower adds something like 30-32kw. The solution to this mismatch (and that of the micro loads of separate room-by-room zoning switching in and out) is ~400lbs of boiler water in a buffer tank with an internal heat exchanger for the water. The load from the shower has also been cut in half (to ~15kw peak) by using a drainwater heat-recovery heat exchanger to pre-heat the incoming flow from the city water. With that much mass and a reasonable temperature hysteresis to the tank controls it's fairly insensitive to boiler oversizing (the mass establishes a pretty good minimum burn time), but even with a "perfectly sized" boiler for heating there is plenty of hot water available for multiple successive showers. If there's an Achilles heel to the system, it would be that I can't fill an 80 gallon spa in one go with a "right-sized" boiler driving the combined system, since tub-filling capacity limited by the mass of the storage- the heat-recovery capacity is only available for simultaneous drain/hot-water flows, like showers, not batch draws.
Fortunately I don't have such a spa in my house.
But if I did, and put in a 30kw boiler in to be able to support that load, the efficiency of the combined system wouldn't be TOO atrocious, since the thermal mass of the buffer keeps it from short-cycling, but it's still less than ideal.