Is it okay to retrofit heat pumps before building fabric?

This article was originally published in issue 42 of Passive House Plus magazine. Want immediate access to all back issues and exclusive extra content? Click here to subscribe for as little as €15, or click here to receive the next issue free of charge

A couple of issues back I discussed the idea that in order to decarbonise our housing stock as rapidly as possible, the dogma of fabric first should be questioned. This led to an interesting question from Lloyd Alter, of Treehugger.com: “If you install a heat pump first and do the fabric second, is the HP not then grossly oversized? I understood that’s a problem with them.”

While this is a valid question, it’s an area where the solutions have evolved with technology and the answer is different to what it would have been if it was five or ten years back. I’m primarily talking about air source heat pumps in this article, but some of the points apply to other types.

The problem that Lloyd alludes to is to do with the risk of “short cycling.” This comes down to a straightforward way you can control a heating system which I like to call “IT Crowd” control; switching it off and on again. This is how conventional fossil boilers and thermostats work – they run, warm the building up, then switch off (fossil boilers can modulate but let’s keep it simple for now).

The controller has a little hysteresis, which means it doesn’t immediately switch back on as the temperature drops, but permits maybe a degree of drop and therefore some period of time before the boiler kicks back in. Due to the mechanics of heat pumps, this leads to inefficiencies – it takes a bit of time and energy to build pressure and temperatures back up to operating conditions, which you don’t get back at the end of the cycle. It can also shorten the life of the compressor as a lot of the mechanical stresses occur on start up.

But as with many things in sustainable energy, its more complicated than that. For a start, how do we size heating systems anyway? We *should* do a full set of heat loss calculations, and produce an “accurate” peak heat load – let’s say 10 kW for a typical-ish, not very well insulated house. If we’re a bit mean we might pick design conditions of 0C outside and 20C inside, giving 500 W/K. Assuming our heat loss factor is correct (spoiler alert, it’s probably not) if we went for a more conservative 24C inside and -6C outside, we’d need a 15 kW heat pump – the size is very sensitive to the conditions we pick as our worst case.

On top of that, it’s not uncommon to add an arbitrary 10 or even 20 per cent to a heat loss even for heat pumps, sometimes disguised as a “warm up factor” but is basically a fudge or comfort margin.

Readers of Passive House Plus will have some inkling of how inaccurate heat loss calcs might be (assuming they are done at all). I doubt that many (if any) installers on typical heat pump installations insist on an airtightness test being done, or carefully document and account for all thermal bridges (I was delighted to hear Nathan at the BetaTeach podcast mention he’s used a co-heating type test to quantify this, but it currently seems unlikely this will be a mainstream service). How are the U-values arrived at? If we aggregate all these inaccuracies, do we tend to over or underestimate?

But more importantly, how much does it matter? After all, looking at the range from one popular manufacturer, the domestic ASHPs come in 6, 8.5 and 11.2 kW. If your heat loss calcs come out at 7.25 kW, you have to choose to be 20 per cent oversize or 14 per cent under.

There are several more elephants in this room of uncertain heat loss. Firstly, there is the vast majority of the year when it’s nowhere near our worst-case scenario of zero or even -5C. A heat pump therefore spends most of its working life “oversized” compared to the actual demand, and we don’t worry about that.

Another lurking proboscidean is hot water; in order to re-heat a cylinder in a sensible period of time, we need a certain amount of power. A one hour re-heat of 200 L would require about 10 kW – significantly more than the peak heat loss from a well-insulated flat for example. There are alternatives to heat pump powered cylinders, but they are generally compromises and the subject for another column.

It turns out that unlike the Irish border we can throw a bit of technology at this issue to make it go away, namely inverter driven compressors. Happily, with a few exceptions, many new heat pumps have this feature. The inverter is a like a dimmer switch to the lightbulb’s toggle switch, allowing efficient operation at part load. This means that the heat pump’s thermal output can be turned down without (or at least, with less of) an efficiency penalty. There’s a lot more to unpack there, and I’ve probably upset a lot of heat pump engineers by oversimplifying. There are a couple of important caveats here – firstly that the hydronic design (i.e., pipe and radiator layout and sizing) is on-point, and secondly that the system is (re)commissioned to a high standard.

Both of these are large rabbit warrens in their own right; many subtle parameters can (and should) be optimised on some modern units. This means that we perhaps shouldn’t be too worried about the fabric retrofit (if it does indeed happen, and to the standard assumed) undermining the performance of the heat pump. It would also not be beyond the bounds of possibility to swap it out for a smaller unit, passing the partially used one on to a larger home.

An alternative approach is to put in an undersized heat pump before a fabric retrofit. This means that on the coldest days, the house might be a little chilly, or it is necessary to top-up with some direct electric heaters. While this has a running cost and carbon implication (although average grid emissions are now very low, the periods of highest demand depend on higher emitting generation), and have an implication on the infrastructure, this would of course be mitigated on completion of the fabric retrofit.

So, should we worry about accidentally oversizing heat pumps in a step-by step retrofit? Perhaps less than we used to. All of this depends on some important assumptions of course: that the heating system is carefully designed, with prior knowledge of the fabric retrofit; that a good quality unit is used with an inverter drive and a comprehensive controller; that the heat pump system is commissioned to a high standard, and re-commissioned after a deep energy retrofit to suit the new load; and that ultimately, that fabric retrofit is completed to a good standard.