gemini://ur.gs/blog/heat-pump-installed/

Today the heat pump was commissioned and brought into service. It took 7 working days in total (Tue - Fri last week, Mon - Wed this week), which was 3 days longer than promised, due to a missing part. Not ideal when there's no heating, but the weather stayed fine, and wife + baby decamped to the grandparents for the first week, to avoid the noise.

The pump itself is whisper-quiet, but ripping out the existing hot water cylinder and replacing it with a bigger one (150L -> 190L) with a beefier heat-exchange coil was definitely not quiet, so decamping was a good decision. I stayed throughout, and it was a bit of a trial.

The one aspect of the install that I'm not very happy with is that all the levers are out in the open; it's a private back garden, but in a few months the baby is quite likely to want to start fiddling with them. I plan to fit a sort-of meter box around the controls at the bottom of the trunking; that should be enough for peace of mind. The grille keeps the fan blades far enough away that even baby fingers are safe there.

ASHP installation is a complex business, and I only understand a small proportion of the considerations that go into it, but it might be of interest anyway - and just writing it out helps solidify my own understanding (and also my understanding of what I don't yet understand!)

The first step in the process was a heat loss survey - a nice person comes to the house and calculates heat loss for the whole property, and then goes room-by-room doing the same thing, using pretty much the method described in my post about U-values.

The room-by-room survey is essential, because it lets you decide whether the existing "emitters" (radiators, under-floor heating if you're lucky) are sufficient to keep that room warm with the new heating system or not. In my case, all the radiators were over-sized given the gas boiler (very common), but 4 were under-sized given the heat pump, so I had to replace them with larger ones.

The radiators emit less heat when attached to the heat pump because the water in them is a lower temperature. The gas boiler would heat the water up to at least 65°C; doing that with a heat pump is basically impossible.

You can run the numbers with different "design temperatures", i.e., a spec for the heat pump. The lower the design temperature, the more efficiently the pump runs, but the less the emitters can heat the rooms. Since you also need a greater volume of water to deliver the same amount of heat to the emitters, pipe sizing becomes an issue too - can the existing pipes deliver the water, or would they need upgrading as well?

Honestly, if I'd had to switch to larger pipes to all the radiators in the house, I wouldn't have gone the heat pump route. That's just too much mess and noise for me.

In the event, the existing 10mm pipework was deemed OK with a design temperature of 50°C. With a higher design temperature (55°C is the practical maximum), I wouldn't have needed to replace two of the radiators, but it's pretty cheap to do and allows the whole system to operate more efficiently.

With a lower design temperature, I would have needed to replace some or all of the pipes, and more of the radiators. I might also have needed to switch to under-floor heating in some areas. I might still do that in future - especially for the kitchen and bathrooms - but definitely not this year. A design temperature of 35°C is not uncommon when you have underfloor heating throughout, for comparison.

The pipes comes with one more kicker - being 10mm, they don't hold a great deal of water. All together, the total volume of the central heating system is quite small (the hot water cylinder is physically separated), so the heat pump can get it up to design temperature very quickly, then has to shut down. The water cools quickly as the heat is transferred to the emitters and/or the hot water tank, and the pump has to start up again to repeat the cycle.

This is called "short cycling", and it's murder for a heat pump's performance. It takes heat pumps a few minutes to get up to maximum efficiency when they're started; fewer, longer cycles means you pay that cost less often. To achieve that, a "volumiser" - which is basically a buffer tank - is installed on the central heating circuit. This adds about 30L of capacity to the system, and also breaks the direct connection between the heat pump and the emitters.

This is something I wasn't expecting - rather than being a "split" system, where the refrigerant is heated outside then pumped inside, the water that goes through my radiators is pumped outside to the heat pump. It does the heat exchange with the refrigerant out there, and pumps the hot water back to the volumiser. A separate pump is responsible for getting the water into the radiators from there. I don't really understand why that's a good thing, but I'm assured it is.

Another job the volumiser does is help to tweak the "delta temperature" of the heat pump. This is the difference in temperature between the cooler water going to the pump, and the hotter water leaving it after being heated. The smaller the delta, the more efficient the pump is, but the more volume you need in the system. Mine happens to be designed around a delta temperature of 7°C; 5°C seems to be the current target. A higher delta also lets you get away with smaller pipes, by reducing the flow requirement; perhaps that was part of why my system is designed this way.

The heat pump is located in the back garden, which is north-facing. This isn't ideal - it's the coldest part of the curtilage, and the pump works more efficiently when it's warmer outside - but none of the other faces were suitable. There's no front garden to speak of, and the east+west faces are too close to the neighbours.

Even in the back garden, it was a close call. To be "permitted development", a heat pump's noise as perceived by said neighbours must be at or below 42dB(A); there's a methodology for it based on the presence of sound-reflective surfaces, distance from neighbouring properties, etc. I **just** scraped in at 42dB(A) - if my garden were slightly smaller, the neighbours would be slightly closer, and I would have had to apply for planning permission, find a quieter heat pump, put up bigger fences... there were possibilities, but not good ones.

In practice, the noise is fine. You can chat next to the pump without noticing it's there, and at the boundaries of the garden, you really have to strain to hear it. As a background drone, you tune it out very quickly. The cold air it blows on you if you stand in front of it is much more noticeable, and will be welcome in summer when producing hot water. It's basically air conditioning for the garden!

Anyway, since the water is going outside, it needs to be protected from freezing. The heat pump can do this by working constantly when the temperature drops, but that's energy-expensive. Instead, the water has glycol (antifreeze) added. This has two drawbacks:

It doesn't flow as easily through the system; air bubbles and unevenly heated radiators are more common

Obviously, you shouldn't drink the radiator water, that's a third drawback. It's better than spending the extra energy on keeping the water above freezing on a cold night, though.

The "domestic hot water" - taps, showers, baths - comes from the 190L cylinder, rather than the volumiser; the glycol rather demands that it be so. This makes it an "indirect" cylinder, I think. The heat gets in simply by circulating the central heating system's water through a coil in the cylinder. I'm very unclear on the plumbing details, except that if the coil cracks, we're in for a really bad time. Here's a picture of the cylinder in an intermediate stage of installation:

There's an integrated 3kW immersion heater; it's wired up to a timer switch that heats the water to 60°C once a week. This is the "legionella purge", and it costs 156kWh/year, or ~0.42kWh/day. Not ideal, especially as the risk is mostly theoretical, as I understand it, but a regulatory requirement. Hopefully I can meet the requirement entirely from excess solar generation; otherwise, regs permit it to be reduced to once a fortnight (~0.21kWh/day average) in some circumstances. This will be a very literal risk/reward conversation, if it happens!

Finally, there's deciding when the heat pump should switch on, and what it should be doing - hot water or central heating. "Smart thermostats" have become very popular in recent years, but are a non-starter for me. I asked for their Luddite option but they struggled to get one of those delivered, so I ended up with a Hive.

- Hardwired basic control unit with a button for hot water and a button for central heating

No doubt, the base station also exfiltrates usage data, so that went in the bin straight away. The other two units work perfectly well without it, and there's no need to sign up for a hive account. The complex unit exposes all the required functionality; right now, I have central heating on manual, and hot water set to run from 1pm-2pm every day.

What I want to be able to say is "do the hot water when the solar panels are likely to supply the greatest proportion of necessary power" - it doesn't matter when in the day that is, since the cylinder stores heat - but the complex unit definitely can't express that. It might be something a non-evil home automation system can express.

Metrics are non-existent at the moment. I'd like to be measuring CoP (coefficient of performance, kW of heat out for kWh of electricity in) over time, at a minimum; again, there are non-evil home automation projects that can collect that. Something to explore another day.

As long as the CoP is greater than the gas:electric price ratio, I'm saving money. If it drops below that, I'm losing money. There are a lot of compromises in the installation, as detailed here; if you were building a new house around a heat pump, it would be very different, and the performance would be higher.

The modelled seasonal CoP (SCoP, average CoP over the year) is 3.57 for heating, and 1.68 for hot water (this includes the legionella purge electricity, which has a CoP of 1; ignoring that, it's 1.75). With standard assumptions on heating and hot water use, that gets a combined SCoP of 2.73. I reckon I can do a fair bit better than that, overall - our hot water usage is pretty low compared to their assumptions - but I won't know without measuring it.

I re-did my energy demand using ~180 days of historical data from the current supplier's website. It includes half of the winter months, and it wasn't a mild winter. We also had a fan heater (mocked up to look like a fire) in the living room that we used a bit in the early days of living here; that has now been removed and consigned to the garage. So even this might be an over-estimate.

A price cap has all sorts of implications for payback periods and so on, but it will be decidedly helpful for most people, and is to be applauded regardless of who orders it. We can quibble about the implementation details, but not plunging ⅔rds of the population into fuel poverty has a lot going for it. I'm still happy to be getting rid of gas and installing solar capacity, whether it makes financial sense or not.

The French approach seems to be a price cap combined with taking at least EDF into public ownership when it complained that being forced to sell energy at a loss was a bad thing. There's a lot going for that approach, especially when the companies in question are producing gas domestically.

We'll find out more about the UK scheme tomorrow, but what's been trailed is some kind of massive bung to energy companies, funded from borrowing. Conventional wisdom holds that this is getting future taxpayers to pay for it, which is at least vaguely progressive, but hey, a sustained period of high inflation could erode the debt significantly before then.

Depending on how badly thought out the minutiae of it all is, there is a very bizarre possibility: if import gets capped, but export does not, I might be exporting generated solar to the grid at, say, 50p/kWh, while importing at 28p/kWh. From Octopus's point of view, it's even weirder - they buy it from me at 50p/kWh, and are then compelled to sell it to my neighbour for 22p less than they paid, with the difference being made up by a credit from the Treasury. It would be a huge solar subsidy, but it would at least have the effect of getting everyone to turn off their PV diverters and maximise grid export.

On another note, I've been graphing Don't Pay UK's sign-up figures over time, using a little HTTP scraper + cron + LibreOffice. Once an hour, I grab the current number of signups and write it to a file, Once a day or so, I copy-paste that into a spreadsheet that has a chart set up with an exponential trend line. Here's how the chart looks:

The deal is that the collective pledge is only put into action if 1M is reached by 1 Oct, and we're a long way short of that. The trendline jumps around from day to day, but it's generally projecting somewhere between 225,000-375,000 by then. Oh well.

As a menacing pressure group, reminding government and industry that we don't actually have to play by their rules, it seems to have done pretty well, though.

Heat pump installed

Installation details

Off-grid in suburbia: errata

Don't Pay UK