Improve efficiency: Heat the hot water when it's cold

[borpin (Brian Orpin)]

@stephen

there are some (ie bradford white → kwick shot)

Wrong side of The Pond

[Timbones (Tim)]

I made a chart showing the effect of the starting temperature of the tank on the COP of the DHW cycle. (Posting in this old thread as it seems directly relevant)

[MyForest (David Bowen)]

As it’s been a bit chilly in the UK recently I’ve entered a new situation where space heating just runs continuously. My old hack to do DHW when there’s a quiet moment is no longer working.

The Ecodan FTC has an option which says to heat the water if the tank temperature drops by a certain delta (between 5 and 15). With that limited control I’d end up heating the water when someone had a shower in the morning. I’d prefer to wait until later in the day when it’s warmer. Of course I could schedule this in the FTC but I have an aversion to trying to set up timers on the user interfaces that have very few buttons.

So I added some code to my algorithm that now interrupts space heating if the hot water is too cold and the DHW schedule says it has priority (via a new crontab).

Consequently, today it:

• did not heat in the morning after the shower at 06:45 when it was -1 outside
• did interrupt space heating at 12:00 when it was +2 outside

As a result the CoP for the DHW was 2.5 versus 1.8 that I see when it’s colder. Still not as good as we’d like, but an improvement caused by delaying heating the DHW until the environmental conditions were better.

[mjr]

The problem I have is spotting the difference between someone having the last shower of the morning and there being someone else wanting a shower (and not a cold one!). Not unreasonably, the inhabitants are of the opinion that the tech is there to fit in with us, not the reverse. So I mess with the probability of a reheat by changing the target temperature through the day, but I just accept that sometimes we’re going to reheat water in worse conditions than needed. We’re still so far ahead of the stock performance that it doesn’t matter.

@MyForest

Of course I could schedule this in the FTC but I have an aversion to trying to set up timers on the user interfaces that have very few buttons.

I’ve programmed far worse. At least this one has a decent display and doesn’t require me to lie on the floor! The benefit of programming a basic schedule into the FTC is that it keeps working even when Melcloud can’t get through. It might matter less if you have a cn105 connection, or at least that melcloud injection proxy.

[MyForest (David Bowen)]

You’re quite right.

Fortunately we’re down to three occupants and the tank can do two showers so I just lose out if my tech fails

[Timbones (Tim)]

My current strategy is to heat the tank at 1pm each day (start of Cosy cheap tariff), but only if the tank has dropped below 25°. My chart above shows that I get better COP below this temperature. Practically I find that if temperature (measured at the bottom) is above 25° then there’s still plenty of hot water for a shower or two, so a reheat isn’t needed.

(I’d really like to measure the temperature in the top pocket as well to refine this further, but that’s for the future)

This isn’t working so great in cold weather though, as there’s an extended period of non-heating: 12-til-1 is waiting for the cheaper electricity, and 1-til-2 is doing DHW. I either need to keep heating on until 1pm, or push DHW to 2pm. Maybe some smarts that decide when it’s best? Or just move it to early morning when it heating is less important for comfort (cos everyone’s in their bed) even though ambient conditions are less good.

[MyForest (David Bowen)]

I solved your problem by switching to DHW at the moment the algorithm was about to stop space heating. If you recall I changed from “Force” to “set a higher target tank temp” approach which has been meh.

The Ecodan itself will switch to space heating when it’s done with DHW or my algorithm steps in and sets the tank target temp back to 10 Celsius.

I guess this doesn’t help because you’re trying to do it on-board in the FTC to keep life simple.

If you set the scheduled timer for DHW (say at 14:00) and the “only heat when it’s dropped 15 Celsius below target” does that allow space heating to run under Cosy from 1-> 2 and then (optionally) do DHW or just keep space heating?

even though ambient conditions are less good

Your problem is the second daytime Cosy which is encouraging you to DHW in the day. As I’m on Intelligent I don’t have that luxury so I do a cheap but inefficient DHW during the night.

Do you think Octopus want heat pump folks doing DHW in the way you are doing it? It sounds like you should. However, if you have 1 hour DHW in that 3 hour window then you only have 2 hours of heat input to make it through to the next morning. Unless you are a smarty like Timbones and you have a battery

[Timbones (Tim)]

@MyForest

I solved your problem by switching to DHW at the moment the algorithm was about to stop space heating.

But if the heating never stops…

@MyForest

I guess this doesn’t help because you’re trying to do it on-board in the FTC to keep life simple.

I am actually sending “Force” command from a cron job, with some simple logic to decide if it wants to or not. Worst case, if internet happens to be out at that exact time, I can to force it on manually via internal controls.

@MyForest

If you set the scheduled timer for DHW (say at 14:00) and the “only heat when it’s dropped 15 Celsius below target” does that allow space heating to run under Cosy from 1-> 2 and then (optionally) do DHW or just keep space heating?

Prohibiting the hot water during the other 23 hours with a large “DHW max. temp. drop” setting would in theory achieve the same effect. However, I found that the Ecodan would heat the water whenever it wasn’t prohibited, regardless of the tank temperature. [@mjr noted the same here]

@MyForest

Unless you are a smarty like Timbones and you have a battery

I do have a little rule that boosts the heating by 0.5° at noon if there’s sufficient battery left, knowing that it’ll be replenished an hour later.

My overall heat pump strategy now is to be mostly using manufacturer’s out-of-the-box controls, with some optimisation tweaks and nudges over the top. If my hacks fail for some reason, the basic operation continues to function.

[Ian_Calderbank]

this is a fascinating one. I have probes top, upper middle, lower middle and bottom so I was able to clearly see the terrible effect of a reheat on a not-very-cold tank. This does seem to be default behaviour in a lot of setups - I rather suspect that the 99% of heat pump installs that don’t have monitoring systems like ours with geeks watching over them must suffer from this.

I’m now optimising for least monetary spend in the following way

• Intelligent Octopus. house battery (40kwh) charges overnight along with 2 cars.
• heat hot water to 55C during early part of cheap overnight period on timer. This takes 1 to 1.5hrs. space heating off in this period is fine, we’re asleep.

although the hot water is heating during the least efficient time for the HP : an example of a really bad one, COP of 1.8 for the HW on 2nd december, -2C, 2 defrosts during the reheat - its also the least cost time. on a warmer day like today (7C) the HW COP was 3, the CH COP was 4. (thanks timbones for the separation code!).
regardless, the factor of low electricity price at night far outweighs the factor of higher efficiency during daytime warmth, and just doing it on timer every night keeps it really simple.

this is done almost entirely using the mfr’s timer controls. the extra tweak I’ve done is a blocking relay (controlled by an ESP from HomeAssistant) on the 3-way valve. By default samsung will move the 3 way valve towards the cylinder if space heating is not running and it has decided the water in the outdoor unit and/or the outside temp has got too cold. And thus steal a load of heat from your HW cylinder to warm the outdoor unit up. I force the 3-way to stay on heating outside of the times specifically allocated to HW, so if this kind of anti-freezing measure kicks in, it takes it from the lower grade heat left in the rads (which is more than enough), and protects the hard won higher grade heat in the HW cylinder.

By not needing to do another HW reheat for the entire day then I can use all the house battery capacity to drive space heating, lighting, washing machine, cooking etc. result is that on days above approx 3C , I do not pull any peak rate electricity.

our usage varies depending on a variety of family circumstances but using the whole 300L is not unusual which is why I go to 55. Avoiding a 2nd reheat during the 0530-2330 peak time (which would have to be served from the battery) is key to allowing the battery to serve everything else for the whole day.

If I was on a different tariff with a 2nd off-peak, I’d probably set it up differently, but IO as a tariff is such a no brainer with 2 cars and a substantial battery.

However : the crappy COP from my DHW reheat to 55 at night is pushing me down the heatpumpmonitor COP league table

[Timbones (Tim)]

@Ian_Calderbank

By default samsung will move the 3 way valve towards the cylinder if space heating is not running and it has decided the water in the outdoor unit and/or the outside temp has got too cold. And thus steal a load of heat from your HW cylinder to warm the outdoor unit up.

(Interestingly, the Mitsubishi does the opposite, switching to CH to do the defrost, then back to DHW to continue heating the tank.)