@davidnolan22 That made me smile but yes just about, I think the smallest units can be from 4 or 5kW and modulate down to 35% so close to 1.8 or 2kW

You can also be a bit smart with the design and programme in rest periods, for example if it such a low KW loss its likely unless the house is tiny it will retain the heat longer than average so run to 22 Deg C and then have a dead band of maybe 3 Deg so cycle at 19 Deg giving you an av of 20.5. this would allow the system to run longer and less cycles. That said it slightly goes against the perfect setup but what can you do?

I suppose you could install a 5kW and create a 5kw loss by ventilating a bit more? this way the system would keep running but then that it literally wasting energy. 

@ashp-bobba Not a chance for a swap unfortunately, this is already a 5 years old system. On the bright side though this is in the middle of continental Europe so we can hit -15C on winter which makes system not ~4x oversized but "only" around ~2.5x. + solar gain is also higher on the summer but I don't know how that's calculated for cooling.

Posted by: @gergoh

↑

@jamespa
Your understanding is entirely correct. It is massively oversized. And the compressor can't seem to be able to modulate to under ~3kW or so, which means balancing demand is quite the challenge. This setup is what came with the property so now I'm trying to cook from what I have. 🙂 I guess I need a way to store heat? Either in the thermal mass of the home or in a tank?

OR perhaps add another heat source for shoulder seasons?

I'm really sorry to hear that.

As others have said this is massively oversized.  It is also massively overcomplicated.  Taking the two together this is, I'm sorry to say, one of the most extreme examples of a poorly designed system reported on this forum.

Storing heat, other than in the concrete slab if there is one, or a swimming pool sized vesse, is not likely to be practical.  If you do have a concrete slab in which the ufh is burried, then maybe, just maybe, you can find a batch heating schedule, perhaps using octopus cosy, that works and has tolerable temperature swings.  If that's a remote possibility it might be worth giving it a go and I'm happy to make suggestions or better still you could try to find someone who already does that and therefore can advise from practical experience not just theory 

Otherwise, unless you are prepared to accept massive inefficiency, I think you need to redesign from scratch, eliminating buffer tank and most of the controls and valves (including mixing valves) and buy a heat pump closer to the right size which is also capable of 'above dew point' cooling if that's what you want to do.   Even in the high season the current system and way of operating it is going to struggle to deliver an efficient heating, in the shoulder season you have no chance.

I should clarify I am not a heating engineer, just a householder with a background in physics/engineering who has been studying this stuff for a couple of years. 

@gergoh 

The 8kw newish  r290 Mitshi machine can modulate really low. might be a good option. 

@jamespa Thanks for telling the cold hard truth 🙂 Even though it sounds bad I'm pretty optimistic right now.
Replacing the heat pump with a smaller model doesn't make much sense for me financially speaking because running the current one inefficiently might be cheaper than the initial cost of a new one. But once the current compressor inevitably dies I will downsize.
Other than that all the rest what you've suggested sounds like easily solvable by just disabling "by software". I can keep all the valve actuators open to match the minimum load of the heat pump, essentially having something like a (semi) open-loop system. I can also override the mixing valve, essentially bypassing it entirely.
Even the dew point is a not a problem since I can change the setpoint via MQTT.
Then if all else fails in shoulder seasons I can always switch to using the AC.
It kind of sounds like to me that following your suggestions I can improve a lot by just having more intelligent control, even though it probably never will be "optimal" in it's current form.
Then when the time comes and the heat pump will be replaced I will let someone re-design the system (and I will have DATA to do so!) and get rid of the buffer tank too.

Does this sound like an OK plan?

> I should clarify I am not a heating engineer, just a householder with a background in physics/engineering

I get it. I'm an electrical engineer / SW developer but I also have to learn this stuff out of necessity. It's kind of fun, tho.