I ran cooling over a few sessions, collected some data, and crunched the numbers.
tl;dr it does kinda work, but the effect is subtle and requires constant monitoring and adjusting.
The Ecodan 11 kW is a big unit can deliver a lot of cooling power, I would guess about as well as it can provide heat in the winter. The limiting factor is the radiator system: my combined emitter surface has a rated output of 18kW at Δ50. These radiators can, in therory, provide 1.1 kW of “cool” when 6° cooler than the room, assuming they can be fully saturated and have active convection. This grows to 2 kW at 9° delta.
However, the cold water runs runs along the bottom of each radiator rather than filling it with convection, so the effective cooling capacity is considerably reduced. (The low-loss header probably isn’t helping either, with inverted temperatures).
[Friday, August 14th] I ran the heat pump for nearly 2 hours. The heatpump drew nearly 2 kW of electricity for 5 minutes, plunging the flow temperature to 12-13° (well below dew point), then staying off for 20 minutes to let the temperature come back up before starting again. 4 cycles in total. (Similar results with other sessions)
This averaged about 370 W of electrical consumption, 120 W for the pumps, so 150 W for the compressor. Assuming a modest COP of 4.0, this equates to about 600W of cooling power into the radiators - about half of the theoretical capacity. Mean return temperature of 18° against a room temperature of 24°. Would need about 1600 W of cooling just to maintain that temperature (ignoring solar gain).
The main problem is that heat pump is too big for the property (even in winter), so it’s difficult for it to get into a nice running state, especially when minimum output is 4 kW, and even radiators can only transfer that power when flow is 15° below the room temperature.
Combined with the requirement to [manually] keep flow temperature above dew point, means the heat pump will always be cycling to provide a small amount of cooling. (Doesn’t help that the Ecodan races to its target as quickly as it can, rather than taking it slow.)
The radiators were cold to the touch and would have been absorbing some heat, but I could not detect a noticeable difference. I can maybe see a reduced gradient in the rising room temperature if I squint. Hardly a scientific experiment though.
Caveat: your property will have different figures for emitter capacity, heat loss, etc so I cannot predict how well cooling will work for you. I would guess that 11 kW is equally oversized for your new build, so cycling is very likely. Your actual mileage may vary.
Note: I had 2 hardware fault events while cooling, which I’ve never during 5 years of heating, so there some something the Ecodan didn’t like doing. I limited my tests to periods when I could keep an eye on it.