I would also look at an air/air install if you've got to make radiator changes, you want cooling or you want to migrate over time (ie keep the gas too). A lot of that is taste though but even if you don't want air/air having an air/air quote to wave at the air/water people never hurts 😉

Whichever way you go the essence is simplicity. Every extra box, widget, magic contraption, valve and whatnot increases the cost and decreases the performance and the reliability. 

@paulc my view after reading the whole thread is that I would pause before committing to this proposal. The individual components may all be capable of working, but they still don’t appear to form one clear, coherent system design.

As @jamespa and others have already highlighted, the biggest unanswered question is why a house with a calculated heat loss of around 5.2kW is being offered an 8kW unit and a plate heat exchanger arrangement. The 8kW label alone doesn’t tell us whether the heat pump is right or wrong... what is important is its actual output at your local design temperature and proposed flow temperature, together with how far it can modulate down during milder weather. Those figures need to be provided from the manufacturer’s engineering data.

The part that concerns me most is the mention of 55C. If that means 55C going to the radiators through a plate heat exchanger, the heat pump may need to produce an even higher temperature on its side of the exchanger. At that point, an anticipated SCOP above 4 becomes extremely ambitious. A heat pump capable of producing 70C does not make 60C operation efficient, it simply makes it possible.

You also can’t say the radiators are "50% oversized" without stating the flow and return temperatures used for that comparison. I would want a room-by-room schedule showing each room’s calculated heat loss and the output of its existing radiator at the proposed mean water temperature. Otherwise, the 50% figure is essentially meaningless.

I suspect the existing system may be more usable than the proposed design gives it credit for. Moving approximately 5.2kW at a 5C system delta T requires roughly 15 litres per minute. Whether the existing 22mm pipework can achieve that depends on its length, fittings, branches and the available pump head, but it is not obviously impossible. That should be calculated and then verified during commissioning before a plate heat exchanger is accepted as necessary. Sadly, most installers just guess or make assumptions.

I would also question whether replacing a perfectly serviceable cylinder is worthwhile in your particular circumstances. You currently heat it cheaply overnight with the immersion and have modest hot water demand. A heat pump cylinder may improve recovery and allow the heat pump to produce more of your hot water, but the financial payback could be very long. The cylinder decision should be justified separately rather than being allowed to dictate a more complicated space-heating arrangement.

Before proceeding, I would ask the installer for one document setting out the proposed heat pump’s output and minimum modulation at the relevant conditions, the design flow temperature, flow rate and delta T on both sides of the plate heat exchanger, the pressure losses, room-by-room radiator outputs, the purpose of the 3kW element, the freeze-protection strategy and the basis of the predicted SCOP.

If they can provide that and the figures stack up, you have something concrete to assess. If they can’t, I would not proceed. At the moment, it feels as though the system is being designed around the Heat Saviour rather than around your house and that is the wrong way round.

all good advice I will follow this up