Welcome and I am sorry you are having this problem.

The obvious answer is to delete the buffer and secondary pump which will remove the problem entirely and almost certainly improve system performance.

Other than that, and assuming it's the pump not the noise of water in the pipes you can hear, relocation of the pump is the obvious solution.  Another possibility is to put flexis between pump and pipework or to construct some kind of isolation platform to reduce transmission of sound. However the latter has the potential to amplify by creating a sounding box.

I am not sure there is much else I can suggest.  The sound from a nearby  droning pump is going to be difficult to suppress entirely when there is no background, so relocation or deletion seem to me to be the only options guaranteed to work.  Can you identify anywhere it could go given the topology of the mapped pipework, or do all the pipes converge in the loft so really it has to go there?

One last thought is to do some research on circulator pumps.  There may be ones that are quieter than others.

I agree with @jamespa that your buffer tank and secondary pump may not be required at all.  I have a 1960s two storey house with 30 radiators.  I have neither a buffer tank nor a secondary pump and the system works fine without them.  I don't know anything about underfloor heating, so that may push you towards a buffer tank, but I don't think that is necessarily the case.

Assuming you keep the buffer tank, the first thing to check is that it is plumbed correctly.  Suppose the flow from the heat pump comes into the top left of the tank.  Then, the flow into the radiators must depart from the top right of the buffer tank.  The return from the radiators comes back to the bottom right and the return to the heat pump goes out of the bottom left.  

There have been some examples quoted on YouTube, where a buffer tank was cross-plumbed - i.e. flow in from heat pump at top left and flow out to radiators at bottom right.  Then return from radiators at top right and return back to heat pump at bottom left.  That would be absolutely disastrous, as water flows are crossing one another in the buffer tank.  It would cause lots of turbulence in the buffer tank and also the two water pumps would be fighting each other to some extent.  It would also make the system very inefficient, due to the mixing of hot and cold water in the buffer tank.

Assuming you don't have this problem, then ideally you want the flow rates in the primary and secondary circuits to be the same or similar.  If not, then you get a lot of water mixing in the buffer tank.  This is known as distortion, which reduces efficiency.  I am not sure whether this can also make the system noisier, but I imagine that it might.

The Renewable Heating Hub video explains potential problems with buffer tanks:

Another potential cause of noise is sludge in the pipes and radiators, due to the oxidation of radiators.  Air in the system can also cause noise.  If radiators have a cold central area at the bottom shaped like a mountain, this is a sign that they are full of sludge.  If you feel confident that you know how to do this, I suggest that you remove the magnetic filter to see how dirty it is and clean it.

If you are using zoning or thermostatic radiator valves, then it is possible that they may be closing and restricting the flow too much.  That reduces efficiency and may also put the pump under too much strain, which might make it noisier.

Three things that caused noise in my system after commissioning.

• Volumiser was gurgling which just turned out to be the air bleed valve not open. Was exposed and could have easily been brushed against.
• Balance valve between the feed and return pipes to the HP. Literally as the engineer was leaving it began to "squeal" because the previous fault altered the system flow.
• One radiator lock valve was partially closed and made an obvious hydronic noise throughout the system.

All made noises throughout the system but first two were easily localised. Last just needed me to check each rad till I found the one with a partially closed valve. My Grundfos pumps from before (gas boiler) and currently in the Daikin HP didn't make noise themselves more from the water travelling around the system. I do have a 60% PWM cap on the HP pump which makes my system silent. I can still get noise circulating when the pump goes 100% PWM during a defrost cycle. Which for me happens maybe once or twice when near zero overnight temps.

Apparently there are ways to reduce shock waves when plumbing systems due to pressure changes maybe that applies to flow rate changes too. Worth pursuing maybe. Not a heat engineer by any means.

I used to  (and still do to some extent) have "water flow" noise in my pipes which made me turn down the circulating pump at times, and I guess this is simply because the water flow in a HP system will be about 4 times greater than that with a gas boiler (due to 4 times lower deltaT, on average), but I reduced this greatly just by fully opening all lock-shield valves. This might seem counter-intuitive, to 'unbalance' your system, but I've not noticed any negative performance issues, and I appreciate the quiet(!)

This is quite simple to try out, and revert, so might be worth a go. Obviously make a note of what the valves are set to before fully opening them, so you can go back if it doesn't help.

Another thing to check is the buffer tank for air - any air ought to accumulate in that, and there should be a bleed valve to release any accumulation.

Finally, are the pipes to the pump supported correctly? There shouldn't be any 'play' in the pipework either side of the pump, and when it's running you should only feel a slight hum if you hold the pump, rather than any real vibration.

Also might be worth mentioning your HP make, in case anyone has experience with a similar set-up/problem.

@jamespa Thanks for sharing thoughts. If I mentioned that the heat pump vendor categorically won't install their systems without a buffer tank (and additional circulation pump) that might tell you who we've gone with! But for avoidance of doubt I may as well mention for the benefit of others that it's an Aira install. I have heard online interviews with some of their senior staff where they've vaguely alluded to the possibility of reconsidering the buffer tank requirement on all installs, but until such a time I'm left looking for answers whilst retaining the buffer.

I can't stress enough that individually everyone involved from Aira's side has been helpful and is continuing to try to find a solution, but the challenge feels more that there is no-one owning the overall issue so I'm left chasing several separate leads.

Having mapped out our pipework I'm pretty sure I can see a way of tweaking the run of a couple of lengths (of easily accessed) pipework, that would in essence allow the pump to move from right next to the loft buffer tank, and instead into a ground floor washroom - the same washroom that all the existing flow/return pipes currently converge in for the downstairs UFH and downstairs radiators, as well as where the return for the upstairs circuit merges with the primary return. It's a job I could probably do myself within a couple of hours, but obviously that's well off the cards for now for any potential warranty reasons 🙂

Being nothing more than a vaguely capable amateur the only real question I have is whether there are any significant limitation on the distance between the buffer tank in the loft, and the location of the circulation pump itself. I understand that the pump needs to be installed before any of the pipework splits to supply the three circuits we have (upstairs rads, downstairs rads, downstairs UFH), but can't see an obvious reason the pump has to be particularly close to the buffer. In fact in my mind I can almost see that location the pump at ceiling height on the ground floor would lead to an additional pressure head on the feed side of the pump that might even reduce any suction effects that cause the pump to work harder - but I could be wildly wrong with that!

I'd happily share my scribblings of the pipe runs if anyone's interested, but without spending an age on drawings it's quite tricky to represent in 2D.

Posted by: @jtg

↑

Being nothing more than a vaguely capable amateur the only real question I have is whether there are any significant limitation on the distance between the buffer tank in the loft, and the location of the circulation pump itself. I understand that the pump needs to be installed before any of the pipework splits to supply the three circuits we have (upstairs rads, downstairs rads, downstairs UFH), but can't see an obvious reason the pump has to be particularly close to the buffer. In fact in my mind I can almost see that location the pump at ceiling height on the ground floor would lead to an additional pressure head on the feed side of the pump that might even reduce any suction effects that cause the pump to work harder - but I could be wildly wrong with that!

I dont think there is any material limitation but you do have to avoid creating areas of negative pressure once gravity is taken into account (which is one reason systems are pressurised).  Is the pump controlled by the heat pump or free running, you would need to take that into account.  Perhaps post the sketch you have done and see if anyone else comments.

Posted by: @jtg

↑

I can't stress enough that individually everyone involved from Aira's side has been helpful and is continuing to try to find a solution, but the challenge feels more that there is no-one owning the overall issue so I'm left chasing several separate leads.

Perhaps if you can point them at one (eg the one you have suggested if you can work out the details) then they may eventually decide its easier to comply than deny!

Thanks @grahamf - I'd not thought about the consequences of the flow/return of the buffer tank being mixed up. It was a completely clean install that side of the system so I'd hope not, but definitely worth ruling out so I'll check that! Very interesting to hear you have a 30 radiator system running off just the pump in the external unit... it makes the buffer tank and extra pump in our 10x radiator and relatively small UFH seem a little excessive. Our UFH is small bore 10mm piping though (and is also likely our index circuit) so I suspect that is having some impact on the amount of work the pump is having to do. 

The noise does feel like a combination of the pump whining and rushing water, and I think that's also being exaggerated/amplified because the pipework in the loft immediately drops into an airing cupboard where the pipes are then clipped to a fairly thin plyboard outer wall. Feels like there would have been less chance of vibration if the pipes were clipped to one of the blockwork walls of the airing cupboard instead but that wasn't considered at install time (and it's now a real squeeze to do any relocating of those pipe drops - that would involved a complete rework and moving of the cylinder)

@jtg while you are checking the plumbing of the buffer tank, make sure the flows are at the top and the returns are at the bottom.  Mars explains in the video I referenced that buffer tanks use stratification to help the system to work well.  Obviously, if flows are at the bottom instead of the top, then stratification will work against you.

I would also check the magnetic filter and clean it.

Thanks @mikefl - I'll maybe have a look at the lock-shield valves as a bit of a test, sounds worthwhile. I'm not wholly sure the whole system was fully balanced in the first place, though that's mostly conjecture as I wasn't around at the time. I heard a bit of a suggestion of 'closing them all off and then opening them all a bit until they got warm', which to be fair they do all get warm... but it didn't quite seem the ΔT based balancing I was expecting!

Just an off the wall thought, with the pump being in the loft is there a possibility any air in the system is getting trapped in the pump housing? 

Extend the primary branch and make sure you have more than 10x the length V width before bends and fittings, so if the branch is in 28mm pipe make sure you have 280mm between changes in direction or fittings causing resistance, this helps a bit. You can attach bellows to each side of the pump if you think some of it is pump vibration. Having a larger impeller running slower can help (so larger pump on speed 1 V smaller pump on speed 3) this will need your engineer to size correctly.  (Note) its known that close coupled joints and bends cause significant noise, the best design is not to have to many close together, nice longer pipes for velocity and less resistance can help. 

Making sure there is expansion space between clips, rubber lined clips can help, if there are 2 bends offset like a C as a pipe travles you may find as the pipes heat up and expand they are pushing harder into the clips transferring more vibration. 

Having the correct sized or set pump for its duty rather than an over sized or over clocked pump and then throttaling the rads back also helps.

Most pups are now self priming but that only work in some cases if when installed it is the right way up or on the correct plain, I have seen many times where the installers do not remove the head and rotate it into the correct position to suit the installation, this is always in the instructions but many do not read them.

Last note: Micro pipework is naturally noisy but before you were running maybe 15l/m through the main pipe, now you will be shoving 30l/m to acheive the same performance. It why so many of us insist on upgrading some of the old pipwork where possible. 

I do not know your installation or its full design but I hope this gives you some ideas to check out.