Changing the 15mm pipe would be very disruptive and no doubt expensive, so I don't want to go there.

As an alternative, I would be happy to compromise on not running at DT5, but using DT10 instead. That would halve the flow rate needed to transport any given amount of power, and may bring my 15mm pipework back into the workable and quiet range. Noise is a serious issue if the heat pump is going to run 24/7.

A system working at DT10 would give a flow temperature increase of 2.5C over DT5 to get the same radiator temperature, since radiator temperature is the average of flow and return.

If I look at the SCOP values for a Vaillant Aerotherm 7kW, they are 4.13 at 40C flow and 3.91 at 45C, hence about 4.02 at 42.5C flow. So for me, with my planned radiator upgrades, running at DT10 instead of DT5 likely gives a reduction in SCOP from 4.13 to 4.02. Currently, I use about 13,000 kWh for heating and hot water. The reduction in SCOP equates to less than £23 per year at standard electricity prices, and much less in practice with cheap rate overnight electricity and solar. Hence, I can't see any economic argument for changing the pipework - the payback period would not be in my lifetime!

Question then is would my pipework be fine with the flow rate back at about 0.167l/s and DT10 to deliver 7kW?

@bobflux 

The hydraulic simulator is very interesting. I don't know if I can give you enough information to do a reasonable model of our heating system, particularly as I don't know which routes the pipes actually take. The only other information that I have right now is that there are 6 radiators upstairs and 7 downstairs, which will become 8 as I plan on adding one.

I have a thermal camera, so I can take a look however, the heat signature may not be visible if the pipe is not up against the floor or ceiling but rather runs along a joist part way up. When I've walked around looking for cold spots on really cold days I've never noticed any pipes except at the radiators themselves. This is in stark contrast to my parent's bungalow which has micro-bore pipe in the internal walls and is really obvious on the thermal camera.

When we first moved into the house, it had single glazing, cavity wall insulation, and about 150mm of loft insulation. If I take all the upgrades out of my heat loss calculations and go back to the U-values for the house as we bought it, then the heat loss works out at about 12800W.

Our gas usage at the time was around 25000 units a year compared to 14400 units now, and we were both out at work 5 days a week, which isn’t the case now.

The existing radiators have an as new output of about 12500W at 70/50/20C and DT20.

In the early years of living in the house, the heating system just about coped on the coldest days, and sometimes we would put the gas fire on to help it out.

It seems reasonable therefore to assume that the pipework can transfer about 12kW at DT20, as per the original state of the house and the gas central heating system.

With all our insulation upgrades the heat loss is now roughly half what it was at about 6200W. To support that heat transfer running at DT10 the flow rate needed would be the same as 12400W running at DT20.

The fact that the heating system worked years ago therefore lends some credence to the idea that the pipework will be able to cope with the flow rate needed for a 7kW heat pump working at DT10, i.e. 0.167l/s or 10l/min.

I'm coming to the conclusion that I will need a 7kW heat pump that can be configured to work at a flow rate of 10 litres/min and hence DT10, instead of 20 litres/min and DT5. Any suggestions for what might fit the bill: 7kW and high temperature output for hot water as it would need to work with the existing tank? (It looks like the Vaillant Aerotherm 7kW is a possibility assuming it can be configured for 50% of its normal flow rate).

Check the minimum flow rate specifications of any heat pump you are considering. Something like a 7kW Vaillant Arotherm specifies a minimum flow rate of 10L/min, but an optimum flow rate of closer to 20L/min (for dT of 5C, obviously). If you fall below the minimum flow rate, most ASHPs will simply throw a flow rate error and switch off until the flow rate is increased. So you may need to operate closer to 12L/min to give yourself a little margin for error.

@old_scientist 

If you fall below the minimum flow rate, most ASHPs will simply throw a flow rate error and switch off until the flow rate is increased. So you may need to operate closer to 12L/min to give yourself a little margin for error.

That's a very good point. There definitely needs to be some margin.

Via the link below, it seems the 7kW Aerotherm has a minimum flow rate of 540 litres / hour, which is 9 litres / minute. Your point stands though, this is still uncomfortably close to the 10 litres per hour max that I really want to be putting down the 15mm pipework.

https://professional.vaillant.co.uk/for-installers/products/arotherm-plus-air-source-heat-pump-7kw-146561.html

Maybe I should just look at 5 kW heat pumps. I have 2.5 kW and 3.5 kW of air-to-air, which performs very well on heating (SCOP 5.2 and that's exactly what it did last winter). If I simply used the air-to-air as a top up on really cold days, a 5 kW ASHP would probably suffice.

At a practical level, you'll also want a heat pump system with fixed rate circulation pumps where you can lock in the desired flow rates, rather than a system that uses PWM to vary flow rate, where the control logic will almost certainly want to increase the flow rate to reduce the dT closer to the nominal 5C specification. Or at the very least a system where you can disable PWM mode of operation to allow a fixed flow rate.

How sound this strategy is, I do not know.

Posted by: @old_scientist

↑

Or at the very least a system where you can disable PWM mode of operation to allow a fixed flow rate.

Or where you can fix the pwm% as you can eg with Vaullant and doubtless others 

Posted by: @rob-of-york

↑

Via the link below, it seems the 7kW Aerotherm has a minimum flow rate of 540 litres / hour, which is 9 litres / minute. Your point stands though, this is still uncomfortably close to the 10 litres per hour max that I really want to be putting down the 15mm pipework.

I'm now confused.  I thought that there was no point in your system where the whole flow would be forced down a single 15 mm pipe.  If there is such a point then even 5kW will be challenging.

@jamespa 

Thanks, you are right. I needed to look at this again.

At worst the flow splits into two parallel 15mm pipe runs. I am convinced that as an absolute minimum the pipework as a whole supports at least 0.148 l/s (535l/hr) flow rate, since before the insulation upgrade it needed to run that flow rate to get enough heat into the house (12.5kW at DT20). However,  the maximum it can cope with must be considerably greater, so I don't need to be that conservative. Coming at this from the perspective of flow speeds, if I want to stay at say 0.75m/s flow speed or below in the 15mm pipework to avoid noise, then the combined flow rate for two 15mm pipes would be 0.22l/s or about 780 l/hr, which is well above the 540 l/hr minimum of the Vaillant. And I'm convinced that around the 540l/hr is definitely fine, so no issue with the heat pump's minimum flow rate. 

Assuming I had the pwm% set to 50% on the Vaillant, that would give a flow rate of 600 l/hr and a flow speed down two 15mm pipe runs of 0.58m/s, which is firmly in the quiet zone. The pwm% set to 50% would equate to running at DT10 at 7kW, however, I don't have an issue with that. The small reduction in SCOP due to higher flow temperature needed to get the same radiator temperatures compared to DT5, is a perfectly acceptable compromise when set against the costs and disruption of pipework changes.

That would seem to work, which is good!

Posted by: @rob-of-york

↑

Assuming I had the pwm% set to 50% on the Vaillant, that would give a flow rate of 600 l/hr and a flow speed down two 15mm pipe runs of 0.58m/s,

That may not be the case, because its not necessarily the case that PWM = 100% means that the flow rate is 1200l/hr as depending on the pump characteristics and your pipework it could be greater.

But your final conclusion

Posted by: @rob-of-york

↑

That would seem to work, which is good!

is very highly likely correct, at some achievable value of PWM%.

Its impossible to give a guarantee because, however much calculation you do, we cant possibly be sure of the level of noise which depends of other factors as well.  However keeping the flow rate below 0.9m/s-1m/s is generally held to be good. 

My feel is that you are well into the right area, the calculations show it is highly likely to work, so you have to choose between (a) 'suck it and see, replace only if necessary' with a high probability of first time success and and low probability of the hassle of replacing pipework, and (b) replacing the pipework proactively -  with a 100% probability of first time success but also 100% certainty of incurring the hassle of replacing pipework. 

Given that you have told us that replacing the pipework is high hassle, the first is preferable, at least in my book, unless there is material extra hassle if you are eventually forced to replace the pipework retrospectively.  Others, including some installers, will disagree; ultimately its your call.

FWIW I faced a similar issue with a part of the domestic water supply in my install, albeit with the 'replacement' hassle not particularly high.  I discussed it with my installer and he opted for (a), with my agreement.  Right decision it turned out.  Sometimes doing the experiment is the best way, there is only so much you can engineer a system when there are relevant unknowns.

@jamespa 

Thanks, its good to have a second opinion on this.

I'm definitely in the (a) category of try it and see and only make changes to the pipework if absolutely necessary as it would be a massive hassle as well as costly to change it.

That said, I'd like to do everything I reasonably can to confirm that there is no need for pipework changes before embarking on any install. The worst thing would be to get a heat pump installed, with no turning back at that point, and later find that the pipework is too noisy or the system won't work properly at low temperatures, and I am therefore forced into changing it.

This winter, I'll see if I can get an idea of the flow rate the system is currently using from measuring the flow and return temperatures to the gas boiler on a cold day when it is providing its full output. Since the boiler is 24kW and the radiators only put out about 12kW at 70/50/20, the radiators will be the limiting factor and the boiler will modulate down to that. Still, if I get a DT that is someway below 20, then that would be good evidence that the flow rate is a non-issue.

Posted by: @rob-of-york

↑

@jamespa 

Thanks, its good to have a second opinion on this.

I'm definitely in the (a) category of try it and see and only make changes to the pipework if absolutely necessary as it would be a massive hassle as well as costly to change it.

That said, I'd like to do everything I reasonably can to confirm that there is no need for pipework changes before embarking on any install. The worst thing would be to get a heat pump installed, with no turning back at that point, and later find that the pipework is too noisy or the system won't work properly at low temperatures, and I am therefore forced into changing it.

This winter, I'll see if I can get an idea of the flow rate the system is currently using from measuring the flow and return temperatures to the gas boiler on a cold day when it is providing its full output. Since the boiler is 24kW and the radiators only put out about 12kW at 70/50/20, the radiators will be the limiting factor and the boiler will modulate down to that. Still, if I get a DT that is someway below 20, then that would be good evidence that the flow rate is a non-issue.

I think trying to deduce flow rate from an instantaneous measurement of deltaT on a system where you dont know the house loss is a stretch to say the least, no more accurate than the calculations you have already done, quite probably less accurate

Id be tempted to buy and splice in a cheap flow rate meter if I really wanted to be certain, then crank up the circulator to find out what the noise level is at 1200l/hr.  However if it were my house and given all the calculations you have already done I would judge the risk that the problem couldn't be solved by some means as negligible, and just get on with it, but again its your call.  In the end you either want to fit an ASHP or you dont and there will be a solution.

Posted by @jamespa 

Id be tempted to buy and splice in a cheap flow rate meter if I really wanted to be certain.

I'll do exactly that, thanks.

I plan on changing all the 55-year old radiators, so a flow meter can go in when the system is drained down for that. Then I'll have complete piece of mind about what flow rates can be achieved, before having a heat pump installed.