@Editor is correct. Under MCS, installers are required to provide a flow temperate and an ESTIMATED SCOP as part of the survey. However,……
A key difference here is that the Heat Geek SCOP is not estimated - it is a minimum performance guarantee. The actual installation design work is done by Heat Geek HQ after it has received the survey results and the customer’s selection of one of three performance options. One of the outcomes of the design work is the calculation of the system’s maximum flow temperature. This is the reason why there is no flow temperature in the survey, and why it appears along with the SCOP on the MCS Certificate. This may be complete rubbish but it is what my research is suggesting.
Space and HW heating calculations appear both in the survey and on the MCS Certificate.
 

@editor 

In principal I agree with most of what you say, most particularly the bit about the nature of the guarantee being key.  However there is one possible (but key) exception/subtly to my agreement:

 

There are two possible and legitimate scenarios where it might not be possible, or at least where there is an understandable and arguably legitimate reluctance, to provide a specified flow temperature etc namely:

 

1) the design is done by a demonstrably good tool that does not reveal its inner workings (AI/machine learning)

2) the design has a degree of uncertainty, or even risk, associated with it

 

I suspect that #1 applies in many cases, including Heat Geeks.  Number 2 definitely applies to all heat pump installations.    Of course #2 is only truly acceptable from a customer point of view if the supplier is prepared to cover that risk/uncertainty by means of a guarantee, making retrospective changes if necessary.  Note that most installers, and the MCS rules instead deal with this risk by insisting on specifying lots of design detail, ie transferring the risk to the customer!  

 

There is a third possible scenario which may also apply namely

3) the design may risk violating one or more arbitrary rules imposed by MCS notwithstanding the satisfactory performance, and providing design details would reveal that thus 'incriminating' the supplier.

Im not saying that this scenario exists in Heat Geek's case, but it very likely did with my (successful) installation.

Finally there are possibly commercial reasons.  At present HG has a commercial advantage with intelligence it has gained from monitoring its past installations.  This allows it to 'trim' specifications (and thus its own costs) whilst offering a performance guarantee.  That commercial advantage is diluted if it reveals too much detail.

 

Any or all of the above may explain a reluctance to provide the design detail which the customer shouldn't need.

Im should be clear that I am neither defending nor attacking Heat Geek, just looking at it another way!  I personally think we should be spending time understanding (and critiquing, albeit with a realistic hat on) the nature of the guarantee given its potential to be a game changer in the industry. 

Posted by: @jamespa

↑

What Heat Geek appear to be doing is pioneering a sea change in the industry and, in my view, we need to give them credit for that.  They are taking some risk in doing so, not least in terms of the MCS design specs, but the risk they are taking is presumably calculated and from a consumer point of view necessary if the industry is to get to where we would surely all like it to be.  

The reason they are doing this is hopefully because they assessed they have enough data to be able to know what they can safely promise. Credit to be given specially when they do not control their installers.
Others using own installers stayed very close to what MCS requires them to do around SCOP.
I am curious how many retrofit installs HG, Octopus, BG have done so far. 
 

Posted by: @jamespa

↑

There are two possible and legitimate scenarios where it might not be possible, or at least where there is an understandable and arguably legitimate reluctance, to provide a specified flow temperature etc namely:

1) the design is done by a demonstrably good tool that does not reveal its inner workings (AI/machine learning)

This is my best guess/hypothesis. HG are innovating in a competitive market, using advanced analytics/AI. There's no MCS requirement for them reveal flow temp.. 

If they did reveal it, consumers could question it (and the poor installer would be middleman..). Like I questioned their heat loss calc until they changed it. 

I agree with @editor that consumers may benefit in case they could scrutinise the flow temp, which would potentially enable reuse of the design, more options.

But then, this industry process does not even allow reuse of a heat loss survey, allowing reuse of any aspect of the design would be HG helping others..

I'm going to throw it out here in the expectation of being shot down in flames.

A perfectly valid way to design a R290 heat pump system for the vast majority of houses that are under 7kW is in 3 stages:

1 Do a few basic checks, if it passes install a 7kW heat pump

2 Run it for a season at whatever ft is needed to heat the house; monitor.

3 Adjust radiators and flow temperature once you actually know something from measurement rather than supposing it from spreadsheet whatiffery, having regard to ROI.

4. Don't tell MCS you did this

 

Apply above to any house where annual gas consumption < 15MWh

Are heat loss surveys worth, for the customer, the paper they are printed on or just a cast iron shield for installers?

I accept the above is overstated, but contend that there is an element of sanity in it.

Is this where Heat Geek are headed?

@jamespa 
Hope someone tries it! As a consumer, I would like a performance warranty, etc and to see some savings, like with HG installs. 
With R290, there's less risk but obvs there's also the dreaded HW tank..
It would be good if installers would also lend a few temperature sensors so the HC and rad tuning could be quicker. 

Roughly how most air/air is done except that step 3 doesn't exist and step 4 can be ignored as the grant has no real value. There's an awful lot of pontificating goes on about sizing and calculations and detailed analysis for air/air but in most countries nobody actually does any of that crap at all.
They measure the floorspace, peer at the materials, take an educated guess and add a bit - like a gas boiler, because being oversized isn't that big a deal.
The MCS are just a liability and I live in hope that EAS2024 coming in on the 15th October will eventually bury them in solar at least.
 

@jamespa It seems to me the 'experimentalists' who use measured data are missing one thing. The calculations are done based on the coldest temperature expected for a given location. The weather being what it is these days, it's quite possible the coldest temperature is not reached in a given year, perhaps it's more like every 5 years or whatever. If the experimental data doesn't cover a suitable cold snap then it will naturally conclude a smaller heat pump is required. Then when you do have a prolonged cold snap you find it's undersized.
Now we can have a conversation about whether we need to cover the 99th percentile coldest or the 95th or 90th or whatever (and whether wheeling out a few electric fires is better than designing for the extremes), but it's not comparing apples to apples.

In defence of the "experimentalists" the calculations are done based upon an enormous number of "assumed values" and "guess a number". They are also often quite mathematically unstable (a small change in a couple of the "assumed values" totally changes the output). In other words the calculations people are just experimentalists too except one set looks at the data and extrapolates, the other looks at the equations based  upon extrapolation from other buildings, vastly simplified and plugging guesses into it.
Neither are particularly good although at least the combination of the two sometimes highlights that the results are nonsensical somewhere.
 
The conversation that we'd be having in a sane industry would be
- Why does sizing matter so much and how are we going to fix it
- Why does being oversized break everything and how are we going to fix it
 
You design technology intended to be installed quickly and cheaply by a man in a van to be idiot proof, because it's cheaper to do that than stop hiring idiots.
 

Posted by: @etchedpixels

↑

- Why does being oversized break everything and how are we going to fix it

I think that's the key question. After all inverter units can modulate down, so what are the limits about how far you can modulate? Is there was way to modulate to say 10% of nameplate without losing efficiency?
If you could bang in a 10kW unit to every home no matter whether the losses are 3kW, 5kW or 9kW, the economies of scale might mean you can buy that unit for the same price as a smaller one. 
If that's not possible, make the units modular so you can just swap pieces over rather than buying a whole new system. If your system isn't big enough, buy some more capacity like upgrading the hard drive in your computer, without having to throw the whole thing away.
 

Posted by: @prunus

↑

@jamespa It seems to me the 'experimentalists' who use measured data are missing one thing. The calculations are done based on the coldest temperature expected for a given location. The weather being what it is these days, it's quite possible the coldest temperature is not reached in a given year, perhaps it's more like every 5 years or whatever. If the experimental data doesn't cover a suitable cold snap then it will naturally conclude a smaller heat pump is required. Then when you do have a prolonged cold snap you find it's undersized.

Thats true only if you do the wrong experiment.  Done correctly (eg by plotting daily energy consumption vs daily mean OAT, or even just measuring annual consumption) it does not matter whether or not the design temperature is actually reached during the measurement year, you can still work out the figures at the chosen design temperature.

Posted by: @prunus

↑

Now we can have a conversation about whether we need to cover the 99th percentile coldest or the 95th or 90th or whatever (and whether wheeling out a few electric fires is better than designing for the extremes), but it's not comparing apples to apples.

There is a figure for this in the MCD design document MIS-2005-D (either 99% or 99.6%,measured on an hourly basis).  The design temperatures specified in are based on this figure.

 

Posted by: @prunus

↑

Posted by: @etchedpixels

↑

- Why does being oversized break everything and how are we going to fix it

I think that's the key question. After all inverter units can modulate down, so what are the limits about how far you can modulate? Is there was way to modulate to say 10% of nameplate without losing efficiency?

If you could bang in a 10kW unit to every home no matter whether the losses are 3kW, 5kW or 9kW, the economies of scale might mean you can buy that unit for the same price as a smaller one. 

If that's not possible, make the units modular so you can just swap pieces over rather than buying a whole new system. If your system isn't big enough, buy some more capacity like upgrading the hard drive in your computer, without having to throw the whole thing away.

 

Being moderately oversized does not break it, the problem is gross oversizing (eg a factor of 2) which causes excessive cycling, leading to materially decreased efficiency.  It also means that the heat pump is predominantly operating at a less efficient part of the compressor operating curve, further reducing efficiency.  Gross oversizing is very easily avoided in almost all cases by doing some simple sense checks on survey results.  Unfortunately MCS has not thought to mandate this.

Unfortunately early day installs overestimated air changes (and sometimes fabric U values ) leading to some spectacularly wrong conclusions.  My own house is a good example, two professional surveys (each taking 3 hours) said the loss was 16kW, whereas in fact its 7kW.  Fortunately a simple sense check dismissed the 16kW brigade.  Things have got better since then, the advice on air changes in particular has changed, there is some recognition that U values of certain materials can be overestimated and organisations like Heat Geek and Octopus have accumulated sufficient experience to spot the issues, and the tools are improving.  Nevertheless I would always advise doing a sense check if the estimated loss comes out at more than 6kW.

The fundamental challenge with increasing modulation ratio appears to be the limits of compressor technology.  Presumably this comes down, at one end, to manufacturing tolerances which place a limit on how slow you can run it (if you run it too slowly refrigerant escapes past the seals) and at the other end to the strength of motors and materials, which limit how hard/fast you can compress it.  I don't know whether any terchnical advances are around the corner which may change this, but I wouldn't rate the chances particularly high given that compressors are a 100+ year old technology.  That said some manufacturers seem to deal qwith oversizing better than others, so I don't doubt that subtle optimisations of the control curve will continue to diminish the negative effects.  Mitsubishi did release a heat pump with two compressors for a while, one 2kW and one 6kW.  However they have since withdrawn it presumably because either it didn't sell or didn't work well.  

Its a moot point whether, for the very typical house that (in fact) is 6kW or less, size actually matters*.  At this level you are installing a 5-6kW heat pump anyway in order to have sufficient oomph for DHW reheat, and because its anyway difficult to find a heat pump with a capacity much less.  

In summary I wouldn't put a great weight on the likelihood of heat pumps that can do a 10:1 modulation ratio emerging soon, but at the same time neither do I think the problem is so serious that it merits a technical change if the consequence of that is an increase in hardware prices.  Some simple sense checks would eliminate altogether the gross oversizing we sometimes see.

 

* the OFGEM standard/typical house consumes 11.2MWh/yr of gas, so its loss is definitely under 6kW