[OpenTRV-interest] Controls and stuff

[Simon Hobson]

John <john at stumbles.org.uk> wrote:

> What I'm trying to do is model the energy requirements of a house and do what-ifs: e.g. what if I increase loft insulation from 50 to 250mm, ...

I think you'll find it's far more complicated and having hourly data will simply add to the volume of data without adding information.
Taking loft insulation for example ...
Not only do you need temperature data, but also rainfall (is the roof wet ?), wind (windchill factor, higher while the roof is wet and which will persist in the wet state for variable times depending on temperature humidity and wind), and insolation. Even in cold weather, on a sunny day (as it's been here at work until an hour or so ago), the loft can get "quite warm" and thus any attempt at detailed calcs is (IMO) bound to fail.

Simple example.
Outside temperature is 1˚C, sun is out, been no rain for a bit and there's no wind - loft is a balmy 20˚C (say) , heat loss is minimal.
Another day, outside temperate is (say) 5˚C, it's windy, overcast so no sun, the loft is cold (5˚C probably*), and heat loss is higher.
If (for simplicity) we say the rooms (ceilings) are at 25˚C, you've 4 times the heatloss on the warmer day !
* If the loft is as ventilated as it needs to be to avoid condensation, there's be wind blowing right through. Chill factor from wind blow-drying the wet roof, and the fact that the rain itself is colder than the air temp, will make the roof itself below 5˚.



Damon Hart-Davis <dhd at exnet.com> wrote:

>> I think (as Tim Small has remarked) that you're describing weather compensation, which is increasingly widely available on modern boilers and is possibly going to be required in a coming revision of Part L of the Building Regulations.

Now, if only the authorities mandated an open standard for high-lvel comms with boilers. There was OpenTherm but that effectively died as there were "proprietary pressures" for it not to succeed. Basically manufacturers could see that if an open standard got going, they wouldn't be able to apply the huge markups (like the, plucking random numbers from the air, "£100 for a 10p thermistor and a box" scale) they can apply for their own closed "solutions".

> I have always assumed that weather compensation is simply the part that adjusts settings (eg flow temperature) based on current external temperature, ie reactively.

Correct. And if deploying openTRV (or anything with similar feedback ability) there's a better way of doing it.
Weather comp is simply applying a fairly crude model that says "if it's warm outside, we need less heat, so the rads can run cooler, so lower the flow temperature setpoint". With openTRV we have feedback of the actual demand - we know the actual actuator position for every TRV. If they are all closed, we know demand is met and can reduce temperature, if there's a valve fully open, we know there's a large demand and need to increase temperature.
So there's room for experimentation to control boiler flow temp so that the valve actuator positions fall within a certain range. My "gut feeling" is that the position of the "most open" valve is the one to use - and control the flow temperature so that (for example) the most open valve is at at least 25% and at most 75% open.
For most conventional systems, there will be several benefits. Firstly energy savings from running the boiler cooler and hence losing less heat through the exhaust, elimination of non-condensing operation when combination of low demand and high bypass ratio gets the return temperature above that required to keep the boiler condensing, and a massive reduction (due to effectively maintaining a higher flow rate) in system noise that normally happens when demand is low so there's a high back-pressure making all the valves "hiss".

> There are opportunities to do more than that with the forecast, getting heat into the house early/late/whatevs if the next little while is going to be especially hot/cold/windy depending on the thermal time constant of your house.  Potentially especially good for those with (say) UFH for example.

Yes, that's the downside with UFH, and stored heat - our office has off-peak storage heaters :-(
I've sort of been working on plans, if I were every (unlikely !) in a position to do any of it, to combine UFH and it's comfort (warm feet) with the fast response of radiators.
One idea I toyed with was running a separate pump etc for each room, simple mechanical TMV set at something giving a comfort temperature for the floor, and having that take it's supply from the passing flow to the radiator. Under high demand, there'll be a good flow, the UFH gets a hot supply, and the rad will be hot - it's supply will be a little lower than the boiler output as the UFH has had a bite from it first. As the TRV closed, the flow reduces, and the effect of the UFH drawing heat will result in the radiator flow temperature dropping. As demand drops further, the UFH will start cooling down as there won't be a high enough flow rate to keep it going. And if demand drops to zero, then the UFH shuts off too as it's no hot supply.
It works in my mental model is all I can say ;-)

> Note that the compulsory compensation suggestion is out for government consultation right now, and I voted against it given the long payback times and the other rather lower hanging fruit (ie better places to spend the money to save energy first), and even against compulsory timers since most people don’t use them:
> 
> https://beisgovuk.citizenspace.com/heat/heat-in-buildings-online-consultation/consultation/
> 
> Guess what I thought *would* be a good solution that policy should support?  B^>

I wasn't aware of that, I'll go and take a look.



John <john at stumbles.org.uk> wrote:

> I think the Regs could do a service by assessing and mandating a usability factor for controls

I agree there. I work in IT so get to "use" all sorts of "interesting" ideas around UIs - I struggle with some heating controllers so I understand why many won't touch them. I know my Mum would not even think of doing anything with hers, and she's fairly tech savvy - a right silver surfer !

> ... the building regs are there to mandate what installers provide: if time control isn't compulsory then cheapskate installers won't fit them and even householders who want and would use them won't have them. And given that a timer costs the square root of sfa there seems no justification for omitting them.


Absolutely.
But the problem with mandating <something> is that the regs tend to be a bit inflexible and not keep up with developments - or they are open to interpretation. An obvious one would be to mandate that "controls be provided to facilitate reasonably efficient operation of the heating" - but then, there'd be an endless argument as to what was "reasonable".




John <john at stumbles.org.uk> wrote:

> Do they break it down i.e. do they estimate the cost to be huge or the savings to be miniscule?

I can't help think that weather comp, in the absence of everything else that should be there, will make a fairly small difference. In effect, the saving is that of running the boiler at a lower exhaust temperature and thus recovering a bit more heat from the exhaust. if the system is bad enough that it goes into non-condending mode when heating demand is low and bypass ratio is high, then it'll make a saving.
Once you've got all TRVs, room stat(s), properly set time controls - I suspect the incremental savings of adding weather comp are very small, and the price "fairly high" for what is effectively a 10P thermistor in a box.

> I'm leaning towards weather comp to generally tailor the heating to demand with a simple timer (left set to a sensible programme and which they'll hopefully just leave alone).

I worked in a (rented) office once where the entire heating system was run by weather comp. The rads were full flow jobs, with output control by a manual flap that restricted air flow out the grill, and with flow temp by weather comp with one internal and one internal sensor. It was ... staying polite ... absolutely rubbish.
Unless you have a controlled environment so you can balance the system, then weather comp alone will not provide even an approximation of sensible control. Better would be to sense system flow, as an indicator of TRV actuator positions and hence demand, and adjust flow temperature based on that - see above.