Wet UFH how to retro fit an electrical heater element?

[readiescards]

It's March, sunny but cold. Yesterday my PV report shows:

System Production: 41.7 kWh

Self-consumption: 25.74 kWh

Export: 15.97 kWh

 

So I'm exporting lots of unused electric (once my UVC is satisfied) and then late afternoon we fire up the log burner and the central heating comes on.  Daft this when I could use the PV electric excess to warm up the UFH  instead of expensive LPG via a Viessmann boiler.

 

Looking through old threads I notice a reference to a 'Willis' external immersion device, never heard of that but a quick Google and it looks the ideal device especially as my SolarEdge PV control system can handle a 3kW immersion element.

 

My local plumber knows nothing about an external immersion device, so hence some questions:

 

1. does it need an additional expansion vessel adding to the system?

2. if I install the device just before the ufh manifold, would extra valving be ESSENTIAL to prevent the ufh pump running the return water through the boiler circuit all the time?

3. I have a BS1820 temp sensor buried in the slab, the control circuit seems simple:

 

if (slab temp <  desired slab temp  and solar PV excess exists)

{

 run ufh pump

 willis unit on & powered by SolarEdge PV diverter wall socket

}

 

since the boiler will auto control itself and will back off if the returned ufh water is up to temp

 

Failure points:

* if the ufh pump failed the Willis unit dual thermostats will cut in and switch it off

* if the slab temp sensor fails the floor could get too hot so an extra temp sensor on the manifold could verify the return ufh water remains within acceptable limits.

 

Have I missed something?

 

Or should I just stick to running round the house when the sun comes out switching on electric towel rails and fan heaters

 

 

 

 

 

 

Edited March 27, 2018 by readiescards
Add solar PV report stats

[Dreadnaught]

@TerryE has commented extensively on such solutions. I suspect he will be along shortly.

[dpmiller]

Another option, which offers some extra control options:

 

https://www.advanceappliances.co.uk/product/151-plug-n-glow/

[ProDave]

A bit of a re plumb, but to me it would make a lot of sense to have a decent sized buffer tank for the heating, and divert excess PV to heating that. Then as well as your tank of DHW, you would have a tank of heating water for the evening.

[Nickfromwales]

You shouldn't be running around at all....it's massively overrated

Havent you already got 2-port zone valve(s) fitted for the space heating and the uvc ? 

If so, just fit the Willis heater into the UFH return, at the manifold ( eg before any other parts of the system tee in / out of the UFH primary pipework ) and then tee the return back into the UFH flow via another 2-port ZV. That way the hot water created by the Willis heater travels the shortest distance and doesn't risk any nuisance tripping of the internal overheat stat as it'll never see the high flow temps of the boiler flow.

The power to the immersion opens the Willis 2-port ZV and as the heating and hot water should be off at the controls ( by design, amended at the time of the Willis heater being installed so they don't have any conflict ) the UFH manifold pump will simply recirculate through the Willis, through the ZV, and back into the manifold blending valve. 

The casualty will be the stat in the Willis which will work a bit of overtime but their cheaper than shoplifting so not a real worry. 

 

8 minutes ago, ProDave said:

A bit of a re plumb, but to me it would make a lot of sense to have a decent sized buffer tank for the heating, and divert excess PV to heating that. Then as well as your tank of DHW, you would have a tank of heating water for the evening.

I think the heating would consume what's being generated pretty much as it's being generated, or near as damnit, so a few quid for a Willis would be where I'd start the research. Buffer tank ( perhaps with DHW uplift coil ) would be my get out of jail card if the Willis wasn't conveying enough of the generated PV to space heating. Having DHW uplift in that buffer would give you oodles more DHW when space heating wasn't needed / significant. 

At this late stage I'd want the retrofit to be as least disruptive as I could make it, hence my first suggestion. I'd even just look at the ground floor and not the whole of house space heating as the ground floor slab should do most of the work anyhoo. ?

[TerryE]

2 hours ago, Dreadnaught said:

@TerryE has commented extensively on such solutions. I suspect he will be along shortly.

 

as @Dreadnaught says I have even described all this in my blog posts, so all you need to do is to read this, and ponder what I say. 

 

Make a point of reading the referenced  posts which go into a lot of detail including my mathematical simulation of the system to determine its operating characteristics so that was I confident that the system would work this way before I implemented it (and the first commission tests validate the model).

 

We have this as our sole source of UFH heating and at the moment we are at a comfortable 22°C with a ½°C ripple over the day using 17½kWh of heating at E7 for a large 4 bedroom passive class house.  Hence my recent post:

But to your specific Qs:

1. Does it need an additional expansion vessel adding to the system? It add an extra ~1ltr of water in the UFH circuit, so No. 
2. If I install the device just before the UFH manifold, would extra valving be essential to prevent the UFH pump running the return water through the boiler circuit all the time?  Again read my posts and have a look at the photos.  I have my Wilis effectively on a 15mm bypass across the UFH loop with a PHE on the other branch.  No valves, but the flow runs roughly 50:50% through the two branches. That's on the warm loop side in a boiler mix-down implementation.  I don't have a boiler, so my TMV is just cranked open.  It is entirely redundant and I've only got one because this was the cheapest Wunda config that suited my needs.  If you read the physics of this then you will see that this isn't an issue.
3. I have a BS1820 temp sensor buried in the slab, the control circuit seems simple ... Have I missed something?  What a bizarre type of sensor.  A lot of people do this and for the life of me I don't understand why.  I've already got this remote sensor which integrates the slab temperature by UFH zone: the UFH flow loops.  I have ds18b20s on the out and return of each loop.  Simpler, easier and a simple way to balance the flow through the loops as well.  OK, so I have lots of ds18b20s measuring just about everything, but at £1.50 each this is hardly breaking the bank.

On your supplementary points:

• If the UFH pump failed the Willis unit dual thermostats will cut in and switch it off.  Eh?? This isn't a failure point; it's an essential backup safety feature.
• If the slab temp sensor fails the floor could get too hot so an extra temp sensor on the manifold could verify the return UFH water remains within acceptable limits.  Yes, which is why you should always monitor your flow and return temperatures and include cut-offs in your control system.  If you are too mean to spend the extra ~£6 on doing this per loop, then by all means do it on the manifold out and return.

I am not sure that you are using for your automation system but your inclusion of Javascript-like pseudo-code suggests that you are reasonably IT literate. 

 

My HA is based on a single RPi3 with SSD running Node Red.  All of the IOs are done by two (soon to be one) ESP8266 modules USB-serially attached to the RPi (so the core DHW + CH system is independent of our Enet Hub + Wifi).  I do use other WiFI-connected Wemos modules for nice-to-have remote sensing.  I run the NodeMCU Lua subsystem on my ESPs and talk native json to Node Red, though I also run and use a MQTT broker on my RPI.  

 

The support that I get on the Lua system is excellent (but that's because I on one of the core developers on the project ).

 

By far the most expensive bits in my system are the 4×Crydom SSRs, but that's so my sparkie could sign off the 240V side and keep my BCO happy.  The rest cost < £100.  Not a virtue; just no point in spending any more.  Yes, it has involved a lot of personal time, but it's my hobby project, and my son-in-law is into all of this shit as well so he's my SPoF backup.

 

Edited March 27, 2018 by TerryE

[Nickfromwales]

The only issue I see is the difference in slabs, whereas @readiescards slab is actually a thinner liquid screed iirc? These dont tend to hold heat for very long at all, and I doubt that chunking heat in would be as effective as a constant regulated flow would be.  

[readiescards]

60mm liquid screed, 6.3m3 over 112m2

[TerryE]

@readiescards Mine is 100mm with extra 200×200 bracers every 2m and a 300×300mm ringbeam and ditto 2 loadbearing  cross-members so you have ½ the thermal capacity / sq m.  how many sq m in your GFL, because your curves are going to be a bit steeper than mine?