advanced controls for heating, solar, HW etc

[dan_aka_jack]

Whilst doing research on efficient ways to heat a house, it's becoming clear that one way to make a big difference is to install "advanced" monitoring and control electronics.  Are there any off-the-shelf boxes that will do all these functions in one box:

1) Take information from room thermostats in every room to control the CH

2) Control the boiler for CH and DHW

3) Control the solar system

4) Control the heat dump

5) Log energy use

6) "learn" the physical characteristics of the house and adapt the heating system to be the most efficient

I assume there are off-the-shelf boxes that will do functions 1-4.  But what about functions 5 & 6?  Has anyone experimented with building their own control circuitry using PICs or embedded Linux boards?

Thanks,
Jack

*edit* Here are some links to embedded Linux systems:

* http://linuxdevices.com/articles/AT2614444132.html

[KenB]

Jack,

Point 3.  Control the Solar System?   There's only one Guy I know who knows how to do that ;-)

Joking aside,  it seems logical to want to gather as much information from each room, in terms of temperature profiles, to make a room by room decision of whether you need to demand heat.  Some south facing rooms will have tremendous solar gain on winter afternoons, for example, and will not need as much heat as those on the northerly side.

I have some ideas on how to economically control each radiator, to make a whole series of micro-zones, and I am developing a product that does exactly this.

A smart controller that learns the typical hours of occupancy and heat demand of a given room could also help to optimise the heat required by each room.

I suspect that the contribution from the solar water panel will decline as you enter the winter months,  anyone who has experience of say a 20 tube Navitron in December, might like to comment to what is likely.

During the winter when the boiler is needed, the heat dump might be useful as a buffer or accumulator. Heating 260 litres of water to 85 C, presents a known load to the boiler, and if it will prevent the boiler from unecessary cycling, all the better.

Remember when a boiler turns on at 6am on a frosty December morning, everything is working against it. The combustion air is cold, the water in the pipes is cold and the house is cold.

Points 5 and 6 are of particular interest. Most of my fuel reductions have been achieved by not heating rooms that have litle or no occupancy.

However, using a 150W PC to process your energy control system is burning 400W of gas back at the power station. It would be more efficient to burn that gas in the home and get the benefit of most of its calorific value.

I see that nPower have announced a third major price rise in electricity and gas.  Its getting like painting the Forth Bridge, there's always some utility raising their prices, and I suspect it can only get worse.


Ken

[dan_aka_jack]

Hi Ken,

Thanks for your reply.  It was actually one of your earlier posts in another thread that got me thinking about "intelligent" control systems in the first place!

When is your product likely to be ready?  Do you need any beta testers?

However, using a 150W PC to process your energy control system is burning 400W of gas back at the power station. It would be more efficient to burn that gas in the home and get the benefit of most of its calorific value.

Sure.  I was thinking more along the lines of a 450mW single-board computer running embedded Linux.

[wyleu]

The energy consumption of the management component should be, and in this day and age, could be minimal. A wireless based web server would seem an excellent way of interacting with it. In effect you get the management for almost free that way and the reliability is considerably better. But as Ken mentions the gathering of data to make judgement is a different matter and for such a situation the old PC will probably be a useful tool in the hands who have an idea to attempt some R&D.
Without trying to steal Ken's thunder, where does he see the boundaries of such a system? For instance would it be controlling relays on the board or sending control signals to something else? Have you considered X-10 for some of these command mechanisms? It saves the hassle of wires running everywhere. I don't know much about wireless comms but I do wonder if we might easily reach a point where there are more collisions than free packets bouncing around the ether when every last device starts to sprout an aerial, especially when your neighbour is doing the same sort of thing I've seen collision based ethernet fail well before the calculated limits in just such situations. Zig-bee seems to learn and handle it all quite well but what of the many more simpler devices that are appearing?
The last time I looked at SCADA it was busy trying to refit a security model in the light of recent (and not so recent now ) political events. So what with manufacturers all trying to develop their own proprietry systems could this not all suffer the same fate as the living room, awash with different remote controls and isolated systems with little or no overall system to bolt it all together. I have made several attempts to do some degree of centralized control but it generally seems to run up against some low level incompatability that didn't meet the marketing departments price point and seem to find it quite unsettling.

Also the isolationism of systems lead to a ring fenced system where measurement and control are combined in one 'system' that is unable or unwilling to publish data for the benefit of other systems. Great if you wish to Veismannize your house but what if you already have kit from different manufacturers? B&O make wonderful systems but look at the price and try connecting other things to it, it quickly ends up as a hotch potch that doesn't get used exept in carefully orchestrated demo's for friends but doesn't actually play much music the rest of the time.

Sorry for sounding a bit down on this, I hope it's simply my lack of research that can be easily debunked here. I am a great believer in intellegent control being a massive contributor to the enormous challenge we all face, I just fear that, as is often the case with complex computer systems, the hardware and software are all there , it's the management decisions that are lacking.

Chris@wyleu.

[KenB]

Chris,

Good points.

It has to be cheap, reliable and easy to use.  It should be "fit and forget" rather than "plug and pray".

In this day and age it needs to be seen to be saving you energy, without being gimicky.

I have seen dual port valves for as little as £18.50, but having a smart controller with a mains relay for each radiator in the house might be a bit of an overkill.

I live in a 1905 semi, and the house next door is unoccupied.  I suspect that I am contributing quite a lot to next door in terms of heating.


Ken

[dan_aka_jack]

Thanks for all your replies!

For a few days, I was seriously considering attempting to build my own heating control system.  The heart of my plan was an off-the-shelf single-board PC running embedded Linux.  Each room would have at least one temp sensor and each radiator would have a motorised valve.  These valves and sensors would be wired to the computer (I share Wyleu's fears about overcrowding the 2.4Ghz wireless band). The system would also control and monitor the boiler, the hot water tank and the solar hot water system.  The single-board PC would be powered from the mains with battery backup and hopefully the whole system would draw well under a watt most of the time.  The single-board PC would have an ethernet port so it would be easy to connect it to a PC for more in-depth monitoring and control.

I've gone off this idea for two main reasons:

1) it'll take AGES to develop and build!

2) I'd worry that a system as "hand made" as this system would scare off potentially buyers for the house.  We're only planning to stay in the house for about 5 years.  If we were planning to stay in the house for more like a decade then I'd think much more seriously about installing such a system.

Instead, I've decided to install a much more simple system which will just monitor temperatures in each room.  Each room in my house has an 8-core cable running to the room from the cupboard under the stairs.  This cable will eventually be used for the alarm system when I get round to installing it but I figure I can use the 3 spare cores to put a temp sensor in each room and connect each sensor to a simple data logger.  My plan is that I'll log the temps for each room and use that information to manually tweak the TRVs on each radiator and to tweak the timings for the boiler.  Does that sound sensible?  I realise that thermostatic radiator valves are far from perfect but hopefully, when combined with decent monitoring, it should be possible to configure a reasonably efficient central heating system.

[wyleu]

I believe there should be a conceptual difference between measurement and reaction.
The measurement of values should be community wide how you react is a seperate data stream.

If you control channel can't control devices how can you do any damage?

[dan_aka_jack]

Good point, Wyleu.

I know this is an "unanswerable" question... but... can anyone suggest how much benefit a highly controlled CH system would have over a basic system where all the radiotor control is just done via TRVs?  Sorry, I didn't phrase that well.  Can anyway give decent estimate for the differnces in energy consumption between these two systems:

1) Basic system.  The boiler is controlled by a timer.  Individual radiators are controlled by thermostatic radiator valves.  If a room is going to be unoccuplied for any length of time then a human will turn the TRV down to zero.

2) Highly controlled system.  Each room has several temp sensors.  Every radiator has a motorised valve.  Each room has its own electronic control device which controls both the radiators and the boiler.

My gut reaction is that the highly controlled system wont actually give much benefit to the basic system assuming that the house dwellers (me and my girlfriend) can get into the habit of turning radiators off when we leave a room.

Also, one simple question: what does the boiler do if all the radiators are turned off?  Will it detect that water wont flow round the circuit and turn the pump off?  Or is there a bypass pipe?  Or will the pump keep trying to push water round the circuit until it dies?!

[KenB]

Dan, Chris & Forum,

The key to making savings in heating fuel can be summarised thus:

1.  Burn the fuel as eficiently as possible - ie a condensing boiler
2.  Scale the output of the boiler such that it is running efficiently for most of the time
3.  Have the means to turn down the boiler power so that it can run efficiently on reduced heating demand

All of these can be achieved with a modern condensing boiler.

4.  Have the means to determine the optimum heat output needed to satisfy the given demand and the prevailing conditions.
5.  Have the means to shut down zones that require no heating
6.  Have a knowledge of patterns of usage, likely ambient temperatures and the heat retention characteristcs of the room

These points can be addressed using a smart controller, individual room theremostats and the means to shut down radiators on a room by room basis.

7.  Aim to achieve the highest standard of insulation.


A historical record of the heat demand would also be useful. For example, if I used 96kWh of gas on December 12th 2005, it would be a reasonable first estimate to assume that I was likely to use the same amount on Dec 12th 2006, and then make a correction for the external temperture.

I reduced my gas consumption from 20500 KWh to 16400kWh between 2005 and 2006. I employed a very manual method of monitoring and control, so that I could understand the basics, and then find an effective means to automate it.

I replaced my old boiler with a modern condensing boiler - one that didn't burn 6kWh per day just keeping the pilot lit!
I added underfloor heating and a new well insulated extension to the north facing back of the house to eliminate draughts.
The Navitron 20 tube panel offset about 1200kWh of gas this summer for hot water
I turned off the heating in unoccupied rooms and made full use of TRVs in seldom used areas
I set the main thermostat in the living room to 17C overnight, and turned it up to 19C when that room was occupied.

I suspect the next step is to improve the insulation in the loft, by fitting cellotex between the rafters.



Ken

[dinitro]

Apart from using TRV I think zoning needs to be mentioned...

Our house has 3 floors, each with programmable stats that control zone valves.
The bedrooms are heated at night, but during the day only the lower floor is heated, with periodic
boosting of the upperfloors just for traffic.

A combination of the programmable stats and zone valves mean that the heating may be off
or on a lower setting.  I use Danfoss TP7000MA.  they have remote sensors. These sensors are placed
at either floor level or placed in a particularly cold room eg. below a cold water tank!

The effect is a macro control of the heating, with micro control by room TRVs.  If you want better spec TRVs
I would suggests using remote bulb ones, the bulb can be positioned 1-2 metres from the rad and so get a better
approximation of the room temp.  This is non electrical.
Another suggestion is to use the honeywell TRVs that are battery controlled TRVs, remote sensors can be used with these.

http://europe.hbc.honeywell.com/products/ecatdata/pg_hr40n1.html

I'm sure the techically minded can hack into the control electronics and have external control through X-10 / linux etc.

dinitro

[dan_aka_jack]

great, thanks for the replies.

[KenB]

Jack, Chris

I bought a PIC starter kit today for around £20 from Farnell.  (Part No 439-0787 £19.15)

It allows you to get started with the smaller 14 pin devices such as the PIC16F684, which have built in ADC channels ideal for temperature sensing.  They also do a signalanalysis board and PC software that will provide a basic oscilloscope,datalogger or FFT analyser - all for anout £13.30  (Farnel Part 815-6255)

Software running on the PC allows you to program this dev-board from a USB port, and get data back from it.

The PIC can easily send data to a 433MHz transmitter module such as the RF Solutions (Farnell Part 722-5647  £5.61).

Last year I got a PIC and 433MHz transmitter to control my boiler via the Danfoss Wireless Thermostat. The PIC generates a wireless message that looks like a demand for heat from the thermostat. When the set temperature is reached, the PIC sends a boiler off packet.

With these basic parts you have the means to get started with a smart heating/solar panel controller system.

This could be extended to an individual thermostat for each room, all communicating back to a master controller or alternatively, each room thermostat makes an individual decision whether to turn the radiator on, via a 2 port valve.

There are other low cost ways to individually control radiators, and anyone who is interested should PM me.

At the end of the day, central heating controllers are fairly dumb, turning the boiler on and off according to heat demand from a £5 thermostat or a £20 time clock.  Add a bit of intelligence to the controller, and you should be able to get an improvement in gas consumption.

I wonder if its possible to persuade a modern condensing boiler to modulate down to a lower power setting, by feeding a control voltage into the temperature control pot on the front panel.



Ken

[dan_aka_jack]

Cool, best of luck with your PIC project.  Please do let us know how you get on.

I've only ever used a PIC once and that was a while ago (for my A-Level in Electronics!)

Two quick questions:

1) Can you get a PIC to talk to an Ethernet network?  I guess not?

2) What sort of programming environments are available for PICs now-a-days?  Do you still have to hand-code the assembler for absolutely everything (power-on-self-checks, drawing a character on the screen, what to do when an interrupt is called etc)?

Thanks,
Jack

[wyleu]

http://www.microchip.com   for all things pic'ish.

Some of them do do ethernet and there's a C compiler available as well. The Dev boards now no longer require you to unplug the chip for programming. All in all a very mature little family.

[Alan]

Hi All

Ref : PIC Check out   http://www.oshonsoft.com/  Brilliant cheap soft ware.
write in basic, simulate and compile to hex.  token programs for pulse width modulation and
Measuring temperature with DS18S20 ( Three wire temp logging )
Have tried lots of head banging stuff. Cant beat this sofware.

Regards

Alan