Monitoring resource usage

My recent use of my Current Cost unit to determine whether or not the insulation on my heat store was efficient enough has rekindled my interest in energy monitoring again. And the fact that my gas-powered central heating boiler has broken down made me wonder how hard it would be to measure the gas consumption of my house in the same way that I currently measure my electricity consumption. And if I’m going to do that, then I may as well look at the water usage too. And it may be interesting to combine those with the internal and external temperature.

Since I started playing with my Current Cost, back in 2007, it has become a Current Cost Classic, and there is a new kid in town, the Current Cost Envi, which does everything the Classic does, but has the ability to also independently monitor several individual appliances. It occurred to me that it would be possible to subvert a couple of those appliance monitors to measure the gas and water usage. I suspect that this is likely to confuse the Envi display unit (which will assume that all inputs are electrical) but the XML that comes out the unit and into my home server will allow me to correctly interpret the data and present it as a series of useful graphs.

So on Thursday I acquired a shiny new Current Cost Envi unit in ghastly white, as sadly my supplier didn’t have any tasteful black ones left. I also had a chat with the friendly folks at Current Cost, and thanks in part to the long relationship that we’ve had with them, I’ve acquired a couple of wireless sensor development boards. These have a single simple digital input, which sends a fixed 0 or 500 signal to the display unit on it’s own channel. Not quite as useful as a proper analog setup (which I’m already thinking I could obtain by hacking a current-clamp board!), but absolutely ideal for counting pulses on (say) a gas or water meter…

So the next part of this project is to dust off my (ancient) electronics skills, and work out how to interface one of these boards to my gas meter in such a way that the gas board won’t have kittens next time they read the meter; ie, contactlessly. A few people have already done this, using a variety of different mechanisms, depending on their meter type. I need to spend some time with my gas meter to determine what’s possible, but hopefully I can do something simple with hall-effect sensors, photodiodes or similar.

Water is looking more difficult; I’ll have to buy and install my own flow-meter into the rising main of the house; not sure how expensive that will work out to be, or how I could calibrate it, but I’m sure there must be suppliers out there with lots of information waiting to be read on the internet.

And if nothing else, that would finally settle the discussion over whether it will be cheaper for me to move to a water meter tariff or not!

Efficient insulation

Four years ago I replaced the hot water system in my house with a heat store, which (at the time) was a fairly revolutionary system with some real advantages:

• mains pressure hot water
• high capacity
• enables more efficient boiler usage
• no cold water tank in the loft
• no hot water cylinder in the airing cupboard
• DIY installable

The heat store is essentially a large tank of very hot water, heated indirectly by the boiler. That hot water is then used to instantly heat cold water from the rising main, to provide the hot water for the house. In some ways it operates on similar principles to the hot water part of a combination boiler, except that the heat store is much more powerful (160Kw), allowing it to cope with multiple hot water outlets with ease. It is also potentially more efficient than a combination boiler, as the boiler can run for long efficient burns while charging up the heat store, rather than short on/off burns of the combination boiler while heating water on demand.

In my case, I installed the system in the loft, as it allowed us to reclaim a lot of cupboard space. The system came factory insulated, and it was claimed that it needed no more insulation, even when installed in outbuildings or loft spaces. But I’ve always had a niggling doubt, wondering if I was losing heat to the loft, especially when the weather is cold. But I’ve never been able to prove it to myself one way or the other.

The failure of our central heating boiler this week has given me the opportunity to prove that the heat store is adequately insulated, because while we’re waiting for the new boiler to be installed, I’ve switched the heat store over to electric operation. And electricity usage is very easy to monitor, especially if like me, you happen to have a Currentcost meter.

The heat store has a 3Kw heating element built in, which for the time being I’ve set to run permanently via a thermostat. Whenever the water in the tank drops below 70c, the heating element will cut in and heat the water. Currently the loft drops to about 2c overnight, so if the insulation is inadequate I’d expect to see the heat store lose heat overnight (when we’re not drawing hot water), and for the heating element to cut in, and a 3kw electric load appear on my Currentcost logs.

But it doesn’t. It looks as though the only time I need to top up the heat in the heat store, is when I draw heat off it to produce hot water for the house. Which is good. Incidentally, I believe that the regular “blips” of power usage that are particularly noticeable through the night are caused by the freezer, when the compressor cuts in.

Click the image for more detail:

12 hour power usage

Toys!

I’ve been monitoring my whole house power consumption for some time using a currentcost meter, connected up to my home server. However, I’ve been meaning for a while now to pick up a meter for measuring the power consumption of individual devices to help me fine tune my power consumption. This week my local Aldi were running a special offer of a (fairly sophisticated) meter for £8. Turns out that it can measure voltage, current, mains frequency, power consumption, and power correction factor, which indicates how efficient a device is. Since then I’ve been wandering around the house measuring things, trying to work out what we leave running all the time that we actually need to turn off.

So far I have been unable to find a single “wall wart” adaptor that uses *any* measurable power when idle, despite the media talking about how we should all go around and disconnect them all. Things like my powered speakers for my laptop however, draw 10w all the time. Ooops.

And while I was in Aldi I also picked up one of these blood pressure monitors too. I wasn’t sure how accurate it would be, but on trying it out, it seems to give approximately the same readings that I’ve seen the nurses record during my hospital visits. So it may well be surprisingly accurate.

However, my view is that I’m less interested in the actual readings, and more interested in any marked changes in the readings; because if I see any significant increase or decrease then I’ll be needing to go visit my GP anyway.

“Measure twice, cut once”, Anon

As many of you will know, over the last 3 or 4 years I’ve become increasingly concerned with what is often referred to as “Green Issues”, and in particular the environment and how we are impacting upon it. I guess it’s one of the results of getting older and having kids – a case of suddenly realising that what we do now is going to affect them, and even their children. Now this isn’t to say that I’m particularly fanatical; I’d like to think that I’m a pretty pragmatic kind of guy. But I do try to do what I can, even though that’s pretty much limited to trying to reduce my personal impact on the planet, and finding ways to reduce the impact of IT on our planet (this latter one because its part of my day job, working in emerging technology for a large multinational IT company).

So I was really interested to come across this little device recently. Although I’ve seen such things before (usually at high prices), this is the first one that I’ve seen that can handle proper 3 phase supplies, and also has a PC interface on it, allowing it to be integrated with a home automation system.

The companys business model is to sell these (hopefully in huge quantities!) to the energy supply companies, who will then either sell or give them to end-user power-consumers in an effort to help them to reduce their power consumption. This model is being helped (in Europe anyway) by the expectation of legislation that will require energy supply companies to find ways to help energy consumers to reduce their power consumption (and hence CO2 emissions). Unfortunately you can’t (as an individual) buy one just yet, but fortunately (for me) the company I work for is involved in a trial using them, and so I was allowed to buy one at the price an energy supply company would pay. I don’t believe I’m allowed to say exactly what that was, but let me just say that it was a pleasantly affordable deal.

Setting it up took about 60 seconds. Hook the current clamp around one of the meter tails (the wires running from the supply companys electricity meter to my fusebox) and turn on the wireless transmitter. Plug the display unit into my server with a serial cable, and then into the mains for power. Power consumption numbers appear on the display unit – simple as that. I must admit that I’ve not done a detailed check on the unit for accuracy, but a few quick empirical tests show that it’s definitely close enough that I’m happy to believe the accuracy claims on their website.

Reading the data from the serial port is a small matter of a little bit of perl. The data itself is encoded in a tagged text format, similar to XML, which naturally makes it trivial to parse in a perl script. So its easy to extract the interesting bits of information, and store them in a round robin database, where they can then be analysed, graphed, or whatever. In my case I’m extracting the instantaneous power consumption figures for the three cable clamps (even though as a domestic user, I only have single phase) and the temperature at the display unit. This means I can combine the room temp of my study (where my server and the display unit are) with the internal temperatures of my server, and see how external temperature affects those, as well as watching how my electricity usage varies over time.

Of course, the trick here is to use this thing to lower my energy consumption. But that doesn’t really mean to stop doing things. We will still need to run the washing machine, the tumble-dryer, the fridge and freezer etc, and we’ll still want to boil water in the kettle. Because although all of those things take a “lot” of power, it’s power that’s being spent on useful work. Where we can (and should) be concentrating is the “resting” power-consumption of the house. The stuff that’s running 24×7, and doing nothing useful, that could be shutdown. And yet, to my surprise, despite having quite a lot of general stuff that does run 24×7, my base resting power consumption is pretty low (well, compared to some of the other people on the trial, anyway!)

Overnight the numbers are under 200w. And that includes my server, my printer, my ADSL/wireless, my 24 port switch, the central heating controls, the alarm, and a nightlight for the kids, plus a surprising number of “standby items” including alarm clocks, radios, 2xTV, set top boxes, various cellphone and laptop chargers, etc etc. And the big ticket items are the ones that it would be difficult/impracticable (for me) to change – namely the server and its associated support networking. So it seems like we’re doing a pretty good job at the moment, though I think I might try unplugging some of the easily accessible stuff, and see if I can half it.

After all, 100w saving doesn’t sound a lot, but when its over a whole year, that’s 876kwh. Nearly £90 of electricity ($170). And there are 25 million households, and a lot more commercial/industrial establishments in the UK who could attempt this …

If you’re interested, you can see near-realtime graphs of all this data here.