Pure Energy
My fascination with energy consumption is a long time coming, and I think I’m now arriving at a new level of need for information. OK, it’s not need, really. It’s intrigue. I’ve taken energy consumption measurements of most every household device that I own and I’ve even been tracking my monthly energy bills, year on year. I recently built a plot of energy consumption (from my electric bills) vs. average monthly temperature (from weather underground), covering the past two years. The plot says exactly what you would expect:
- Our bills are largely a function of the outside temperature, proving that our heat pumps dictate our bills.
- Our comfort zone, with neither heat or A/C running, is somewhere between 70′F-80′F. And it would seem we spend 3-4 months a year at the sweet spot in Portland, which ain’t too bad.
- The snowstorm of Jan’09 was a mighty cold one!
- Taking a 15-day vacation and lowering the heat (from 68′F – 60′F) can make an impressive dent in power savings (see Dec’09-Jan’10)
This is pretty cool information but, to me, it only makes me want more detail. With no action on my part, I may be getting more detail in the very near future. PGE recently installed a new ‘SmartMeter’ on my house. Information is limited, but PGE claims “By 2010, customers with smart meters will be able to access detailed information about their power consumption”. I spent some time looking at SmartMeter solutions in other cities around the country, and I found at least a few that offered hourly energy logs on the energy company websites. Combine this with detailed temperature data from the weather underground or a home thermometer, and you’ve got a great map of daily energy consumption.
All that will be pretty cool, but it still isn’t enough to really understand where your energy is going.
- I can’t measure and differentiate high current devices like my heat pumps, dryer and microwave from low current devices such as my refrigerators or computers.
- I don’t have visibility into device duty cycles, for example, a refrigerator or tankless water heater
- I can’t realistically measure the power consumption of my lights with dimmer switches (by the way, Lutron has an interesting demo on how to estimate power with their switches)
- I don’t have visibility into seasonal usage models as most household energy readings are taken during a specific season.
The Front End
Now, I could accept this lack of visibility and just stop here. But for me, it just whets my appetite⦠what more information is lurking in the day to day or even second to second fluctuations within my home? Can I split the total energy readings into their constituent components? What about if I get a solar panel, can I measure the input energy?
At this point, I feel no other option than to nerd out on this and take it to its logical conclusion. What is that? Well, I have to say that I like the way Brainside Out has framed it.
Google’s making a presence, too, with their PowerMeter application. Right now, it looks like they are teaming up with some utilities (hopefully PGE soon!) as well as device manufacturers such as TED. The data is pushed to the Google servers and then summarized using simple plotting. See this excellent summary at Engadget of an integration of a TED system with Google’s power meter.
Whole House Data Collection
OK, enough dreaming of how cool my world could be. How am I going to actually get the data to supply these GUI options? I have read about The Energy Detective (aka TED) in the past, but I have hesitated in purchasing it because it seems to be only a partial solution. If I’m right about what content PGE will be deploying this year with the SmartMeter, the TED is almost obsolete. Specifically, TED only gives you whole house energy monitoring versus time (or, you could configure it to read a single breaker), so its value versus the SmartMeter is sample frequency, probably one second resolution versus 1 hour resolution. Thankfully, there is something better out there, the Brultech ECM-1240
According to the specs, it is a 7-channel power meter than can measure any number of circuit breakers. Furthermore, if you want to measure than more seven separate breakers / loads, you can chain multiple ECM-1240′s [still trying to figure out how many can be chained]. Connect them to your data collection server via ethernet, RS232, USB or wireless. Now that is sweet! I counted the breakers in my circuit panel, and at a total of 40 breakers, I don’t think I’ll be measuring every single one, but with careful selection of the breakers that I want more information on and then one channel connected to the total energy consumption, I should be able to get a very detailed set of data streaming into my server. And if you really want to geek out, you can connect natural gas meters, water meters, and solar power inverters into the Brultech channels and achieve whole house telemetry. All that for about $200 for the basic model. Just to see what you can do, check out this web site where a Brultech is currently online.
Luckily (?), my current 1963 Pushmatic electrical panel is a fire hazard and I have been told I need to replace it as soon as possible. With a little planning and a good contractor, that Brultech should be running live soon.
Device Specific Data Collection
There is still the matter of power telemetry for general household devices. While the per-breaker readings are useful, they don’t really distinguish between my Airport Express and my stereo, or my refrigerator and instant water heater. For those, I have been using the Kill-a-Watt, but I think it has run out its useful life for me. The Kill-a-Watt is great for instantaneous power, power factor, voltage, etc. readings. Its primary value to me is in measuring energy consumption over very long windows of time. Howevwer, it doesn’t give you any insight to duty cycles or personal usage models. For example, it averages a refrigerator’s on/off duty cycle into a single number and it filters out the natural bursts of my tankless water heater’s activity. The next logical step is the Watts-Up .net which has local logging capabilities, up to a total of about 1.4 days of 1Hz sampling of a single data point. If you really want to geek out, and I do, the device can sample up to 18 data points every second and stream them to your computer over ethernet for infinite recording capabilities. My plan:
The entire Watts Up setup is a little spendy ($235 for the Watts Up .net
The Future
Unfortunately, I can’t be buying a Watts Up for every device in the house and rolling the numbers up into my real-time logging software. But don’t you think that this is where the future is going? Education is the first step to energy consumption reduction, and it is my opinion that for the U.S. to become energy savvy, we need this kind of telemetry on every household device. Powerline communication is easy and effective. It seems reasonable to me that, with some standards in place and government regulation, we could have all devices sending power meter readings to a home server for daily roll-ups of dollars spent on each device. Perhaps such a thing already exists. Had I seen the energy consumption of my basement refrigerator three years ago, for example, I would have gotten rid of it. But instead, I let it run for three years consuming three times the energy that a cheap Energy Star model would consume.




You know, at some point you’re going to be saving so much energy that you’re going to have to question the amount of energy wasted on energy-measuring devices. (Sort of like how using Windows Task Manager uses CPU cycles to determine how many CPU cycles are being used.) Just saying.
If you look at the actual Y-axis on my energy vs. temperature plot, you will see that we have a lot of room to save energy!
Just for you, I’ll have to conduct a study of energy consumption with and without task manager and report out here. My bet is that the data is mostly already being logged and the additional microseconds of activity are not large energy consumers. Much, much larger is looking at a web site with flash content or running iTunes in the background.
Oh, you know I think this is all fascinating, though I’m happy to let you work it all out first. I still do think it’s funny (at some level) that you’ll have to have several extra devices consuming power in order to measure your power (and thus lower your power consumption). But yeah, a Watts Up is not a dryer.
Anyhow, Task Manager itself reports that it uses 1% of my CPU when it’s running. It’s not going to bog down my computer, but it’s not nothing.