Saturday, August 20, 2016

measurement and control

INDEX to the series

I have been learning about my home electrical usage through my smart meter data and finding it not so smart without some added software. Nevertheless, I have been making use of all this new data and making improvements (reducing the variable portion of my electric bill) while making plans for next steps some of which involve more measurements.

Adding the timers to the water heater, fridge and freezer to move them off my peak rate times was an obvious move but a bit bothersome. The timers don't always work. I had trouble with a wireless type so went with plug in autonomous timers. The cheap Chinese ones I bought on Ebay for $10 each sometimes do not switch as they are supposed to. I went electronic but it would be simpler just to have mechanical timers and cheaper. Try to find three prong versions (properly grounded). The timer for the water heater had to be wired in.

I have been thinking about remote controlled AC sockets. Particularly interested if they monitored the load as well as controlled it, something like the Belkin WeMo.

What if these devices mounted directly in my electrical panel as snap in replacement approved devices that communicate wirelessly? That doesn't exist yet. Maybe with the Internet of Things (IoT)?

Controlling my appliances could be done currently through some type of home automation system, via software, in a browser. That would be nice. I am looking around for suitable systems.

About the monitoring, I still don't have appliance level data from my smart meter or any meaningful breakdown. There are software tools (some of which have their own hardware and sensors) like Plotwatt, Bidgely, Navetas, eyedro and Neurio.

Some, like Plotwatt, are being marketed to the utilities as a service they would provide, whether added to my bill or not, is playing out now in the market. I am looking at my account on-line at Hydro One Networks and I can see many tools to help me understand my electrical usable but no GreenButton links as yet or any appliance level information, just general.

There are systems marketed for the DIY types like TED (The Energy Detective) and OpenEnergyMonitor.

But using my smartmeter is not the only way to understand my house electrical and to monitor my appliances. It might not even be the best or only way. It will probably be some combination of tools that will guide me to what I need to know.

Different tools work different ways. Some plug in between my appliance and the wall. Others involve the installation of a small sensor in my electrical panel.

Of the plug in kind, the Kill A Watt has been popular for a number of years. It measures the energy used at the outlet and stores some totals like kWh per day, week, month, year since plugged in and if you enter rates, it will give you a cost. It does not connect to anything so you can't get the data out of it and when you unplug it resets. Not very useful with tiered rates but the kWh measures could be useful. It is limited to regular plug in 120 VAC appliances up to 1500 watts so you can't use it for a water heater, clothes dryer or airconditioner or a furnace, in fact any of your larger appliances. But still a useful tool. I have had one for years. It does not measure low power very accurately, like below 10 watts.

These and other types of plug in meters may be available at your library http://georgesworkshop.blogspot.ca/2014/12/your-library-has-tools.html

Understanding home electrical use is part of the broad category of Noninvasive Load Monitoring (NILM) - good background at Wikipedia.

Some interesting academic links:

https://web.eecs.umich.edu/~prabal/pubs/papers/campbell14gemini.pdf

http://www.ucl.ac.uk/carb/pubdocs/CP-DMU-09-IEECB08-NoninvasiveMonitoring-2008-NRB-AJW.pdf

http://digitalassets.lib.berkeley.edu/techreports/ucb/text/EECS-2012-152.pdf

Thanks for your interest.

George Plhak
Lion's Head, Ontario, Canada

INDEX to the series

Friday, August 19, 2016

heat

INDEX to the series

I hope the basement will be warmer this year because I have rebuilt the two basement windows in the house. There had been significant cold air coming in from the outside before this repair.

I am also planning to cut off heat to the sun room. It has a lot of big windows. A wonderful room in the summer. A wonderful room in the winter also but it must cost me a lot to heat it. I don't know how much but I would not be surprised if heating the sun room was 1/4 of my heat cost including fuel and electricity. Calling it a sun room in February here in Lion's Head is a gross misnomer. There is not much sun here in February.

I have replaced the operable sun room windows with nice thermopane but the large ones are single panes. This room has a bad air leak noticeable when the wind blows a certain direction and the door to the outside does not seal very well at the bottom.

So this winter, I plan to temporarily partition the sun room, including the forced air duct that feeds heated air to to its three floor vents. That duct is large (16x8 inches) and has no damper (air control valve) in it now. I would like to add one. A quicker fix is to block the three floor vents. There is no return air duct. Return air just flows back into the main house through a six foot wide open doorway.

I am not going to heat that room this winter. I will modify my behavior by doing without a whole room for the winter except for cold storage. I don't use the front door although I keep the snow clear from the door as an emergency exit so the partition will need to have a door.

Conservation is key to reducing my electrical usage. If I need less heat, I will use less electricity even though I use oil primarily. In my furnace, there are two hungry AC motors and the ignition which run while the furnace runs. I have not seen it on the smart meter yet. I'll be firing it up for a test run in the next few days.

I had installed a programmable thermostat a couple of years ago. My strategy was to go between two heat levels. One as low as I could stand for night (about 13C) and another for the daytime (about 18C). Thinking about my huge rate differential in the study leads me to a different strategy. Should I minimize during the peak times by keeping the same temperature (low) all day while creating a warmer room in the house (the work area or the sleep area) with a local portable electric heater?

With better insulation in the basement, no heat in the front room, it might even be possible for me to ride out the six hours of peak without running the heat? Like I do with the water heater, the fridge and the freezer. If that were possible, I could then have the night time temperature whatever I wanted, at least in electricity terms.

I have become quite interested in the technique of the Moody wall which I could use upstairs?

Thanks for your interest

George Plhak
Lion's Head, Ontario, Canada

INDEX to the series

Tuesday, August 16, 2016

insulated refrigerator freezer 2

INDEX to the series

Another chart from last night on the superinsulated refrigerator freezer.

I want to know, using my smart meter data, if the refrigerator's standby energy use is any different (hopefully less) with the "blanket" of insulation I added to the freezer compartment. My previous test was inconclusive because the ambient temperature was different from the one measurement I made before I added the blanket. I wish now I had taken more measurements.

The standby electrical usage of a thermal box type applicance, like a refrigerator, a freezer or even a hot water heater varies mainly with the temperature of the air around it (the ambient temperature) and the insulation.

In order for me to make a meaningful comparison, all things (controllable variables) should be kept the same. The ambient temperature is not really a controllable variable but it is from an experimental point of view.

So I am taking snapshots of my smart meter data and reading each manually. I see a graph of my usage at any time and I can pick out the appliance. I can measure it very accurately. But with the software I have now, I cannot tabulate or totalize usage by appliance over say a month or a year. So I am estimating to get an idea for ranking, highest to lowest. I am also learning a lot about how my appliances behave as electrical loads and explaining as I go (the INDEX). I can "see" my electricity use with my smart meter data, but just how useful is it?

(click any pic to enlarge) I should compare measurements at the same ambient temperature. I could make a graph of standby usage with temperature for each of these type of electrical loads. If I took a number of readings at different temperatures, I would be able to plot how the appliance usage varies with ambient temperature. Better efficiency would tilt the slope of the line. Better thermal boxes would not vary as much with ambient temperature. It is really ALL ABOUT INSULATION for thermal boxes. The more the better.

The nights here are still pretty warm. About the same as for the previous test 20C so this will qualify as a repeat for comparison. And to shake down the method. There were a few mistakes this time also.

Not as careful as I should have been, I see this morning that I left the Exterior Moisture Control ON last night. This switch activates a heater in the door gasket. I want that OFF! I had thought I left it OFF. I am going to put a piece of tape over the switch so that it stays OFF.

One thing I notice about the chart is that the fridge seems to be coming ON for shorter periods each time across the chart. I remember that my bedtime snack was a small dish of ice cream, so I had opened the freezer door about an hour before the chart. Its possible that the fridge was still recovering. I should make sure next time that the fridge has not been opened for at least several hours before the measurement.

The chart covers from about 22:30 to 01:30. The width of the peaks are 0.92cm, 0.80, 0.76 then 0.72. I am using cm as an analog for time. Time on the chart I measure to be 1.92cm = 1 hour.

The peaks also seem to drop down slightly, the first two are higher than the last two. I see that the base load drops about 25 watts at about midnight. I am not sure what turned OFF in the house. I was asleep. This is my whole house electrical usage so the fridge is riding on top of the base load.

Let's do some calculations anyway.

Here is the same chart in CorelDraw with the measurements added.

For the first two peaks, the fridge is using (275-125=) 150 watts and for the last two (250-100= 150 watts so the instantaneous power is the same as the last tests.

For the first peak, run time is 0.92 and cycle time is 1.84 so the duty cycle is (0.92/1.84=) 0.5. The second is (0.80/1.69=) 0.47. The third (0.76/1.65=) 0.46. I cannot calculate the fourth since I don't see the complete cycle. Doing the same calculation as before (power x time x duty cycle) I calculate for the three peaks 1.8, 1.69 and 1.66 kWh per day or 1088, 617 and 604 kWh per year. Quite a wide spread and more usage than before.

I know the insulation does not make the refrigerator use MORE power. Something about my measurement or the method is not right. And there was that dish of ice cream.

There must be an easier way to do this. I should let a computer monitor my energy use!

I am not discouraged.

Thanks for your interest

George Plhak
Lion's Head, Ontario, Canada

Update - The metal frame of the fridge where it touches the door gaskets is now covered with dew (condensation) so I see that the switch for Exterior Moisture Control really is activating heaters in the door frame. The dew isn't usually there. I wonder how much energy that takes and when the heaters come on? I think I remember reading that the heaters are activated by a time delay relay in the fridge that starts when the door is opened using the same switch that turns on the fridge light.

INDEX to the series

Saturday, August 13, 2016

the ontario grid - ieso and sme

INDEX to the series
I've found some interesting reading at the "Independent Electricity System Operator" or ieso.ca. This is a screen shot of their home page this morning.

I see that this morning, here in Ontario, Canada, over half of my electrical energy is coming from nuclear. None is coming from solar. A tiny sliver from wind. I can see, almost real time, the power source breakdown for the Ontario grid. I can also see the demand prediction for the day.

IESO runs the Ontario grid. They also run the smart meters. They also operate the wholesale market for electricity. The price shown is probably the wholesale rate. I'm guessing that if I buy electricity in megawatts and if I was a customer of IESO, I could get this wholesale rate? I think IESO is the "middleman" between OPG (Ontario Power Generation) and the local power companies. You can see a list of them here. In fact, IESO call them "distributors" I think these are the "customers" of IESO and I am a customer of Ontario Hydro Networks, the one I get my bill from. I don't know much about the electricity market yet but this is a good place to start.

IESO is a Crown corporation owned by the province. It has a president and Board and lots of reports and an engagement process with its "stakeholders". They also run Save on Energy programs. More study required by me.

IESO also runs the smart meter data repository through an appendage imaginatively named the "Smart Metering Entity" or SME sme-ieso.ca. This smart meter data repository also has a great name, the MDM/R (Meter Data Management and Repository). Seriously. This is where my smart meter data is kept.

Yes, SME have their own website. Their explanation of what they do:

"The Smart Metering Entity (SME) operates the Meter Data Management and Repository (MDM/R). The Government of Ontario designated the Independent Electricity System Operator (IESO) as the SME in 2007 to support its Smart Metering Initiative (SMI). The MDM/R processes smart meter consumption data from MDM/R service recipients in Ontario to support Time-of-Use (TOU) billing. The SME also provides information about Smart Metering to businesses, education leaders, and consumers."

Is that clear?

The whole smart meter system is a secure system. If I want my data, I have to be tied in to some third party company, like Plotwatt. I authorize them to access my data at the MDM/R and then I can get my data through that third party. The third party analyses my data. They do disaggregation for me. Supposedly I get to choose which service I want to use to understand my data. Maybe I can have more than one? I can, they tell me, turn off the authorization at any time.

Interestingly, even though I am not getting the disaggregated information (analyzed and organized by appliance) from Plotwatt, the third party I am "partnered" with, I can actually download my actual real meter data. How exciting!

This is the TRENDS tab in Plotwatt this morning. This chart does seem to be working for me on Plotwatt. At the green arrow, I can see the meltdown that happened last week when I had to reset my Blueline sensor to get re-attached data wise to my electrical meter. I lost about a day and a half of data but that doesn't really matter to me at this point. I am sure that the power company didn't loose my data since they get it directly through their own network. I'll get billed for that day and a half.

At the yellow arrow are two little buttons. One of them downloads the supposedly analyzed appliance disaggregated results but it is not any better than the last time I showed you. I don't seem to be getting that information. The other button (the one with the down arrow) downloads the raw data to my computer. It actually tells me that the data is in GreenButton csv format.

So this (click the pic at left to see the graphic or click this link to get the file) is my data downloaded this morning. A 2 MB file that I've put on my own server so you can click the link to download it. I have removed the id information.

I see my GreenButton data as a spreadsheet. One measurement every minute of the USAGE (in Kwh), going back about a month and ending the moment I downloaded it. 41,199 data points. I could make a graph but I get pretty nice graphs already with Blueline Energy Cloud. It's clear why I need the "interpretation" to get totals by appliance or find another way to measure the key appliances directly to get some kind of totals more easily?

I'll end with The Empowered Customer a pdf from the Blueline site, one of their News from Dec 2015 which talks about the American experience with how customer access to their data can help reduce demand by being more aware of power use in detail. There are quite a few interesting claims being made. Americans call it GREEN BUTTON CONNECT.

That's all for now. Thanks for your interest.

George Plhak
Lion's Head, Ontario, Canada

INDEX to the series

Wednesday, August 10, 2016

the old water heater

INDEX to the series
This is a picture of my utility electric use last night from my smart meter. I was preparing to measure the refrigerator again after superinsulating but the night wasn't cool enough for a comparison to the previous measurement. So I'll tell you about my water heater instead. (click any pic to enlarge)

Last night, everything was OFF but the refrigerator, the water heater and the baseline load. I turned OFF the chest freezer for the night since its signature looks so much like the refrigerator that it makes it difficult to take detailed readings from the graph. I just turned it back ON. It stays frozen solid at least six hours since I don't open the lid. I check the temperature inside periodically to make sure.

On the picture, I can see the clear signature of the water heater, that big square spike at 1am. The water heater uses 3kW which is a lot. It always uses 3kW but the time it stays ON varies. It goes ON for a shorter time if I do not use hot water. About every six hours or so for a few minutes. I hadn't used any hot water since the last time it came on so this energy is not from my use of hot water, but just to keep it hot and ready. Yes there are on-demand water heaters, mine is not.

A water heater has a very distinctive spike energy signature. If I multiply the width of the spike times the height, I can see the energy it is using very, very accurately, every minute. Mathematically, my energy use is the AREA under the curve.

The energy to keep the water hot depends primarily on my temperature set point and my hot water use. Energy also varies with the temperature around the appliance and the amount of insulation it has. I'll zoom in on the water heater graph in a moment.

The marching square wave down at the bottom is the refrigerator. It is like a backwards water heater I suppose, with the same kind of control system. It used to be called bang-bang meaning that it was either ON or OFF, nothing in between. An insulated box that gets cold instead of hot and keeps it that way using a standby amount of electricity by going ON and OFF until you use it by opening a tap or opening a door. Then it uses more energy by staying ON for longer before going OFF. Like the refrigerator and the freezer, the amount of energy can vary with the temperature around it. An air conditioner would also be a bang-bang energy load. I don't have one of those. It is also called "duty cycle" control. The English used to have a cute term "mark-space-ratio". It is also called "hysteresis control". Whatever the name, control is accomplished by ON-OFF-ON-OFF and that is how it looks in the smart meter data, like a square wave that varies in width (time ON). This type of control is very common in appliances and has been around since Edison. It is actually quite efficient.

Because these appliances make nice clean signatures it is possible for me to pick them out by eye on the graph when several of them are added together. This is what my computer should be doing for me. I am having to do it myself but learning in the process and telling you about it.

Some of my other appliances don't look at all like a simple bang-bang. The dishwasher and clothes washer have complicated signatures as the various motors, heaters and relays inside go ON and OFF. Little bang-bangs on their own but they add up to a complicated mess as far as the appliance signature. The messy signature is something I can recognize as the dishwasher but it is not so easy to calculate the area under the curve since it is so irregular.

It should also be possible to add up the amount of energy each appliance uses and keep track of it. With what I have, it would be too tedious for me to do manually on an ongoing basis. For some reason, I am not being provided with the appliance level data by the study. It's a behavioral study. They want to see what I do? Hmmm.

What I am doing is taking advantage of the summer to study my appliance energy characteristics using this new data from my smart meter that I could not see before. I expect that the summer will be easier to understand since I won't have the furnace and other winter loads messing up the measurements. I will put my appliances in a list from highest to lowest use, studying their signatures and taking some detailed measurements while noting the test conditions, like the temperature. That way when I make changes I can have some careful measurements to compare to.

All my appliances will need to be replaced at some point. All are over 20 years old. I would personally like to quantify any improvements I can make before I go blowing large wads of cash on new machines.

Overall, if I can reduce the energy use of the main appliances and keep some of them from coming ON during my peak rate period where I can, I should see a real reduction in the energy portion of my bill.

One of the first things I did was to install a timer on my water heater to cut off power to it during my peak rate period. I have since done the same for my refrigerator and chest freezer. These appliances now do not come on during my peak period. Essentially I am trying to almost leave the grid during my peak.

The Blueline hardware and software which reads my smart meter is a pretty impressive tool and I hear that is about to be improved with a browser and mobile update. What's also becoming apparent is that add-on tools (like Plotwatt) which pick out the appliance level data are value added products that I (we) should expect to pay for as a service. I haven't found one as freeware. No idea of the cost. I can access my raw data but its pretty boring. I will need some tool to disaggregate my data.

Back to that water heater. This is my old water heater. A new one is beside it and ready to be installed.

The outside temperature at the Blueline sensor was 18C at 1am last night. I can read this from another part of the Blueline software. I will take the outside temperature to be the same as the temperature in my basement. This won't work in the winter when the basement temperature will be cooler but not as cold as outside at the meter. I'll have to take a reading of the basement temperature then if I want comparable the standby energy readings. I will have the new water heater in place by then hopefully. I won't get to test this one again so I want a good reading of it's standby energy use for comparison.

I have added two dimensions to the graph. I want to know how long the water heater was ON fairly accurately. I could use a ruler but I am using a graphics program, CorelDraw. The 2.45cm dimension is the length of time that the water heater was ON. The 5.13cm length is the distance from 12.50 (ten minutes to 1am) to 1:15 or 25 minutes. So the water heater was on for 2.45/5.13x25= 11.94 minutes or 0.199 of an hour. Energy is measured in kilowatt-hours so I want hours, not minutes.

I can also see, thanks to the Blueline software, the energy use exactly at any point on the graph by moving my cursor along the graph line. The box that pops up shows me that the energy use at the instant shown by the blue dot is 3.135KW. I am going to subtract a little bit for the baseline and the refrigerator that looks like it has come on within a minute or two of the water heater so I am going to ignore that part. I will take the water heater as 3kW which is it's nameplate rating also.

So the energy used by my water heater during that spike is (3kWx0.199hour=) 0.597kWH or just over half a kilowatt hour. During my off peak, this amount of energy costs me ($0.11x0.597=) $0.065 (six and a half cents) but during my peak rate period, this same energy would cost me ($0.54x0.597=) $0.32 (thirty two cents). This is for one heating cycle, normally the water heater would come on about every six hours. There is a pretty good chance that at least two cycles will occur during my peak, two during my off peak, if I had not added the timer.

By having all the water heater limited to off peak, I figure that I am saving almost $350 per year on my study special billing rates, more than enough justification for a $20 timer. Your payback won't be anywhere near that but I'll bet that you will find savings if you look at your water heater. Some are easier to do than others but the timer should be a no-brainer unless you use a LOT of hot water, and then there may be even better ways to save.

I don't think I posed my rate structure but I talked about it, 11 cents off peak, 54 cents on peak and my peak time is different. I do not pay delivery and other charges for the duration of this study. For me, it is all about timing - stay off the peak!

You will notice of course that my peak/non-peak differential is different from yours but then I am on this study for a year. They gave me a special incentive (a very punishing peak rate) and all this data. They want to see what I will do? I feel like a rat in a maze sometimes. It truly is not easy to SEE where your electricity goes, even when you can see it in a graph.

Do look at the disaggregation article. I think that watching this market develop is going to be interesting.


I mentioned the temperature set point on the water heater. The fridge and the freezer have them as well. For the moment I am leaving the set points as they are. The subject of set points will have to be another discussion. Safety obviously is a key issue.

Thank you for your interest.

George Plhak
Lion's Head, Ontario, Canada

INDEX to the series

Wednesday, August 03, 2016

flexplan/blueline/energycloud/plotwatt update

INDEX to the series
Not a useful error message to wake up to. And there is a typo!

This is in my browser this morning. I am trying to access my electrical energy usage from last night.

I was using Firefox, so I tried Chrome. I tried my phone on a different network. No. EnergyCloud is DOWN. Perhaps to a webmaster the message might mean something? Who knows where EnergyCloud is? It's in the cloud. But the service is down. I've tried over the last hour.

Time for an update to McMaster anyway so I wrote them a letter which is the basis of this article. Time for some reflection and assessment.

FlexPlan - The McMaster University study of Hydro One customers I agreed to participate in, funded by the Ontario government. More info about this project as part of the Ontario Smart Grid Fund and the DeGroote school of McMaster..

Blueline Innovations - based in Newfoundland, maker and seller of the PowerCost monitor and EnergyCloud. The Blueline hardware was provided and installed at no charge by a McMaster/Hydro contractor, but I could have done it myself had I bought it.

EnergyCloud - a service of Blueline Innovations that allows me to access my usage information using my browser. The raw data is the actual kW used by my house recorded every minute. The graphs I have found to be interesting and useful.

PlotWatt - a software company in Durham, NC whose product energy management software supposedly learns from my usage via EnergyCloud and makes recommendations to me by email, supposedly, after it "learns" for about a month. They are tied somehow to Blueline as a "partner". I do not pay for their "service" which has been a disappointment so far.

Hello from Client ID xxxxxx

This morning I cannot access the Blueline Innovations software with my browser. I have tried both Firefox and Chrome. With both I get an error (pic attached). I kept trying for the last hour. I do enjoy looking at the graph of my usage.

While I am talking about Blueline, the mobile ap is difficult because it keeps requiring sign in and my password is not remembered. This may be caused by my phone. When I log in, the information the app contains is really not useful. Anyway, it says that EnergyCloud has stopped when I do the login.

Plotwatt does shows my usage and trends right now but STILL shows NOTHING about my appliance usage. It's been over three months now and you said it was "learning"? The only thing I can see is the downloadable .csv file attached. This is not useful information?

I find the advertising on the Plotwatt site misleading. This is not what I have. And to a certain extent on the Blueline website. I have no appliance specific information?

Thanks to your program, I now know much more about my electricity usage. But not due to the software. The real time info graph is nice to have but I have to figure it out from there.

I hear nothing from McMaster except your direct relies to my questions. How is the program going? I see that the funding was over $1.3M.

George Plhak



This is the appliance spreadsheet I downloaded this morning from PlotWatt. Down the left side are the days of the last month and across the top are some categories that Plotwatt decided, not me. Notice the "learning" column.

This "data" does not mean much to me as relates to what I think my usage is. I certainly have other major appliances, like a water heater, that are not shown. I told the software specifically what appliances I had in the setup. It should know.

When I look at the PlotWatt website, it seems that their main focus is restaurant energy management although they do make the pitch at a utility level that having customers know more about their usage is a good idea and the idea is compelling.

So far, for me at least, the idea is not delivered on.

I'll let you know what I hear from McMaster.

Update Aug 4 - I am able now to log into EnergyCloud but there is no data. The main graph is blank.
Update Aug 5 - Excellent tech support from John at Blueline. I now have data again. They had a melt down in the Cloud or something. My Blueline sensor and bridge needed to be reset and John helped me do that. Short discussion with John. I am not supposed to get the detailed appliance data as part of the study so he says that I will not see it, even after three month of "learning". Go figure.

Thanks for your interest.
George Plhak
Lion's Head, Ontario, Canada

INDEX to the series