72% of our usage is grid supplied, even though we produced more than we used!

In the last year (April 1, 2012 to March 31, 2013), we used 6,046 kWh of electricity and produced 8,576 kWh for a net surplus of 2,530 kWh. That’s awesome right?

But 72% of the electricity we used was supplied by the grid when solar could not supply enough power to cover the need at that moment, like at night or on a cloudy day*.

Or another way to look at it, although we produce more than we use, most of what we produce we don’t really use directly. We only use 28% of what we produce. The rest goes back to the grid to pay back what we used when the sun wasn’t shining and to build up a surplus for a rainy day.

So even with all that sun, we still draw a lot of power from the grid that requires coal and other nasties to be burnt to serve our electricity needs.

It makes sense, most of our heavy use, hot water for showers, cooking, washing dishes, all occurs early in the morning or evening when the sun is not at it’s brightest or best angle. The more we time our usage to occur when the sun is shining, the less we demand of the grid.

There has to be a common industry term for this? Anyone know? Percentage of power supplied by the grid as compared to total usage when solar or other renewable is in the mix?

To me this seems like a much more important number to track if you have solar and are concerned with your direct carbon producing footprint.

* In order to find how much energy we used from the grid I added up all the usage values on an hourly basis that were greater than what was being produced by solar. For example from 5-6am, total demand was 1000Wh. The sun was just coming up and the system was only producing 200Wh. That means the grid was supplied 800Wh during that time. 80% is grid supplied for this hour. Now do that for every hour in a year. Hint, it helps if all your energy values are stored in a database.

Estimating heat energy for 2012 – Revised

Now that we have January – March 2013 circuit-level usage values, I thought I’d go back and revisit my original 2012 heat estimate using a different method.

I had estimated January – March 2012 heat energy based on a linear regression analysis of our April – December heat values. There are a number of problems with this approach. Mainly that heat pumps use more energy the colder it gets outside, and secondly the amount of passive heat we gain from the sun can significantly reduce the amount of energy required for heat.

This time I used a less formulaic approach to estimate heat energy usage. I simply calculated the kWh/HDD per month for 2012 and 2013, and compared the values.

First lets look at the first 3 months of 2013. We recorded 3,239 HDD, a 20% increase from 2012 to 2013. We used 746 kWh for those 3 months. If we divide 746 kWh by 3,239 HDD we get 0.230 kWh per HDD.

Now let’s try the same for the first 3 months of 2012. We recorded 2,107 HDD and I estimated 327 kWh for heat energy. 327 / 2,107 = 0.121 kWh per HDD. That is a 128% difference from 2013. Something is clearly off.

Since 2013 was colder and less sunny, I would expect our 2013 efficiency to be less because heat pumps become less efficient at lower temperatures. So I manually adjusted the 2012 kWh values so that the kWh/HDD percentage was similar to the 2013 values, then I lowered it a bit to take into account the warmer temperatures and increase sun in 2012. Did I mention this wasn’t very scientific?

What we get is closer to 620 kWh for heat energy for the first 3 months of 2012. This is roughly a 90% increase from my earlier estimate. It also means that a 20% increase in colder weather roughly equals 20% more heat energy usage.

Looking at Q1 performance again, that means out of the 445 kWh increase in 2013, 65% of that increase was due to heat energy, 33% was water heating and everything else was 2%. That sounds a little more realistic.

First quarter 2013 performance

Q1 2013 summary: 20% more cold, 22% more usage and 15% less sun.

Now that we are starting to collect our second year of performance data, I’ll be comparing our new data to the same period last year. Year-over-year comparisons should be more illustrative than comparing to previous months. I’m also switching to a quarterly reporting period to look for larger trends, but I will continue to post data monthly at netplusdesign.com

**See April 7 Update at end of post.

In Q1 last year our net usage was 228 kWh. This year is was 953 kWh, a 300% increase! In March 2012 we actually generated a surplus, but not this year.

Overall, here’s how the first quarter compares to 2012.

2013	Usage	Solar PV	Net usage or (surplus)	Avg. daily usage	HDD
Total	2,499	1,546	953	28.1	3,238
Jan	881	478	403	28.4	1,189
Feb	812	449	362	29.0	1,067
Mar	806	618	188	26.0	983

2012	Usage	Solar PV	Net usage or (surplus)	Avg. daily usage	HDD
Total	2,054	1,824	228	22.6	2,701
Jan	873	369	504	28.2	1,125
Feb	666	597	69	23.0	957
Mar	515	860	(345)	16.6	619

All values in kWh (except HDD which is base 65°F).

Seeing all the increases, I wanted to understand if our 22% increased usage (445 kWh) was due mainly to the 20% increase in colder weather (HDD), or if other factors were leading to our increased electricity usage.

Broadly speaking, hot water, heat and all other circuits (mainly plugs loads and lighting) each accounted for about 30% of the total energy for the 1st quarter. Out of the three, hot water is the only one we can accurately compare because we still lack circuit-level data for Q1 2012.

For the last three months we’ve averaged 74.8 gallons of water per day, up 46% from first quarter 2012. That’s an extra 600 to 800 gallons of water per month, or nearly 24 extra gallons per day. This includes a 25% increase in hot water usage. Although we don’t have circuit-level data on the hot water heater, we do have monthly hot water usage data. We can use this to estimate the energy required to make hot water in Q1 2012.

Our hot water heater used an average of 275.8 watts to heat one gallon of water in Q1 2013 (water heaters have to work harder in the winter). We used an extra 533 gallons of hot water in Q1 2013. 533 * 275.8 = 147 kWh. That’s 33% of the overall usage increase right there.

As for the cause of the increase? To remain happily married I will only say two words, goats and mother-in-law. I don’t think I should say anything more on that topic.

Now let’s look at heating.

Determining the increase in heat energy is not easy to answer because we only have estimated usage values for heat in Q1 2012. Estimated values for heating have a much higher degree of uncertainty. Using this quarter’s values to check the accuracy of my 2012 estimate shows that my estimate may have been off quite a bit. It’s hard to know how much it may be off because I don’t know an easy way to factor in heat gain directly from the sun shining through the windows. Below are the heat pump electric usage values for Q1 2012 and 2013.

Heat (kWh)	2012*	2013	change
Total	327	746	128%
Jan	149	282	89%
Feb	119	270	127%
Mar	59	194	229%

* 2012 values estimated (see Estimating heat energy for 2012).

All values in kWh.

Based on the values in this table, we used an additional 422 kWh or a 128% increase in energy use for heating in 2013. This would almost completely use up the 445 kWh overall usage increase! So I know this can’t be correct. For example, if I use the HDD formula, 0.2261 x 1900 HDD(base 50) + 0.756, I get 430 kWh, which is far below the actual recorded 746 kWh. Which means my estimated heat values for 2012 are off. But we know that heating has increased some amount due to the colder temperatures and cloudier skies. We just can’t determine how much it increased.

So let’s look at all other circuits. Again we can’t really compare apples-to-apples yet, but there are a number of changes that have occurred this year. As I mentioned, we now have a barn full of goats, chickens and other critters. On the coldest days we used heated water buckets to keep the water from freezing. In March we had an egg incubator running for 23 days and and we’re using grow lamps to start our garden seedlings. As of March 30 we’re also now using a 250W heat lamp in the chic brooder. This will last for a few weeks before we switch to a smaller wattage, pushing our energy usage up 4 to 5 times it’s usual base load for about 5-6 weeks. Soon the incubator will be in use again. We’re going to try hatching and raising some turkeys this year.

In summary, Q1 usage is up compared to last year. Some of the reasons for this are fairly obvious, others remain difficult to tease out of the data. Hopefully next quarter will offer better comparisons and we’ll finally be able to compare apples-to-apples with circuit-level data. It should be easier to see where usage has changed and why. If we understand why, then we have a better chance of improving performance over time.

April 7 Update: Using a different method I estimated  that heat energy most likely did increase 20%. That means out of the 445 kWh increase in 2013, 65% of that increase was due to heat energy, 33% was water heating and everything else was 2%. That sounds a little more realistic.

You can see heat maps and detailed charts of temperature and electrical usage at netplusdesign.com. View hourly, daily and monthly values for solar, usage, net usage, temperatures and HDD.

335 days and 6 slices of data

The image represents 11 months of daily data from February through December 2012. There are 6 vertical slices in order from left to right; Generation, Usage, Air Source Heat Pump, Dishwasher, All other and Heating degree days. Gray boxes indicate no data for that day.

Estimating heat energy for 2012

April 7 Update: Using a different method I estimated  that we used 903 kWh for heat energy in 2012, that’s 16% of our total energy use. That would have cost us about $117 (using $0.13 per kWh).

We don’t have a full 12 months of data for our heat, but using heating degree days (HDD) and performance thus far we can estimate that we used 591 kWh (+/-20%) for heat in 2012. That’s 11% of our total energy use for the year.

Here’s the math…

Our estimate relies on heating degree day measurements. This discussion assumes you are comfortable with HDD. If not, we recommend this excellent article.

First we had to determine the optimal base temperature for our HDD. After some trial and error using a quick prototype, we determined that 50 degrees (F) most accurately predicted heat energy usage. This results in the following formula to calculate kWh.

kWh = 0.2261 * HDD + 0.756

To start, we know the ASHP used 283 kWh during the time period March 16-May, and September-December. We just need to estimate January through mid-March.

We started recording temperatures in February so let’s start there. There were 522 HDD (base 50F) in February.

119 kWh = 0.2261 * 522 HDD February + 0.756

For comparison, there were 518 HDD (base 50F) in December. We used 108 kWh in December for heat. That’s about 208 Wh/HDD. We could assume that we would have used about the same amount of energy in February, although we generated twice the amount of electricity in February, which means we would have used a lot less energy for heat because we were getting heat from the sun. But let’s use the more conservative estimate for now, 119 kWh.

We don’t have our own temperature data for January but we can use degreedays.net to find the HDD at any base temperature for January in our area. Albany International Airport has the closest matching temperatures. Using a base temperature of 50F degreedays tells us there were 654 HDD (base 50).

149 kWh = 0.2261 * 654 HDD January + 0.756

For comparison we could multiply the 208 Wh/HDD measure from December above times 654 HDD to get 138 kWh. This number is lower but since the temperatures were cooler in January than December we know the ASHP would have to work harder to make heat, so the higher number makes sense. January was about 23% cooler than December. If we add 23% to 138 kWh, we get 170 kWh. Not very scientific but if gives a sense of margin of error.

As for March, we only have the second half of the month’s data for the ASHP, which used 19 kWh. There were 90 HDD from March 16 through the end of the month. That’s 211 Wh/HDD. As a check we can plug 90 HDD into the formula to get the estimated kWh.

21 kWh = 0.2261 * 90 HDD + 0.756

It’s pretty close.

There were 174 HDD from March 1 to March 15.

40 kWh = 0.2261 * 174 HDD + 0.756

Again for comparison, 211 Wh/HDD (from above) * 174 HDD = 37 kWh. Pretty close, it was twice as cold and about twice the amount of energy use.

40 (3/1-3/15) + 19 (3/16-31) = 59 kWh total for March.

Now we can estimate our total heat energy for 2012.

283 (3/16-12/31) + 149 (Jan) + 119 (Feb) + 40 (for the missing part of March) = 591 kWh for 2012. That works out to $77 for heat, not counting delivery charges. 591 kWh represents 11% of our total energy use for the year. That seems about right since the ASHP represents 9% of our total energy use for the time period we have circuit level data.

There is at least a 20% margin of error for the estimate and that’s not even taking into account space heat energy contributed by the sun.

Polynomial curve fit analysis

I’ve also been looking at a polynomial fitted curve to better estimate kWh based on HDD. This makes some sense because heat pumps are more efficient at higher temperatures and less efficient at lower temperatures. A linear regression analysis would not be able to capture that type of operating behavior.

* Note: Electricity supply cost calculated using the last bill we payed for electricity, May 2011.

We’re net positive for 2012!

Year Summary

We’re net positive for 2012! We used 5,601 kWh and generated 8,856 kWh for a net surplus of 3,256 kWh. We used an average of 15.3 kWh per day.

We’re within 6% of our projected energy use of 5,950 kWh for the year. It also looks like we used about 80% of the estimated 2,440 kWh for hot water energy. It’s too early to tell how we’re doing against our heat energy estimate since we missed recording the primary heating season. We should know better by mid-March.

But we can calculate the current BTU/SF/HDD, which is a common measure to compare the efficiency of houses of different sizes and climates. First we need to know the total heating degree days (HDD) and heat energy usage for the period we have data. We recorded 5,885 HDD (base 65F) for 2012. 3,434.800 HDD were recorded when we had circuit level data for the ASHP, which used 277.331 kWh for heating in that period.

Now we convert kWh to BTUs,  277.331 kWh * 3412.14163 (kWh to BTU conversion factor) then divide by 1408 SF and divide again by 3,434.800 HDD (March 16-Dec 31, 2012) = 0.196. This is not a completely accurate number as it does not include Jan 1 through March 15, but soon we will have a full year of circuit level data and a more representative number.

We used an average of 56 gallons of water per day. 34% of our total water usage was hot water.  For the time period we have circuit level data that works out to 258 watts per gallon of hot water. I’ve found some estimates on the web that say we should be using about half that wattage to heat our water, so we’re going to do a bit more investigation.

December Summary

In December, we used 668 kWh and generated 302 kWh for a net deficit of 366 kWh. We had nine consecutive surplus months. This is the first month since February that we used more than we generated.

We used 152 kWh for heat, a 227% increase over November. Dividing 152 kWh by 979 HDD we get 155 watts per HDD. That’s a 182% increase from November. Not only are we using the heat more, but the heat pump has to work harder because it’s colder outside. Heating accounts for 22% of the month’s total usage. Hot water accounted for 32% of the energy for the month. We used 65.9 gallons of water per day, up 7% from the previous month.

The largest energy increase in the last few months has been in the All Other category, followed by the water heater, ASHP and the stove. Most of this makes sense. We’re spending a lot more time indoors. The holidays bring visitors, more showers, more laundry and lots and lots of cooking!

We generated 302 kWh, roughly half of November’s production. It was down 24% from the estimated December value of 399 kWh. A very cloudy and snowy month. In fact we only generated 116 watts the last 5 days of the year due to snow blanketing the collectors.

	Usage	Solar PV	Net usage or (surplus)	Avg. daily usage	HDD3,4
Total	5,601	8,856	(3,256)	15.3	5,885
Jan1	873	369	504	28.2	1,1255
Feb2	666	597	69	23.0	957
Mar6	515	860	(345)	16.6	619
Apr7	925	378	(538)	12.9	535
May	309	867	(558)	10.0	170
Jun	325	939	(614)	10.8	97
Jul	319	970	(651)	10.3	17
Aug	315	1,018	(703)	10.2	31
Sep	292	883	(591)	9.7	168
Oct	433	511	(78)	14.0	357
Nov	498	615	(117)	16.5	830
Dec	668	302	366	21.6	979

All values in kWh (except HDD).
¹ January values based on meter reads.
² February values based on TED data.
³ Heating Degree Days, a measure of how many outside degrees in a day it is below a base target inside temperature, 65F.
⁴ Calculated from our HOBO outdoor weather monitor hourly data, unless otherwise noted.
⁵ January HDD data downloaded from degreedays.net, Station ID: KALB (Albany International Airport).
⁶ March values based on meter reads. (TED died March 1st, eMonitor installed March 16, 2012)
⁷ Values starting in April are based on eMonitor data.

Next month we’ll start including information gathered from our two additional temperature/humidity sensors. We will now be able to see how temperatures differ from floor to floor.

You can see heat maps and detailed charts of temperature and electrical usage at netplusdesign.com. View solar, usage, net usage, temperatures and HDD for all of February and circuit-level data for 16 days in March and the full months of April through December.

Happy 2013 everyone!

Home Performance App Upgrade

I’ve updated our home performance app at, NetPlusDesign.com/home_performance.

Screenshot from new app.

At the beginning of the year I put together a simple online tool that displays our energy and temperature data on a daily and hourly basis. It gave me an easier way to track and visualize our energy usage on a month to month basis without a bunch of spreadsheet work. The heat maps provide an easy way to quickly find days of high energy use and see what else was going on that day. This helps me understand why usage might be high on a particular day.

Now that we’re coming up on twelve month’s data, I wanted an easier way to see across months and drill down from the macro to the micro scale. So I built some new views that would allow me to slice by time period (year, month) and dice by aspect (energy used, generated, heating degree days or a summary across all of them).

It’s a bit slow, I’m just a hobby developer, but give it a chance and let me know what you think. I’ve released all the code on Github under the MIT “Do what ever you want with my code, just don’t expect it to be bug free” License.

In a few days I’ll upload December’s data. Happy Holidays.

November performance

We used 498 kWh and generated 615 kWh for a net surplus of 117 kWh in November. This is our ninth consecutive surplus month. We’ve generated a surplus of 3,622 kWh so far this year, guaranteeing we’ll be net positive for our first year of operation.

The month proved to be sunnier than average. We generated about 63% more than predicted, despite the fact that it was about 12% cooler than average for the month. Usage was up 19% over the previous month. Lot’s of guests, cooking and laundry for the Thanksgiving weekend period.

We used 46 kWH for heat. That’s about 55 watts per HDD. That accounts for 9% of the month’s total usage. Compare that to hot water which used 38% of the energy for the month. We used 61.7 gallons of water per day, down about 9% from the previous month.

I’ve been updating some of the data reporting measures. This month I’ve moved to reporting all heating degree days at a base temperature of 65 degrees rather than 68. This should make it easier to compare performance to other homes that use 65 as a base temperature.

Month	Solar PV	Usage	Net usage or (surplus)	Avg. daily usage	HDD3,4
Jan1	369	873	504	28.2	1,1255
Feb2	597	666	69	23.0	957
Mar6	860	515	(345)	16.6	619
Apr7	925	378	(538)	12.9	535
May	867	309	(558)	10.0	170
Jun	939	325	(614)	10.8	97
Jul	970	319	(651)	10.3	17
Aug	1,018	315	(703)	10.2	31
Sep	883	292	(591)	9.7	168
Oct	511	433	(78)	14.0	357
Nov	615	498	(117)	16.5	830

All values in kWh (except HDD).
¹ January values based on meter reads.
² February values based on TED data.
³ Heating Degree Days, a measure of how many outside degrees in a day it is below a base target inside temperature, 65F.
⁴ Calculated from our HOBO outdoor weather monitor hourly data, unless otherwise noted.
⁵ January HDD data downloaded from degreedays.net, Station ID: KALB (Albany International Airport).
⁶ March values based on meter reads. (TED died March 1st, eMonitor installed March 16, 2012)
⁷ Values starting in April are based on eMonitor data.

You can see heat maps and detailed charts of temperature and electrical usage at netplusdesign.com. View solar, usage, net usage, temperatures and HDD for all of February and circuit-level data for 16 days in March and the full months of April through November.

Temperature and energy use observations

As I continue sifting through our performance data, I’ve been focusing lately on outdoor temperatures and their relationship with our heating and cooling energy use. These are a few of my observations.

We have 9 months, 274 days, 6,576 hours of temperature data, from February 1 to October 31. Of those 6,576 hours, 4,635 outdoor temperature hours were below 68 degrees F.  Roughly 70% of the recorded time period has been below 68 degrees F.

Breaking out the temperatures by 10 degree buckets you can see a rough bell curve. The majority of hours were in the 60-70 degree F range. The next three months of data will add more hours to the 60 degree and below buckets.

If we overlay air-source heat pump values on the same chart (below), we can see the power used for heating and cooling based on the temperature ranges. The  dip in the 50-60 degree range is likely to indicate our base building temperature.

The dip is where we shift from heating to cooling or dehumidification in our case. I really should plot the right side with interior humidity ranges rather than temperature ranges. I’m guessing the hottest days were more dry and less likely for us to use the AC.

We used the majority of our heat energy for temperatures below 54F degrees. This makes sense, but when I looked at the hours the heat was on in the two temperature ranges I saw something I didn’t expect.

The heat was only ON for for a small proportion of the hours that temperatures were below 54F degrees. It was off more than it was on. After I mulled this over it made sense too. A lot of spring, summer and fall nights can get quite cool, but the insulation and mass of the house keeps it warm through the night. Plus, we generally keep the heat turned off in the summer for obvious reasons.

But it made me curious what types of temperature swings we were experiencing monthly.

There is quite a range of temperatures throughout the months, March being the extreme due to a freakishly warm spell. The dotted line is 68F. This shows we’re getting some cool nights all through the summer.

That’s it for this week.

Notes

Number of kWh used for heating below 54F = 77.81 kWh

Number of kWh used for heating between 54F and 68F = 14.131 kWh

Number of hours heating was ON when temperatures were below 54F = 195 hrs

Number of hours heating was ON when temperatures were between 54F and 68F = 80 hrs

Number of hours heating was OFF when temperatures were below 54F = 1,351 hrs

Number of hours heating was OFF when temperatures were between 54F and 68F = 1,958 hrs

Heat was ON 6% of the total hours that temps were below 68F. In that time we used 91.941 kWh for heat, an average of 334 Wh/hr.

Heat was ON 14% of the total hours that temps were below 54F. In that time we used 77.81 kWh for heat. 82% of total heat usage, an average of 399 Wh/hr.

Heat was ON 4% of the total hours that temps were between 54F and 68F. In that time we used 14.131 kWh for heat. 18% of total heat usage, an average of 177 Wh/hr.

Temps were above 68F = 1,941 hrs

Temps were below 68F = 4,635 hrs

Temps were between 54F and 68F = 80 hrs

Hours temps were below 54F = 1,351 hrs

Correlating HDD with heating energy use

I spent a few of the Sandy storm days indoors and decided to indulge my inner nerd. I’ve been tracking all our house performance data using monthly spreadsheets, which can make it difficult to filter data or look for trends across months. So I downloaded mySQL and imported all our energy and temperature data into the database. This is the first in a series of posts on what I find as I sift through the data.

A while back I stumbled upon this article, Linear Regression Analysis of Energy Consumption Data. At the time I was learning about how to calculate HDD days based on our own temperature data captured by our outdoor HOBO temperature and humidity sensor. Now that our data is much easier to query I decided to give the linear regression analysis a try.

The main idea is that if we plot energy used to heat our house on one axis and HDD on the other (and we’re extremely lucky) we should be able to see a fairly consistent pattern. Meaning we could use the results of the analysis to predict how much energy will be required in the future if we know the probability of HDD for a future time period.

The R² value of 0.1134 in the chart at the top of the page tells us that we don’t have a strong correlation between energy use and HDD. The closer the R² value gets to 1, the stronger the correlation. So what could be the reasons that we don’t have a stronger correlation?

One reason may be that an air source heat pump uses different amounts of energy to make heat depending on the outside air temperature. It has to work harder at lower temperatures to generate the same amount of heat.

It may also have something to do with how the HDD values are selected. I started with daily snapshots, but there could be 23 HDD in a 24 hour day, and the ASHP only operated for a few of those hours, either because we turned it off at night or we experienced a day with a lot of sun. So I switched to hourly snapshots and discarded any hours where the ASHP was not drawing power (the result is the chart at top of page). This however, did not result in a closer correlation.

My guess is that we don’t use the heat in a consistent manner so we’d have a difficult time getting a strong correlation. Plus, the way the heat pump itself works may also make it difficult to correlate.