Overall Performance

Lifetime Generation: 25070 kWh (as of 7/27/2014)

Actual Monthly Statements           2008  Energy Generated (kWh) Energy Consumed (kWh) Generated Consumed (%) Jan-Feb --- --- -- Feb-Mar --- --- -- Mar-Apr --- --- -- Apr-May --- --- -- May-Jun --- --- -- Jun-Jul 438 472 93 Jul-Aug 392 529 74 Aug-Sep 419 467 90 Sep-Oct 317 371 85 Oct-Nov 308 310 99 Nov-Dec 178 369 48 Dec-Jan 128 466 27 Summary 2180 2984 73

[+] 2 0 0 9 [–]  2009  E Gen (kWh) E Con (kWh) Gen Con    (%)   308 369 83 322 392 82 394 361 109 460 360 128 409 313 131 480 543 88 486 265 183 345 351 98 301 353 85 232 305 76 166 340 49 140 431 32 4043 4383 92

[+] 2 0 1 0 [–]  2010 E Gen (kWh) E Con (kWh) Gen Con    (%)   231 340 68 272 341 80 445 299 149 437 326 134 460 401 115 437 606 72 407 502 81 471 553 85 391 339 115 336 447 75 206 906 23 179 936 19 4272 5996 71

[+] 2 0 1 1 [–]  2011 E Gen (kWh) E Con (kWh) Gen Con    (%)   190 857 22 230 691 33 376 616 61 362 456 79 404 462 87 401 498 81 443 505 88 493 476 104 358 332 108 281 524 54 224 706 32 205 653 31 3967 6776 59

[+] 2 0 1 2 [–]  2012 E Gen (kWh) E Con (kWh) Gen Con    (%)   208 419 50 304 368 83 450 438 103 356 375 95 580 376 154 485 520 93 410 489 84 463 426 109 403 352 114 258 415 62 243 482 50 205 542 38 4365 5202 84

[+] 2 0 1 3 [–]  2013 E Gen (kWh) E Con (kWh) Gen Con    (%)   152 398 38 274 406 67 410 439 93 401 465 86 362 367 99 423 528 80 437 388 113 451 521 87 395 350 113 236 593 40 217 570 38 93 197 19 3851 5522 70

[+] 2 0 1 4 [–]  2014 E Gen (kWh) E Con (kWh) Gen Con    (%)   176 458 38 421 467 90 361 506 71 398 463 86 468 402 116 392 452 87 --- --- -- --- --- -- --- --- -- --- --- -- --- --- -- --- --- -- ---- ---- --

• This is a parallel-metered system, so my utility company records the actual amount of energy generated in each monthly statement period, regardless of how much energy I consume. (My meters are read in the middle of the month, so each statement period covers some of one month and some of the next month.)
• The Energy Generated and Energy Consumed values listed in the table are as reported to me by the utility company in my monthly statements.
• Generaged / Consumed is a simple ratio, indicating what percentage of my energy usage was generated by the solar panels.
• The lifetime value is as shown on the meter, which records all energy generated since the system went online on 6/09/08.
• We installed a geothermal heating/cooling system on 7/16/10. This is a more efficient way to heat and cool a house; although the summertime months will show lower consumption, the wintertime months will show higher electrical consumption, but this is offset by the lower natural gas consumption. The net effect of this energy trade-off is tracked on the geothermal site's Performance page.

Typical Performance

Some days I read the two meters at the same time each morning to gather some data regarding a typical day. This is useful for learning the effects of things such as daily weather variation.

• The air conditioner was a huge energy hog in the summer, so we replaced it with a geothermal system in late July 2010. This reduces our electricity usage for summer cooling, and it reduces our gas usage for winter heating, but it increases our electricity usage for winter heating. We're still adjusting the system for optimal winter performance, but the winter 2010-2011 numbers are much higher for this reason.
• In the autumn the amount of sunlight we're receiving on a sunny day is still pretty high, but there are much fewer sunny days than in the summer, so average generation is noticeably lower.
• In the winter we still generate a decent amount of energy on a sunny day, but there are a lot fewer sunny days. Also, we generate little or nothing for a day or two after each snowfall until the panels melt clear.
• The amount of sunlight received in spring should be about the same as we receive in autumn, but we generated much more electricity in Spring 09 than Autumn 08, because the inverter calibration problem was fixed in November 08.
• The average energy consumed in Summer 09 was lower than Summer 08, because we were out of town for several weeks, so we turned off our air conditioner during some of the hottest days.
• Summer 10 was hotter and rainier than normal, so generation was low and consumption was high.

Daily Performance

Other than weather variations, daily performance is affected by shading. When the sun rises/sets far enough to the north, the dormer in the center of the roof shades the west side in the early morning and the east side in the late afternoon. When the sun is in the west, the mature maple tree in the backyard shades the whole roof. When the sun rises/sets farther south, the dormer and maple tree aren't an issue, but the pine tree in the neighbor's backyard provides some afternoon shade.

The LCD display on the inverter shows a graph of the power generated over the course of the day as a function of time. The bottom of the scale is marked in hours from 4 am to 10 pm, the left side of the scale is marked in kW from 0 to 3.6 kW, and the area under the curve is the total energy generated in kWh. If the roof received no shade, this curve would be a nice bell shape that is symmetrical about 1 pm (Daylight time) or noon (Standard time).

For comparisons of similar installations near my house see:

• The Urban Ecology Center has no shade but is tilted only about 20 degrees above horizontal (a little less than my roof).
• Curt's house has no shade from trees, but has a very similar dormer in the middle of the roof. His tilt angle is about the same as mine, and he has the same make/model of panels and inverter.

Every few weeks we have a beautifully sunny day without any clouds, so I take a picture of the LCD readout to see how the graph changes over time: (click on any of these thumbnails for a larger image)

Summer 2008

You can see two irregularities in these graphs: (1) the tops are flat (limited to about 2.4 kW), instead of having the typical peak of a bell curve, and (2) there are notches in the sides. The first problem has been investigated by Tron Melzl (a field engineer for Magnetek, the inverter manufacturer), and a potential solution is being tested over the next few months. The second problem is the expected effect of early morning and late afternoon shading. Click here for detailed analyses: July Analysis August Analysis September Analysis.

Autumn 2008

You can see the inverter calibration in early November appears to have corrected the flat-top problem; although I should wait until the more consistently sunny days of March to be sure. You can also see changes in the shading as autumn progresses: (1) the sun rises/sets farther to the south, so the dormer isn't making morning/afternoon notches, and (2) the sun is lower at noon, so the neighbor's tree (southwest of my house) is making a noticeable notch between noon and 2pm. Click here for detailed analyses: September Analysis October Analysis November Analysis.

Winter 2008/2009

At the start of winter (sun's lowest point), the early-afternoon-tree-shading notch is more significant, and a new morning notch has appeared: the sun now rises far enough to the south that the neighbor's roof shades the bottom row of panels for a couple hours early each morning. (It's not a problem in the late afternoon, because his house is a little southeast of mine.) However, late afternoon shading has appeared from pine trees near his garage, because the sun is setting so far to the south. Later in the winter (as the sun returns northward), these notches become less noticeable. Click here for a detailed analysis: December Analysis.