Effectiveness of Virtual Tracking?

cdre
cdre Solar Expert Posts: 78 ✭✭
I've had a tough time trying to prove virtual tracking is effective.... I'm working through my off-grid system design/philosophy and realized that virtual tracking may be useful for me given that 2/3 of my loads occur when the sun is up (home office with IT and A/C).

Did a ton of research and ran across many articles which referenced west-facing panels being helpful for grid-tied systems since homeowners will have more of an opportunity to use what they generate in the afternoon/evening than they do on mid-day oriented panels. Other than that, I ran across this forum entry that has been referenced here on the forum before. Beautifully flat curve. Did some experimenting in PVwatts and I cant get anything nearly as consistent and flat as that.

I ran a few scenarios based on my location in Belize. All data below is for 5kW systems with the cited orientations (azimuth and tilt). I parsed out the data for each day - the spreadsheet reflects the total monthly average for each 1 hr window with sun (0500 - 1900). Ran my experiment on January as it is my weakest month. The data below focuses on maximizing early morning generation (expectation that my daytime loads will be coupled with bulk battery charging early).

I ran a number of scenarios but can't find any that generate significantly more electricity early or late enough to really be worth it. Trying to understand why... Is atmospheric interference at lower elevations (early in the morning/late in the evening) more significant in driving the peak to the center of the day than the angle at which the sun's energy hits the panels?

Has anyone else had any quantifiable results with virtual tracking?

Attachment not found.

Comments

  • cdre
    cdre Solar Expert Posts: 78 ✭✭
    Re: Effectiveness of Virtual Tracking?

    I should state that none of the curves above reflect a combination of 1/2 E 1/2 W vs a full South array as gave the flat curve in the referenced post. One could combine an extreme west and extreme east curve to make a relatively flat curve, but based on what I'm seeing from PVwatts, the left and right ends of the curves wouldn't be significantly different from that of the south facing curve. The flatness would just be created by chopping off the top 25% of your mid-day charging.
  • cdre
    cdre Solar Expert Posts: 78 ✭✭
    Re: Effectiveness of Virtual Tracking?

    I parsed the data for June. Longest days of the year, so figured there would be more of an impact. Looks like reasonable results at 70 degrees off axis (110 and 250) and a higher angle of 40 degrees. Shift the power curve about an hour to the left and right without significantly compromising total power production. Being that daytime (and nighttime) A/C loads will be higher in the summer. There is actually a little bit of an advantage there.

    Still won't generate the amazingly flat curves seen on the previously referenced post, but anything I can do to shift loads from battery directly to PV would be great.

    Attachment not found.
  • cdre
    cdre Solar Expert Posts: 78 ✭✭
    Re: Effectiveness of Virtual Tracking?

    One more. Here's models of a 10kW array. Curves for 10kW south then splitting between 110 and 250 with a 40 degree tilt. All in all looks pretty counter productive... At least on those alternative angles.
    Attachment not found.
  • cdre
    cdre Solar Expert Posts: 78 ✭✭
    Re: Effectiveness of Virtual Tracking?

    And another. This one includes 10kW pointing E (not split between E & W) vs 10k S. Then I threw more PV at the problem with 10kW E and 5kW W.
    Attachment not found.
  • pleppik
    pleppik Solar Expert Posts: 62 ✭✭✭✭
    Re: Effectiveness of Virtual Tracking?

    I read the first page in that thread you linked to, and one thing I noticed is that the OP is in Australia and reports his latitude as about 31 degrees south, whereas Belize is about 17 degrees north. That's a lot closer to the equator. I don't know if that makes a difference or not in how flat you can get the power curve, but it might.

    Another thing I noticed is that the OP is running two arrays with radically different orientations into a single MPPT. This sort of configuration means that most of the time you're not harvesting the maximum power from both arrays. If this is what's causing the flat curve, then you'll never see it in simulated data because the simulation tools assume you've got efficient MPPT. It may be that the OP happened to hit upon some magic combination of array orientations into the single MPPT which happens to give a really flat output curve.

    For what it's worth, I have two arrays oriented SE and SW (because of the orientation of my home, not any desire to flatten the power curve) at 45 degrees latitude, and the maximum output of the arrays is about two hours apart. The overall generation curve is not particularly flat (take a look here, but you'll have to go back to October for good data since we've had a terrible run of cloudy weather here), or even that much flatter than a single array facing south would have been. I have panels with integrated MPPT, so each panel is always generating the optimal power.
  • cdre
    cdre Solar Expert Posts: 78 ✭✭
    Re: Effectiveness of Virtual Tracking?

    Thanks for your reply. I was messing with it a little bit more today. Did some models where I pointed the panels directly at the azimuth angle with a tilt of 40 degrees and with 60 degrees. Still couldn't really get a reasonable solution that didn't significantly compromise total produced power. Looks like pointing south is ultimately the best solution for me.
  • MarkP
    MarkP Registered Users Posts: 51 ✭✭
    I most often hear virtual tracking referenced with respect to grid tied systems, tailoring generation to match late afternoon consumption either so you don't have to sell power cheaply at noon and buy it back expensive in the afternoon, or just because the local utility will pay more for it then since it is their peak too.  In my case i am off-grid.  I end up throwing a lot of potential power away because I have nowhere to store it.  I do have batteries but more batteries would be expensive.  Instead I want to generate power earlier in the morning and later into the evening, reducing load and wear and tear on the batteries.  Also I get lots of clouds here.  I would rather spend any additional money on more panels rather than more batteries or fancy trackers.

    Being off grid it is wise to have some spares, which I have not had until now.  I have been running with all 15 panels pointed SSE, 9 fairly flat and 6 at about 45 degrees.  All the panels were on the same charge controller.  For many reasons I have to move the panels and now is the time to implement a scheme I had been contemplating for some time.  I purchased 2 more charge controllers.  I am building a gable roof over my 10,000 gallon catchment tank.  The ridge line will run North-South.  The six 285 watt panels will go on the east facing roof at 45 degrees.  The six 235 watt panels will go opposite on the West facing roof also at 45 degrees.  I don't have a roof facet facing south so I will have to build a frame of some kind but it will also be at 45 degrees although not being part of the roof I can experiment with the slope some.  Each bank of panels will feed into its own Conext MPPT 60 150 charge controller.  I have the ComBox so I can watch the production of each bank relative to each other.  At about $500 per charge controller it is perhaps a rather expensive experiment but they do sort of count as spares.

    At my latitude of 19.5 degrees North the sun is actually past vertical for a few hours each summer and definitely rises and sets well north of east and west in the summer.  The South facing panels won't see much sun then but the East panels should start cranking early in the morning and the West panels will produce late into the evening.  I don't expect to see either more or less production per day because I only produce what I use.  However I expect to see the state of charge of the batteries improve as the panels will be carrying the load more hours of the day.  Furthermore, the new roof will have room for a couple of thousand more watts of panels for cloudy days.
    15 Panels (about 3,000 watts), Schneider Conext 60-150 MPPT Charge Controller, Schneider Conext 4048 Inverter, 8 x 6-volt Costco GC-2 Batteries.