Author Topic: Primitive logging with a Kill-A-Watt meter  (Read 3546 times)

0 Members and 1 Guest are viewing this topic.

Offline MadScientist267

  • Impossible Condition Curator
  • Moderator
  • Hero Member
  • *****
  • Posts: 1514
  • Karma: +44/-4
  • Rules? What rules?
Primitive logging with a Kill-A-Watt meter
« on: April 22, 2014, 08:57:32 pm »
I've got a very primitive logging system in place now to keep track of the dependence on the grid in the current configuration...

Bunch of verifications going on this time around... It's now been a whole year the solar has been up top, and the system's strengths and weaknesses have become more or less apparent. The graphs below agree with what I've more or less kept track of in my head for some time now, and so before I make the final dive into the big project, I wanted some confirmation.

The data for these is collected on a best-effort basis and I've made it as simple as I possibly can to put them in and make it all mean something in a big picture. Its based on the data from a Kill-A-Watt meter, using only 2 numbers per entry; Watt-hours and elapsed time. Periodically collecting them isn't all that big of a deal, just have to remember to do it :P

So far there are 3 plots that can be generated, and I'm looking at doing a fourth. They are:

1 - Average power deficiency
2 - Watt-hour deficiency, broken down on a per-hour basis
3 - Watt-hour deficiency, broken down on a daily basis

The fourth will be just a plot of the rising watt-hour count. It hasn't been implemented yet because what I'm really after is how deep I'd be digging into the batteries had I let them go down further.

The way this whole thing works is, everything on the van is completely 12V now. The main inverter never gets used (except to run the soldering iron on occasion) and the "big" lights (also on an inverter, albeit a tiny dedicated one - a "convenience" factor story for another thread).

The grid charger is always on, and is set for ~12.3V to allow approximately 25% DoD of the batteries. As this point is reached, the load seamlessly and gradually shifts over to the charger, the grid taking up the slack. This arrangement was to allow for a hybrid approach with minimal cycle damage to the batteries. It roughly equates to how much a day of full sun (at this time of year) can put back in the following day.

What you see below is what the charger had to pick up to meet these criteria.



2979-0

The average power deficit over the last 10 days. The dot in the far upper left is to establish the maximum "comfort zone limit" that the new system would be able to reasonably tolerate (50W). The "best-effort" aspect of the logging is pretty evident in this shot... The code that generates the graph is set up to insert zeroes for the missing data points and ignore them for calculation. As you can see, things started out pretty sunny for the first few days then the sky got less than cooperative, followed by once again sufficient sun the last couple of days.



2981-1

This is the watt-hour usage broken down by hour. More than evident here is the best-effort thing... As well as the search for the worst case scenario. I wanted to see what it looked like for this layout in hard numbers on paper, not just in my head. The highest peak was actually obtained when I was helping out the panels in mediocre skies - and beyond. I went ahead and brought them to a low boil from grid. This went on past any solar activity, and the inefficiency of doing so is rather obvious. The forecast for the next day wasn't great, and all this wasn't done to see if I can damage a battery, so I topped them up. Did a normal drain again that following night, and the joke was on me... plenty of sun since, as you can see. Thanks wunderground... ::)

You can also see where the grid was just beginning to pick up last night before I went ahead and shut everything down. There are a couple of small blips before this where I bumped the charger up just long enough to verify the batteries were full. The dot at far left is the current battery's total capacity by label (100% DoD, 2820Wh). The new setup will be precisely twice that ;)



2983-2

This is the watt-hour by day breakdown. It is simply the cumulative of each day extracted, calculated, plotted, and reset to zero for the next 24 hour period. Like the upper left dot for battery capacity, the "0" dot at hour 11 is artificial (establishes zero as the code's interpretation of the data doesn't inherently plot one to establish it naturally).



I've worked out that the supply is approximately 80% efficient overall, with the sweet spot being at moderate load (about 120W), and from there I'm of course able to work out what the minimum requirements of the new system will be. Everything should stay flowing without risk to the bank if I can keep the average deficit below 50W with the current layout.

So far, so good. :)

That's it for now, I'll put up some newer graphs after some more time goes by...

Steve

Wanted: Schrödinger's cat, dead and alive.

Offline MadScientist267

  • Impossible Condition Curator
  • Moderator
  • Hero Member
  • *****
  • Posts: 1514
  • Karma: +44/-4
  • Rules? What rules?
Re: Primitive logging with a Kill-A-Watt meter
« Reply #1 on: May 03, 2014, 08:40:33 pm »
Update -

Hit 21 days, here are the plots after some tweaking to the plotting code...

I've dispensed with the hourly plot. The data was there but difficult to interpret. Added a total kWh plot instead. The set is much more informative now at a glance:

Total Wh:
2991-0



Average deficiency (still need to play with this code a little, the (0,50) point is still just a hack, just haven't gotten back around to set the scale the proper way haha):
2993-1



Daily deficiency:
2995-2

Unlike before, pretty obvious now where the two "paranoia points" were and I kicked the charger into "fill-er-up" mode ;)

Good news is, everything is pretty much on target. I've been a little lax with consumption, but I'd rather read a little high and find out my actual usage down the road is a little less, than the other way around. :)

Steve
Wanted: Schrödinger's cat, dead and alive.