Friday, January 18, 2008

The Need for Automatic Control in Off-Grid Solar PV Systems

To get as much as I can from my off-grid system I manually connect loads to it, watch as battery voltage declines, and remove the loads before the battery voltage falls to an unacceptably low level. Since I’m not always home, and because I’m not always monitoring voltage while I am at home, I often fail to use the available energy efficiently. I also run the risk of over-discharging the batteries, possibly causing irreversible damage to them. I’ve outlined a plan to automate my system previously, and I’ve made significant progress toward implementing that plan. It’s an ambitious project, but well worth the effort. Read on for additional justification, and a progress report.

I now have a Transfer Switch installed and functioning. When I turn on the inverter, loads attached to the circuit are powered by it. When I switch off the inverter, the loads are automatically switched to grid-supplied power instead. I’m using my refrigerator and a chest freezer as the test loads. Installing the Transfer Switch was “Step 1” of my plan. Automating the switching on and off of the inverter based on battery voltage is the next step.

Recently, I switched on the inverter with the loads attached in the afternoon as I often do. I monitored the battery voltage at regular intervals, intending to switch off the inverter before bedtime. Unfortunately, I fell asleep without switching the inverter off. I woke up at 4:00am to find that my refrigerator and freezer had no power. This wasn’t supposed to happen; the Transfer Switch was supposed to connect the loads to grid-supplied AC if the inverter stopped functioning. I found that battery voltage had dropped to 11.7 volts. The Transfer Switch was “chattering”. Analyzing the situation, this is what I determined: When battery voltage dropped to a point where the inverter could no longer function, the Transfer Switch disconnected the load, connecting it instead to the grid. With no load on the batteries, battery voltage quickly increased to the point where the inverter could once again function. This, in turn, caused the Transfer Switch to reconnect the inverter to the load. This cycle kept repeating at a rapid pace, which resulted in the Transfer Switch chatter. Eventually the breaker for the grid-supplied AC tripped. The tripped breaker didn’t stop the cycle, it just prevented grid-supplied AC from getting to the load. I had to manually intervene by turning off the inverter and resetting the breaker.

The incident described above demonstrates an interesting battery characteristic. When a load is applied, voltage drops. When the load is removed, battery voltage increases, even when the battery is deeply discharged. It is for this reason that I decided to use low and high set-points in my control circuit. And to facilitate differences in battery bank sizes and type of batteries, I’ve decided to make those set-points adjustable. I’ll make them adjustable to one-hundredth of a volt, which is much better control than some commercially available equipment provides. Once batteries are disconnected, they will not reconnect until they’re once again fully charged. After I’ve determined the appropriate settings, I should not experience a problem like the one I experienced recently.

I’ll get many of the parts needed for the project tomorrow (Saturday), and will hopefully have a portion of the circuit built and tested by the end of this weekend. Check in on me later for a progress report.


1 comment:

d said...

Yep, it's quite interesting getting things stable. I had to have my lower set-point higher than I expected and my higher set-point lower than I expected, and keeping the two far enough apart to avoid oscillation was tough until I added I nice big power-supply cap upstream of the inverter.

In my case I'm drawing ~20W and my cap is ~68mF (milliFarads), so you'd need a much bigger beast, possibly in proportion to the load size.