On Thu, 13 Jun 2019 10:25:59 -0700 (PDT), trader_4
wrote:

On Wednesday, June 12, 2019 at 8:43:28 PM UTC-4, wrote:
On Wed, 12 Jun 2019 12:17:50 -0700 (PDT), trader_4
wrote:

On Wednesday, June 12, 2019 at 2:51:36 PM UTC-4, wrote:
On Wed, 12 Jun 2019 10:31:06 -0700 (PDT), wrote:



I found this that explains a lot. We're on the right track of thinking
about what happens and that it's kind of a paradox as to how to detect
the loss of the "grid". They call this problem "islanding". This
talks about the ways they detect it, it's complicated and even includes
active action by inverters to deliberately try to destabilize the
grid so that it goes off kilter and all the other inverters can more
quickly detect it. It looks like they use voltage changes, freq changes
and power factor changes as possible parts of the algorithms and
that it's not a trivial problem. I think this gear is more for a larger
commercial array, but the same principles should apply to all inverters.


https://www.solectria.com/blog/anti-...-pv-inverters/

that is interesting

it is also what makes the frustrating situation for people that spent a lot of money for a solar system, and they still have no electricity when the grid is down, even when the sun is out.

m

Hence my question, could you disconnect from the grid and use a small
inverter or a little Honda to clock your solar inverter and rock on.

I think if you read that description I provided of how they detect
the loss of the grid, the answer is almost certainly not. They are
looking for stability in voltage, freq, power factor and it's
unlikely a typical home generator is going to fool it. Especially
with variable load, which you'd have when a fridge cycles on, etc.
Of course that description looks like it's for a bigger solar install than
a home one, but the home ones have to comply with regs and must use
similar techniques. Most interesting in that was that the inverters
described there actually deliberately screw with the grid when they
think they may be in an island situation, the idea being that you
can't perturb a true grid, but if the "grid" is an island with just
some other solar sources, you can and then other inverters will see
the deviation and shut down too.


My generator was pretty stable, looking at it on a scope and the
frequency doesn't drift that much when the engine grunts looking at a
frequency counter.


"That much", is likely a world of difference compared to the normal
grid and would be instantly detected.


An inverter generator should not change at all.

Maybe.


I think the author was just trying to dumb down his talk. When he says
"screw with the grid" what he really means is these things do not have
an internal clock and without a stable clock coming in they just go
nuts. It detects that condition and shuts down, probably in a cycle or
two.
.

He didn't say "screw with the grid" and he actually explained what he
meant.


"For the active anti-islanding function, we use a technique called Slip Mode Frequency Shift. This varies the reactive power output of the inverter. The goal of this protection method is to destabilize an islanded feeder by trying to influence the frequency. This diagram, from the IEEE 1547 testing procedure, shows how we test the anti-islanding function in the inverter.

We intentionally force the frequency out of spec and push against the grid, so we can properly detect and dissolve the island. We actually turn the anti-islanding function off, create an island, turn the anti-islanding function back on, and test the ability to disturb the island within two seconds. In our tests, we actually do it much faster."

That is still flowery rhetoric for saying an unclocked inverter will
go nuts. I think if it got a stable clock it would run.