OCPD operation in islanded systems with low available fault current

[tallgirl]

I was responding to a post where Jon stipulated a fault that would draw more than 12 times the inverter rating. No listed inverter is going to maintain the set output voltage at 12 times its current rating.

Perhaps pvnoob's situation with commercial is different, but in the resi work I do I'm near certain that any fault that isn't too much to shut down the inverters is in the thermal (not magnetic) range of any OCPDs. In other words, longer clearing times no matter if the source is utility or microgrid. It's faults that are over the inverters ability to keep outputting voltage where the behavior will be different.

We may be coming at this from opposite sides.

Since inverters aren't going to have the available fault current in the magnetic portion of the curve, perhaps it's time to re-evaluate the delay period in the thermal part of the curve. In just math terms, that's how I'd express what I'm trying to say.

 

[pv_n00b]

I've been doing some Etap models with limited available fault current and I'm finding that it's going to be dependent on the rating of the OCPD. If I have a 750kVA inverter I'll get about 1,100A of fault current. So small branch circuit CBs are still going to trip fast with no problem. But as I move up through feeders to the single OCPD on the inverter the trip time increases up to 5 hours on the largest CB. So at the branch level, I'm good. At the inverter level if there is a fault downstream of the inverter CB but upstream of the distribution section feeder CBs I'm okay since even if the CB won't open the inverter will shutdown so same result.
But if there is a fault between a feeder CB and a branch panelboard I want the feeder CB to trip and for the inverter to stay online, and those 100-300A CBs are just not going to trip fast enough with the limited fault current before the inverter stops playing. I'm going to need to find an alternative for the inbetween.

 

[310 BLAZE IT]

When designing PV+BESS system that offers an island operation mode, how are people dealing with fault clearing by OCPD when the available fault current from the PV and BESS are much lower than what is available from the utility? For instance, if I have a 100A feeder CB in a panel and there is a fault on the feeder it is protecting the utility might provide 10kA of fault current and trip the CB very fast. But with the utility disconnected and the feeder powered only from the PV and BESS it might see only a few hundred amps of fault current. So it will eventually trip but it will take a long time and maybe the PV or BESS inverters shut down before the CB clears the fault. How can I restore that nice fast trip time when I have such a low available fault current?

We use differential protection, it is extremely expensive requiring fiber and SEL relays but necessary in order to clear the fault during islanding mode, even with a generator+PV+BESS. Another option is to use a specialized battery which has high enough fault current, however this limits the battery and PCS application.

We have 800A power circuit breakers dialed down to 200A going to a 100kW PV array so hopefully there is a molded case solution in the future

 

[pv_n00b]

I have seen the "put a relay on every CB" solution and was trying to avoid that due to the cost. It's a tough nut to crack. I had not thought of using derated power CBs but I can see how that would work. Just crank down the short time and instantaneous so it's not inverse. No overload capability.