The Grid Is Out. Is the AC Side Dead?
Core Knowledge Application Domain 2 · Task 2.CA storm has taken the utility supply out across the area, and a technician is sent to a commercial rooftop site that has battery storage. Standing at the switchgear, a colleague says the AC side must be dead, because a PV system cannot run without the grid. When is that reasoning correct?
Reveal answer and explanation
Correct answer: A) It holds for a grid-tied string or central inverter, which cannot run without the utility, but not for a multimode inverter, which can keep backed-up circuits live
The colleague's rule is right about the inverter they have in mind and wrong about the site they are standing on. A plain grid-tied inverter, string or central, needs the utility present to run. Take the grid away and it stops, which is both a safety requirement and a fact about how it works. That is the system most people picture when they say a PV array cannot run without the grid.
A multimode inverter is built to do the opposite in exactly this situation. The JTA lists it beside string, micro and central as a distinct type, and the distinction is what happens during an outage: it disconnects from the utility and forms its own island, supplying a backed-up section of the building from the batteries and from the array. Everything downstream of that transfer point can be live while the street outside is dark, and it can stay live for as long as the sun and the batteries hold out.
Anti-islanding is not the counter-argument it looks like. What it requires is that an inverter must not energise a utility line that the utility believes is dead. A multimode inverter satisfies that by opening its connection to the grid first and only then energising the backed-up loads, which are on the other side of that opening. The requirement is about not backfeeding the network, not about the building being dead.
The O&M point: the crew that installed this site knew what they were installing. You arrive at a site from a portfolio, after a storm, in a hurry, and the only thing you actually know is what you have confirmed on the spot. Identifying the inverter type is not trivia to be recalled for an exam. It is the step that decides whether the general rule your colleague just quoted applies to the switchgear in front of you, and a rule that is right nine times out of ten is the most dangerous kind to carry onto an unfamiliar roof.
Why the other options are wrong
- B) It holds for any inverter, because anti-islanding protection is required of all of them
Anti-islanding is real and it is required, and this reads the requirement as a broader promise than it makes. It forbids energising the utility's conductors when the utility is not there, which protects line workers and anything else connected to that network. It says nothing about whether a building can be supplied from its own generation behind an open disconnection point. A multimode inverter complies with anti-islanding and powers the backed-up loads at the same time, because those two things happen on opposite sides of the break.
- C) It holds only while it is dark; once the sun is up the inverter restarts whether or not the utility is present
This has the dependency backwards. A grid-tied inverter does not wait for darkness, it waits for the utility: no grid means no output regardless of how bright the day is, which is why a grid-tied site is dark during a daytime outage. Sunrise restarts an inverter that has the grid and is merely idle overnight. It does nothing for one that is missing the reference it needs to run at all.
- D) It never holds, because the array stays energised whenever there is light on it
This is correct about the DC side and it is answering a different question. The array is indeed energised whenever light falls on it, and that is worth never forgetting, but the claim being examined was about the AC side of the switchgear. Carrying a true fact into the wrong part of the system is its own hazard: it leads people to treat every AC conductor as live and every DC conductor the same way, which sounds cautious and in practice flattens the distinctions that tell you what to isolate and what to verify.
References
- NABCEP OMAT JTA v.2026.5 — Domain 2, Task 2.C: "Inverters (e.g., string, micro, multimode, central)"
- NABCEP OMAT JTA v.2026.5 — Domain 2, Task 1.D lists "Multimode (i.e., grid-tied with energy storage)" as a system type in its own right, which is the definition the correct answer rests on. No standard is cited for anti-islanding behaviour; it is described here only in the general terms the JTA itself uses for codes and standards.