Microgrids: Running When the Grid Above You Does Not
The question comes up after every major outage. The panels are on the roof, the sun is shining, and the house is dark. The equipment is working correctly, and the reason is a safety rule that most owners were never told about - which is also the starting point for understanding what islanding actually requires.
Why Your Solar Shuts Off in a Blackout
A grid-connected inverter is designed to follow the network. It measures the voltage and frequency present and injects current synchronised to them. It does not create the reference; it borrows it.
When the grid fails, that reference disappears, and the inverter has nothing to synchronise with. It could in principle continue generating into the local wiring, and this is precisely what standards prohibit. If it did, it would energise a line that a repair crew has isolated and tested as dead. Anti-islanding protection requires the inverter to detect the loss and disconnect within a fraction of a second, and it is enforced everywhere for that reason.
The result is the familiar situation: working panels, a sunny day and no power. The safety logic is sound, and what it reveals is that grid connection and independent operation are different capabilities that happen to use overlapping hardware.
Making the system islandable means adding what the grid was providing. A hybrid inverter with a battery can disconnect the house from the network at a defined point and then establish its own voltage and frequency on the house side. Once that reference exists, the solar can resume, and the installation runs as a very small island. The additional cost is the transfer switch, the battery and a more capable inverter.
What a Microgrid Actually Is
A microgrid is a section of network with a defined electrical boundary, containing its own generation and load, able to operate either connected to the main grid or separated from it, and to move between those states under control.
Four things are needed. A point of common coupling - a single switch where the boundary is drawn and separation happens. Generation within the boundary sufficient for at least the critical load. At least one resource able to form the grid, setting frequency and voltage rather than following them. And a controller that manages the transition, balances generation against load while islanded, and resynchronises before reconnecting.
Resynchronisation is more delicate than separation. Closing a switch between two alternating systems that are out of phase produces a violent current surge that can damage equipment on both sides. The controller must match voltage, frequency and phase angle within tight limits before closing, which takes time and reliable measurement.
The balance problem inside an island is also much harder than on a large grid. A continental network absorbs the loss of a large load without measurable effect. A microgrid of a few hundred kilowatts feels every motor start. With almost no inertia, frequency moves quickly, so the controller must act in milliseconds - which is why battery storage is effectively mandatory in any microgrid with significant renewable generation.
Where They Are Worth Building
The first category is where an outage causes harm. Hospitals have run islandable systems for decades, though traditionally as diesel standby rather than as a microgrid proper. The difference is that a microgrid runs continuously in parallel with the grid and can take over seamlessly, rather than starting after the lights have already gone out.
Military bases adopted the architecture for the same reason and a different threat model, and universities and large campuses followed because a defined boundary happens to align with an ownership boundary. Several US campuses have operated microgrids through regional outages, which is where much of the operational experience comes from.
The second category is where the grid is weak or absent. An island community running on diesel has a microgrid already - adding solar and storage lowers its fuel cost without changing the architecture. Remote mines and processing sites are the same case with different economics.
The third is the largest by population. Village-scale systems serving a few hundred connections in sub-Saharan Africa and South Asia are microgrids in every technical sense, and they are the fastest-growing route to energy access because they reach a service level that supports refrigeration, milling and small workshops. The engineering is the same as a campus microgrid; the difference is that here it is the only grid there is.
What Islanding Rewards
A microgrid values generation differently from a national grid, and the difference is worth being precise about.
On a large network, a source that stops producing is covered by everything else, and its variability is diluted in a national aggregate. Inside an island there is no aggregate. Whatever is generating is the entire supply, and a gap must be filled locally by storage or a generator that someone has to fuel.
This inverts several of the usual priorities. Predictability outranks peak output, because a controller can plan around a known profile and not around a surprise. Local siting is inherent rather than advantageous. And the avoided cost of the alternative is high: the comparison is not against a wholesale market price but against diesel delivered by road to a remote site, which is expensive and logistically awkward.
This is the setting in which continuous generation is worth most. In a well-connected European grid, constant output competes against cheap flexibility that already exists. In an island with no neighbour to lean on, output that simply continues removes the need for the storage and fuel that would otherwise cover it. Any new source should be evaluated in both settings, because the same technology can be marginal in one and decisive in the other.
Frequently asked questions
Why does rooftop solar stop working during a blackout?
Because grid-connected inverters follow the network's voltage and frequency rather than creating them, and standards require them to disconnect within a fraction of a second when the grid fails. This anti-islanding protection prevents them energising lines that repair crews have isolated and tested as dead.
What do you need to keep running when the grid fails?
A transfer switch to separate cleanly at a defined point, a battery, and an inverter capable of grid-forming - establishing its own voltage and frequency reference on the isolated side. Once that reference exists, the solar can resume and the installation runs as a small island.
What makes a microgrid different from a backup generator?
A microgrid operates continuously in parallel with the main grid and can separate seamlessly when needed, rather than starting up after the lights have gone out. It also manages its own frequency, voltage and balance while islanded, and resynchronises under control before reconnecting.
Why is reconnecting harder than disconnecting?
Because closing a switch between two alternating systems that are out of phase produces a violent current surge that can damage equipment on both sides. The controller must match voltage, frequency and phase angle within tight limits before closing, which requires time and reliable measurement.
Where are microgrids most valuable?
Where an outage causes harm - hospitals, military sites, data centres - and where no grid exists. The largest category by population is village-scale systems in sub-Saharan Africa and South Asia, which are microgrids in every technical sense and the fastest-growing route to energy access.
