Perth recorded an operational minimum demand on its grid last spring that would have been unthinkable a decade ago: the sort of midday trough where rooftop solar does so much of the lifting that scheduled generation barely needs to turn up. If that happened on the east coast, AEMO could lean on an interconnector, shuffle some of the surplus to a neighbouring region, or at worst shed it into Tasmania via Basslink. In the South West Interconnected System, there is no neighbour. There’s no cable running anywhere. The isolated SWIS grid is exactly that: isolated, full stop, and every operational problem it has gets solved inside its own fence line or not at all.
I’ve spent a fair bit of time looking at the NEM’s five interconnected regions and the various projects (VNI West, Marinus, the Queensland-NSW links) built specifically so one state’s surplus can bail out another’s shortfall. Western Australia doesn’t have that option, and it never will in any form resembling today’s technology. The SWIS runs from north of Geraldton down through Perth and across to Albany and the Esperance networks beyond it, servicing roughly a million customers, and the nearest meaningful grid is the North West Interconnected System up near the Pilbara: itself a separate island, serving the iron ore and gas export operations, with no link to the south. Western Australia’s electricity system is two islands, not one, and neither talks to the mainland NEM at all.
Why isolation changes the engineering
Let’s be careful with the terminology here, because it matters more in an island system than almost anywhere else in Australia. Power, measured in megawatts, is what the grid can deliver at an instant: the capacity to meet a sudden need. Energy, in megawatt-hours, is the volume delivered over time. On the NEM, a shortfall in instantaneous power can sometimes be patched by an interconnector pulling MW from next door for the twenty minutes it takes a peaker to start. In the SWIS, AEMO Western Australia, which, confusingly to outsiders, also operates the WEM, the Wholesale Electricity Market, as a separate structure from the NEM – has no such patch available. Every megawatt of firm capacity sitting in reserve has to physically exist inside the fence.
That shapes how the system operator thinks about reliability. The WEM runs its own capacity mechanism, a Reserve Capacity Mechanism that contracts for capacity years ahead rather than relying purely on an energy-only spot price the way the NEM historically has. It’s a structurally different market design, built for a structurally different problem: you cannot borrow your way out of a shortfall when there’s nowhere to borrow from.
The mix underneath Perth
The SWIS generation mix has moved a long way from where it sat even five years ago. Coal is still present: Synergy’s Muja and Collie stations, and the Bluewaters plant, have carried peak demand for decades, but closure dates have been locked in and brought forward more than once, echoing the pattern east-coast readers will recognise from how coal exits are tracking against replacement builds nationally. Gas remains a substantial firming source, partly because Western Australia sits on its own gas reservation policy that keeps a slice of North West Shelf and onshore production for domestic use: a very different story to the east-coast shortage anxieties covered in pieces like whether the gas industry has been written off too early.
Rooftop solar is the real disruptor. Perth has one of the highest per-household rooftop solar penetration rates of any city in the world, and on a mild sunny day in October the SWIS sees minimum operational demand fall to a level that would have described a small country town’s draw fifteen years ago. That’s the same dynamic described in the rooftop solar grid stability discussion – except in an island system, there’s no spillover region to dump the surplus into when demand craters at midday. AEMO WA has had to build curtailment schemes and emergency solar management settings into standard operating procedure years ahead of most of the NEM region having to take it seriously.
Batteries doing the island’s heavy lifting
Big batteries have arrived in the SWIS faster, proportionally, than almost anywhere else in the country, precisely because there’s no interconnector to lean on instead. Western Power and Synergy have both brought large-scale storage online, and the state government’s own Collie battery projects are explicitly framed as replacement firming capacity for the exiting coal units, not a bolt-on extra. It’s the same calculus laid out in the firming maths comparing gas peakers against big batteries, but with the stakes raised: if the SWIS gets its firming build wrong, there’s no east-coast style emergency import to cover the gap. The system either has the MW sitting in reserve, or it has a problem.
Worth noting too that grid-forming capability matters more here than almost anywhere in the NEM. As coal units retire and system strength thins out, the sort of inverter technology discussed in why the grid needs grid-forming inverters isn’t an optional refinement for the SWIS. It’s close to existential. An island grid with declining synchronous generation and no neighbour to borrow inertia from has far fewer places to hide if frequency control goes wrong.
A worked example on reserve margin
Here’s a simple way to see why the isolation changes the arithmetic. Suppose peak summer demand on the SWIS sits around 4,000 MW on a stinking February afternoon – aircon load stacked on top of everything else. AEMO WA’s reserve capacity target typically wants a margin on top of that forecast peak, built from a mix of thermal, gas, battery and demand-response capacity contracted years in advance through the Reserve Capacity Mechanism. On the NEM, a similar reserve shortfall in, say, Victoria can sometimes be patched in real time by drawing extra MW through an interconnector from NSW or South Australia if those regions have spare capacity running. In the SWIS, that reserve margin has to be entirely self-funded and self-built before the event, because on the day, there is no neighbouring region with spare MW to call on. It’s the difference between a household with a credit card and one that has to keep its emergency fund in cash under the floorboards. Both can cope with a bad month. Only one can borrow if it’s wrong about how bad the month will be.
Esperance, and the edges of the island
It’s easy to talk about the SWIS as if it’s just greater Perth, but the system stretches a long way south and east: down past Bunbury and Albany, out towards Esperance on the south coast, where the network thins to something closer to a long, lightly-loaded spur than a metropolitan grid. Those fringe-of-grid towns face their own version of the isolation problem nested inside the larger one: voltage support and reliability get harder the further the line runs from Perth, and Western Power has had to use standalone power systems – essentially a small local microgrid with battery and solar, physically disconnected from the main network – for some of the most remote customers rather than keep stringing wire. It’s a neat, practical admission that at a certain distance, running a pole-and-wire connection stops making sense next to a local battery and solar system. I’d argue that’s actually the right call for the handful of sites where it applies, even though it runs against the old universal-service instinct that every property gets a wire. A grid, even an isolated one, doesn’t have to mean every connection is physically continuous with it.
What’s coming for the isolated SWIS grid
The state’s own Whole of System Plan, produced by AEMO in its role as WEM operator, sketches a future SWIS leaning harder on renewables, storage and demand-side participation, with coal gone by early next decade. The practical test, as it so often is in this industry, will be whether transmission and storage actually land on the timeline the planning documents assume: a gap that’s tripped up REZ rollouts in the eastern states, documented in pieces like NSW’s coal exit outrunning its REZ build. Western Australia doesn’t have the luxury of blaming an interconnector dispute or a neighbouring state’s slow approvals process if its own build runs late. Whatever reliability problem turns up on a stinking February afternoon in Perth is one the SWIS has to solve entirely by itself, with whatever it managed to build in time.
That’s the quiet discipline this grid imposes on everyone who plans it. No borrowed capacity, no imported inertia, no neighbour’s spare battery to call on at 6pm. Just the island, and whatever it built.
– Anjali Rao, Grid & Storage Correspondent
Photo by Amirreza Taqavi on Unsplash