About / How it works

1 What this is & why

The WA Electric Vehicle Network is a chain of DC fast-charging sites strung along Western Australia’s highways — from the Nullarbor across to the Kimberley. For long-distance EV drivers, one dead charger can strand a trip. This dashboard exists to answer a single honest question at each site: if I turned up right now, could I actually get a fast charge?

It does that by watching each charger’s live status and its recent real-world charging sessions, then flagging trouble early — before a driver arrives to a nasty surprise. It is deliberately failure-first: a charger that is quietly crippled (delivering a trickle instead of a fast charge) is treated as a problem even if it still reports “available”.

2 Data sources & their limits

Everything here is built from two public sources, combined:

• Chargefox live status feed — the network’s own public, anonymous status feed. It tells us each station’s reported state (available, in use, offline, out of order) and, during a session, roughly how much power is flowing. This is the backbone of every status on the dashboard.
• PlugShare driver check-ins — the reviews and check-ins EV drivers leave after they charge. We use these as a real-world sanity check: they can confirm a charger is genuinely dead, or soften an “offline” flag when drivers report it actually worked.

How often it updates

We regularly check each site so an active fault is caught quickly.

What the data can’t see

• The feed reports a power reading, not what a specific car could have drawn. It is a ceiling, not a promise.
• It carries no battery state-of-charge, no per-car maximum, and no precise session timestamps — we reconstruct sessions from the readings themselves.
• Some charging is effectively invisible: a driver topping up from a roadhouse power point (roughly 3.3 kW) or an on-site generator doesn’t show up in the charger feed at all.

3 Status categories

Every site carries one headline status. It is driven by the charger’s reported state and its measured behaviour — not by operator notes (see section 9). When more than one applies, the more serious one wins, in the priority order shown.

The key idea: status follows the charger’s reported state and its real measurements. A site can read “available” in the raw feed yet still be flagged degraded or down here, because the charging data shows it isn’t delivering.

4 DC reliability score

Each site shows a reliability score — a single percentage answering “over the last month, how much of the time was this actually good for a fast charge?” It is a track record, separate from the live status: the status tells you about now; the score tells you whether this site has been dependable.

• Measured over the trailing 30 days.
• “Up” means genuinely usable for DC charging — capable of delivering at least 20 kW. A charger that’s technically “available” but only trickling below that floor does not count as up.
• While the site has less than 30 days of history, it reads “N days so far” so you know the sample is still building.

5 Last confirmed working charge

Beyond “is it up now”, each site shows when a real car last got a proper fast charge here — the single most reassuring signal on the page: proof, from actual charging data, that the hardware recently did its job.

The bar for that is what we call an “effective DC charge” — a session that actually proved the charger working properly:

• Reached a solid peak power — at least 25 kW on a Tritium unit, or at least 50 kW on a Kempower unit;
• Delivered real energy — at least 5 kWh; and
• Ended cleanly, rather than cutting out mid-charge.

The last-charge line can show one of three things:

• “Last confirmed DC charge: X ago” — an effective charge happened recently. This is the reassuring case: a real car pulled full power here and finished normally.
• “Last DC charge: X ago — degraded (~46 kW)” or “(below full speed)” — a car did charge, but the session fell short of an effective charge. We say “degraded” when the charger itself was capped below its rating (for instance a 150 kW unit stuck at ~46 kW), and “below full speed” when the car simply couldn’t pull full power. Either way the site is being used — it just isn’t running at full speed — which is very different from a dead charger.
• “No confirmed effective DC charge in the last N days” (N being how far back our records go) — we simply have no record of an effective charge here. That is not proof the charger is broken: a quiet remote site may just not have been used, and some charging is invisible to us (a roadhouse power point or generator — see Data sources).

6 Charger slowness vs car slowness

A slow charge isn’t always the charger’s fault. Plenty of things about the car legitimately cap the speed, and it would be unfair — and misleading — to blame the hardware for those. The dashboard is capability-aware: before it ever flags a charger as slow, it rules out the ordinary car-side reasons.

We do not blame the charger when the low power is explained by:

• A high battery — the natural taper. Every EV slows down as it fills; above about 80% state-of-charge, that slowdown is the car protecting its battery, not a charger fault.
• A small battery. If the numbers imply a vehicle with under 25 kWh of usable capacity (a plug-in hybrid or electric motorbike), its modest draw is the car’s limit, not the charger’s.
• A low per-car DC maximum. Some cars simply can’t accept high DC power. If the same charger has proven it can deliver strongly for other cars, a slow session is read as car-limited.

Only once those are excluded — and the charger has not recently proven it can pull strong power — will a slow result count against the charger.

7 Committed nameplate & derating

Each charger has a rated (nameplate) power — the speed it’s built to deliver. The common Tritium units on this network are rated at 50 kW; the Kempower units at 150 kW or 75 kW depending on model. We compare real sessions against that committed figure.

A derating watchdog watches for a charger that has quietly dropped below its rating. If a unit that should do 50 kW settles more than a couple of kW under nameplate across sessions, that’s treated as degraded — a charger sliding below spec, even while it still reports “available”.

8 Suggested fault diagnoses

When a charger misbehaves in a recognisable way, the dashboard offers a suggested cause. Different faults leave different fingerprints in the power data — a coolant-pump problem looks nothing like a comms dropout. These are pattern reads, offered to help operators triage faster.

To stay fair, fault suggestions name a symptom, never a company — the wording is deliberately vendor-neutral.

9 Operator notes

Chargers sometimes carry a public note from the operator (“Type 2 available”, “pull forward”, and so on). We show these for context only.

10 Energy & time wasted

For operators, the dashboard estimates the real cost of underperformance — roughly how much charging time and energy has been lost because a charger wasn’t delivering. Crucially, this only counts loss that is the charger’s fault.

• Slow DC charging is measured against what a healthy unit should deliver (50 kW for a Tritium; a conservative 100 kW baseline for a proven Kempower). The car’s own high-battery taper is excluded — anything above about 80% state-of-charge doesn’t count against the charger.
• Forced onto AC — when a DC charger is offline or so crippled it’s effectively unusable (below about 11 kW) and a driver has to fall back to slow AC, that lost time is counted at the actual AC power they had to settle for.

You can view these totals over a Recent window, This Month, or This Year, and they can be scoped to one operator. Because they’re reconstructed from public readings, every figure is clearly marked “est.” — an estimate, not a meter reading.

11 Operators

The network’s 49 sites are run by two operators, and knowing which matters for reading a fault:

• Horizon Power — 27 sites. Regional and remote WA. Many of these run on off-grid stand-alone power systems (SAPS) — local solar and battery — or on small isolated town grids. When one of these goes offline, the power supply itself is a genuine suspect, so we tier the diagnosis accordingly.
• Synergy — 22 sites. Connected to the main South West Interconnected System (SWIS) grid. An outage here is almost always a communications dropout rather than a power failure.

You can filter the dashboard by operator, and the reliability and waste figures can be scoped to a single operator so each can see their own fleet.

12 Routes view

The Routes view re-arranges the same live data into travel corridors — the chargers laid out in the order you’d actually hit them driving a highway, so you can read a whole trip at a glance. Each stop gets a simple trip-usability colour:

Corridors are laid out in travelling order, and short dead-end detours off the main run are marked as spurs so you can tell a there-and-back side trip from a through-route.

13 Map view

The Map view plots every site geographically, each as a coloured pin at its real location — so you can see the whole network at a glance and judge how far apart your charging options really are. Pins use the same status colours as the rest of the dashboard (see Status categories), and the most serious ones are drawn on top so a problem site never hides behind a healthy neighbour. Tap a pin for a popup with the site’s name, region, current status and operator, plus a link straight to its full detail page. A legend in the corner lists the statuses currently on screen with live counts, and you can filter by operator (Horizon Power / Synergy) from the header. It reads the exact same live feed as the list and Routes views, and refreshes every few minutes.

14 Nullarbor traffic estimate

The Routes view carries an experimental estimate of how many EVs are crossing the Nullarbor — the long, remote Eyre Highway run where charging stops are few and far between.

How it works, in plain terms:

• By looking at the sequence of charging sessions along the corridor, we try to chain together the stops that look like the same car crossing — matching on battery size (within about ±15%) and travelling in a consistent direction and timeframe.
• That gives a firm observed lower bound — crossings we can actually see — which we then extrapolate upward (using capture assumptions of 50%, 70% and 90%) to estimate the true total, because we can’t see every car.
• Driver check-ins along the route feed in as extra presence evidence.

15 Charging now

The dashboard shows a live count of how many chargers have a car plugged in and charging right now across the network — a quick pulse of real-time activity. Tap that count to switch to a charging-first sort, floating the sites with live sessions to the top so you can see where the action is.

16 Honesty & caveats

The short version of everything above:

• We are an independent project, not affiliated with any charger operator or app.
• We’re only as good as the public data we read — if the underlying feed is wrong or stale, so are we.
• Fault causes are suggestions from patterns, not confirmed operator diagnoses (unless explicitly marked confirmed).
• Energy and time-wasted figures are estimates, reconstructed from readings — not meter data.
• The Nullarbor traffic count is experimental and almost certainly an undercount.
• Some charging is invisible to us (roadhouse sockets, generators), so “no confirmed charge” is never proof a charger is broken.

Our aim is to be useful and honest: to flag genuine trouble early without crying wolf, and to be upfront about what we can’t see. If a charger here looks worse than your own experience of it — or better — trust the road, and let the operator know.