From "Ring Main Unit → Box Transformer → Main Station," How Does a Cellular WiFi Router Clear the Substation's "Last Mile" of Communication?
——A Communication Engineer Who's Visited 200 Stations Drew a "Communication Breakdown Map"
Prologue: I've Visited 200 Stations — I Drew a Map
Twelve years in the communication business. I've visited no fewer than 200 substations.
110kV, 35kV, 10kV. Urban, suburban, mountainous. Newly built, aging, under retrofit… I've seen too many stations.
But if you ask me to draw a map — not the electrical single-line diagram, not the network topology — just "where communication has broken down" — I can draw you a map that's shockingly consistent.
The break points are always in the same places.
Main station to ring main unit? Broken. Ring main unit to box transformer? Broken. Box transformer to terminal sensor? Broken.
Not occasionally broken. Repeatedly broken. Broken in every possible way.
You think it's the equipment? No. You think it's the solution? Not entirely.
The real problem is — along the communication chain from main station to terminal, every segment has a different "temperament," and you've been using the same approach to deal with all of them.
Today's article won't talk about concepts or pile up specs. I'm just going to walk along this chain, segment by segment, and tell you exactly what's hard about each leg — and how to clear it at the lowest cost.
Main station to ring main unit — the distance isn't far, tens to hundreds of meters. Most people think: what could possibly go wrong on this leg?
The problem is precisely the "thinking nothing's wrong."
Because this leg is too "routine" — so routine that nobody bothers to monitor its communication quality. The ring main unit's DTU and FTU are running, data is being transmitted, the main station backend can see it — everything looks fine.
But do you know how many times I've seen this on-site?
The backend shows "communication normal," but when O&M staff arrive on-site — the ring main unit door is open, and the DTU inside is flashing red.
Dataisbeing transmitted. But it's "incomplete" data.
What does that mean? The communication link isn't completely down, but the packet loss rate is absurdly high. 30 out of 100 messages are lost. The main station backend is barely holding on with caching and retransmission. On the surface it looks "normal," but the data is already inaccurate.
Can you make accurate decisions based on inaccurate data?
The core pain point of this leg isn't "down" — it's"half-dead."
And what causes "half-dead"? Usually two things:
First, poor line quality.Many of the communication cables in old stations were laid haphazardly back in the day — no shielding, no protection, connectors oxidized. Over time, signal attenuation becomes severe.
Second, the exit equipment can't hold up.Many stations use ordinary commercial routers at the main station exit. The moment the temperature rises, they throttle. The moment electromagnetic interference hits, they drop packets. They're not unusable — they just can't withstand the substation environment.
How to fix it?
The solution for this leg isn't complicated — swap the exit equipment for industrial-grade, and raise the link quality.
You don't need to re-lay cables. You just need to go to the main station communication cabinet and swap that "half-dead" commercial router for a cellular WiFi router that can actually take a beating.
For example, some people use the USR-G806w from USR IoT — wide-temperature operation, metal enclosure, multiple watchdogs. Drop it in the communication cabinet and basically forget about it. It doesn't fix the aging line problem, but it guarantees — even if the line quality is a bit poor, the data still gets out as completely as possible.
This leg doesn't need a major overhaul. Swap in one reliable device, and you go from "half-dead" to "rock solid."
If main station to ring main unit is the "neglected short hop," then ring main unit to box transformer is the real "last mile" — and the hardest mile.
Why? Because this leg's characteristics are: many points, long lines, harsh environment, nobody to maintain.
In one power supply area, there might be a dozen or even dozens of box transformers, scattered across every corner of the city. Each ring main unit to each box transformer is an independent communication link.
You can't pull a fiber to every box transformer — the cost won't allow it. You can't send someone to maintain every box transformer — you don't have enough people.
So most stations go with: 4G wireless communication.
Sounds great, right? No cables to pull, no trenches to dig, just insert a card and you're online.
But the reality is —
4G signal in the city really isn't as stable as you think.
Box transformers in basements? One bar of signal. Box transformers behind tall buildings? Signal comes and goes. Box transformers in industrial zones? Electromagnetic interference so bad the 4G module just "quits."
The most outrageous case I ever saw: at one station with 12 box transformers, 7 of them had 4G communication that was "intermittently interrupted" — not completely down, but dropping every few hours for a few minutes each time. The backend data looked "mostly continuous," but if you looked closely, it was all fake data filled in by interpolation.
You dare use that data for condition assessment? You dare use it for fault prediction?
The core pain point of this leg isn't "no communication." It's"there's communication, but it's not reliable."
And "unreliable" is scarier than "nonexistent." Because with "nonexistent" you know to go fix it. With "unreliable" you don't even know where the problem is.
How to fix it?
The solution for this leg comes down to one word:backup.
You can't rely on just one 4G card. What you need is — when wired goes down, wireless takes over. When wireless is weak, a backup network kicks in. When the primary network fails, it auto-switches.
This isn't high-tech. This is the most basic communication redundancy design. But in actual projects, fewer than 30% of stations actually do it.
How? At the communication node of the ring main unit or box transformer, place a cellular WiFi router that supports automatic multi-network switching. Primary path goes 4G. Backup path goes wired. Both down? It switches to a backup carrier.
The USR-G806w that some people use supports exactly this logic — 4G/3G/2G full-network, automatic wired-to-wireless switching, built-in eSIM so no card needed. Drop it in the box transformer's communication box and basically forget about it.
You don't need to assign a communication engineer to every box transformer. You just need to give every box transformer a "can't-die" communication node.
That's the real solution for the "last mile" — not making the signal stronger, but making the signal"unbreakable."
This is the last leg of the entire chain, and also the most easily overlooked.
Most people think communication ends at the box transformer. The smart gateway inside the box transformer collects the data, sends it to the main station, done.
But have you ever asked yourself one question — are the data from those temperature sensors, partial discharge monitors, and oil level gauges inside the box transformer actually accurate?
I've seen too many cases like this on-site:
The backend shows transformer temperature at 65°C — normal. But O&M staff go on-site and measure it — 82°C.
The backend shows switchgear partial discharge as normal. But in reality, the sensor's communication module crashed three days ago. The data was "made up" by the gateway itself.
The data reached the main station — but the data is fake. This is more dangerous than no data at all.
Because with no data, you know to go investigate. With fake data, you get misled and make the wrong call.
The core pain point of this leg isn't "communication down." It's"communication is there, but the data isn't trustworthy."
The usual cause: the sensors' own communication modules are too weak, with poor interference resistance. In the substation's electromagnetic environment, they constantly "talk nonsense" — either they don't send data, or they send the wrong data.
How to fix it?
The solution for this leg isn't about "adding devices." It's about"making data verifiable."
You need to add a "data gatekeeper" at the box transformer layer — not a gateway, but an industrial communication device that can validate, filter, and flag upstream data.
Its job isn't to "forward data." It's to"judge whether the data can be trusted."
Signal abnormal? Flag it. Data spike? Alert. Communication interrupted? Auto-switch to backup link and report.
It doesn't produce data. It guarantees the data's authenticity.
Get this layer right, and every piece of data your main station backend receives is "verified." Only then do you dare make decisions based on that data.
Let me break this chain down for you. You'll see —
| Chain Segment | Core Pain Point | Root Cause |
|---|---|---|
| Main Station → Ring Main Unit | Half-dead, incomplete data | Exit equipment can't hold up |
| Ring Main Unit → Box Transformer | Communication exists but isn't reliable | Lack of redundancy and auto-switching |
| Box Transformer → Terminal | Data arrives but isn't trustworthy | Missing data validation layer |
Three legs. Three completely different problems. But the solution approach is actually the same —
At every key node, place a reliable piece of industrial communication equipment. Let it take the hit for you — the environment, the interference, the uncertainty.
You don't need to tear up the entire chain and rebuild it. You don't need a single power outage. You don't need to touch a single power cable.
You just need to place the right device at the right spot.
One at the main station exit. One at the ring main unit. One at the box transformer. Three devices, three lines of defense, and the whole chain is connected.
The USR-G806w that some people use is built for exactly this job. 4G full-network, automatic multi-network switching, wide-temperature operation, built-in eSIM. Drop it in the communication cabinet and forget about it. It's not expensive, but it can make your "last mile" truly connected.
After visiting 200 stations, my biggest takeaway is this —
Substation smart transformation isn't held back by "not enough equipment." It's held back by"unstable communication."
The online monitoring, smart terminals, AI analysis that you spent a fortune on — they're all waiting for a stable communication link. When the link is up, they come alive. When the link is down, they're all just decorations.
And this link — from main station to ring main unit, from ring main unit to box transformer, from box transformer to terminal — every segment has its own temperament. Every segment needs to be handled on its own terms.
Don't try to use one solution for everything. Don't think pulling one cable solves all problems.
Follow the chain. Look at it segment by segment. Reinforce it node by node.
When communication is stable and data is accurate —that'swhen your smart transformation truly lands.
This "last mile" is worth walking carefully.
