Old Station Retrofit: Most Afraid of "Pull One Thread and the Whole Thing Falls Apart"? Cellular Router Lets You Change Only the Network, Not the Station Itself
——An Old Engineer with 15 Years of Station Retrofit Experience, Sharing Some Words from the Heart
This article is written for everyone who "doesn't dare to touch old stations."
I know what you're thinking.
You have a batch of old stations to retrofit. The dispatch center has set targets. The boss has slammed the table. But you open the proposal and see — full station blackout? Impossible. Outage window? Two days. Construction crew? Just those few guys.
You stare at the station wiring diagram for half an hour, then close your laptop and light a cigarette.
It's not that you don't want to retrofit. It's that you don't dare.
The four words "old station retrofit" sound like a technical problem, but they're actually a psychological one. You're not afraid of the technical difficulty — you're afraid that if you touch one wire, the whole station goes down.
Today's article won't preach any big theories. It'll just talk about one thing: how to turn old station retrofit into a minimally invasive surgery — "change only the network, don't touch the station body."
Over the years of doing station retrofits, I've seen too many people trip over the same spot.
It's not that the technology isn't good enough — it's that the mindset collapses first.
You take over an old station that's been running for ten, fifteen years. The secondary circuits are a tangled mess. Some labels are already unreadable. There are tons of places where the drawings don't match the actual site. You stand in front of the cabinet, multimeter in hand, and the thought in your head isn't "how do I retrofit?" — it's:
"If I touch this wire, will the cabinet across the way trip?"
This fear is real, and it's completely reasonable.
The secondary system of an old station, after years of operation, has accumulated all kinds of temporary wiring, patch configurations, and bypass modifications. It's long since stopped looking like what's on the drawing. You think you're working from the drawing, but in reality, you're "opening a blind box."
So many projects end up with plans revised again and again, schedules delayed again and again, budgets exceeded again and again. It's not because it's hard — it's because everyone is using a "full station overhaul" mindset to deal with a problem that only needs a "network upgrade."
It's like someone who just has a cold, and you want to give them a full body checkup, a blood transfusion, and surgery — it's not that it can't be cured, it's that it's unnecessary, and the risk is enormous.
What's the most brutal reality of old station retrofit?
You don't have the luxury of "working at your own pace."
A new station can take its time — design, procurement, construction, commissioning — three months, six months, no problem. But an old station can't. An old station is carrying load. Users can't lose power. The grid can't lose power. The window you can争取 might just be 24 hours from a planned maintenance outage, or even 12 hours.
In those 12 hours, you need to complete:
If any single step runs overtime, everything behind it falls apart.
The worst case I ever saw: a 110kV substation retrofit. Because the network configuration had a problem during on-site commissioning — the GOOSE messages didn't match up — it ate up an extra 8 hours. How did it end? The O&M team stayed up all night, manually checking every single point table, changing them one by one, testing them one by one. By dawn, everyone's eyes were bloodshot.
After that, I figured out one thing: for old station retrofit, any work that can be done off-site must absolutely not be done on-site.
You think swapping in new equipment is the end of it? Too naive.
What's sitting in your old station are protection devices from ten years ago, measurement and control units from five years ago, online monitoring added three years ago… They each speak a different "dialect." Some use IEC61850, some use IEC60044, and some use the manufacturer's proprietary protocol outright.
You want them to "talk" on the same network? You have to add protocol converters, add gateways, add middleware. The more you add, the more failure points you create, the higher the latency, the greater the chance of something going wrong.
Even worse, these middleware devices themselves need power, need configuration, need maintenance. You set out to simplify the system, but you end up making it more complex.
A colleague of mine once said something I thought was incredibly insightful:
"The biggest trap in old station retrofit isn't that the equipment is old — it's that you're trying to use 'addition' to solve a problem that needs 'subtraction.'"
What does that mean? You don't need to replace every device. You just need to build them a "bridge" — a stable, reliable, protocol-agnostic communication bridge.
Once the bridge is built, even old devices can speak the new language.
This layer of fear is rarely spoken aloud, but it's the most lethal one.
Have you seen a station after a retrofit like that? The equipment is new, the screens are lit, the data is there — but every few days the communication drops, the backend goes offline regularly, and the O&M staff spends their days running back and forth between "restart the router" and "check the network cable."
This kind of retrofit is worse than no retrofit at all.
Because it gives the boss an illusion — "We've already gone smart." Then when something goes wrong, you're the first one to get blamed.
So when you're choosing a solution, what you really care about isn't "how advanced is this equipment?" — it's:
"Can it let me sleep at night?"
You don't want a fancy system. You want a system that, once installed, you can forget exists.
Enough about fear. Time for the prescription.
The core idea is eight words:Network first. Station body untouched.
How to do it? I've distilled my years of experience into four steps. Take them and use them directly.
Don't rush to tear out the old and install the new. Sit down first. List every device in the station that needs to communicate, and draw a communication requirements map — who needs to talk to whom? What protocol? How much data? What latency requirements?
Once that map is clear, you'll know exactly what to add and what not to touch.
Most of the time, you'll discover that 80% of the smart requirements in the station can actually be solved with just a stable network. You don't need to touch the secondary circuits. You don't need to touch the protection devices. You don't need to touch a single power cable.
This is the most critical step.
You know better than I do how brutal the communication environment in an old station is: high temperature, high humidity, intense electromagnetic interference, cramped spaces, no air conditioning… Ordinary commercial equipment in this environment crashes every other day and dies every five days. You simply can't rely on it.
What you need is industrial-grade communication equipment specifically designed for this kind of environment — wide-temperature operation, metal enclosure, EMI protection, fault self-healing.
For example, at the remote communication link stage, some people use the USR-G806w cellular router from USR IoT — 4G full-network compatible, automatic wired-to-wireless switching, built-in eSIM so you don't even need to insert a SIM card. You put it in the cabinet and basically forget about it. It doesn't replace any of your on-site equipment. It just lets your on-site data "get out" reliably.
You don't need to change a single wire in the station body. You just need to put a dependable router at the key nodes.
That's it. That simple.
I've emphasized this point repeatedly, but I'll say it again.
All network configuration, VLAN assignment, GOOSE tables, routing policies — do it all in the office. On-site, you only do three things: "connect, power on, verify."
You can use configuration templates provided by the manufacturer, or use SCD files to pre-generate all parameters. The less work you do on-site, the lower the risk.
For old station retrofit, you must assume "something will definitely go wrong."
So in the network architecture, there must be backup links. If wired goes down, wireless takes over. If the primary path fails, the backup path kicks in. Don't put all your eggs in one basket.
This isn't overkill. This is life insurance.
I've been doing station retrofits for fifteen years. My biggest realization is this —
The essence of old station retrofit isn't "replacing with new." It's "extending life."
You don't need to turn an old station into a new station. You just need to let it hold on for another ten, twenty years on its existing foundation — long enough to reach the next major retrofit cycle.
And the key to this "life extension" isn't equipment, isn't software — it's communication.
When communication works, data flows. When data flows, smart transformation has a foundation. When the foundation is solid, you can finally sleep at night.
So stop torturing yourself with a "full station overhaul" mindset.
Find communication as your single entry point, pick the right rugged device, lay down the network as your "base layer" — everything else will follow naturally.
You don't need to "pull one thread and have the whole thing fall apart." You just need to "change one network and bring the whole station to life."
That's the smartest way to do old station retrofit.
