Tuesday, 2:10 in the afternoon. Thirty-eight degrees in the shade, a logistics park outside the city, and a containerized 500 kVA set screaming its high coolant temperature alarm for the third time that week.
The site crew had already done the obvious. They’d washed the radiator. Twice. They’d ordered a new radiator, in fact — a big one, with a bigger price tag. And I almost let them install it.
Then I walked up to the container, put my hand in front of the radiator inlet louver, and felt what the radiator was “cooling” with.
It felt like standing behind a hair dryer.
That radiator was fine. The coolant was fine. The box itself was eating its own heat, and no radiator on earth can fix that.
Here’s the pattern that gives this fault away. The generator starts the day happy. It carries load all morning without complaint. Then, somewhere in the early afternoon — the hottest part of the day, when the building needs it most — the coolant temperature creeps up, and up, and the alarm goes off.
People blame the coolant. The thermostat. The water pump. Sometimes the load. Honestly, I’ve seen crews replace all four before someone thought to check the air.
A radiator doesn’t cool with magic. It cools with the temperature difference between the coolant and the air passing through it. Shrink that difference — feed the radiator air that’s already hot — and the whole cooling system starts losing. At some point it loses faster than the engine makes heat, and you get an alarm at 2 P.M. on a day the set ran fine at 9 A.M.
Inside a canopy or a container, the radiator fan is pushing an enormous volume of hot air out through the discharge louvers. That air has to go somewhere. And air, like water, takes the easiest path available.
On one site I worked, a containerized set sat tucked into a corner between two warehouse walls — maybe three meters of clearance on the discharge side. The hot air came out, hit the wall, curled along it, and rolled right back into the intake louvers on the same end of the box. The generator was running in its own weather system.
The fix was almost embarrassing. We didn’t change a single part on the machine. We turned the container so the discharge faced open air and added a simple duct extension. The alarms stopped that same week.
Look, a discharge that exits horizontally into a wall, a fence, or another piece of equipment is a recirculation loop waiting to happen. The air doesn’t care about your drawings. It cares about pressure, and the intake side of a running radiator fan is a low-pressure zone that will happily suck back whatever the discharge just threw out.
Even with open space around the box, plenty of enclosures put the intake louvers and the discharge opening too close together — sometimes on the same wall, sometimes around the same corner. The designer assumed the hot air would rise and drift away. On a still, hot afternoon, it doesn’t rise fast enough. It lingers, and the fan pulls it back in.
This is the one I see most often in retrofit canopies — a quiet enclosure added to an older open set, with louvers cut wherever they were easiest to cut. The generator ran for years as an open unit in free air. Then someone boxed it in, the intake started seeing discharge air, and suddenly the “old reliable” set can’t survive August.
The old foreman on that job disagreed with me, by the way. He wanted a bigger radiator. Said the engine was getting old and needed more cooling. Maybe. But I asked him one question: if the engine were the problem, why did it only happen after the canopy went on? He didn’t have an answer. The canopy came partly off, the discharge got its own duct to the roof, and the “old engine” ran another two summers without a peep.
Silent canopies are wonderful things. They’re also, by design, full of restrictions — acoustic baffles, tortuous air paths, foam-lined turns that absorb noise. Every one of those restrictions costs airflow.
Push the restriction far enough and the radiator fan can’t move the air the engine needs. The fan stalls, or close to it. Airflow collapses. And here’s the cruel part: the air that does get through is the hottest air in the box, because the fresh air can’t get past the restrictions but the discharge can still slosh around inside.
I once opened a canopy where someone had added an extra layer of acoustic foam “to make it even quieter.” The set overheated within an hour of full load. The foam did its job beautifully — it just also did the radiator’s job, badly. We pulled half the foam out of the discharge path and the temperatures dropped like a stone.
If your set got quieter and hotter at the same time, that’s not a coincidence. That’s your airflow telling you exactly what happened.
People forget the engine itself inhales. The combustion air intake on most sets draws from inside the enclosure or from right next to it. Hot intake air is less dense — less oxygen per breath — so the engine makes a little less power and runs a little hotter doing it. It’s a small effect on its own. Stacked on top of a radiator that’s already breathing warm air, it stops being small.
Roughly speaking, every 10 degrees Celsius rise in intake air temperature costs you a few percent of output. I won’t put a hard number on it because engines differ, but the direction is never in your favor, and it compounds. Hot air in, less power out, more heat rejected into the same box. The loop feeds itself.
That said — don’t go drilling holes in the canopy on day one. A separate combustion air intake ducted to outside air is a real fix, but only after you’ve fixed the radiator airflow. Priorities matter.
You don’t need instruments worth thousands. A cheap infrared thermometer does it.
Run the set at a decent load — at least half, ideally more — on a warm day. Let it stabilize for twenty minutes. Then measure two things: the ambient air temperature a few meters away from the box, in the shade, and the air temperature right at the radiator inlet louver, in the intake airstream.
If the inlet air is roughly the same as ambient, your airflow is fine. If the inlet air is running 10 degrees Celsius or more above ambient, the enclosure is recirculating its own heat. Period. I’ve run this test dozens of times, and I’ve never seen a 10-degree delta that didn’t trace back to a short-circuited air path.
One more tell: the fault is load- and weather-dependent in a specific way. Fine at night, fine in the morning, fine at light load — alarms only on hot afternoons at real load. A bad thermostat doesn’t care what time it is. Recirculation does.
And please — resist the urge to “fix” it by opening all the canopy doors. I know it feels right. More air, right? Wrong. The fan and the louvers were designed as a pressure system. Open the doors and the discharge finds the shortest path back to the fan instead of going out through the duct. I’ve seen door-opening make the temperatures worse, not better. The box is a system. Treat it like one.
The real fixes are boring, which is why nobody wants to do them.
Give the discharge its own path to the roof or well clear of the intake, with a duct long enough that the air can’t curl back. Put the intake low and on the opposite end from the discharge, so the box breathes in a straight line instead of a circle. Add baffles between the discharge stream and the intake side so the two airstreams can’t meet inside the enclosure. Keep real clearance — open air, not a wall — in front of the discharge.
None of this is expensive. A sheet-metal duct and some baffles cost a fraction of the radiator that logistics park almost bought. What costs money is the sequence everyone follows: replace the coolant, replace the thermostat, replace the radiator — and then finally look at the air. Do the two-minute test first and you skip straight to the answer.
A quick word on monitoring, since most of you won’t be standing in front of the louver with a thermometer every afternoon: a simple temperature sensor at the radiator inlet, trended over time, catches this fault before the alarm does. That’s one of the few places where remote monitoring earns its keep on a standby set — inlet air temperature drifting upward month after month is the enclosure slowly choking, and you’d never see it on the coolant gauge until the hot day arrives.
For the deeper background — what overheating looks like from the engine’s side, and how to lay out cooling air for a whole generator room rather than just a box — the pieces on overheating warning signs and room ventilation design cover that ground. And if you’re choosing between a boxed set and an open one in the first place, the containerized vs open-type comparison and the enclosure retrofit guide are worth a read before you sign anything.
Next time your set runs hot on a summer afternoon, don’t order a radiator. Walk up to the box, hold your hand in front of the inlet louver, and ask yourself one question: is this the air the radiator was designed to breathe? If it feels like a hair dryer, the box is the problem. And the box, thankfully, is the cheapest thing on the site to fix.
Email：[email protected]
Work：Monday to Friday,9:30-18:30
