We spend thousands of hours monitoring equipment during steady-state operation. We watch trends. We take vibration readings. We feel good when the line is flat.
But that's not where a lot of the critical damage starts.
I've worked on enough hot pumps and compressors to know this: a lot of the chronic bearing and seal problems trace back to what happens during start-up and shutdown. I see it all the time—guys treat a start-up like a hurdle to clear. If the pump survives those first few seconds, they think they're done.
They're not. The damage has already started.
What Equipment We're Talking About:
This is especially critical for hot, high-speed rotating equipment—boiler feed water pumps, centrifugal compressors, steam turbines, process pumps operating above 200°F. Equipment with hydrodynamic bearings or precision rolling-element bearings running in demanding service.
If you're working on this kind of equipment, the first minutes after you hit start are the most unforgiving period for bearings and seals compared with when it's just running.
What's Actually Happening (That You Can't See):
When a machine is sitting cold and idle without lift or turning gear, clearances contract and oil films relax back toward static conditions rather than full hydrodynamic separation.
You hit start. You're asking that machine to go from zero to 3,600 RPM in seconds. You're asking physics to catch up while the rotor is already spinning.
Here's what's happening inside the casing:
For hydrodynamic bearings, the oil wedge is building to full film. At 0 RPM, there is no full hydrodynamic separation. For those first few rotations, you're in boundary or mixed lubrication—some load carried by fluid film, some by metal-to-metal contact until speed and oil flow create full film pressure.
For rolling-element bearings, start-up combines high load, low speed, and marginal lubrication. When that rides on top of impact, misapplied load, or marginal fits, you can get raceway indentations and early fatigue. The damage might not show up as vibration for months.
Thermal shock: The rotor heats up faster than the casing—differential expansion. The shaft grows axially while the housing is still cold. If your clearances were tight, they just got tighter.
Pipe strain: Connected piping is heating up and pushing on nozzles. The casing is distorting. Your perfect alignment is moving.
It's chaos trying to find order.
Why the Machine Looks Perfect—and Still Gets Damaged:
This part trips up young technicians every time.
You hit start. Vibration is low. Temperatures stabilize. Pressure looks good. You sign the permit and walk away.
You think you won.
Damage doesn't always scream. Sometimes it whispers.
In a worst-case cold-start scenario, you can wipe enough babbitt off a bearing land to speed up fatigue and cut bearing life. You might crush a seal face, creating a microscopic leak path. You might lightly damage a race from shock loading.
The machine will run. The DCS screen will show green. But early-stage damage, especially localized or microscopic wear, won't show as significant vibration change until later in the degradation curve.
The failure shows up six months later. Everyone blames the seal manufacturer.
No one looks back at the start-up.
A Start-Up I Got Wrong (2003):
I learned this on a boiler feed water pump overhaul.
We had just finished a major rebuild. Operations were pushing hard—they needed water yesterday.
The procedure called for a two-hour warm-up cycle to let the casing temperature equalize and manage differential expansion. We gave it forty-five minutes.
I put my hand on the casing. It felt warm. I thought, "Good enough."
We started it up. It ran smoothly. Vibration was beautiful. I went home feeling like a hero.
Three months later, the inboard bearing failed catastrophically. The shaft was scored deeply.
The RCA showed multiple contributors. The compressed warm-up was one factor. The differential expansion likely caused the wear ring to touch the impeller eye just enough to create a rub. That rub created localized heating. The oil degraded. The bearing died.
I saved an hour on Tuesday. I lost the pump in May.
Start-Up Behaviors That Quietly Kill Equipment:
I see these in many plants:
Starting before the oil temperature is at spec. If the manual says 100°F minimum, 90°F is not close enough. Viscosity at operating temperature strongly influences film thickness and bearing life. Cold oil might have higher viscosity, but it doesn't flow or distribute properly.
Rushing the soak period. Thermal growth takes time. You can't rush heat transfer. Controlled warm-up periods exist to manage internal clearances and prevent rubs.
Loading immediately. Ramping a compressor to full load before metal temperatures stabilize is asking for a rub or seal contact. Staged ramping of speed and load is in the procedure for a reason.
Trusting permissives blindly. Just because the light is green doesn't mean the machine is mechanically happy. Instrument logic can only see what is sensed. Many failures happen with all permissions satisfied.
Treating it as an Ops business. Start-up is a mechanical event, not just an operational one. Improper operating practices—including rushed start-ups—are consistently listed as a major root-cause category for pump and turbine failures.
Why This Is a Leadership Problem:
We blame operators for pushing buttons too fast. We blame technicians for not catching it.
This is a leadership failure.
When a Plant Manager stands in the control room tapping their watch, they're actively encouraging a bad start-up. When we prioritize tons per hour over asset health during that first hour, we're making a conscious trade between short-term output and long-term equipment life.
Technicians know when a machine needs more time. They know when the warm-up isn't done. But if the culture says "Production is King," they stay quiet.
Organizational culture, leadership pressure, and misaligned incentives are core drivers behind poor maintenance practices and premature failures. Rushed start-ups are rarely accidental. They're enabled.
What I Do Differently Now:
After that boiler feed pump disaster, I changed my protocol.
I verify oil temp personally. I don't look at the DCS screen. I use a local gauge or temperature gun. I confirm we're within the OEM temperature and viscosity range.
I stand at the coupling during ramp-up. I use a handheld vibration analyzer and an ultrasonic probe. On lightly instrumented machines, you often pick up the sound or feel of a rub before it shows clearly in routine vibration data.
I never rush the soak. If the procedure says 2 hours, we wait 2 hours. Differential expansion is real. I don't care who's yelling.
I watch motor amps carefully. If the current flickers or hangs high during start, something is binding or misaligned. Amps are a fast, simple indicator of mechanical load.
I stay for the first hour. I never walk away until thermal equilibrium is reached and vibration has stabilized under normal load.
The Real Cost:
The repair bill is just the surface cost.
The real cost is the erosion of reliability and trust in your assets.
In many demanding hot or high-energy services, repeatedly shocking equipment at startup can significantly reduce effective life, turning otherwise long-life pumps into chronic bad actors. You create a cycle of rework that buries your maintenance budget.
You lose trust in your equipment.
Your 5-Minute Start-Up Rule (Use This Next Week):
Pick one machine that's coming online. Follow this checklist for the first 5 minutes:
Stay at the machine. Don't walk away.
Confirm oil temp is within OEM range (verify locally, not just DCS).
Listen for rubs. Use ultrasound or just your ears at the coupling.
Watch starting and running amps. Look for spikes, flickers, or high hang.
Do not load fully until temperatures and vibration stabilize.
Time the soak period. Don't guess. Use a clock. Give the machine the full warm-up time specified in the manual.
It feels like you're wasting time. In reality, you're buying extra life and stability from that machine.
Start-up tells the story. Failures just finish it.
See you next Friday.
— The Maintenance Edge
What's your experience with start-up failures?