Startup is one of the highest-risk phases of a steam turbine outage. Vibration behaviour during the run-up tells you more about the actual condition of the machine than most static inspections will. Understanding what causes startup vibrations — and how to distinguish the different types — is essential for anyone responsible for outage completion and restart.

Why Startup Vibrations Deserve Specific Attention

The run-up from turning gear speed to operating speed typically takes 20–60 minutes, depending on machine size and startup procedure. During this window, several phenomena occur simultaneously: the rotor heats up, oil film stiffness changes, bearing behaviour evolves, and the machine passes through one or more critical speeds. Vibration that seems tolerable at low speed can escalate quickly as speed increases.

What makes startup vibrations particularly demanding is that they can be caused by multiple factors at once, and some of those factors interact. A rotor with mild thermal bow, slightly tight clearances and residual unbalance will behave very differently than one with only a single issue. Getting the assessment right requires understanding which type of vibration you are actually seeing.

The Main Causes

Thermal bow — rotor sag

A rotor that has been stationary for any length of time develops a thermal bow due to gravity and differential cooling between the upper and lower halves of the casing. This bow creates a 1× vibration (once per revolution) that typically reduces as the rotor heats up evenly during the run-up.

If a rotor has not been on slow-roll turning gear for a sufficient period before startup, thermal bow can be severe enough to cause a rub. Standard procedure requires a minimum turning gear period (typically 4–8 hours after shutdown, depending on machine size) before restart. If this is missed or cut short, the risk increases significantly.

Field note

Thermal bow typically shows as 1× vibration during slow roll that decreases steadily as the rotor heats up. If high 1× vibration persists at slow roll and does not reduce after 20–30 minutes on turning gear, investigate before attempting a run-up.

Bearing clearance and alignment

After any bearing work — replacement, shimming, re-babbitting — clearance settings and alignment directly affect vibration during startup. Clearances that are too tight reduce the oil film's ability to dampen vibration and can cause contact between shaft and bearing. Clearances that are too wide increase the risk of oil whirl instability at lower speeds.

Coupling misalignment is another common post-maintenance issue. Even small misalignment after coupling reassembly produces characteristic 2× vibration, which often increases with speed and load. Confirm alignment before startup whenever coupling work has been performed.

Oil film instability — oil whirl and oil whip

Oil film instability is one of the more serious startup vibration causes and is worth understanding in detail. In a journal bearing, the oil film rotates at approximately half of shaft speed. Below a threshold related to bearing stiffness, this film can become unstable and drive vibration at sub-synchronous frequencies — typically 43–48% of running speed. This condition is called oil whirl.

Oil whirl is problematic in itself, but if allowed to continue as speed increases, it can lock at the first critical frequency of the rotor and become oil whip. Oil whip is largely independent of running speed and can persist above the whirl onset speed. It is associated with high amplitudes and is potentially destructive if not addressed.

Warning

Sub-synchronous vibration that appears during run-up and does not reduce as speed increases is a serious warning sign. If amplitude grows rather than stabilises, stop the machine and investigate bearing condition and clearances before continuing.

Residual rotor unbalance

Any work involving rotor components — blade replacement, seal replacement, coupling changes, or removal and replacement of any rotating elements — requires that the rotor be rebalanced before startup. Residual unbalance shows as 1× vibration that increases with speed squared (it is proportional to ω²). High 1× vibration at operating speed that was not present before the outage is a strong indicator of residual unbalance.

If balancing was performed during the outage, verify that the trim balance run was conducted properly and that the final balance condition is documented before the run-up. If the machine was not rebalanced and rotor components were disturbed, this needs to be addressed before startup.

Rub on seals or stationary components

A rotor that contacts labyrinth seals, diaphragm bores, or other stationary components during startup produces characteristic vibration signatures — often a combination of 1× and higher harmonic content, sometimes with sub-synchronous components if the rub creates a temporary thermal bow. A rub can develop and worsen rapidly.

Signs of a rub include a sudden change in 1× phase angle (not just amplitude), combined with changes in vibration character that evolve over time as the rotor heats. If you suspect a rub, stop the machine immediately. Continuing to run through a rub causes progressive damage and significantly worsens the situation.

Critical speed crossing

All steam turbine rotors have one or more critical speeds — running speeds where the forcing frequency matches the natural frequency of the rotor, producing a resonance peak. Passing through a critical speed produces a temporary peak in vibration amplitude that then decreases as speed moves away from the critical. This is expected behaviour and does not indicate a problem.

The concern is when the critical speed amplitude exceeds limits (set by the machine's amplification factor and mode shape), or when the machine cannot pass through the critical (the vibration remains high and does not come back down). The amplification factor at the critical speed is a key diagnostic parameter that should be recorded during each startup and compared to historical baseline.

Data to Collect During Startup

Good startup data is the foundation for any post-startup assessment. Collect and record the following during every startup:

  • Vibration amplitude and phase angle at each bearing — 1× filtered as minimum, broadband overall if available
  • Speed (rpm) vs. time — so you can reconstruct the run-up profile
  • Bode plot (amplitude and phase vs. speed) if your monitoring system supports it
  • Bearing metal temperature at each bearing
  • Lube oil supply pressure and temperature
  • Time from last shutdown to turning gear removal — to assess thermal bow risk

If the machine does not have permanent monitoring installed, portable vibration instruments should be connected before startup following any significant bearing or rotor work.

When to Stop, When to Continue

What you observeLikely causeAction
High 1× at slow roll, decreasing steadily as rotor heatsThermal bowContinue — monitor for further reduction
1× increases proportionally with speedResidual unbalanceMonitor against alarm limits; stop if limits are approached
Sub-synchronous vibration at ~45% of running speedOil whirl onsetInvestigate bearing condition and clearances before continuing
Sudden change in 1× phase angle, combined with sub-synchronous contentRubStop immediately
2× component that increases with speedMisalignmentAssess against limits; do not continue to full load without investigation
Transient amplitude peak followed by reduction as speed increasesCritical speed crossingNormal — record amplification factor
High amplitude at critical speed that does not reduceHigh amplification / other issueDo not attempt to coast through — stop and investigate

Practical Summary

Startup vibrations are rarely a sign of a single isolated problem. In most cases, multiple contributing factors are present simultaneously. The most important things to have in place:

  • A documented pre-startup checklist that confirms turning gear period, alignment checks, oil system condition and bearing clearances
  • Vibration monitoring in place before the run-up starts — not only at the first sign of a problem
  • A defined hold point and shutdown criterion, agreed before startup, rather than improvised during the run-up under time pressure
  • Someone responsible for watching the vibration trend during the run-up and with authority to call a stop if needed
Key point

Keep notes during the startup. Write down what you saw, at what speed, and at what time. If the machine needs to be shut down for investigation, this information will be critical for diagnosis. The startup data window closes the moment the machine is back at operating speed.