Steam Turbine Inspection: What to Check, When, and What the Findings Mean

There is a version of turbine inspection that consists of walking around the machine with a torch, signing a form, and going home. It happens more often than the industry admits, and it is worse than doing nothing, because it produces a document that says the machine was inspected.

A real inspection produces numbers. Clearances measured against recorded limits, deposit samples with a chemistry result, photographs of specific stages with a scale in frame, and a written comparison against what the same inspection found last time. The value of an inspection is not the looking. It is the record it leaves for the engineer who has to decide, two years from now, whether the machine can run another cycle.

This article sets out what a steam turbine inspection should actually cover at each level, what the common findings mean, and where inspections most often go wrong. It sits alongside our steam turbine maintenance schedule, which covers when these inspections fall due, and our guide to steam turbine overhauls, which covers the work that follows from them.

The Three Levels of Inspection

Inspections are not a single activity, and confusing the levels is how scope disputes start halfway through an outage.

Running inspection is what can be assessed with the machine in service: vibration signature, bearing temperatures, performance parameters, gland leakage, oil condition, and the behaviour of the governing system under load change. It requires no shutdown and it is the only inspection that sees the machine doing its actual job. Its limitation is that it infers internal condition rather than observing it.

Borescope or minor inspection opens access points without lifting the casing. A borescope through the inspection ports gives a view of blade condition, deposits, erosion and any foreign object damage in the accessible stages. Bearings can be lifted, clearances measured and the thrust assessed. This is the workhorse inspection, and on a machine with good running data it is frequently sufficient to justify another cycle.

Major or open inspection lifts the upper casing and exposes the full rotor and stationary path. Every stage becomes visible, the rotor can be lifted for runout and balance checks, diaphragms can be examined and measured, and the full clearance set can be taken. This is the only inspection that sees everything, and it is expensive enough that its scope should be planned rather than discovered.

The mistake worth avoiding is treating these as a ladder that must be climbed in order. A machine with a clear vibration change and rising bearing temperature does not need a borescope inspection to confirm there is a problem. It needs the casing off.

Before Anything Opens: The Baseline Review

The most productive hour of any turbine inspection is spent before a single bolt is undone, reading the history.

Pull the vibration trends for the past two years and note where they moved and what else was happening at the time. Pull the oil analysis history and look at particle count, water and acid number as trends rather than as individual results. Pull the performance data and calculate whether heat rate or steam rate has drifted. Pull the operating log and count the starts, the trips and the hours.

Then read the last inspection report and extract every measured value in it. Those numbers are the comparison set for everything you are about to measure. An inspection with no predecessor to compare against produces a snapshot; an inspection with one produces a rate of change, and the rate of change is what determines how long the machine has.

Finally, confirm what spares exist and what their lead times are. This is not a formality. The scope of an inspection should reflect what could realistically be repaired within the outage window, because discovering a nine-month lead time item on day four of a two-week outage changes the entire plan.

The three inspection levels and the limit of each. A machine with a clear vibration change does not need a borescope to confirm there is a problem. Source: MIMAH engineering practice.

The Bearing Inspection

Bearings are where most turbine inspections earn their money, because bearing condition is both readable and consequential.

Journal bearing babbitt should be examined for wiping, cracking, fatigue pitting, and embedded debris. Wiping indicates the oil film failed, which points at oil supply, oil temperature, load, or alignment. Fatigue cracking in a regular pattern points at dynamic loading. Embedded debris means the filtration is not doing its job, and it will recur unless the source is found.

Clearances should be measured with the bearing in place using lead wire or a dial indicator lift check, and recorded to two decimal places. The number to compare against is both the manufacturer's limit and the last recorded figure. A bearing at 0.30 mm against a 0.35 mm limit is acceptable today; a bearing that was 0.22 mm two years ago and is now 0.30 mm is wearing at a rate that will exceed the limit before the next inspection, and that is a different conversation.

The thrust bearing deserves disproportionate attention. Thrust wear is progressive and largely silent until it is not, and thrust failure is among the fastest routes to a destroyed machine because it allows axial contact between rotating and stationary components. Measure axial float, inspect every pad, and check the pad thickness against the drawing rather than against how the pads look. Discoloured or unevenly loaded pads indicate a thrust distribution problem, which usually means either a misaligned coupling or an unexpected axial force from the steam path.

Oil deflectors and seals should be inspected while the bearing is open. Oil leakage past a deflector is a housekeeping annoyance; oil ingress into the gland area is a genuine problem.

Alignment, Coupling and Foundation

Alignment is checked cold and interpreted against the expected thermal growth, and this is where a lot of inspections quietly go wrong: a machine aligned perfectly cold may be badly aligned hot, and the target offsets have to come from the manufacturer's thermal growth data rather than from an assumption of zero.

The coupling should be inspected for wear, fretting at the hubs, bolt condition and, on flexible couplings, the condition of the flexing elements. Fretting corrosion at a hub fit is a sign of relative movement that should not be occurring.

Foundation and baseplate inspection looks for cracked grout, loose hold-down bolts, soft foot and any evidence of movement. These problems develop over years and are usually visible long before they show up in vibration data. Soft foot in particular is worth checking properly with the bolts released one at a time, because it distorts the casing and produces alignment readings that cannot be corrected by shimming the obvious way.

Piping strain is the related item and the one most often missed. Disconnecting the main steam and exhaust connections and measuring how far the flanges move tells you whether the piping is pulling the casing out of position. A machine that will not hold alignment usually has a piping problem, not an alignment problem.

The Steam Path

Where the casing is open, or where borescope access allows, the steam path inspection is looking for four things.

Deposits on blades and nozzles reduce efficiency and change the flow area. The pattern matters: deposits concentrated in the early stages usually indicate carryover of dissolved solids from the boiler, while deposits further along point at different volatility. Sample them. Deposit chemistry identifies the upstream problem, and a steam path cleaned without fixing the source will simply foul again.

Erosion at the later stages is caused by water droplets in wet steam, and it appears as leading edge thinning and roughening on the last-stage blades. It is progressive and it eventually threatens blade integrity. Erosion at the early stages is a different mechanism, usually solid particle erosion from oxide scale shed by superheater or reheater tubing, and it produces a characteristic pitted appearance on the nozzle and blade leading edges.

Mechanical damage from foreign objects is usually obvious and usually traceable. The important part of the finding is not the damage but the source, because a machine that has ingested one object will ingest another unless somebody works out where it came from.

Cracking, particularly at blade roots, shrouds, tenons and in the rotor itself at stress concentrations, is the finding that changes everything. Any suspected crack indication needs non-destructive examination rather than a visual judgement, and the method depends on the location and material. This is not an area for interpretation by torch.

Diaphragm condition, nozzle block erosion, and the state of the labyrinth seals complete the picture. Seal clearances that have opened up are a direct efficiency loss and a common reason a machine has drifted off its performance baseline.

What each bearing finding actually points at. The finding is the symptom; the cause is upstream and will recur unless it is found. Source: MIMAH engineering analysis.

Valves, Governing and Protection

Stop valves, control valves and non-return valves should be stroked, inspected for seat condition and stem wear, and checked for deposits that would impair closure. A control valve with a worn seat leaks steam past it, which affects both efficiency and the ability to control speed on a load rejection.

The governing system deserves functional inspection rather than just visual inspection. Response to a step change, stability at part load, and the behaviour of the servo system all indicate whether the machine will handle a disturbance safely.

Overspeed protection is inspected and tested as a system. The mechanical bolt or ring, if fitted, should be examined for the condition of its spring and its freedom of movement, and the electronic channels should be verified individually. The turbine trip block and its associated solenoids and pilot valves need proving, and a trip test that exercises the whole chain is worth considerably more than a signal injection.

What the Findings Actually Mean

The hardest part of turbine inspection is not gathering the findings but weighing them, and the weighting depends on rate rather than on state.

A clearance inside limits with a low rate of change supports running to the next scheduled interval. The same clearance with a high rate of change does not. A single erosion finding at one stage is monitored; the same erosion progressing measurably between inspections is a repair decision. A deposit layer is cleaned; a deposit layer that returned within a year is a water chemistry investigation.

The findings that stop everything are relatively few: any confirmed crack indication in a rotating component, thrust bearing damage, evidence of rubbing between rotating and stationary parts, and any indication of water ingestion into the steam path. Each of these has a mechanism that will not resolve on its own.

Everything else is a judgement built from the rate of change, the consequence of failure, the availability of spares, and the operating regime the machine faces next. That judgement is far easier to make well when the inspection recorded numbers rather than adjectives, which is why the recording standard matters as much as the inspection scope.

Recording the Inspection

An inspection record that will still be useful in five years needs a consistent structure, and ISO 14224 provides one: a taxonomy for equipment, failure modes and maintenance data that turns individual reports into analysable history. Where a plant is building a condition monitoring programme around these inspections, ISO 17359 sets out the general framework for how condition monitoring should be structured.

The practical minimum for each inspection is every measured clearance with its limit and its previous value, photographs with scale and location, deposit and oil sample results, the alignment figures as found and as left, the protective device test results, and a written statement of what changed since last time. That last item is the one most reports omit and the one every future reader needs.

The short list of findings that do not wait for the next convenient outage. Each has a mechanism that will not stabilise on its own. Source: MIMAH engineering practice.

Where Inspections Go Wrong

Recording condition as adjectives. Good, satisfactory and acceptable are not data. The next engineer cannot compare against them.

Inspecting without the previous report. Without a comparison set, an inspection can only find gross defects, and gross defects are usually already known.

Treating the borescope as a substitute for measurement. A borescope shows surface condition well and tells you nothing about clearances.

Scoping the outage around the budget rather than the risk. Deferring the thrust inspection because the outage is running long is a decision that occasionally ends a machine.

Not sampling deposits. The deposit is evidence of an upstream problem, and cleaning it away without analysis discards the only sample you were going to get.

Failing to check piping strain. Alignment that will not hold is usually being pulled out by pipework, and no amount of shimming fixes that.

Frequently Asked Questions

How often should a steam turbine be inspected internally? Borescope or minor inspection typically falls on an annual or biennial cycle depending on service severity, with major open inspection at the overhaul interval. The right answer for a specific machine comes from the manufacturer's requirement adjusted for start count, steam purity and operating history rather than from a general rule.

Can a borescope inspection replace opening the casing? For a machine with stable running data and no adverse findings, often yes, and it is far cheaper. It cannot replace an open inspection where there is a vibration change, a performance loss, evidence of water ingestion, or any suspected crack indication.

What is the most commonly missed inspection item? Piping strain, followed by soft foot. Both distort the machine, both are checked in minutes, and both cause problems that get misdiagnosed as alignment or balance faults for years.

What findings justify stopping the machine immediately? Confirmed cracking in a rotating component, thrust bearing damage, evidence of rub between rotating and stationary parts, and signs of water ingestion. Each of these has a failure mechanism that does not stabilise on its own.

Inspection Is a Comparison, Not a Snapshot

The single idea that separates a useful turbine inspection from an expensive one is that condition on its own means very little. The machine is not trying to tell you where it is. It is trying to tell you where it is going, and that only becomes visible when this inspection is laid against the last one in the same units.

Across four decades of turbine and rotating equipment work, from overhauls in Nigeria to root cause investigations at White Nile Sugar in Sudan, the inspections that changed outcomes were rarely the ones that found something dramatic. They were the ones that found a clearance moving faster than it should, on a machine everybody thought was fine.

Planning an inspection or weighing up what one found? Talk to our engineering team. We will scope it against your operating history and give you findings in numbers, with a clear statement of what changed and what it means for your next running cycle.