Glossary term

Oil Analysis

Diagnostic method that tests lubricant condition, contamination and wear debris to support machinery health, maintenance and release decisions.

Branch: Mechanical Engineering
Glossary type: method
Content: Glossary term
Updated: Jun 25, 2026
Revision: v1.0.0 · reviewed

Definition

method

Oil analysis is a diagnostic method that tests lubricant condition, contamination and wear debris to judge machinery health and maintenance action.

Oil analysis uses laboratory and field measurements such as viscosity, water content, particle count, wear metals, additive depletion, acidity, oxidation, nitration, fuel dilution, soot, ferrous debris and microscopy. It is useful only when the sample location, sampling method, operating hours, lubricant type, makeup oil, filter changes, alarm limits and machine symptoms are tied to an engineering decision.

Oil analysis is the diagnostic use of lubricant samples to understand machine condition. It can show whether the lubricant is still fit for service, whether contamination is entering the system and whether wear debris suggests an active failure mode.

The value is not the laboratory report by itself. The value is the engineering decision that follows: continue operation, resample, filter, change oil, inspect a bearing, reduce load, stop the machine, revise lubrication practice or release the machine after maintenance.

Engineering Role

Oil analysis is common in gearboxes, turbines, compressors, engines, hydraulic systems, stern tubes, pumps, mills and large rotating machinery. It complements vibration monitoring, temperature trends, shaft centerline plots, visual inspection and maintenance records.

A useful oil-analysis program controls:

sampling point and sampling method;
machine operating state and hours on oil;
lubricant grade, batch, makeup oil and top-up history;
filter type, filter changes and bypass events;
alarm limits and trend limits;
maintenance action linked to each abnormal result.

Without this context, oil analysis can create false confidence or false alarms. A dirty sample bottle, stagnant drain sample, recent oil change or unrecorded top-up can make a trend meaningless.

What Oil Analysis Measures

Common tests include viscosity, water content, particle count, ferrous debris, elemental wear metals, additive elements, total acid number, total base number, oxidation, nitration, soot, fuel dilution and microscopy.

Each test answers a different question. Viscosity asks whether the lubricant can still build the intended film. Water and particle count ask whether contamination is attacking the film, additives, surfaces or filters. Wear metals and debris morphology ask whether machine parts are wearing normally or producing abnormal particles. Acidity and oxidation ask whether the oil chemistry is degrading.

No single number proves machine health. A bearing can fail with acceptable viscosity if hard particles enter the contact. A gearbox can show high wear metal after a maintenance event without ongoing damage if the debris is residual and falling.

Viscosity Change

A simple viscosity-change screen is:

\displaystyle \Delta \nu\%=\frac{\nu_s-\nu_r}{\nu_r}100

where $\nu_s$ is sample kinematic viscosity and $\nu_r$ is the reference or new-oil viscosity at the same test temperature.

If a gearbox oil should be:

\nu_r=46\ \text{cSt}

and the sample result is:

\nu_s=58\ \text{cSt}

then:

\displaystyle \Delta \nu\%=\frac{58-46}{46}100=26.1\%

That change is not automatically a shutdown criterion, but it is large enough to demand cause analysis. Possible causes include oxidation, contamination, wrong oil, water, soot, thermal degradation, shearing, fuel dilution or mixing with another lubricant.

Water and Contamination

Water can reduce film strength, corrode surfaces, deplete additives and support microbial growth. A simple mass screen converts parts per million to water mass:

m_w=C_w10^{-6}m_o

where $C_w$ is water concentration in ppm and $m_o$ is oil mass.

For:

C_w=800\ \text{ppm}

and:

m_o=50\ \text{kg}

the estimated water mass is:

m_w=800(10^{-6})(50)=0.040\ \text{kg}=40\ \text{g}

The action depends on lubricant type, machine criticality, free water evidence and source. A one-time water result after washdown is different from a rising trend that follows every shutdown.

Particle and Filter Evidence

Particle count is useful only when sampling location and cleanliness target are controlled. A filter beta ratio is:

\displaystyle \beta_x=\frac{N_{up,x}}{N_{down,x}}

where $N_{up,x}$ and $N_{down,x}$ are upstream and downstream particle counts larger than size $x$ . Approximate capture efficiency is:

\displaystyle \eta_f=1-\frac{1}{\beta_x}

For:

\beta_x=200

the efficiency estimate is:

\displaystyle \eta_f=1-\frac{1}{200}=0.995=99.5\%

If downstream particle counts rise while the beta ratio should be high, the review should check bypass, wrong filter element, damaged seals, poor installation, overload, sampling error or contamination entering after the filter.

Wear-Debris Trend

Wear metals are more useful as trends than as isolated values. A simple rate is:

\displaystyle r_{Fe}=\frac{C_{Fe,2}-C_{Fe,1}}{\Delta t}

where $C_{Fe}$ is iron concentration and $\Delta t$ is operating time between samples.

If iron rises from:

C_{Fe,1}=35\ \text{ppm}

to:

C_{Fe,2}=95\ \text{ppm}

over:

\Delta t=250\ \text{h}

then:

\displaystyle r_{Fe}=\frac{95-35}{250}=0.24\ \text{ppm/h}

The diagnosis still needs particle type, machine material map and symptoms. Fine rubbing wear, cutting wear, fatigue spalls, corrosion debris and maintenance residue do not imply the same action.

Diagnostic Use

Oil analysis is strongest when it is combined with other evidence. Rising iron plus increasing bearing temperature and high-frequency vibration is more serious than rising iron alone. Water plus falling viscosity and emulsified oil points to a different action than high silicon and hard particle contamination. A stable oil report does not clear a machine if vibration, orbit or temperature evidence says a fault is active.

Useful decisions include resampling, filtration, dehydration, oil change, bearing inspection, gearbox borescope, root-cause review, load restriction or planned shutdown. The action should be proportional to machine criticality and trend confidence.

Validation and Release

A defensible oil-analysis record states sample point, sample method, operating hours, oil hours, machine state, lubricant grade, makeup oil, filter history, lab method, result uncertainty, alarm criteria and action taken.

Release should be withheld when the sample is not representative, viscosity has shifted without explanation, water is above the accepted limit, particle count is rising after filtration, wear-debris morphology indicates active damage, the lab result conflicts with temperature or vibration symptoms, or the machine has no baseline for comparison.

Common Mistakes

Do not compare samples taken from different locations as if they were one trend. A drain sample, live-zone sample and filter-debris sample answer different questions.

Do not treat elemental wear metals as a complete particle-size picture. Spectrometric methods can miss large particles, while ferrography and filter patch inspection may reveal severe debris that dissolved-element trends understate.

Do not clear a machine from oil analysis alone. A good report can coexist with misalignment, looseness, overload, oil whirl, poor preload, sensor error or a fault that has not yet generated debris.

Limits

Oil analysis is a decision-support method, not a direct view of every contact. It depends on sampling discipline, lab method, dilution by oil volume, filter capture, debris transport, machine operating mode and historical baseline.

The practical goal is to connect lubricant evidence to a failure mode and an action. A result is useful when it changes a maintenance or release decision with enough confidence to justify the risk.

REF

Disciplines