Oil analysis separates three questions: lubricant health, system contamination, and machine wear. One result rarely answers all three. A representative sample, known test method, unit, applicable criterion, trend, and operating context turn a laboratory report into maintenance evidence.
Changing oil after an alert may temporarily reduce contaminants and wear debris while leaving the water ingress, dirt source, or degrading component untouched. The maintenance response should address the mechanism, capture the as-found condition, and confirm the outcome with a comparable follow-up sample.
Sampling location defines the observation
A bottle filled from the bottom of a stopped reservoir describes settled material. A sample taken downstream of a filter describes a different condition from fluid reaching lubricated components. Trending requires a consistent location, operating state, stabilization time, temperature, and collection procedure.
The PNNL operations and maintenance guide recommends sampling from an active low-pressure line upstream of filtration where the system design permits, using the same point each time. The final arrangement depends on the system and approved procedure. Clean containers, proper flushing, identification, and controlled transport are also parts of measurement quality.
Three evidence families
| Family | Question | Examples |
|---|---|---|
| Lubricant | Can the fluid still perform its lubrication function? | viscosity, acid number, oxidation, and additives |
| Contamination | Has a system boundary been breached? | water, particle count, silicon, and mixed fluids |
| Machine | Is wear material present and which mechanism is compatible? | iron, copper, chromium, morphology, and ferrography |
Elevated iron does not identify a component by itself. Elements may originate from several alloys, external contamination, or recent maintenance. Large particles may be underrepresented by some spectrometric methods. Test selection should reflect the machine, lubricant, criticality, and failure hypothesis.
Program architecture
- Register equipment, reservoir, fluid, volume, sampling point, and collection method.
- Select tests by function and failure mode instead of assigning one package to the entire plant.
- Establish a baseline after commissioning or another verified condition.
- Test new-oil batches when they form part of the decision reference.
- Build limits from specifications, trend, and validated operating experience.
- Define when to repeat, confirm, notify, or intervene.
- Connect the report with the order, material, as-found condition, and follow-up sample.
Competency is another control. ISO 18436-5 addresses qualification and assessment of personnel performing laboratory-based lubricant analysis for machinery condition monitoring. Collector, analyst, reliability engineer, and decision authority may be separate roles, with traceability among them.
Completed case: gearbox RED-22
All values are instructional and do not represent a customer, benchmark, or PM Run performance. RED-22 drives a feed conveyor, operates 20 hours per day, and contains 85 liters of ISO VG 220 oil. A permanent sampling point is located on the active return line upstream of the filter under the plant design.
Baseline and sampling condition
| Field | Case definition |
|---|---|
| Operating regime | 72% to 80% load for at least 40 minutes |
| Oil temperature | 58 °C to 64 °C at collection |
| Flush volume | volume specified in the point procedure |
| Viscosity at 40 °C | internal range from 198 to 242 mm²/s |
| Water | internal alert above 300 ppm |
| Particles | historical reference 17/15/12 under the adopted code |
| Iron | stable range from 28 to 45 ppm over six samples |
These limits belong to the teaching case. Viscosity comes from the approved specification. The water criterion reflects the system, oil, and local experience. Cleanliness is interpreted with the test method, location, and application.
Result that opened the investigation
| Test | Previous baseline | August 12, 2026 |
|---|---|---|
| Viscosity at 40 °C | 218 mm²/s | 205 mm²/s |
| Water | 180 ppm | 820 ppm |
| Particles | 17/15/12 | 21/19/16 |
| Iron | 38 ppm | 142 ppm |
| Copper | 7 ppm | 18 ppm |
Viscosity remained inside the case range while water, particles, and iron changed together. The laboratory noted a cloudy appearance. Planning did not open a generic oil-change order. The response rule first required confirmation of identity, sampling point, container, flushing, regime, and recent maintenance.
Sample confirmation
A second sample was taken 24 hours later at the same point and regime using another bottle from a controlled batch. Results were 790 ppm water, 21/19/16, and 151 ppm iron. Repetition reduced the likelihood of bottle contamination. History showed intensive external washdown three days before the first sample.
Inspection found a saturated breather, a damaged cover seal, and water-entry marks. Particle size and quantity justified complementary evaluation. Vibration showed a moderate increase around gear-mesh frequency but did not cross the plant's independent intervention trigger.
Decision and planned order
The risk matrix combined water, iron trend, criticality, and available redundancy. The line could transfer duty to another gearbox within 36 hours. Engineering decided to inspect RED-22, restore contamination boundaries, assess gears and bearings, recondition or replace the oil according to the as-found state, and conduct a controlled test.
The notification recorded results, methods, units, point, operating condition, trend, and hypothesis. Cause remained unconfirmed. The linked order covered isolation, controlled draining, opening, deposit sampling, breather, seal, filter, internal components, cleaning, refill, and testing.
As-found condition
Inspection confirmed free water in the lower region, a saturated breather, and a deformed cover seal. One bearing showed micropitting and ferrous particles compatible with the finding. Gear teeth did not have damage requiring replacement. The approved assessment concluded that filtration would not restore the required condition within the available window, so the oil was replaced.
The bearing, breather, and seal were replaced, and the washdown method was changed. The order recorded the new-oil batch and volume. Cause fields were updated after inspection: water entered through a degraded boundary and the washdown practice contributed. Bearing wear was captured as a confirmed consequence.
Test and verification samples
The operating test checked temperature, sound, vibration, and load. An initial sample after circulation checked service cleanliness but did not enter the long-term trend. The comparable 100-hour sample showed 190 ppm water, 18/16/13 particles, and 46 ppm iron. After 30 days, iron reached 35 ppm and cleanliness returned to 17/15/12.
The verification window lasted 90 days with samples at 100 hours, 30 days, and 90 days. The instructional target required water below 300 ppm, iron returning toward baseline, no recurring cloudiness, and complete operating context for every sample.
SAP PM handoff
Structured values may be assigned to measurement points under the company's architecture. The complete report remains linked to the technical object and event according to document policy. The notification records the deviation and triage. The order organizes intervention, materials, lubricant batch, as-found condition, disposal, and test.
The RED-22 sampling plan gained the fixed point, operating condition, test panel, and post-maintenance rule. The washdown task was revised to protect the breather and seal. History now connects contamination, damage, action, and confirmation.
PM Run can mobilize the route, support planning, and return field confirmations to SAP PM. It does not collect the sample, conduct laboratory testing, automatically interpret the report, provide a sensor, or produce a predictive diagnosis. Plant professionals retain those responsibilities.
Building applicable limits
Limits can come from fluid specification, equipment engineering, laboratory capability, and the trend of a stable population. An absolute alert protects a property. A relative alert detects change. A multivariable rule can elevate priority when water, particles, and wear increase together.
The criterion should name method and unit. Karl Fischer water, crackle screening, and relative saturation answer different questions. Viscosity at 40 °C cannot be compared directly with a value at 100 °C. A cleanliness code requires compatible counting and reporting standards.
Controls after maintenance
Recent repair can generate a temporary run-in signature, disturb deposits, or introduce assembly contamination. The verification plan distinguishes a cleanliness check from the sample used to assess stable wear. Both need labels and timing.
If the first post-work sample remains abnormal, the response considers incomplete flushing, residual debris, active wear, wrong fluid, or another ingress path. Repeated oil replacement without that investigation can hide the signal and consume the evidence.
Decision failures
- Changing sampling points between campaigns.
- Using an unsuitable container or touching the inner cap.
- Sampling immediately after top-up without recording the intervention.
- Ignoring temperature, load, and stabilization time.
- Changing oil and closing the event without removing the source.
- Inferring one component from a single wear metal.
- Mixing laboratory methods in one trend without a series-break record.
- Treating the laboratory report as an automatic work order.
Program indicators
The program follows valid-sample rate, time from result to decision, confirmed alerts, contamination recurrence, interventions with follow-up samples, and failures with an available trend. Rejected bottles and unsuitable points form a separate quality queue.
Alert rate alone does not measure success. A decline can indicate improvement or poor sensitivity. Governance combines coverage, sample quality, confirmation, and maintenance outcome.
Choosing the right discipline
Predictive maintenance combines techniques and decisions into a condition strategy. Industrial equipment inspection structures routes, criteria, and anomaly handling. Oil analysis addresses representative sampling, selected tests, and interpretation of the lubricated system.
Report coherence review
Before triage, the analyst checks whether the results form a physically possible account. High water and cloudiness are compatible observations. Rising iron and larger particles may converge on wear, but spectroscopy and particle counting observe different populations. A conflict calls for a complementary test rather than an average across methods.
RED-22 viscosity remained within its case range, which reduced concern about severe loss of that property in the sample. It did not cancel the water risk. Each test controls one dimension, while the overall condition combines function, criticality, trend, and corroborating evidence.
State-based response plan
| State | Minimum evidence | Case response |
|---|---|---|
| Questionable sample | identity or collection inconsistency | repeat before changing the asset |
| Isolated alert | one parameter outside its rule | confirm and review context |
| Converging alert | two compatible signals | notify and assign priority |
| Confirmed condition | inspection or complementary test | plan correction of the mechanism |
| Post-maintenance | completed work and circulation | sample in the defined window |
The matrix prevents two extremes: ignoring every change until functional failure or opening intrusive work for every alert. State changes as evidence arrives, and the responsible authority remains recorded.
Laboratory and method governance
The technical agreement with the laboratory defines method, quantification limit, unit, turnaround, preservation, and rejection rules. A method change appears in the report and trend. The laboratory may recommend investigation, while plant authority owns priority and work release.
Split samples can support interlaboratory comparisons in selected cases. Differences are evaluated against repeatability, preparation, and equipment rather than choosing the convenient number. Systematic divergence receives a corrective action and follow-up check.
Return-to-trend criterion
The 100-hour result confirmed initial cleanliness while still reflecting recent work. The stable series resumed only after two consecutive samples met operating-context, water, and wear-metal criteria. This rule kept run-in behavior from entering the historical baseline without identification.
Handling a rejected sample
A bottle does not enter the condition indicator when identity, volume, seal, or preservation fails. The quality record retains the reason because silent deletion would hide a route weakness. Planning schedules another collection at the first valid operating condition.
In the exercise, a parallel sample arrived without time and temperature. It was excluded from water and metal confirmation. The event led to a mobile-form rule requiring point, regime, and timestamp before submission. Effectiveness was checked across the next twenty collections.
Economic use without weakening risk control
Test-panel cost is compared with consequence, oil volume, outage opportunity, and value of information. A small noncritical reservoir may favor scheduled replacement, while a large critical system can justify detailed analysis. Economic selection still respects environmental, safety, and governing technical requirements.
Technical references
- PNNL, Operations and Maintenance Best Practices Guide, oil analysis section, accessed August 23, 2026.
- ISO 18436-5:2012, official page, accessed August 23, 2026.
- NIST Handbook 158, sampling procedures, accessed August 23, 2026.
Frequently asked questions
Does a red report require immediate oil replacement?
The response depends on the test, magnitude, trend, criticality, and approved rule. A questionable sample may require repetition first.
Is new oil a clean baseline?
A delivered batch may differ from the internal target. When the decision relies on that reference, sample and document the batch.
What closes the anomaly?
Source correction, documented as-found condition, and a follow-up result within the approved criterion. Fluid replacement without confirmation leaves the mechanism open.
To coordinate plans and orders over the corporate system, review PM Run's operational layer for SAP PM.
