Validating DOCA Sensor Readings Against Gravimetric Oil Sampling
Reliable oil monitoring in high-purity compressed air requires a defensible reference method. An online optical sensor can provide continuous information about oil contaminants in liquid, aerosol, and vapor forms, while gravimetric sampling offers a laboratory-based measurement for comparison.
Validation therefore means more than checking whether two numbers appear similar. It requires a defined measurand, controlled sampling conditions, synchronized measurements, and statistical analysis that accounts for uncertainty, concentration range, and the physical form of the oil.
For the DOCA sensor, this process can demonstrate how optical readings perform across pharmaceutical, hospital, automotive, chemical, textile, electronics, and clean-room applications.
Define the measurement objective
The first step is to state exactly what the sensor and reference method are measuring. Oil may be present as deposited liquid, suspended aerosol, or vapor. A gravimetric result from a filter or impinger may represent one fraction, whereas an optical sensor may respond to particles or droplets passing through its measurement zone.
The validation protocol should therefore specify the reporting unit, such as milligrams per cubic metre, the sampling duration, pressure and temperature conditions, and the oil fraction under investigation. If vapor-phase hydrocarbons are included, a sorbent-based or solvent-extraction reference may be needed instead of a simple membrane filter.
ISO 8573-1 can be used to define compressed-air purity classes, while the relevant parts of ISO 8573 provide guidance for sampling and contaminant measurement. The selected procedure should match the intended application and the oil type used in testing.
Build a controlled test arrangement
A suitable test rig should generate stable compressed air with a known and adjustable oil load. It should include clean-air operation, several contamination levels around the expected working range, and conditions that represent the intended installation. Pressure, flow, relative humidity, temperature, and oil type must be recorded throughout each run.
The DOCA sensor and the gravimetric sampling point should be installed as close together as practical. Their inlet geometry, tubing length, residence time, and flow resistance should be documented. Long or narrow tubing can cause oil deposition and create a difference between the material reaching the sensor and the material collected by the reference sampler.
Before each test, the system should be flushed until background readings stabilize. Blank samples and zero-air measurements help identify contamination from the compressor, test lines, filters, solvents, or laboratory handling.
Collect matched reference samples
Gravimetric sampling normally captures oil on a preconditioned filter, collection medium, or vessel. The medium is weighed before and after sampling under controlled temperature and humidity. The mass increase, corrected using field blanks, is divided by the sampled air volume:
[ C_\text{oil}=\frac{m_\text{sample}-m_\text{blank}}{V_\text{air}} ]
The balance resolution, filter conditioning time, sampling flow, and uncertainty of the flow meter should be included in the measurement record. At low concentrations, electrostatic effects, moisture uptake, and handling residues can be comparable to the oil mass itself.
Replicate samples are essential. They reveal the repeatability of the gravimetric procedure and help distinguish sensor disagreement from reference-method variability. Where vapor is relevant, the protocol should verify collection efficiency and breakthrough rather than assuming that all hydrocarbons are retained.
Synchronize sensor and sampler data
An optical reading should be paired with the correct gravimetric sample interval. Record the sensor’s time stamp, averaging period, response delay, alarm state, and calibration status. If the instrument reports a rolling average, the reference result should be compared with the equivalent average over the same period.
Sampling duration should be long enough to collect a measurable oil mass without allowing the process conditions to drift. Several repeated runs at each concentration provide stronger evidence than a single long sample. Tests should include both increasing and decreasing contamination levels to identify hysteresis, surface conditioning, or memory effects.
The comparison must also account for differences in particle size and oil distribution. A sensor that responds strongly to aerosol droplets may not show the same behavior when the same total oil mass is present mainly as vapor.
Analyze agreement and uncertainty
A useful validation dataset includes the gravimetric concentration, DOCA reading, replicate number, operating conditions, and uncertainty estimate for every run. Plotting sensor response against reference concentration can reveal linearity, curvature, saturation, and a practical detection limit.
| Validation feature | Gravimetric reference | DOCA optical measurement |
|---|---|---|
| Measurement principle | Collected mass determined by weighing or chemical analysis | Optical response associated with oil contamination |
| Output | Time-integrated concentration | Continuous or near-real-time reading |
| Main strengths | Traceable reference and established sampling practice | Fast trends, alarms, and process visibility |
| Main limitations | Laborious, delayed, and sensitive to sampling losses | Response may depend on droplet size, oil type, and optical conditions |
| Best validation use | Establish reference concentration and uncertainty | Assess agreement, response time, and operational stability |
Regression analysis should be supported by a bias assessment, such as a difference-versus-mean plot. Report slope, intercept, confidence intervals, repeatability, reproducibility, and the range over which the relationship is valid. Correlation alone is insufficient because two methods can correlate well while maintaining a significant systematic bias.
Uncertainty should combine reference sampling, balance resolution, flow measurement, blanks, repeatability, and sensor variability. Acceptance limits need to be set before testing and linked to the application’s purity requirement.
Control factors that affect optical response
Oil composition can change refractive index, viscosity, droplet formation, and light scattering. Testing should use representative compressor oils or process contaminants, with the oil identity and preparation method recorded. Humidity and water aerosols should be evaluated because condensation or mixed droplets may influence the optical signal.
Routine checks should include zero stability, response to a known challenge, recovery after contamination, and sensor cleanliness. If the instrument is intended for continuous industrial operation, validation should also examine vibration, pressure cycling, temperature changes, and extended measurement drift.
Practical validation recommendations
- Use a traceable gravimetric or chemically verified reference procedure appropriate to aerosol, liquid, or vapor oil.
- Place the sensor and sampling inlet together and verify their flow rates independently.
- Perform blanks, replicates, zero checks, and multiple concentration levels.
- Match sensor averaging windows to the gravimetric sampling periods.
- Report bias and uncertainty separately from correlation and repeatability.
A well-documented comparison gives the DOCA sensor a clear performance profile rather than a single unqualified accuracy value. The results can support calibration decisions, application-specific acceptance criteria, and evidence for industrial deployment.
The DOCA Project can use this validation framework to connect laboratory measurements with real compressed-air conditions. Its technical findings, test results, and application data provide a foundation for demonstrating how continuous optical monitoring complements established gravimetric oil sampling.