By David Placzek and Dr. Lukas Hafner, Deepfluid

Typically, oil condition monitoring is performed through standardized laboratory tests at specified intervals. A representative sample is taken on-site from the system being monitored and analyzed in the laboratory under controlled conditions. This allows for a detailed analysis of numerous parameters that reflect the condition of the oil, such as viscosity, density, and air release behaviour in accordance with DIN ISO 9120.

However, when the sample is pulled from the equipment, it must travel some distance to the lab. During this transit, the oil sample may lose some characteristics that defined the system in which it was operating. While this does not corrode the integrity of the sample, it may not define an accurate representation of system conditions.

In the laboratory, it is not possible to correlate the oil’s interaction with the system’s behavior, which is characterized by constantly changing process conditions such as pressure, temperature, flow rates and air-contents. As these process conditions change, the measurable properties of the oil also change proportionally, and these properties directly determine the efficiency and service life of both the system and the oil. Comprehensive monitoring of the system’s condition can therefore only be achieved through laboratory analysis in conjunction with field measurements.

This approach allows for direct measurement of how the oil interacts with the equipment and generates data points that were previously unthinkable. This enables operators to make predictions that can extend the service life of the oils and make plant operations more efficient or less prone to errors.

What Gets Measured

It is well known that, during the operation of hydraulic systems and transmissions, air is inevitably though unintentionally mixed into the oil. The air content alters the oil’s properties by creating a multiphase mixture, thereby influencing measurable operating parameters in both the short term (efficiency, NVH, temperature) and the long term (oxidation, additive depletion, oil aging).

The Deepfluid bubble profiling technology combines an intelligent vision module and an intelligent LED system. This captures real-time images of the fluid as it flows through the device. Through the use of computer vision-based image processing, each air bubble is identified, sized and classified on a continuous basis. This allows trends and patterns to be recognized and established. No on-line calibration and constant re-calibration is required for this equipment, and it can work across various types of oils with different viscosity ranges and colors or aging-states.

The Deepfluid optical approach evaluates bubbles within a defined size range of 8 to 500 micrometers and generates time-resolved information such as:

• air content,
• bubble-size distribution,
• bubble count,
• bubble-population dynamics,
• oil-air contact surface / interfacial area, and
• transient air events.

Figure 1. Same Air Content. Different Bubble Behavior. Two fluid states can show the same volumetric air content while differing in bubble-size distribution, bubble count, oil-air interfacial area, and release tendency. Air content alone does not fully describe an oil-air dispersion.

Until now, measuring air content has been possible primarily through indirect analytical methods. In this approach, the conductivity of the oil, excluding air content, was referenced to the conductivity of the oil-air mixture during operation. This allows for the analysis of air content percentages under constant conditions. The biggest problem with this measurement is the change in the oil during continuous operation of the system, since water content, particle content, temperature, and additive content are constantly changing, making continuous measurement during operation impossible.

As shown above in Figure 1, the traditional method of measuring the air volume does not accurately depict what is happening in the oil. The air volume of 0.65% only measures one aspect of the oil. With the direct measurement by Deepfluid, users can get deeper insights and explore another dimension of oil condition monitoring by measuring the bubble diameters during operation and compare it with the same technology in a lab-based air-in-oil analysis. Based on this information, short term behaviour (density change, viscosity change, lubricant film thickness, Air-Intake, Air-Release-Behaviour, thermal conductivity and NVH) as well as long term response (oxidation, additive depletion, mechanical robustness, risk of pitting) can be detected and their respective influence targeted.

A key feature is the availability of so-called Evidence Snapshots. Each calculated measurement point can be linked to an optical image of the fluid at that moment. Engineers can review the underlying image, verify the detected bubble population, and relate an unusual value to the physical condition on which it is based.

This creates point-level traceability between the calculated metric and the visible evidence.

Evidence Snapshots do not replace numerical specifications for repeatability, accuracy, or measurement uncertainty. They add transparent verification and support more informed technical discussion between lubricant developers, test engineers, component specialists, and reliability teams.

Figure 2. From Optical Evidence to Quantitative Bubble-Level Data. Direct optical measurement links calculated air-in-oil metrics to the underlying fluid image. Evidence Snapshots provide point-level traceability between air content, bubble-population data, and the recorded physical condition.

The optical approach has also been demonstrated with visually challenging fluids, including dark, aged, and soot-loaded engine oil. As with any optical method, application limits must be understood. However, Deepfluid’s bubble-level analysis is not restricted to transparent new oils.

The objective is not to replace conventional oil analysis, pressure, temperature, vibration, or standardized air-release and foam testing. It is to add direct evidence about the dispersed air phase and its dynamics.

AIR as a Practical Lab-to-Field Framework

The Deepfluid AIR Framework turns insights from bubble-level evidence into three entirely new practical engineering questions and metrics:

AIR is not an abstract research model. It is a practical structure for planning tests, defining measurement windows, comparing fluids, evaluating component and design variants, analyzing operating states, and verifying corrective actions.

Figure 3. The AIR Framework: Intake, Retention, and Release. The AIR Framework structures air-in-oil behavior as a time-resolved sequence. In controlled testing, aeration duration, temperature conditioning, measurement intervals, and recovery phases can be defined; the same logic can be applied to operating events in testing and field environments.

In the laboratory, air release no longer has to be viewed only as a single endpoint under one fixed condition. Individually defined aeration durations, automated temperature conditioning, and time-resolved optical measurement make it possible to run a fully automated AIR test.

Such a sequence can establish a bubble-level baseline before aeration, follow the bubble population during a defined Air Intake phase, quantify Air Retention after the air supply stops, and measure the Air Release curve over time. The same workflow can connect air content, bubble-size distribution, and bubble-population dynamics with subsequent foam formation and foam decay.

This makes it possible to compare different fluids, additive concentrations, temperatures, or aeration durations within structured, automated test campaigns. The laboratory therefore moves closer to application-related questions without giving up controlled and repeatable conditions.

In testing and field operation, the same AIR logic can be applied to defined operating windows. A cycle may begin at a stable baseline, follow an increase in air content during a load, speed, pressure, or temperature change, quantify how much air remains dispersed, and measure recovery afterward.

The resulting bubble-level metrics can be related to operating data such as temperature, pressure, speed, load, flow, efficiency, vibration, or noise.

The conditions are not identical across lab, testing, and field environments but the measurement logic is.

One Shared Data Language – From Lab to Field

Particularly for OEMs, it is important to be able to transition the product from the lab to the testing phase, then into the field. In the lab, they can control the operating conditions, study the air intake and release, dispersion and formulation effects. Afterwards, they can relate bubble behaviour to operating conditions, design changes and system response. Finally, they can execute in the field and track the changes over time to support any root cause analysis for the future and confirm improvements. This is a movement from controlled conditions in the lab to dynamic conditions in testing to finally real-world conditions in the field.

Typically, conventional sensors will give parameters such as a change in dielectricity, a foam tendency, some noise or vibration and an oil condition change just as a result, without knowing the root cause. However, with Deepfluid, they are able to actually make physical behaviour visible and directly explainable. The size and shape of a bubble can be seen, classified and quantified. This allows for the actual oil-air contact surface area to be determined, and this can be trended over time to establish patterns.

What Recurring Patterns Can Bubble-Level Data Reveal?

Direct optical measurement does not identify a root cause on its own. Its practical value lies in revealing repeatable physical patterns that can be compared with operating conditions, representative baselines, and similar systems.

Examples include:

• An increasing population of small bubbles under steady load may be consistent with continuous air ingress or churning.
• Recurring air-content spikes synchronized with pump starts, pressure drops, or speed changes may point to an event-related source of Air Intake.
• A shift toward larger bubbles following a load or pressure transition may reflect bubble expansion, coalescence, or the beginning of Air Release.
• A slow return to baseline after an operating event indicates that air remains retained in the fluid-system combination or is released only gradually.
• Similar Air Content with different bubble-size distributions, bubble counts, or oil-air interfacial areas shows that the physical state of the dispersion is not necessarily the same.
• Different AIR profiles under comparable operating conditions can help distinguish normal system behavior from a machine-, component-, or environment-specific deviation.

Before-and-after measurements add another practical dimension. By repeating the same operating cycle after a change to a seal, reservoir, component, fluid, or control strategy, engineers can verify whether the intervention altered Air Intake, Retention, or Release behavior.

These observations should be treated as investigation signals rather than automatic diagnoses. Their meaning becomes clearer when bubble-level evidence is evaluated together with pressure, temperature, load, speed, flow, vibration, noise, and a representative baseline.

One Portfolio for Controlled and Dynamic Evidence

The Deepfluid portfolio applies this methodology through three connected solutions.

Figure 4. One Measurement Logic from Lab to Field. From Lab to Field does not mean that laboratory, test-rig, and operating conditions are identical. It means that the same Air Intake, Retention, and Release logic—and the same bubble-level metrics—can be applied across different environments and compared within one engineering workflow.

Deepfluid Air-in-One Lab Analyzer

The Deepfluid Air-in-One Lab Analyzer combines air release, foam, and time-resolved bubble behavior within one integrated and automated workflow.

Defined aeration, fluid handling, automated temperature conditioning, optical measurement, foam observation, data transfer, and cleaning can be connected into repeatable test sequences. A fully automated AIR test can capture the complete progression from baseline through Air Intake and Retention to Air Release.

Throughout this sequence, the system measures the physical development of the oil-air dispersion rather than only recording a final release time or foam volume, including Air Content, Bubble Size and Bubble Size Distribution, Bubble Count and Bubble Population, Oil-Air Interfacial Area, and time-resolved Intake, Retention, and Release behavior.

The Air-in-One Lab Analyzer also supports automated test campaigns. Lubricant and additive developers can compare fluid candidates, formulation variants, additive packages, antifoam concentrations, temperature profiles, or aeration durations using the same test logic and evaluation structure.

This makes it possible to investigate not only whether a fluid meets a defined air-release or foam specification, but also why different formulations produce different Air Intake, Retention, Release, and foam responses. Engineers can examine how the bubble population develops before visible foam forms, how much air remains dispersed after aeration stops, and how temperature or formulation changes influence the subsequent recovery.

The approach is not intended to replace standardized ISO or ASTM air-release or foam tests. These methods remain essential for reproducible specification checks and lubricant qualification. The Air-in-One Lab Analyzer adds a complementary, process-aligned R&D perspective that goes beyond a single pass/fail value.

It enables lubricant developers, test engineers, and technical decision-makers to investigate the mechanisms behind Air-in-Oil behavior, compare formulations under application-related conditions, and develop a more complete understanding of Air-in-Oil Contamination before it becomes a field troubleshooting issue.

Standardized tests confirm whether a requirement is met. The AIR workflow helps engineers understand how the result develops—and how that behavior translates from Lab to Field. It is not a replacement for the standardized tests but rather a tool to understand Air-in-Oil Contamination from an R&D perspective.

Deepfluid Optical Inline Sensor

The Deepfluid Optical Inline Sensor transfers the AIR Framework into dynamic test and operating environments.

Instead of recording only a single air-content value, it can capture complete Intake, Retention, and Release behavior over defined time windows. Engineers can observe when a bubble population begins to form, how rapidly it develops, which size classes dominate, how much air remains dispersed after the operating state changes, and how quickly the system returns toward its baseline.

This adds a time-resolved view of transient events. A load change, speed ramp, thermal transition, pressure drop, component-switching event, or start-stop cycle can be evaluated as a complete AIR sequence rather than as an isolated data point.

The sensor can be used in representative inline or bypass configurations, subject to application-specific review. All device configurations have also been developed for demanding high-pressure applications above 150 bar.

When bubble-level data are combined with temperature, pressure, speed, load, flow, efficiency, vibration, or noise, the measurement supports direct comparison between fluids, components, machine variants, and operating conditions. It can also be used to verify whether a design change or corrective action altered the measured Air Intake, Retention, or Release behavior.

The sensor does not automatically diagnose a root cause. It records how the dispersed air phase responds to a defined event, operating state, or intervention and provides evidence for a more focused engineering assessment.

visiQ by Deepfluid

visiQ by Deepfluid provides the common comparison and reporting layer.

Devices and measurement sessions can be assigned to projects, enabling structured data management across development programs, test campaigns, and field investigations. Evidence Snapshots remain linked to the corresponding measurement points and operating context.

The platform supports machine-to-machine, system-to-system, component-to-component, fluid and formulation, and before-and-after comparisons.

For example, the same machine and lubricant can be evaluated under different ambient conditions or load cycles. Conversely, different components can be compared under the same operating profile, or similar machines can be benchmarked across locations. This helps engineering teams distinguish more systematically between fluid-related, component-related, system-related, and environment-related differences.

visiQ also supports automated reporting, reducing the effort required to compile recurring test results, before-and-after comparisons, and structured project summaries.

Case Study – From Abnormal Air-in-Oil Data to a Targeted Seal Investigation

Deepfluid’s direct measurement system can be used not only to analyze the interaction between oil and air during operation but also to monitor the overall system behavior and the reliability of the installed components.

The leak-tightness of hydraulic circuits is essential and critical for the safe operation of these systems. Leaks can allow air and particles to be drawn into the system under negative pressure and oil to be forced out of the system under positive pressure. The presence of air significantly alters operating behavior by changing viscosity, density, fluid level, and lubricating film thickness. These factors can lead to damage such as pitting, scuffing, and micro-dieseling.

If an operator or service technician frequently inspects an application’s oil tank—either directly or through large sight glasses—high air content can be detected by significant cloudiness in the oil.

However, if the application’s oil tank is located in a hard-to-reach position, operates autonomously, or is only accessed during shutdown, such extreme conditions are detected very late, in the laboratory—if at all—before costly damage occurs. This is the case with the operation of wind turbines. Although speed, torque, temperature, particle content, and potential oil leaks are detected, suction-side air ingress, for example, cannot be detected.

Figure 5A. Recurring Air-in-Oil Deviation During Gearbox Operation. A synchronized view of air content, pressure context, and bubble population illustrates how recurring events can support a targeted investigation of possible suction-side or sealing-related air ingress.

In the case study presented, a defective shaft seal was detected through direct measurement on a supply pump for the injection lubrication system of a wind turbine, based on an iteratively and periodically occurring very high air content and loud noise. A minor issue that can have serious financial consequences.

Risks posed by excessive air content and their costs, using a 2.5 MW turbine as an example:

• Higher operating temperature, which requires additional cooling
• Increased cooling capacity (between 4.9 kW and 9.4 kW) due to reduced thermal conductivity (0.14 W/(mK) → 0.125 W/(mK)), costing between 7k€ and 10k€ per year
• Change in friction conditions in conjunction with increased cooling capacity: 43k€–82k€ per year
• Risk of faster oil aging due to accelerated oil oxidation and thermal oil oxidation: 1 additional oil change (24k€–60k€)
• Total mechanical failure of the main gearbox renders the entire system uneconomical.

In this component, there was persistently high air content which we are able to identify and link to abnormal ingress of air into the lubrication system. However, we also saw the pressure drop during pump operation. This was an indicator for suction-side or sealing-related air ingress. We also noticed a dense bubble population which indicates critical oil-air dispersion under operating conditions. There was a deviation from a similar gearbox indicating that this was a system-specific malfunction rather than normal behaviour.

If these were not identified at this early stage, the equipment would run the risk of micro-dieseling, cavitation, oxidation and temperature increase. This would lead to mechanical damage in the gearbox or oil supply components, eventually leading to reduced lubrication reliability and accelerated wear.

The 360 Approach – One Shared Data Language Across the Lubrication Value Chain

At Deepfluid, they have adopted a 360 approach where they can assist all the stakeholders involved in the lubricant industry as it relates to the oil being in the equipment. It connects formulation development, laboratory testing, component testing, system validation, field operation, maintenance, troubleshooting, and verification of corrective actions.

With the 360 approach, various stakeholders can be involved to ensure that the lubricant is fully assessed in different situations, from the testing and development of the lubricant to its actual application in the component then finally to the end user by ensuring they get the results they need.

An additive supplier may investigate formulation effects. A lubricant manufacturer may compare air-release and foam behavior. A filter or seal supplier may study aeration or air ingress. An OEM may correlate bubble behavior with efficiency, thermal management, or NVH. An operator may investigate an abnormal field deviation. An external oil laboratory or research institution may provide controlled reference analysis.

The questions differ, but the underlying air-in-oil metrics can remain comparable.

The 360 approach ensures that all aspects are taken into consideration for the oil, from the lab testing to the field development, with all the stakeholders involved. This guarantees that the final product is reliable and the user should have a valuable experience.

From Measurement to Action

Start measuring the air in your system. The most useful air-in-oil measurement is not the one that produces the largest number of parameters. It is the one that supports a better engineering process.

• Define a representative baseline.
• Detect a meaningful deviation.
• Interpret it together with fluid, component, and operating context.
• Investigate the most plausible mechanism.
• Verify whether the intervention changed the measured behavior.

Many practitioners only view air as an issue when they see foam. By this time, it is too late and damage has already occurred to the system. Even small bubbles can have system consequences, as shown in the diagram below.

Figure 6: System-level consequences which can be detected by the presence of small bubbles through Deepfluid’s technology

Many operators are not aware of the impacts of air-in-oil and quite often, it is labelled as something else. However, it usually shows up as a foam problem, unexpected NVH, control instability, temperature problem, cavitation problem, pump problem or an oil problem. The key is to monitor these effects in different settings.

Starting with studying air release, dispersion and formulation effects under controlled conditions in the lab. Moving to the testing phase where bubble behaviour is related to operating conditions, design changes and system response. Then finally to the field where changes can be tracked over time to support root-cause analysis and confirm any improvements.

This is the benefit of using the Deepfluid technology as it can capture data from the various phases to bring about actionable insights to improve the reliability of operating systems.