⚡ electrical engineering

Surface Degradation as a Leading Indicator of Marine Rotating-Equipment Deterioration: A Critical Review of Corrosion, Roughness, Tribocorrosion and Condition Monitoring

This paper proposes a surface-aware condition-monitoring architecture for marine rotating machinery by synthesizing literature and re-analyzing experimental data to demonstrate how measurable surface degradation, such as increased roughness and corrosion, serves as a leading indicator of performance decline and functional failure.

Original authors: Nsini Udo

Published 2026-10-07

📖 8 min read🧠 Deep dive

CC BY 4.0

Original authors: Nsini Udo

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the deep, salty waters of the ocean, the metal hearts of ships and offshore platforms face a constant, silent war. These machines, which include the massive pumps that circulate water for cooling or ballast, are built to spin and push fluid, but they are also constantly battered by the chemistry of the sea. Saltwater is not just wet; it is a corrosive agent that attacks metal, while the movement of the fluid itself wears down surfaces, and tiny particles in the water act like sandpaper. For decades, engineers have monitored these machines by listening to their vibrations or watching their power consumption, looking for the loud, obvious signs that something is wrong. By the time a pump starts to shake violently or lose its ability to push water, however, the damage is often already severe, requiring expensive repairs or sudden shutdowns. The question has long been whether there is a quieter, earlier warning sign hidden in the very skin of the machine, a subtle change in the metal's surface that happens long before the machine fails.

A researcher named Nsini Udo has taken a fresh look at this problem, proposing that the condition of a machine's surface is not just a victim of the environment, but a leading indicator of its future health. Instead of waiting for the machine to scream for help through vibration, the new approach suggests we should look at the microscopic landscape of the metal itself. This involves understanding how the smooth, manufactured finish of a pump part slowly transforms into a rough, pitted terrain as it is exposed to salt and motion. The researcher argues that this transformation is the first step in a chain reaction: as the surface becomes rougher and develops tiny pits, it disrupts the smooth flow of water, which in turn reduces the machine's efficiency and eventually causes the vibrations that signal a breakdown.

To test this idea, the researcher did not build a new machine or run a new experiment from scratch. Instead, they performed a careful re-examination of data from a controlled laboratory study conducted previously. In that original study, a single centrifugal pump made of stainless steel was run for 200 hours in a tank of synthetic saltwater that mimicked the ocean. The researchers in that original study had measured everything they could: how much water the pump moved, how high it could push that water, how much electricity it used, and how much it vibrated. Crucially, they also measured the physical texture of the pump's metal parts at the beginning and at the end of the test. The new analysis takes these existing numbers and weaves them together to see if the story of the surface matches the story of the machine's performance.

The results of this re-examination tell a clear story of gradual decline. Over the course of the 200-hour test, the pump's ability to do its job slipped noticeably. The amount of water it could move dropped from 1.50 liters per second to 1.38 liters per second. The height to which it could push the water fell from 12.0 meters to 10.7 meters. Perhaps most telling was the drop in efficiency; the machine went from converting 70.6 percent of its electrical energy into useful work to only 58.0 percent. At the same time, the machine began to shake more, with vibration levels rising from a gentle 0.20 millimeters per second to a more agitated 0.75 millimeters per second. These are not tiny fluctuations; they represent a significant loss of performance in a machine that was running under controlled conditions.

Simultaneously, the surface of the pump underwent a dramatic transformation. At the start of the test, the metal was relatively smooth, with a surface roughness measurement of 0.8 micrometers. By the end of the 200 hours, that number had more than tripled to 3.4 micrometers. While this might sound like a small number, in the world of fluid dynamics, it is a massive change. The metal was no longer just slightly worn; it had become a landscape of tiny hills and valleys. The analysis also revealed that the damage was not uniform. The surface was covered in an increasing number of pits—tiny craters caused by corrosion. The density of these pits jumped from 12 per square millimeter at the 50-hour mark to 68 per square millimeter by the end. The average size of these pits grew from 45 micrometers to 132 micrometers, and the deepest ones reached a depth of 210 micrometers.

The core finding of this work is that these two stories—the roughening of the surface and the weakening of the machine—are happening in lockstep. The researcher suggests that the roughness and the pits are not just side effects of the corrosion; they are likely the cause of the machine's struggle. As the metal surface becomes rougher, it creates more friction for the water flowing past it, much like a rough road slows down a car. This extra resistance forces the pump to work harder to move the same amount of water, which explains the drop in efficiency. The uneven removal of metal also likely upsets the balance of the spinning parts, leading to the increased vibration. The data shows a strong link between the amount of material lost to corrosion and the drop in efficiency, suggesting that as the metal disappears, the machine's performance follows.

However, the researcher is careful not to claim that this is a simple, solved law of nature. The study explicitly states that while the surface changes and the machine failures happened at the same time, the analysis does not prove that the surface changes caused the failures in every possible situation. There are other things that could make a pump vibrate or lose power, such as a blockage in the pipe or a change in the speed of the motor. The analysis acknowledges that converting a measurement of surface roughness into a prediction of how much water resistance it creates is still an uncertain process. The relationship between the roughness measured on a metal sample and the "equivalent sand-grain roughness" used in fluid mechanics models is not a fixed number; it depends on the shape of the pits and the flow of the water. To account for this, the researcher tested several different possibilities, showing that while the exact number is unknown, the range of uncertainty is manageable and can be used to create a safety margin for predictions.

The paper proposes a new way of thinking about how we monitor these machines. Instead of waiting for the vibration sensors to alarm, the researcher suggests a system that treats the surface condition as a vital sign. This would involve a layered approach where environmental data, such as salt levels and temperature, is combined with periodic checks of the metal's surface texture. These surface checks would then be linked to the continuous stream of data from the machine's performance, such as flow rate and power usage. By connecting the slow, steady changes in the metal's skin to the immediate behavior of the machine, engineers could potentially spot a problem long before it becomes a crisis. This approach would allow for maintenance to be planned based on the actual health of the material, rather than a fixed schedule or a sudden breakdown.

The researcher also highlights that this method requires a shift in how we measure and interpret data. A single number describing the roughness of a surface is not enough to tell the whole story. Two surfaces can have the same average roughness but look very different under a microscope; one might have deep, sharp pits while the other has shallow, wide scratches. These different shapes affect the water flow in different ways. Therefore, a complete picture requires looking at the depth of the pits, their density, and the overall shape of the damage, not just a single average value. The study suggests that future monitoring systems should use a combination of these measurements to build a more accurate picture of the machine's health.

While the data comes from a controlled laboratory setting with a single type of pump and a specific saltwater mixture, the implications for the real world are significant. The marine environment is complex, with varying temperatures, salinity, and the presence of marine life that can stick to surfaces. The researcher notes that the current findings are a proof of concept, a demonstration that the link between surface degradation and machine performance is real and measurable. The next steps, according to the paper, involve testing this idea in the field, using high-resolution 3D imaging to map the surfaces of real pumps, and running computer simulations to verify how different types of roughness affect water flow.

Ultimately, this work offers a new perspective on the life of a machine. It suggests that the story of a pump's failure is written in the microscopic details of its surface long before the machine stops working. By learning to read these early signs, engineers may be able to keep marine equipment running longer, safer, and more efficiently. The study does not promise a magic solution or a perfect prediction tool, but it provides a clear, evidence-based path forward. It shows that by paying attention to the surface, we can understand the machine, turning a hidden process of decay into a visible, manageable indicator of health. The work stands as a call to integrate the science of surface materials with the engineering of machine monitoring, creating a more complete and proactive approach to maintaining the vital equipment that keeps the maritime world moving.

Technical Summary: Surface Degradation as a Leading Indicator of Marine Rotating-Equipment Deterioration

1. Problem Statement
Marine rotating machinery operates at the intersection of aggressive surface chemistry, repetitive fatigue, and fluid-dynamic loading. While surface degradation phenomena—such as corrosion, roughening, pitting, erosion, wear, and tribocorrosion—are known to affect hydraulic losses, clearances, mass balancing, and vibration, they are traditionally treated as "receiver-end" inspection variables. Current condition monitoring (CM) systems predominantly rely on indirect indicators like vibration, temperature, pressure, and power. There is a critical gap in the explicit translation pathway that links measurable surface state information to rotating-equipment health. Consequently, surface degradation often serves only as a historical diagnosis during shutdowns rather than an active, predictive leading indicator.

2. Methodology
This work employs a critical review and a transparent secondary analysis of previously published data. The study does not present new experimental measurements but synthesizes existing literature on corrosion, surface metrology, tribocorrosion, and fluid mechanics with a specific dataset: a 200-hour controlled saline-water centrifugal pump test involving a 316L stainless-steel pump exposed to 30 ppt synthetic saline water.

The analysis follows a four-level evidence hierarchy to maintain research integrity:

Level 1: Direct measurements (flow, head, vibration, surface roughness $R_a$ , pit metrics).
Level 2: Mathematical transformations (efficiency calculations, relative changes, dimensionless roughness).
Level 3: Mechanistic inference (linking roughness to hydraulic loss and mass removal to vibration).
Level 4: Prospective modeling (uncertain transfer functions, e.g., $R_a$ to equivalent sand-grain roughness $k_s$ ).

A key methodological component is the sensitivity analysis of the conversion from arithmetic roughness ( $R_a$ ) to equivalent sand-grain roughness ( $k_s$ ) using the relation $k_s = \alpha R_a$ , where $\alpha$ is treated as an unknown transfer coefficient (tested at 0.5, 1.0, and 2.0) rather than a fixed physical law.

3. Key Results
The secondary analysis of the 200-hour pump test revealed the following concurrent trends:

Surface State: Arithmetic roughness ( $R_a$ ) increased from 0.8 $\mu$ m to 3.4 $\mu$ m (a 325% increase). Pit density rose from 12 to 68 pits/mm², mean pit diameter increased from 45 to 132 $\mu$ m, and maximum pit depth reached 210 $\mu$ m.
Functional Performance: Input-based efficiency dropped from 70.6% to 58.0% (a 17.8% relative decrease). Pump head decreased from 12.0 m to 10.7 m, and flow rate decreased from 1.50 to 1.38 L/s.
Dynamic Response: Vibration amplitude increased from 0.20 to 0.75 mm/s.
Material Correlation: A statistically significant correlation ( $R=0.87$ ) was observed between cumulative mass loss and efficiency in the source data, though the author cautions this is configuration-specific and not a universal law.
Sensitivity Analysis: The conversion of $R_a$ to $k_s$ introduces significant uncertainty. At 200 hours, the equivalent sand-grain roughness ( $k_s$ ) ranges from 1.7 $\mu$ m to 6.8 $\mu$ m depending on the assumed transfer factor $\alpha$ . This uncertainty envelope is comparable to or larger than the precision of typical CFD outputs, highlighting that high-resolution simulation cannot compensate for fuzzy boundary conditions.

4. Key Contributions
The paper proposes a Surface-Aware Condition-Monitoring Architecture consisting of five integrated layers:

Environmental/Exposure: Monitoring salinity, temperature, dissolved oxygen, and fouling.
Surface Condition: Measuring roughness, pit profiles, coating state, and electrochemical response.
Functional Response: Tracking flow, head, power, and efficiency.
Dynamic Response: Monitoring vibration, acoustic emission, and shaft displacement.
Decision Support: Integrating evidence into indicators and trigger points for maintenance.

The work introduces a Surface Degradation Index (SDI) that combines normalized surface indicators (e.g., $R_a$ increase, pit density) with uncertainty quantification ( $SDI \pm U_{SDI}$ ). It also outlines a validation roadmap requiring matched-surface hydraulic experiments, high-resolution 3D metrology, and CFD validation before universal thresholds or Remaining Useful Life (RUL) models can be claimed.

5. Significance and Claims
The paper modestly claims that surface degradation is a candidate leading-indicator state variable rather than a proven universal predictor. Its primary contribution is the development of an evidence-based contextual architecture that links material evidence (surface state) to functional state indicators.

The author emphasizes that:

Surface degradation should be viewed as a multiscale process connecting environmental exposure to component morphology and machine performance.
Arithmetic roughness ( $R_a$ ) alone is insufficient; a vector of surface descriptors (including pit density, depth, and texture) is required.
Correlation between surface changes and machine performance does not imply causation; confounding factors like operating point shifts, fouling, and leaks must be normalized.
The proposed framework provides a reusable template for "degradation-aware digital twins," where surface state acts as an explicit interface between environmental stress and machine response.

The study concludes that while the data is consistent with the hypothesis that surface degradation precedes functional deterioration, robust predictive capability requires further experimental validation involving matched natural and artificial roughness, high-resolution 3D metrology, and field data sets that correlate surface inspections with continuous machine monitoring.

More like this