When I was working in a forging plant, we had some very large hydraulic systems. When the system went down for any reason, we lost a minimum of $10,000 per hour in production. When a large hydraulic system’s oil goes bad, it can easily take eight hours to change out or flush. Ask yourself, can your company afford to lose $80,000 in production for a single incident?
Oil is the life of any hydraulic or lubrication system. When it is fresh and healthy, the system operates as intended. As it ages and begins losing its properties, issues within the system occur to the point where your system can fail.
A common practice for a hydraulic system is to take an oil sample monthly or even quarterly. This sample is then sent off to be analyzed. The analysis takes a few days to a week to get back, and it requires a highly skilled person to read the analysis and determine if there is an issue. Trending an issue might take months; all the while, the system is losing performance.
In this paper, we will discuss what type of data should be collected, why that data is important, how the data should be collected, what should be done with the collected data, how to analyze this data, and how to apply the analysis to your system.
What Should Be Collected
There are numerous factors to consider when trying to determine what is affecting a system's fluid, including:
- The ambient temperature
- The atmosphere the system is operating in
- How often the system turns over in the tank
- The pressure the system is operating at
All these outside influences affect the fluid.
For instance, systems operating in a humid environment can ingest moisture from the air. This moisture will cause water ingress into the fluid, often leading to oxidation, an increase in the fluid's total acid number, and damage to the system’s components. Any of these conditions will cause additional problems within the system and affect the components and overall system performance.
Now, the real question is, what data should be considered critical to know?
Understanding a fluid's breakdown mechanism will help direct you to the factors that should be monitored:
- Wear debris (ferrous and non-ferrous)
- Contamination
- Moisture content (percent saturation)
- Electrochemical changes (TAN, conductivity, viscosity)
- Temperature
Every system will become contaminated. Components will wear, seals will decay, the system will ingest contamination from the outside atmosphere, and the operation of a system will create its own contamination. All this contamination is carried through the system by the fluid. Using a monitor that provides real-time data will help trend daily events and anomalies. Knowing how healthy a system is will allow you to understand better what is going on in that system.
We should also know and be able to trend the fluid's temperature. Since temperature directly affects viscosity and contributes to oxidation while influencing other additive properties in oil, it is important to know a fluid's temperature during operation.
Another factor to consider is the water content. Along with being incompressible and non-lubricating, water increases oxidation and is a catalyst for acid formation. Water in a fluid will result in accelerated wear of bearing surfaces, breakdowns in seals and pumps, and an overall decrease in system performance.
Finally, trending changes in the electro-chemistry of oil should be considered. With the changes in the refining of oils over the past years, an oil’s conductivity – or ability to carry an electrostatic charge – has become a matter of grave concern. This static charge can cause damage to the system and is dangerous for operators and maintenance personnel. Correlations can be made from trending fluid conductivity and comparisons of the results to known degradation curves. Analytics can transfer all the data collected into a readable report for you to use.
Why This Particular Data
Each of the above parameters helps tell the story of degradation. By trending the data, correlations can be made between these parameters and the system's health, allowing a very personal understanding of what is happening inside the equipment.
So why is this specific data important? Let’s look at the contamination monitor data first. Data from the contamination monitor will indicate the size and type of contamination in a system. Combined with the metallic counter, it’s possible to correlate the amount of non-metallic and metallic particles in a system.
Knowing the amount and type of particles can help determine what is occurring within a system. Most non-metallic particles come from seal wear, additive breakdown, or the ingestion of particles from outside the system. Metallic particles result from system component degradation.
Considering that the average free clearance in a common cylindrical roller bearing is .010” to .012", knowing the size and quantity of particles within that range will reveal bearings are being damaged. Understand that when the particle size exceeds the film thickness, the result is microscopic damage to the components. This damage usually results in an increase in metallic particles.
Additionally, finer particles work in a system like a lapping compound, which can have a polishing effect that breaks down hard surfaces and increases component degradation rates.
The effects of temperature and water content have a coupled – or related – influence in a system. Water can enter a system through condensation and ingression. The amount of condensation can be related to the system’s temperature fluctuation. All systems must breathe.
If a system is operated in a cool environment for one shift, the oil temperature will rise significantly during operation and cool down during the non-operating time. Since the system is breathing relatively cool air, which would be heavier in moisture content, this moisture will be in the system's headspace. As the hot fluid and the cool air in the headspace interact, the moisture in the air will condense, enter the fluid, and begin to oxidize. When the fluid heats up again, the oxidation rate will increase due to the fluid’s moisture and temperature.
The dielectric constant measures the oil’s ability to carry electrical potential energy. As the oil becomes contaminated or undergoes a chemical change due to degradation, the dielectric constant will change. Monitoring this value can determine the oil’s rate of change. Additionally, since we have the dielectric constant value of the base oil and the base oils do not have the same constant, we will get an indication of whether the correct oil has been added to the system.
Deploying Sensors for Real-Time Data
The current method of collecting commonly used in industry is to obtain an oil sample from a system and send it to a lab. The lab processes the sample to obtain data, which is then available for analysis. The lab places the results on a graph for trending and offers a good or bad estimation.
This method is flawed due to the extended timeframe required to obtain trending data. Since most samples are generally collected monthly, the data will be at least three to four months old before any type of trend analysis is available. By that time, significant damage could have occurred to the system. This also leads to reactive maintenance, which is not an efficient way to maintain production. This is where applying sensor technology can help collect real-time data on systems.
This can include online contamination monitors, metallic contamination sensors, and similar devices that measure temperature, moisture, and conductivity. As technology continues improving, the necessity of collecting precise, real-time data will only increase. By predicting a system's health, a facility can plan for downtime and more efficiently utilize maintenance resources and capital.
What To Do with the Data
Now that we know what to collect, what do we do with this data? First, we must understand that the data we are collecting using sensors in real-time is more consistent. Since the data was collected using instrumentation, we eliminated the human factor.
When we use oil analysis, a person must take the oil sample and repeat the same process each time. The sample is taken from the exact same place, in the exact same manner, in a tiny sample bottle, all the while trying to maintain the same oil flow, temperature, and ambient temperature. The possibility of an error or variation occurring is very significant. This sample is then tested by a person in the lab, affording another opportunity for human error.
By removing the human factor, a sustainable and very reliable method for collecting data is possible. Measuring flow, ambient temperature, and time allows one to determine any outside factors that may have caused an issue with the sample data.
As mentioned above, most systems' current data collection method is a reactive, long-term type of trending analysis. Using the sensors above, the data can be collected while the unit operates. Since the data is trended as the unit is operating, we can more closely monitor the data trends. This allows us to track the degradation of the oil by monitoring many of the critical factors we mentioned above in real-time.
If the data is stored as it’s collected, a graphic of the results can be created. The results can then be compared against a representative sample previously analyzed in a lab. There are now many possibilities for analyzing the data.
We can use the trends, interpolate, and determine when the oil will reach its useful life. By determining an approximate end of oil life based on our data, we can proactively plan the oil change out to limit downtime and possible machine failure.
Trigger points can also be installed to set alarms for immediate reaction. Since the triggers will have a timestamp, the issue at the trigger point can be compared with what occurred in the machine operation. This allows problems to be pinpointed and to determine ways of preventing those issues in the future.
Finally, analyzing the data can help determine other system factors. This includes selecting the proper fluid and additive package and understanding if the fluid is lasting as anticipated.
Using this trending analysis and reviewing the whole system, we can develop more reliable designs that will last longer and improve uptime.
How Does Analysis Predict Oil Life?
The sensor technology mentioned above can be used to develop a series of equations that predict an oil’s remaining useful life. Several equations, such as the Arrhenius Reaction Rate, provide a foundation for this equation-based analysis.
By using the collected data to determine various factors, as indicated in Figure 1, the fluid’s actual life loss at the time of data collection can be determined. The data can then be used to systematically calculate the oil's lifespan.
Each data point, such as temperature, water, contamination, and electroconductivity, influences the fluid’s life. Using this data to create equation factors can help determine when an oil will reach the end of its useful life. The data can also be used to create graphs and determine trends. When compared to the analysis results, it can determine when the oil will begin causing problems within a system.
Moving Forward from Here
Predicting maintenance intervals and reducing downtime should always be a priority, and it takes diligence and patience to achieve this goal. Collecting reliable data to develop trends that will serve as a baseline takes time. Once the baseline is completed, system data can be compared with production output, seasonal changes, and/or any other factors required to determine the trends.
Coupling this data with quarterly or yearly oil analysis from an outside lab will allow for a reliable prediction of required maintenance to improve equipment uptime and drastically reduce reactive maintenance and downtime. The comparison of the lab data is equally important to validate the trended data.
Today’s technology lacks the capability to provide online elemental spectroscopy, which would analyze fluid health more concisely. Based on technology and economics, this is the path forward. Like the evolution of cell phones, this will also be possible in time.
The choice now becomes yours. You need to evaluate the cost of your system, the cost of changing out oil, the cost of downtime, and the labor costs to maintain and repair equipment. The technology is available and ready for use. All you need to do is make it happen.