Fast forward 55 years to 2015 and following our fifth South African democratic elections it seems poetic that once again the winds of change are blowing through South Africa but this time it is not a political revolution that is set to change the landscape of our beautiful country but a green revolution in renewable energy.
South Africa's Integrated Resource Plan (IRP), under the leadership of the Department of Energy (DOE), envisages renewable energy contributing to 42% or 17.8GW of the country's new generation capacity by 2030. One of the ways they plan to achieve this is with 8.4GW of wind generated power.
The DoE’s Renewable Energy Independent Power Producer Procurement Programme (REIPPP) has already overseen the completion of four successful bidding windows with the fourth bidding window having been extended to include additional allocation under the DoE’s expedited procurement process.
The approved bids from all four bidding rounds will result in the construction of several wind farms over the coming years that will collectively house in excess of 1 000 wind turbines.
The estimated life span of wind turbines is about 20 years, compared to conventional steam turbine generator units that have averaged 40 years. The failure rate of wind turbines is about three times higher than that of conventional generators and this has historically been attributed to constantly changing loads experienced by the wind turbine, as a result of environmental variants.
Due to these highly variable operational conditions, the mechanical stress placed on wind turbines is incomparable in any other form of power generation and they consequently require a high degree of maintenance to provide cost effective and reliable power output throughout their expected 20 year life cycle. The wind turbine gearbox is the most critical component in terms of high failure rates and down time.
These premature gearbox failures are a leading maintenance cost driver that can substantially lower the profit margin of a wind turbine operation as they typically result in component replacement. Despite significant advancements in gearbox design, they remain an operation and maintenance cost driver due to the very high associated repair costs coupled with a high likelihood of failure through much of the wind turbine’s life cycle.
Ensuring long-term asset reliability and achieving low operation and maintenance costs are key drivers to the economic and technical viability of wind turbines becoming a primary renewable energy source in South Africa.
Oil analysis, along with other condition monitoring tools, offers the potential to effectively manage gearbox maintenance by detecting early damage as well as tracking the severity of the damage. It is for this reason that most OEM’s recommend routine oil analysis as part of an effective maintenance strategy.
Routine oil analysis is one of the most widely used predictive and/or proactive maintenance strategies for wind turbines and utilises a test slate that evaluates the condition of the in-service lubricant and helps evaluate the condition of internal mechanical components.
We will look at what oil analysis can measure in terms of the first three functions of oil analysis, which are: to detect abnormal wear, oil degradation and contamination. Active monitoring of the above provides early warning of abnormal operating conditions that can lead to catastrophic failures in a wind turbine if not corrected.
Detecting abnormal wear
The fundamental concept behind monitoring wear appears uncomplicated: trend the metal wear rates for sudden increases that indicate a change in the system’s health.
Wear metal generation rates are often described as following a bath tub curve. The bathtub curve represents wear generated over the lifetime of a typical wearing component, with elevated wear levels during bedding-in, followed by prolonged periods of relatively constant wear levels, followed by the onset of severe wear and an exponential increase in metal generation leading to eventual failure at the end of the component’s life.
While this is a sound theory, wear debris generation is a complex phenomenon. Wear rates can increase and decrease throughout the life time of the gearbox due to several factors such as operating loads, lubricant quality, fault progression, etc. Even during fault progression, wear rates are highly mutable depending on the microstructural material properties of the wind turbine gearbox components e.g. cylindrical roller bearings.
Commercial oil laboratories employ varying techniques when it comes to detecting (quantifying and classifying) wear particles in oil, each with its own strengths and limitations. The most widely used and OEM requested laboratory techniques will be described below.
Spectrometric analysis
The spectrometer is used to determine the presence and concentration of different elements in the oil. These are measured in PPM (parts per million). The measured elements are usually divided into three broad categories: wear metals such as iron, contaminants such as silicon and oil additives such as phosphorus.
Wind turbine gear oil analysis usually requires close monitoring of iron and copper wear rates as these metals are most commonly used in the construction of internal gearbox components. In terms of wear metals detected in the oil, the iron wear rate is usually the highest reading because almost everything in a gearbox is made from different steel alloys. Sources of iron include bearings, shafts, and gears while copper wear usually originates from bronze alloy bearing cages.
Unfortunately the spectrometer can only measure very small particles, usually less than eight microns in size. The instrument cannot ‘see’ larger particles that might indicate a severe wear situation is developing.
Ferrous debris monitor
The ferrous debris monitor provides a measure of the total ferrous content of the oil sample and from this measurement the total amount of ferrous (iron) debris can be determined irrespective of the particle’s size.
Wear metal particles, detected by spectroscopy, are typically less than eight microns in size. These small particles can be generated by rubbing wear or fretting corrosion. Larger particles are generated by more severe wear modes such as fatigue wear, pitting and spalling. These larger ferrous particles, present in the used oil sample, can be detected by using this method. The PQ index is not an actual concentration measurement but it can be compared to the iron (ppm) reading obtained from the spectrometric analysis.
If the PQ index is smaller than the iron (ppm) reading, then it is unlikely that particles larger than eight microns are present. Alternately, if the PQ index increases significantly while the iron reading remains consistent, then larger ferrous particles are being generated and further analysis into the cause of the elevated PQ should be performed.
In terms of wear particles, their morphology and quantity provide direct insight into overall gearbox health. An MPE is performed by filtering the oil through a membrane patch of a known micron rating and any debris present is examined under a microscope. The membrane patch is examined for wear, contamination and colour.
An MPE can provide clues to the source of the debris and the potential seriousness of a problem that may be causing it. The individual particles themselves are not categorised but instead observations are recorded for trending purposes using a size and concentration reference matrix.
Analytical ferrography
Analytical ferrography involves observing and categorising particle size, shape, colour and surface texture under magnification.
Evaluating the concentration, size, shape, composition, and condition of the particles indicates where and how they were generated. The particle’s composition indicates its source and the particle’s shape reveals how it was generated. Abrasion, adhesion, fatigue, sliding, and rolling contact wear modes each generate a characteristic particle type in terms of shape and surface condition.
Particle composition is broken into categories that include: ferrous wear, white metal, copper, and fibres. Ferrous particles can further be identified as steel, cast iron, dark oxides, or red oxides (rust). A skilled analyst can also determine if metallic wear particles are caused by cutting wear versus rolling or sliding wear. Wear debris monitoring has been demonstrated to be an effective means of detecting gear and bearing fault initiation.
The main particle types related to the fatigue process encountered in wind turbine gearboxes are: laminar micro-particles (micropitting), laminar particles, chunky fatigue particles and spheres.
Analytical ferrography can be a powerful diagnostic tool in oil analysis. When implemented correctly it provides a tremendous amount of information about the machine under operation.
It is often said that there are three key requirements for maintaining the condition of wind turbine gear oil: keep it cool, keep it clean and keep it dry. In truth this applies to any mechanical system but these requirements in the context of this paper relate to the monitoring of oil degradation and contamination of wind turbines.
Detect oil degradation
The different modes and severities of oil degradation are dependent upon the oil type, application and exposure to contaminants. Oil degrades over time due to its ability to react with oxygen in the atmosphere. This process is known as oxidation. Oxidation causes the viscosity to increase and acids to form in the oil. The rate at which this occurs can be increased by high operating temperature and the presence of contaminants.
In wind turbine gearboxes, oxidation also results in metal corrosion, varnish formation, foaming and/or air entrainment, poor water demulsibility and filter plugging.
A lubricant has many functions to perform and these can be categorised into four fundamental groups: reduction of wear, removal of contaminants, removal of heat and act as a structural material. All these functions are negatively impacted if the viscosity of the oil falls outside of the intended viscosity range i.e. too high or too low. If the viscosity is not correct for the load, the oil film cannot be adequately established at the friction point. Heat and contamination are not carried away at the proper rates, and the oil cannot sufficiently protect the component. A lubricant with the improper viscosity will lead to overheating, accelerated wear, and ultimately failure of the component. It is for this reason that viscosity is considered the most important physical property of a lubricant.
Trending of viscosity data is important as deviations from the norm may indicate base oil degradation, additive depletion or the use of an incorrect lubricant.
When the oil’s viscosity increases, it is usually due to oxidation or degradation typically as a result of extended oil drain intervals, high operating temperatures, and the presence of water, the presence of other oxidation catalysts or the addition of an incorrect lubricant.
Decreases in oil viscosity are attributed to degradation of the viscosity index improver (VII) additive in the oil as a result of shear or due to the use of an incorrect lubricant during the refilling and topping-up procedures.
A low viscosity (<15% of new KV) is generally considered to be more problematic as this results in a reduced film thickness and the consequent propagation of fatigue cracks associated with micropitting. Micropitting is a surface fatigue phenomenon resulting in superficial damage that appears in high rolling contacts and is characterised by the presence of small pits on the tooth surface. They first appear in the rolling zone of the gears and then progress towards the root (dedendum) of the gear.
Micropitting causes tooth profile wear (deviations in the shape of the tooth), which increases vibration and noise, concentrates loads on smaller tooth areas increasing stress on gear teeth and shortening gear life.
Another technique employed to detect base oil oxidation is Fourier Transform Infrared (FTIR) analysis. FTIR analysis effectively measures the concentration of various organic or metallo-organic material present in the oil. When oil is oxidised, the hydrocarbon oil molecules can become restructured into soluble and insoluble oxidation by-products as a result of the sequential addition of oxygen to the base oil molecules. FTIR measures the accumulation of these by-products.
FTIR produces an infrared (IR) spectrum that is often referred to as the 'fingerprint' of the oil as it contains specific features of the chemical composition of the oil. The IR spectrum can be used to identify types of additives, trend oxidation and nitration by-products that could form as a result of high operating temperatures and thermal degradation caused by aeration and/or foaming. These are all important indicators of the lubricant’s ability to perform its basic functions as detailed earlier on in this article.
The usefulness of FTIR in determining oxidation is dependent on the base oil used to formulate the lubricant. Synthetic lubricants often contain ester compounds which have a significant peak in the infrared spectra area where the oxidation level for mineral oils is measured.
Despite this, several studies performed by international oil analysis laboratories have shown a correlation between the FTIR oxidation reading of synthetic oils and other degradation parameters measured in the oil. In synthetic oils the oxidation value by FTIR in itself is not necessarily useful but trending it in conjunction with other parameters can be quite revealing.
With rigorous particle contamination control bearing life can increase substantially, resulting in greater gearbox reliability, uptime and energy production, extended warranty periods and a higher return on investment.
Steven Lara-Lee Lumley, senior diagnostician at Wearcheck Africa
[PULL QUOTE] In wind turbine gearboxes, oxidation also results in metal corrosion, varnish formation, foaming and/or air entrainment, poor water demulsibility and filter plugging.
