Identifying Bearing Failure at an Early Stage
Identifying Bearing Failure at an Early Stage
Detecting wear, imbalance and misalignment of rotating parts within machinery is critical to its health and overall performance. This can be achieved by implementing a variety of proven techniques. Vibration analysis, for example, uses accelerometers to detect potential problems with industrial equipment caused by incorrectly aligned, loose, or unbalanced rotating parts.
These techniques tend to be most effective during the later stages of the wear cycle, when damage has already begun to occur. In the early stages of wear, however, when vibration signals are of low intensity, it can be difficult to separate the wear signature from underlying and background machine frequencies.
Instead of waiting for wear rates to progress to a later stage – at which point the machinery’s performance is likely to be declining while the potential for unscheduled machine downtime will be increasing – maintenance and production engineers can take advantage of a signal processing technique called acceleration enveloping.
This enables engineers to overcome the limitations of conventional velocity spectrum measurements and detect the failure of, for example, rolling element bearings at the earliest possible stage. Then, the rate of wear can be monitored and maintenance work planned accordingly.
In practice, what tends to occur is a defect in a rolling element causes repeated impact events that generate resonant frequencies in the surrounding machine surfaces, causing it to ring. Although the amplitude of the ringing signal decays between impacts and becomes part of the overall vibration signal of the machine, it nonetheless affects the natural resonance response of the machine at the impact frequencies.
Using a high performance accelerometer, acceleration enveloping progressively filters out unwanted parts of the vibration spectrum until the signal of the bearing defect can be isolated from the noise around it. The signal is then clearly identifiable.
This information can be easily collected from the accelerometer using a data collector, ready for review and interpretation by a specialist. An informed decision then can be made on whether or not maintenance work is required immediately or can be planned as part of routine schedules.
Achieving Successful Results
While acceleration enveloping in many ways is the ideal option for detecting bearing failure, it has a number of potential limitations that must be taken into account before being implemented. The first consideration for plant engineers is the suitability of each machine, because acceleration enveloping isn’t fit for use with any and all machines. The technique detects faults involving repetitive, metal to metal interactions, which means anything that masks this, such as gaskets or dampers, may reduce its effectiveness.
However, where an application is deemed to be suitable, several factors help to ensure better results. First of all, in order to measure the low-level signal, accelerometers must be selected carefully and in the proper frequency range to suit the needs of the particular machine or application.
Once specified and ready for use, accelerometers should be correctly mounted on a flat, clean surface in close proximity to the component being monitored to guarantee consistent results. Poor mounting reduces the reliability of results and can make collected data redundant, preventing the correct decisions and appropriate actions from being taken.
Figure 1: Accelerometers must be mounted securely on a clean and solid base and as close to the component being monitored as possible
Once accelerometers have been installed and calibrated, data readings should be taken at regular intervals over a period of time to allow accurate trend analyses to be produced. This allows a steadily deteriorating condition to be identified, for example.
It is important to understand that the information provided is not a simple yes/no answer and requires some skill and experience to interpret. For example, the amplitude of a worsening condition can actually reduce over time as the imperfection becomes slightly smoother.
The potential benefits of acceleration enveloping are clear, but it would be unwise to rely on the technique alone. Implementing it as part of a wider monitoring and analysis regime can be far more effective, helping plant engineers to safeguard the health, performance and productivity of all the assets under their care.
Enveloping in Action
According to the Global Wind Energy Council (GWEC), 268,000 wind turbines were in operation at the end of 2014, with an average of 8,000 separate components per turbine. Of these, a large number are associated with the drivetrain, which is considered the major cause of extended downtime. Wear in gearboxes and bearings, in particular, is known to cause problems. Regular vibration monitoring can prevent these issues from occurring, eliminating the need for expensive repairs.
The complexity of a typical wind turbine, however, does present a challenge for vibration monitoring. For example, the main turbine, gearbox and generator often have more than 15 rolling element bearings installed, while the gearbox incorporates a series of stages, each with multiple gears. These components generate unique vibration signatures with different amplitudes and frequencies, which can be difficult to isolate from each other and can be masked by noise from surrounding systems.
This is where acceleration enveloping can play a crucial role, enabling vibration analysts and maintenance engineers to separate vibration signatures and identify the changes in signal conditions, which can indicate increasing wear.
To be effective, acceleration enveloping requires the use of multiple accelerometers fitted to all key rotating parts. These include the main bearings; planetary, intermediate and high speed gear stages; the generator, both inboard and outboard bearings; and ideally, the nacelle traverse and axial movements.
In each case, several critical factors must be considered. In particular, each accelerometer must be mounted securely on a clean and solid base, and as close as possible to the component being monitored. Normally, standard M8 mountings are used.
It is also important to collect data consistently to enable any change in operating conditions or trends over time to be accurately identified at the earliest possible stage.