While vibration monitoring (VM) has long been in use for the purpose of condition monitoring (CM) onboard ships, the technique is not without limitations. VM involves the detection of vibration over a wide range of frequencies. The detected signal can have a strong contribution from high amplitude low frequency sources such as resonances, out of balance and misalignment. Bearing faults tend to produce low amplitude high frequency vibrations, The result is a low signal-to-noise ratio. Relatively complex signal processing and data interpretation is required to identify a machine with faulty bearings from its vibration signature.
For a number of years, CM and predictive maintenance specialist Parker Kittiwake has been developing a solution to this problem based on monitoring acoustics emissions (AE), a technique based on the detection of high frequency sounds produced by impacts and friction within defective bearings and gears. AE was first developed in the 1960s, initially as a method for detecting crack propagation in materials, but its application to CM is comparatively recent.
Cristiano Garau, Parker Kittiwake’s lead application engineer explains: “Our technique is based on detecting these high frequency sounds which are generated by, for example, a defective bearing or lack of lubrication. The sound travels through the structure of the machine and is picked up by our Acoustic Emissions sensor. It is the algorithm that we use to process the sensor output that differentiates us from our competitors.”
“We supply the AE technology in a form that is very easy to understand for the end-user. The AE signature is analysed in terms of two parameters: Distress and dB level – Distress is a proprietary parameter that measures transient activity [impacts and frictions] and gives an indication of how smoothly the bearing is running. It is very easy to understand and simple alarm levels can be set. For example, from 0-10 Distress you have a good bearing and from 10-15 you have a bearing that’s suspicious. Obviously, you always measure from trending rather than the absolute values and anything over 15 suggests a bearing that may have developed a problem.”
The dB level (also expressed on a 0-100 scale) is the mean level of the AE signature and comes into play when more stringent monitoring is required and serves a comparative trending parameter. Typically, doubling the rotational speed increases the dB level by 12dB. A sharp increase of 25dB often indicates that failure is imminent.
By moving the detection frequency into the AE range, it is possible to reject the low frequency vibration due to normal running. The result is a dramatic increase in signal to noise ratio. There is a clear difference between the AE signature from a damaged bearing and one in good condition. This has lead to the development of trending parameters that are highly effective yet easy to understand.
“Like VM we use a surface mounted sensor to detect waves generated within the structure of the machine,” says Parker Kittiwake technical specialist Neil Randall, who has been heavily involved in refining the AE technique. “But we are detecting at around 100KHz, this is considerably higher than an accelerometer, which generally works up to 10-20KHz. Up at 100KHz what we’re actually detecting are the individual impacts that are occurring within the bearing as discrete events. With vibration monitoring what you tend to be doing is monitoring the response of the structure to the repetitive nature of those individual impacts.”
“It’s a subtle difference but it leads to an important advantage for AE in one particular application area; slowly rotating bearings. Both techniques can be used for monitoring bearing condition at speeds above say, 60rpm. But as you slow things down, with VM you reach the lower end of the accelerometer’s detection capability, which tends to be a few hertz. There comes a point at which you will no longer be able to detect the very low frequencies generated if the impacts are being repeated very, very slowly. Because we’re monitoring at a high frequency and detect a burst of activity as an individual event it doesn’t matter to us how long it takes for the next event to occur, so there’s no limit to how slowly you make the rotation.”
Although the testing of AE with slow-rotating machinery has so far been restricted to land-based machinery, Parker Kittiwake has recently published the results of its work on four ro-ro vehicle carriers owned by Gram Car Carriers (GCC): the Viking Costanza, Hoegh Caribia, City of Oslo and Viking Odessa. The Japanese-built sister ships, constructed between 2009 and 2010, were each equipped with two 6-cylinder medium speed MAK diesel engines, coupled to the input shafts of a RENK gearbox linked to the propulsion shaft. Each gearbox also had an auxiliary output shaft linked to the main generator, with two auxiliary generators for use in port for cargo loading/unloading. Within their first year of operation, three of the vessels developed catastrophic alternator bearing failures at considerable cost to GCC. In 2015, one of the vessels again suffered a catastrophic bearing failure and checks on the three sister ships revealed they were also developing similar failures.
The precise cause of these failures – whether it was incorrect materials, misalignment issues, alternator bearing design flaws, vibrations or vessel structural problems – was difficult to determine, and even DNV GL, who were consulted, were unable to reach any definitive conclusions.
GCC decided its best recourse was online condition monitoring and, given the particular characteristics of the vessels, opted to try Parker Kittiwake’s Alternator Bearing Monitoring System based on AE sensors. Garau explains: “The engine room was too noisy. It had a very complex footprint in terms of vibration to analyse. AE created the opportunity to mask out all those noises and concentrate just on the asset that was failing, i.e. the bearings.”
Viking Costanza, which had recently been fitted with a new set of alternator bearings, was chosen as the trial vessel. The Alternator Bearing Monitoring System logs and trends the data streamed by two AE Sigma sensors installed on the drive end (DE) and non-drive end (NDE bearing), as illustrated. The first three months provided valuable data on the stress that normal operational events were putting upon the bearings and the effectiveness of the lubrication, prompting GMM to install the Alternator Bearing Monitoring System on the other three vessels.
Hoegh Caribia and Viking Odessa showed similar results, but there was a deviation with City of Oslo, which was showing a Distress reading of well over 20dB for NDE, 10dB for the DE, as well as higher dB level readings than the sister ships. Even after the bearings were re-greased, this pattern continued. A month later, GCC decided to change the NDE and the DE bearings a few weeks later. Inspection of the DE bearing revealed significant damage and the new bearing saw a 16dB drop in the dB Level. However, the data indicated that the mechanism causing the damage was still present. A close inspection confirmed that the relative position of the gearbox and alternator had changed. Laser realignment was carried out and the alternator was re balanced using vibration analysis, which finally brought the City of Oslo’s Distress readings in line with the other vessels.
The shipping industry has until recently been slow on the uptake of new technology, but Garau praised GCC’s forward-thinking approach. “Initially the system was used to identify faulty bearings and manage their replacement. It is now being used on an ongoing basis to ensure continued smooth running of the vessel. The chief engineers have recently added another level to the project and are now using the system to identify operating conditions that put the bearings under excess stress. Their aim is to extend the life of the bearings by minimising the time spent under these conditions.”
Take up of vibration analysis has traditionally been restricted by the expertise needed to interpret collected data, but the flexibility of Parker Kittiwake’s AE system, and the ability to highlight changes with just two parameters may offer an empowering solution that facilitates a proactive approach to maintenance. Randall says: “Certainly in terms of our sales of portable instruments to general industry, one of the biggest advantages is that it makes condition monitoring accessible to people at the operations level.”