Signalling change

Maintenance is essential within the aerospace industry. However, because cost-effectiveness is also high on the list of priorities, to have an optimum maintenance strategy requires walking a fine line between doing too little and too much. For example, insufficient or no maintenance typically leads to equipment failure, and in the aerospace sector this could have catastrophic consequences. Conversely, 'erring on the side of caution' and over-maintaining equipment is a costly business. Plus, there are no guarantees that failure will still not occur. Cost-effective failure risk reduction is the overall goal, and helping those responsible for maintenance walk the aforementioned fine line is condition-based predictive maintenance (CBPM). Here, the monitoring of critical components has many benefits, including better scheduling of maintenance, better supply chain intelligence (i.e. what parts will be needed and how soon) and squeezing as much life out of components as it is safe to do. CBPM is not new to the aerospace industry though. Indeed, it could be argued that, as a principle, it is at the heart of Health and Usage Monitoring Systems (HUMS), as used on rotary and increasingly, fixed wing aircraft. In addition, continuous monitoring should afford sufficient time for remedial action to be taken in the event of rapid degradation of a component, and the system can even be set up to instigate action automatically if necessary. Clearly better The majority of failures on jet engines occur because of blade defects, either developing over time or as the result of an impact (such as a bird strike or debris thrown up from the runway). The earliest warning of turbine failure tends to take the form of anomalous blade vibrations, so monitoring for these holds the key to better maintenance and the possibility of raising alarms whilst in service. During the manufacture of a turbine, a common method of assessing blade vibration relies on the use of optical probes mounted in the blade casing assembly. Recording the time of arrival of each blade at a certain point in the cycle and comparing this with previous cycles enables the detection and analysis of anomalous vibration patterns. Results are clear and accurate. However, as a clear optical path needs to exist between the sensor and the blade tips, contamination from dust and exhaust gases rapidly degrade signal quality in the short term, while in the long term the sensor lenses become scratched and require regular replacement. Although this is inconvenient, it is manageable for the purpose of taking measurements on the ground. Clearly, however, the technology is inappropriate for continuous monitoring of engines whilst in service. Thankfully, blade tip monitoring can also be achieved using other non-contact methods including the use of capacitive, high frequency pressure and eddy current transducers. In 2002, a series of trials were conducted by QinetiQ at its turbine test facility at Shoeburyness. The trials compared the three technologies and concluded that eddy current probes showed the best promise for further development. Eddy current sensors are most commonly used for non-contact proximity and displacement measurements. Accuracy is high and ruggedised versions are often used in contaminated environments. In QinetiQ's trials, bench tests were initially conducted on a standard off the shelf eddy current sensor. It was found that several factors influenced the quality of the probe's output including the type and thickness of any material between the sensor and the blade tips, and the material and size of the pocket within which the sensor was housed. Following the initial trials, QinetiQ partnered with UK-based eddy current probe OEM Monitran. The goals were to optimise sensor resolution and to develop signal conditioning circuits that produce sufficiently clear waveforms for accurate tip timing. Both goals were met, and resulted in the launch of Monitran's Turbine-Tip Timing (T3) module. Together, QinetiQ and Monitran had solved the quality issue. Waveforms were clear and timing accuracy on a par with the well established optical techniques. However, what about the quantity of data? Speaking volumes Irrespective of sensor type, it has traditionally taken several hours to process the data from even a single minute of at-speed turbine operation, making realtime health monitoring impossible, particularly where an emergency shutdown may be necessary. Monitran has now addressed this problem with its digital signal processing (DSP) unit, which is capable of performing the necessary calculations fast enough to indicate ‘live' health, and the first units are shortly to be deployed on industrial power generation (IPG) turbines in the UK. The DSP unit has five channels. One is reserved for a traditional accelerometer, to be placed on the turbine housing so that overall vibration levels may contribute to the health profile. The remaining four channels are for tip timing signals, typically from T3 eddy current sensor signal conditioning units. A minimum of four channels per turbine is recommended following QinetiQ's early research which revealed that decay of a signal indicating blade deterioration can occur within a single turbine revolution. Typically the sensors might be arranged in opposing pairs along two orthogonal diameters of the engine casing. The availability of realtime blade health monitoring opens up many possibilities. For example, traffic light style indicators showing green for healthy, amber for advisory and red for shutdown are a classic path to deskilling and automating the monitoring function. Moreover, integration of the unit into an avionics system (such as the aforementioned HUMS) could facilitate data recording. Constant recording would no doubt be impracticable because of the volume of data involved, but even a few revolutions of data harvested at regular intervals, or perhaps at each start-up, would be invaluable. This would remove the need for the scheduling of detailed turbine blade analysis as part of a maintenance routine. Such investigations are costly, require the capture and analysis of terabytes of data by highly specialised personnel and afford little reassurance against failures between inspections. In the event of an incident, the DSP's 1,000 blade pass per channel buffer enables subsequent analysis of pre-trigger blade health. As mentioned, the technology is already making inroads into turbine monitoring for the IPG sector, and is seeing use in ground-based analysis in the aerospace industry. If the current rate of progress continues then we may be only a few years away from in-service, realtime, flight certified turbine blade health monitoring. www.monitran.com