During the 2005 ASA conference in Boca Raton, I had the honor and pleasure of speaking before the general assembly. My topic was Emerging Technologies. A portion of the presentation was allotted to the promising field of “Predictive Maintenance” (PM). By the way, a very good primer to this blog is my blog on Non-Routine, Unscheduled Maintenance, of which PM is meant to mitigate. Simply put, PM systems should be able to predict the failure of an aircraft part before it happens, and will be tied to a specific part on a specific tail number. Cool, eh?

In fact, the concept is not new. What makes the technology more so promising today is that Boeing and Airbus are investing in aircraft designs and systems with this concept in mind, but more on that later. Past concepts in PM were pioneered by the airlines and engine manufacturers. A typical scenario follows. Most engine designs include provisions for monitoring an engine’s EGT, or Exhaust Gas temperature; a critical measure of the performance of the engine, among others. The information is presented to the flight crew for monitoring. Of course there are lower and upper limits for EGT for a given engine operating profile. This data is tapped, digitized, and then transmitted during flight by a system called ACARS (Arinc Communication And Reporting System). ACARS transmitted information is then received on the ground and sent to the airline’s Maintenance Operations Control center for analysis. A graph for that engine may reveal that the EGT during the last week is starting to

creep toward the upper limit; an indication of a deteriorating condition, that if left un-addressed could cause an interruption in that aircraft’s schedule, or worse, an unexpected engine change, diversion, or cancellation. At this time, the airline would dispatch mechanics to the aircraft at an overnight maintenance station to perform a borescope inspection, and to take the appropriate action ahead of the failure, an infinitely more manageable situation than an unexpected failure. Such is the existing state of PM, but the concept is starting to be applied to other systems besides engine installations, by Boeing and Airbus.

The question is, what data is available on board the aircraft for transmission to the ground for PM analysis? An early consideration is the data already being collected by the Flight Data Recorders (FDR’s). Although some of it may be of the type usable for PM programs, it was not designed specifically for that purpose. Aircraft systems are currently being designed to output information that are suitable for PM programs. The nexus for these systems is Boeing’s Aircraft Health Monitoring System (AHM), and Airbus’s Aircraft Condition Monitoring System (ACMS).

Notice that both manufacturers are attempting to quantify the estimated savings of the systems, this because currently these systems would require an investment by the airline to make it work. If you do the math, for a large airline the savings could indeed be significant. These systems are so new however, that a solid track record from an operator has not been clearly established to correlate the savings estimates.

PM: What’s needed?

• Aircraft systems designed to output information on it’s internal condition
• Aircraft to satellite, to ground station transmission of the data
• Software systems designed to process the data

Currently, Airline Maintenance Engineering operations have departments to monitor “Reliability” trends. For example, if component removals for a certain ATA chapter start to trend upward, flags are raised to attract attention to the causes. This is a classic statistical monitoring system. Reliability programs can predict that during a given period you’re going to have so many removals, but it cannot tell you when, or to what components it will happen. Enter PM programs as described. In fact, in the next 10 years or so I predict that artificial intelligence will be implemented in this field. Such systems would not only be able to predict a failure, but based on the current aircraft utilization and flight schedules, will tell you at what time and airport the failure is likely to occur.

To summarize, PM programs can provide:

• On a micro scale, early warnings of deteriorating equipment tied to tail numbers or engine serial numbers
• On a macro scale, the resultant data can more accurately predict future maintenance requirements
• On both scales, PM allows maintenance planners to more proactively and accurately plan and position spare parts
• Allocate required maintenance technicians more efficiently
• Allocate tools and test equipment more efficiently

As PM systems mature and are integrated into airline routine maintenance operations, it will be interesting to see what will be the FAA’s stance on how airlines should address predicted failures. Today, existing failures are either fixed or deferred in accordance with the Airline’s approved MEL (Minimum Equipment List) programs. If a PM program indicates you’re going to exhibit a failure in 3 days, how will it be tracked, and documented? I suppose there will be a pioneer, lead airline that will blaze the trail on what works or doesn’t work, thus establishing the model, and perhaps the basis for the FAA posture for their written ‘guidance’ on the issue.

I predict that Predictive Maintenance will proactively predictively predict maintenance; this postured on the position just postulated previously by this un-perplexed person.