How Predictable are Mechanical Malfunctions?

Maintenance and technical issues precipitated 26 percent of all aviation accidents in 2003. Were some or all of these accidents predictable?

According to a 1978 study, the practices of setting operating limits or undertaking repeated checks of equipment wear conditions only identified 11 percent of mechanical problems. This means that almost 90 percent of mechanical malfunctions went undetected at the time, were therefore unpredictable, and resulted in unscheduled maintenance.

If that result is applicable today, absolute adherence to all manufacturer documents such as AMMs, CMMs, SBs, SLs, and regulator ADs do not reduce the amount of unscheduled maintenance.

So, if operational limits and repeated equipment checks aren't effective in helping you reduce unscheduled maintenance time and costs, what is? Maintenance magazines, formal meetings, and informal gossip help share information.

FAA Service Difficulty Reports and Mechanical Malfunction Information Reports (collected and disseminated by the Helicopter Association International or HAI) are another good source of information.

These two databases--FAA SDRs and HAI's MMIR--are readily available on the web and provide a wealth of data on mechanical malfunctions and aircraft operating problems. The SDR data include reports from Canada, Australia, and the United States, a diverse and significant international data source. This information is timely. Both the SDR and MMIR databases describe mechanical malfunction problems before service bulletins or airworthiness directives are issued. However the websites are not conducive to detecting patterns and trends or identifying priorities. The research can also become time consuming.

A newly patented software product called JetStat has been developed for early identification, risk assessment, and prioritization of mechanical malfunctions in jets and turboprops. HeliStat is the equivalent product for helicopters.

JetStat is based on a simple premise--effectively analyze the available data to reduce unscheduled maintenance. There is a huge amount of valuable information concerning mechanical malfunctions sitting virtually undisturbed in the SDR and MMIR databases. It is undisturbed because until now there has been no effective way to analyze the huge volumes of information.

The number of SDR reports is now approaching 1.4 million and growing at 3,500 reports per month. For an individual, or even a team of individuals, to effectively analyze the data and prioritize the problems would be nearly impossible. What has been lacking is a tool to mine for the gold that exists within the databases, to change them from historical records to actionable, predictive information by which to reduce unscheduled maintenance events.

This goldmine of data has to date been underutilized because the focus has been on reporting individual events, rather than looking for meaning in the forest of events. JetStat connects the dots through a systematic analytical assessment of all the data to identify statistical trends or anomalies. When identified, these trends provide maintenance departments with an early warning system of problems with a high probability of recurring and provide the opportunity to take pre-emptive action.

The system analyzes the database at least weekly to identify areas statistically different from the norm. These unusual items are worthy of further investigation. Once identified, an alert is issued to JetStat users. The outcome is presented in a threat matrix report issued to the user, which indicates the seriousness of the risk and the likelihood of it occurring within the user's fleet.

The service is available through an e-mail subscription service using SDR and MMIR publicly available data. It can also be made available on a standalone basis to individual companies desiring an in-house system on which to run private legacy data for quality control purposes, track supplier quality, competitor comparisons, or to identify and act on internal trends and/or anomalies.

The following are two examples of SDRs for early identification and prevention of potential problems.

Example of early warning:

Cracks from rivet hole in 0.125 inch stamped attachment flange completely across flange separating it into two pieces. (Note) there is no danger of stabilizer separation from the aircraft as the flange is still riveted to the inboard rib in five other places, the stabilizer is clamped to the spar in four places and there is a stop assembly to prevent separation as well. This problem has been with us a long time but what brings this one to light is we have found six cracks on five aircraft in two months. Recommend crews be more vigilant on inspections to detect cracks before they have a chance to grow.

Example of preventive procedure:

During a routine inspection, the left aft wing attachment fitting was found to have a loose attaching bolt. This problem has been typical on a fleet of six aircraft of this model. The parts in question are: P/N-401-187 bolt (AN5-7A), P/N 66762-00 (plate-aft spar), and P/N 62448-02 (fitting-lower left fuselage). Only the left side fitting has been found loose on any of these aircraft. The manufacturer has been advised of this problem and has inspected the subject aircraft. The only preventive maintenance at this time is repetitive torqueing of the bolt, very often. -- By Dr. Alex Richman

A four-week free trial of this service is available by contacting Dr. Alex Richman at 902-423-5155 or arichman@algoplusaviation.com.