3 min read
In the past couple of decades, airline maintenance organizations and MRO shops have found many ways to increase efficiency and effectiveness, but one area of aviation that still has the capacity for optimization is the management of recurring defects.
The Challenge With Managing Recurring Aircraft Defects
Repeat defects are costly in terms of maintenance or customer satisfaction and—in some rare instances—can lead to tragic accidents and loss of life. It’s not that pilots and Aircraft Maintenance Technicians (AMTs) don’t report defects, but rather that the problem appears to have been resolved when in fact it has not, and symptoms return after a period of normal operation. Maintenance and Engineering (M&E) teams have difficulty in tracking and resolving such defects; i.e., which defects are related, where they occur, how often they occur, and whether the cause is ultimately resolved.
For many airlines, there is room for improvement in their M&E operations because they still rely on outdated or inefficient processes or technologies for managing recurring (chronic) defects. They may use an application built into their M&E record-keeping software or work from a dedicated report of potential recurring defects from the IT department. Such legacy systems or processes are time-consuming to use and less effective than they ought to be because they are highly prone to errors—mainly incorrect ATA coding of defects, and failure to recognize related defects that occur over larger time spans.
2 Common Problems With Outdated Legacy Systems
That said, it is important to address common problems with such legacy approaches to fully understand the extent of recurring issues. Here I highlight two major challenges: excess time/labor and human error.
Excess Time and Labor
The fact is, it’s very time-consuming to manually sift through via Excel or other static technology data in hundreds of pages of Maintenance Reports (MAREPs) and Pilot Reports (PIREPS). Experts spend a lot of their valuable time accommodating errors in the data they are given. First, they must manually sift through large data reports where a large percentage of what the report contains is noise—not valid repeat defects—that must be eliminated. Then, they must validate and clean up the valid recurring defects and assign them to be resolved. This includes searching for related defects that were not included in the original report. The next day, the process repeats, but rarely has their work from the previous day been preserved, so they find themselves reviewing the same defects again and again.
Human Error
Second, legacy processes or systems are prone to human error, because
- they rely on manual or human review and
- the excess workload increases the potential for missed alerts and inaccurate results.
A defect can have incorrect or varying ATA codes assigned to it. Maintenance technicians use their judgement to assign an ATA code to a problem, and often the assigned code is not the same as the code assigned to other related defects. Typically, the IT department takes a list of all repeating ATA codes and sends it to a maintenance analyst. If the IT reports rely on the code assigned by maintenance technicians to identify possible recurring defects, then that repeat issue could be overlooked.
By relying on ATA codes to report on specific types of defects the M&E department acts on reports that are rife with errors, full of false alerts or missing valid alerts because of miscoding. Acting on incomplete or inaccurate information is not a good thing!
The challenge for automation is to be resilient to erroneous ATA codes and be able to classify the defect based on what is written; i.e., the natural language. That’s why successful maintenance operations deploy technology that can look at the actual text, not the code, in logs to find recurring defects.
How A Sophisticated Recurring Defect Solution Can Help
Veryon Defect Analysis is a unique technology solution that goes above and beyond simple text mining and ATA code matching; it analyzes the text written by the pilot or mechanic and, based on what was written, will associate defects. It also has large dictionaries that enable it to handle slang, abbreviations, misspellings, airline-specific terminology for both findings related defects and eliminating false alerts. Because Veryon Defect Analysis searches text thoroughly via algorithms, natural language and fuzzy logic, it finds many more defects. However, it presents very few false alerts, so that M&E teams can spend less time distracted by false alerts and more time fixing chronic defects.
But Veryon Defect Analysis does much more than just identify recurring defects; it’s a one-stop-shop where M&E teams identify, manage and validate the corrective actions seen on repeat defects.