Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!usc!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!cbnews!cbnews!military From: wallfesh@drcvax.af.mil (SANDE WALLFESH) Newsgroups: sci.military Subject: Re: Aircraft Maintenance Message-ID: <1991Feb26.012441.6605@cbnews.att.com> Date: 26 Feb 91 01:24:41 GMT Sender: military@cbnews.att.com (William B. Thacker) Organization: AT&T Bell Laboratories Lines: 58 Approved: military@att.att.com From: "SANDE WALLFESH" Irv Chidsey (chidsey@smoke.brl.mil) writes: > [Quotes from Loyd Enochs (lenochs%drcoa1.decnet@drcvax.af.mil) on systems > for trend and failure analysis deleted.] > > Fifteen-twenty years ago I had a friend working on a similar > program for tracking maintainance of jeeps & trucks to get a handle on > what their replacement life might be. Data had to be transcribed from > paper records to punched cards. Stock numbers and mileage were a big > problem, either ellegible or wrong. At that time they had not succeeded > in tracking a vehicle to junking before the records were so messed up > they had to abandon the effort. That sounds like it was quite an effort! Irv's friend would have undoubtedly benefitted from barcode scanning, small handheld computers (CALS compliant, of course), and other technologies that weren't available in the days when punched cards reigned. For instance, many artificial intelligence techniques have moved from the laboratory to the depot (and to other places!) where they can save time, trouble, and paperwork. Recently, I received an announcement for an upcoming conference, Artificial Intelligence Applications for Military Logistics. The conference features a session just on maintenance diagnostics and it includes: - Knowledge Based Systems Role in Maintenance Diagnostics - A Data Fusion Approach to Integrated Diagnostics - A Poor Performance Item Expert - Expert Systems for Military Hardware Diagnostics - Integration of Neural Network and Expert Systems for Diagnostics and Prognostics - Application of Multimedia Expert Systems to the Test Environment Knowledge Based Systems Role in Maintenance Diagnostics, presented by WPAFB, is going to discuss reducing "retest OK" and "can not duplicate" problems with model based systems technology. During the conference's session on technology, there's a presentation on Applications of Neural Networks to Onboard Maintenance Diagnostics. It will describe a method for employing neural nets to address intermittent in-flight failures which often result in "can not duplicate" or "retest OK". Solutions to these problems, they say, can eliminate up to 50% of the maintenance activity for high performance fighters. Sande P.S. Readers of sci.military might be interested in "Artificial Intelligence and National Defense: Applications to C3I and Beyond," edited by Stephen J. Andriole, published by AFCEA International Press in 1987. Saundra K. Wallfesh | "If it's stupid but works, it isn't Dynamics Research | stupid." -- Amphibious Warfare wallfesh@drcvax.af.mil | Review, Spring 1989 Brought to you by Super Global Mega Corp .com