Lower supportability cost is a major factor in procuring new aircraft and in affording the aircraft currently in inventory. This is a challenge in both the military and commercial sectors. This paper presents several initiatives at McDonnell Douglas Aerospace (MDA) that address lower supportability costs. These include: quality and optimization techniques to provide a means to value and allocate diagnostic methods; analysis of flight data using machine learning to characterize and understand intermittent faults; development of a compact fuzzy logic engine for aircraft diagnostic applications; diagnosis of aircraft subsystems using neural networks. These applications and the underlying technologies are discussed. Research has shown that these new technologies can enhance current data analysis techniques and provide fresh insight into current diagnostic deficiencies and future support requirements. Once accurate support costs are determined for individual subsystems, an informed decision can be made on whether to add additional diagnostic capabilities to that subsystem, based on the projected benefits. This analysis might also show that only specific types of faults should be targeted by the additional diagnostics. Certain types of fault classes, such as intermittent wiring, have proven to be especially difficult to troubleshoot with current diagnostic methods and have led excessive customer support costs. MDA has demonstrated in the laboratory that new software techniques using neural networks and fuzzy logic can be used to provide more accurate diagnostic solutions.<>
Application of emerging technologies to improve supportability
Proceedings of National Aerospace and Electronics Conference (NAECON'94) ; 1165-1172 vol.2
1994-01-01
532993 byte
Conference paper
Electronic Resource
English
Application of Emerging Technologies to Improve Supportability
British Library Conference Proceedings | 1994
|Improved Supportability Analysis
NTRS | 1996
|Artificial intelligence supportability
Tema Archive | 1987
|