Neural Network Applications Laboratory Implemented at the San Antonio Air Logistics Center

Abstract

A neural network applications laboratory was installed and implemented at the San Antonio Air Logistics Center, Kelly Air Force Base, Texas. Implementation was a part of a technology insertion thrust to apply technology for continuous product / process improvement in aircraft and engine maintenance operations. Applications laboratory capabilities were established by integration of knowledge, computer equipment, software tools, example methodologies, selected specific application procedures, reference materials and personnel training. The system is hosted on two SUN Microsystems workstations with support peripheral and an Ethernet link to the MicroVAX system that hosts the Computed X-ray Tomography (CAT) system. Commercial and public domain software tools were provided to perform routine data preparation and image processing tasks. Public domain neural network software tools were provided to enable user application. Three test examples were developed using test / image data from the computed X-ray tomography (CAT) system in the “Nondestructive Testing Section, Kelly Science and Engineering Branch at Kelly Air Force Base”. The system was validated with computed X-ray tomography data, but is applicable to a variety of nondestructive evaluation data that is collected and stored in electronic format. The system was fully functional on 15 June 1993.