A Neural Network-based System for Automated Detection of Wire Insulation Chafing from TDR/FDR Data

This proposal is to develop an automated system for identification and characterization of chafing in the aircraft wiring. Chafing degrades wiring integrity and can lead to shorts and/or damaged data pathways. Shorts have been associated with deadly in-flight fires. Damaged data pathways have caused engine shutdowns, loss of control, and other safety hazards. Currently recognizing and measuring chafing requires an expensive and time-consuming process of visual inspection. We propose development of an automated system for identification and characterization of chafing from Time Domain Reflectometry (TDR) or Frequency Domain Reflectometry (FDR) data. At the end of Phase I, we will demonstrate the feasibility of using neural network technology to enable automated chafing identification from reflectometry data. We will also design and prototype an automated chafing identification system suitable for use in portable inspection equipment for aircraft wiring.