NASA Logo, National Aeronautics and Space Administration

+NASA Home

+Ames Home

+Intelligent Systems Division

Wiring Health Management
Intelligent Systems Division Banner

Physics Based Methods for Wire Fault Detection

The Electrical Wiring and Interconnect System (EWIS) in any vehicle is a critical, and sometimes overlooked, electronic subsystem where relatively minor issues can grow and eventually lead to serious safety problems like smoke, fire, and loss of critical system functionality. While technology for detecting hard faults (i.e., opens, shorts, and arcing) is available, this ability only enables mitigation after a serious electrical issue occurs, rather than preventing it from occurring in the first place.

Objective

This research seeks to help prevent the onset of serious wiring system failures through the development of advanced physics-based models and algorithms for detecting chafing, a common precursor to failure, in frequently used types of shielded impedance controlled cable (e.g., coax and twisted-shielded pair). The contributions from this work will advance an understanding of the fundamental principles, and overcome current technology barriers by addressing the issues listed below:

  • Changing impedance along the length of some cables causes large reflections that mask chafing fault signatures, especially in field environments.
  • Prior baseline measurements are required for the nominal un-faulted cable, which can then change over time in an operational environment.
  • Expert knowledge is needed to interpret the measured fault signatures, which are often buried in noise. While this knowledge can sometimes be automated using data driven supervised machine learning methods, this approach does not leverage the underlying physics of the problem.
  • Fault detection signal processing is usually based on ad-hoc methods such as classical correlation (matched filter), which do not correctly account for frequency attenuation with cable length or the fact that a single fault can produce multiple reflection “signatures” within a single measurement waveform.
  • The traditional impedance discontinuity approach and unknown velocity of propagation lead to a limited ability to map the fault reflection “signature” back to the physical fault size and location.

Approach

The above limitations are better understood and overcome through the development of physics-model based optimal fault detection algorithms that use time or frequency data to automatically estimate the model parameters, which include the physical fault location and size, and quantify error in a Bayesian probabilistic framework (including velocity of propagation). Under this paradigm, baseline measurements are not necessary and users no longer need to interpret raw time signatures, since the model provides these capabilities. Furthermore, the rigorous physics-based approach to fault detection in a single cable enables the characterization of the best possible fault detection trade-space that answers the question: how far away and how small can the fault be detected for a given cable? And answering that question in the context of physics, as opposed to field studies (which are important but also difficult to control, expensive, and time consuming), leads naturally to better cable and wiring system design principles. These capabilities all make steps towards enabling important aspects of condition based wiring maintenance with applications to commercial, general, and military aviation (not to mention a wide array of non-aerospace related fields).

Support

This research is supported by NASA's Aviation Safety Program as part of the Vehicle Systems Safety Technologies project, and is conducted by members of the Physics Based Methods Group.

Technical Reports

S. Schuet, D. Timucin, and K. Wheeler. Shielded twisted pair cable model for chafe fault detection via time-domain reflectometry. NASA Technical Report NASA/TM-2012-216001, National Aeronautics and Space Administration, March 2012.
[PDF] [NTRS]

Journal Articles

S. Schuet, D. Timucin, K. Wheeler, “Physics-Based Precursor Wiring Diagnostics for Shielded-Twisted-Pair Cable,” IEEE Transactions on Instrumentation and Measurement, vol. 64, num. 2, pp 378-391, February 2015. [IEEE Xplore]

S. Schuet, D. Timucin, K. Wheeler, “A Model-Based Probabilistic Inversion Framework for Wire Fault Detection using TDR,” IEEE Transactions on Instrumentation and Measurement, vol. 60, num. 5, pp 1654-1663, May 2011. [PDF]

S. Schuet, “Wiring Diagnostics via l1-Regularized Least Squares,” Sensors Journal, IEEE, vol.10, no.7, pp.1218-1225, July 2010. [PDF]

Conferences and Presentations

D. Timucin, S. Schuet, K. Wheeler, "Detection of chafe faults in shielded-twisted-pair cable," Aircraft Airworthiness and Sustainment Conference, Baltimore MD, April 2-5, 2012

S. Schuet, D. Timucin, K. Wheeler, “Modeling and Detecting Wire Faults,” NASA Aviation Safety Program Annual Technical Meeting, St. Louis MO, May 10-12, 2011.

K. Wheeler, D. Timucin, S. Schuet, “Verification of Small Hole Theory for Application To Wire Chafing Resulting in Shield Faults,” Aircraft Airworthiness and Sustainment Conference, San Diego CA, April 18-21, 2011.

D. Timucin, S. Schuet, K. Wheeler, “Electrical Wire Chafe Fault Detection Analysis,” Aircraft Airworthiness and Sustainment Conference, Austin TX, May 10-13, 2010.

S. Schuet, D. Timucin, K. Wheeler, “A Model-Based Probabilistic Inversion Framework for Wire Fault Detection using TDR,” IEEE International Instrumentation and Measurement Technology Conference, Austin TX, May 3-6, 2010.

S. Schuet, K. Wheeler, D. Timucin, M. Kowalski, P. Wysocki, “Model Based Inference for Wire Chafe Diagnosis,” Presented at the Aging Aircraft Conference, May 4-7, 2009. [PDF]

S. Schuet, K. Wheeler, D. Timucin, “Understanding Wire Chafing: Model Development and Optimal Diagnostics Using TDR,” Aviation Safety Technical Conference, October 23, 2008.

Software

Physics Based Wire Fault Detection Toolbox

Team

Stefan Schuet
Kevin Wheeler
Dogan Timucin
David Nishikawa

First Gov logo
NASA Logo - nasa.gov