TY - GEN
T1 - A Bayesian acoustic emission source location algorithm
T2 - 11th International Conference on Applications of Statistics and Probability in Civil Engineering, ICASP
AU - Schumacher, Thomas
AU - Straub, Daniel
PY - 2011
Y1 - 2011
N2 - The Acoustic Emission (AE) technique offers the opportunity to monitor infrastructure components in real-time. AE are stress waves initiated by sudden strain releases from within a solid due to over-loading. Witha network of sensors mounted to the structure's surface, stress waves can be detected and located. The main issue of traditional source location algorithms is that sources of uncertainty and variability in the material properties are essentially treated deterministically using mean values. A Bayesian framework for a probabilistic source location algorithm using Markov Chain Monte Carlo (MCMC) simulation whereby all model parameters are described with probability density functions was proposed by the authors earlier. The advantage of Bayesian networks is that they offer the flexibility to include additional observed data for enhanced prediction of future events. In the current study, an extension of the first basic model that was presented by the authors earlier is proposed that uses additional observed information in form of the maximum signal amplitude. It was found that optimal point predictions based on the extended model were more accurate than the predictions from a traditional source location algorithm as well as the basic probabilistic model.
AB - The Acoustic Emission (AE) technique offers the opportunity to monitor infrastructure components in real-time. AE are stress waves initiated by sudden strain releases from within a solid due to over-loading. Witha network of sensors mounted to the structure's surface, stress waves can be detected and located. The main issue of traditional source location algorithms is that sources of uncertainty and variability in the material properties are essentially treated deterministically using mean values. A Bayesian framework for a probabilistic source location algorithm using Markov Chain Monte Carlo (MCMC) simulation whereby all model parameters are described with probability density functions was proposed by the authors earlier. The advantage of Bayesian networks is that they offer the flexibility to include additional observed data for enhanced prediction of future events. In the current study, an extension of the first basic model that was presented by the authors earlier is proposed that uses additional observed information in form of the maximum signal amplitude. It was found that optimal point predictions based on the extended model were more accurate than the predictions from a traditional source location algorithm as well as the basic probabilistic model.
UR - http://www.scopus.com/inward/record.url?scp=84856723342&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84856723342
SN - 9780415669863
T3 - Applications of Statistics and Probability in Civil Engineering -Proceedings of the 11th International Conference on Applications of Statistics and Probability in Civil Engineering
SP - 91
EP - 98
BT - Applications of Statistics and Probability in Civil Engineering -Proceedings of the 11th International Conference on Applications of Statistics and Probability in Civil Engineering
Y2 - 1 August 2011 through 4 August 2011
ER -