Automatic Fault Identification in Sensor Networks Based on Probabilistic Modeling

This work proposes a mechanism able to automatically cat- egorize different types of faults occurring in critical infrastructures and especially water distribution networks. The mechanism models the rela- tionship exhibited among the sensor datastreams based on the assump- tion that its pattern alters depending on the fault type. The first phase includes linear time invariant modeling which outputs a parameters vec- tor. At the second phase the evolution of the parameter vectors is cap- tured via hidden Markov modeling. The methodology is applied on data coming from the water distribution network of the city of Barcelona. The corpus contains a vast amount of data representative of nine net- work states. The nominal is included for enabling fault detection. The achieved classification rates are quite encouraging.

NTALAMPIRAS Stavros;  GIANNOPOULOS Georgios; 

2016-09-16

Springer

JRC91498

978-3-319-31664-2,   

0302-9743,   

http://link.springer.com/chapter/10.1007%2F978-3-319-31664-2_35,    https://publications.jrc.ec.europa.eu/repository/handle/JRC91498,   

10.1007/978-3-319-31664-2_35,