Sensor Fault Identification Using Autoregressive Models and the Mutual Information Concept

Kraemer, P.; Fritzen, Claus Peter

doi:10.4028/www.scientific.net/KEM.347.387

Paper Titles

Fatigue and Fracture Evaluation of a 70 Year Old Steel Bridge
p.359

The Structural Stability Analysis of a Container Crane According to the Boom Shape Using Wind Tunnel Test
p.365

Measurement of Acoustic Emission Signals: Influence of the Couplant
p.375

A Numerical Determination of Acoustic Emission Sensor Response in Plates Using Dispersion Curves
p.381

Sensor Fault Identification Using Autoregressive Models and the Mutual Information Concept
p.387

Design and Numerical Simulation of an Optic Fibre Sensor for Damage Assessment of Structures
p.393

Broadband Source-Based Interrogation Scheme of Fiber Bragg Grating Sensors for Structural Health Monitoring
p.399

Energy Harvesting for Autonomous Sensing
p.405

Crack Detection in Beams by Means of the Driving Force Parameters Variation at Non-Linear Resonance Vibrations
p.413

HomeKey Engineering MaterialsDamage Assessment of Structures VIISensor Fault Identification Using Autoregressive...

Sensor Fault Identification Using Autoregressive Models and the Mutual Information Concept

Abstract:

This paper presents a combined approach for sensor fault identification looking for changes within one channel on one hand and for changes between the different channels on the other hand. The first method is based on the identification of autoregressive (AR) models from the reference time signals for each sensor channel separately. The reference models are then used for the prediction of the future sensors signals. The statistical properties of the residuals between this prediction and the true measurement allow a statement about changes of the sensor signals. The second method is based on the concept of mutual information between two signals X and Y from two different sensors. Mutual information or transinformation measures the information about the channel X that is shared by Y. This requires a certain redundancy of information represented in the different sensor signals. It can be seen that the mutual information changes as soon as a sensor fault occurs because the sensor fault information is not present in the other sensor signals.

Info:

Periodical:

Key Engineering Materials (Volume 347)

Main Theme:

Damage Assessment of Structures VII

Edited by:

L. Garibaldi, C. Surace, K. Holford and W.M. Ostachowicz

Pages:

387-392

DOI:

https://doi.org/10.4028/www.scientific.net/KEM.347.387

Citation:

P. Kraemer and C. P. Fritzen, "Sensor Fault Identification Using Autoregressive Models and the Mutual Information Concept", Key Engineering Materials, Vol. 347, pp. 387-392, 2007

Online since:

September 2007

Authors:

P. Kraemer, Claus Peter Fritzen

Keywords:

Autoregressive Models, Mutual Information (MI), Sensor Fault Identification, Structural Health Monitoring (SHM)

Export:

RIS, BibTeX

Price:

$41.00

Permissions:

Request Permissions

Add to Cart

References

[1] R. Dunia, S. J. Qin, T. F. Edgar and T. J. McAvoy in: AiChe Journal, Vol. 42, No. 10 (1996), pp.2797-2812.

[2] J. Kullaa in: Proc. ISMA2006 (on CD-ROM), Leuven (Sept. 18-20, 2006).

[3] G. Kerschen, P. De Boe, J. -C. Golinval and K. Worden in: Proc. 2nd Europ. Workshop on SHM 2004, (July 7-9, 2004), pp.819-827.

[4] D. L. Mattern, L. C. Jaw, T. -H. Guo, R. Graham and W. McCoy in: 34 th Joint Propulsion Conference, Cleveland, Ohio, USA (July 12-15, 1998).

[5] K. Worden in: Proc. IMAC XXI, Orlando, USA (2003).

[6] H. Sohn and C. R. Farrar in: Smart Materials and Structures, Vol. 10 (2001), pp.446-451.

[7] H. Sohn, C. R. Farrar, N. Hunter and K. Worden in: Transaction of the ASME, Vol. 123 (2001), pp.706-711.

[8] A. Kraskov, H. Stögbauer and P. Grassberger in: Physical Review E, 69 (2004), 066138I03.

[9] I. Trendafilova and W. Heylen in: Europ. COST F3 Conference on System Identification & SHM, Madrid, Vol. 1 (2000), pp.147-156.

[10] G. E. P. Box, G. M. Jenkins and G. C. Reinsel: Time Series Analysis, Forecasting and Control, 3. Edition, Prentice Hall, New Jersey (1994).

[11] M. B. Priestley: Spectral analysis and time series, Probability and mathematical statistics, Academic Press, Harcourt Brace Jovanovich Publishers (1992).

Paper TitlePages

Sensor Optimal Placement for Bridge Structure Based on Single Parents Genetic Algorithm with Different Fitness Functions

Authors: Han Bing Liu, Chun Li Wu, Jing Wang

Abstract: An investigation into sensor optimal placement for bridge structure based on Single Parents Genetic Algorithm with different fitness functions has been carried on. Five fitness functions are designed from several aspects on linear independence, orthogonality and energy of mode. The two-step method is proposed to determine the number of sensors firstly and then sensor position. An example of a large span arch bridge has proved the following facts: the Single Parents Genetic Algorithm is quite suitable to sensor optimal placement for bridge structure. Fitness functions designed by effective independence index and MAC and BHM combined index are more desirable than other evaluation indices. Two-step method used to determine the number of sensors and position is very effective.

115

Research on Monitoring of Machine Tools Based on Wireless Sensor Network

Authors: Zhong Qi Sheng, Liang Dong, Chang Ping Tang

Abstract: This paper discusses the structure of wireless sensor network (WSN) and the key technologies for the monitoring of machine tools. Multi-sensor is used to monitor the acoustic emission and vibration signal during the manufacturing process of machine tools. Vibration signals and acoustic characteristics are extracted by using wavelet analysis. Based on the fusion of BP artificial neural networks and multi-sensor information, the monitoring of machine tool is carried out in the environment of wireless sensor network.

616

Inertial Sensor Fault Diagnosis Based on an Improved Gain Principal Component Analysis Algorithm

Authors: Qing Hua Li, Yi Wang, Yang Pang

Abstract: An improved gain principle component analysis(PCA) algorithm is proposed for detecting the small deviation fault of the inertial sensor data. During calculating process of the Q and statistics, different gains are set to improve the small deviation fault detecting capability of some important variables. And the filtering technology is applied to reduce the noise of the sample data and emerge the misjudgment phenomenon. Numeric example result shows that the proposed algorithm can achieve fault diagnosis effectively compared with the conventional PCA algorithm.

Application of Integer-Coded Genetic Algorithm to Optimal Sensor Placement

Authors: Bo Ying Qin, Xian Kun Lin

Abstract: In the dynamic testing, the sensor positions have a major influence on the quality of the experimental modal parameters of a tested structure. In order to dispose sensors to reasonable degrees of freedom (DOF), and reflect adequately the dynamic characteristics of tested structure, the sensor positions must be optimized. In this paper, taking the combination of MAC matrix and Fisher information matrix (FIM) as optimization criteria, the integer-coded genetic algorithm (IGA) was applied to optimal sensor position problem (OSPP). The effect of optimization criteria and optimal method to optimal sensor positions were discussed. According to the results, the following conclusion is obtained: using MAC and FIM as optimal criteria, introducing the IGA into the OSPP, the optimal sensor positions can ensure the better linear independence of the mode shape vectors and the better estimation of the experimental modal parameters. Comparing with three existing optimal sensor placement methods, which are Guyan, effective independence (EI), and cumulative method based on QR decomposition (CQRD), their results of the optimal sensor positions indicated that the IGA is better than them.

1114

A New Approach to Analyzing Gas Mixtures Based on Improved BP Neural Network Algorithm

Authors: Teng Li, Xiao Mei Yuan, Shi Liang Yang, Xin Hui Zhang

Abstract: A new approach is presented for analyzing gas mixtures by transforming the problem into a pattern classification one to reduce the effect of the poor repeatability of sensor response on the prediction of gas concentration. The aim of numerical simulation is to determine how successfully the approach using the combination of artificial neural networks with multi-sensor arrays can analyze multi-component gas mixtures. The results indicate that the new approach is realistic for gas mixture analysis, and numerical simulation is a powerful tool to determine the architecture of a network. By constructing improved BP neural network algorithm and basic BP neural network into sensor array signal processing and extracting 6 component as the input of neural network, Our investigation results indicated that recognition results obtained from improved BP neural network algorithm more accuracy than the results obtained from basic BP neural network.

548