Unsupervised Structure Damage Classification Based on the Data Clustering and Artificial Immune Pattern Recognition
This paper presents an unsupervised structure damage classification algorithm based on the data clustering technique and the artificial immune pattern recognition. The presented method uses time series measurement of a structure’s dynamic response to extract damage-sensitive features for the structure damage classification. The Data Clustering (DC) technique is employed to cluster training data to a specified number of clusters and generate the initial memory cell set. The Artificial Immune Pattern Recognition (AIPR) algorithms are integrated with the data clustering algorithms to provide a mechanism for the evolution of memory cells. The combined DC-AIPR method has been tested using a benchmark structure. The test results show the feasibility of using the DC-AIPR method for the unsupervised structure damage classification.
Keywordsstructural health monitoring unsupervised structure damage classification data clustering artificial immune pattern recognition
Unable to display preview. Download preview PDF.
- 1.Kolakowski, P.: Structural Health Monitoring - a Review with the Emphasis on Low-Frequency Methods. IPPT Engineering Transactions 55, 239–275 (2007)Google Scholar
- 2.Sazonov, E., Janoyan, K., Jha, R.: Wireless intelligent sensor network for autonomous structural health monitoring. Proc. SPIE - Int. Soc. Opt. Eng. 5384, 305–314 (2004)Google Scholar
- 4.Hart, E., Timmis, J.: Application areas of AIS: the past, the present and the future. Applied Soft Computing Journal, 191–201 (2008)Google Scholar
- 13.Structural Health Monitoring Benchmark Problem, http://mase.wustl.edu/wusceel/asce.sshm/benchmarks.htm
- 17.Chen, B., Zang, C.: Artificial immune pattern recognition for damage detection in structural health monitoring sensor networks. In: Proc. of SPIE: Smart Sensor Phenomena, Technology, Networks, and Systems, San Diego, California (2009)Google Scholar