Real-time Nondestructive Structural Health Monitoring using Support Vector Machines and Wavelets

We present an alternative to visual inspectionfor detecting damage to civil infrastructure. We describe areal-time decision support system for nondestructive healthmonitoring. The system is instrumented by an integratednetwork of wireless sensors mounted on civil infrastructuressuch as bridges, highways, and commercial and industrial facilities. To address scalability and power consumption issuesrelated to sensor networks, we propose a three-tier systemthat uses wavelets to adaptively reduce the streaming dataspatially and temporally. At the sensor level, measurementdata is temporally compressed before being sent upstream tointermediate communication nodes. There, correlated datafrom multiple sensors is combined and sent to the operation center for further reduction and interpretation. At eachlevel, the compression ratio can be adaptively changed viawavelets. This multi-resolution approach is useful in optimizing total resources in the system. At the operationcenter, Support Vector Machines (SVMs) are used to detectthe location of potential damage from the reduced data. Wedemonstrate that the SVM is a robust classifier in the presence of noise and that wavelet-based compression gracefullydegrades its classification accuracy. We validate the effectiveness of our approach using a finite element model of theHumboldt Bay Bridge. We envision that our approach willprove novel and useful in the design of scalable nondestructive health monitoring systems.