Sensor placement optimization for guided wave-based structural health monitoring

Abstract

Sensor placement optimization is an attempt to reduce the cost and enhance the detection performance in structural health monitoring (SHM) systems. This paper aims at studying sensor placement optimization for SHM systems. The attention is paid to lamb wave or guided wave-based SHM (GWSHM). By using detection theory and Bayes risk framework the expected cost (loss) of decision making or Bayes risk for SHM system is minimized and the optimal detector is derived. The global detection and false alarm rate are used for quantifying the detector performance. In this framework the sensor coverage, directionality and probabilities of damage occurrence are all accounted for. The effect of cross-correlation among actuator-sensor pairs is then considered by presenting an appropriate model for covariance structure. Applying the genetic algorithm, the global false alarm rate is minimized for a target global detection rate and different levels of correlation. In addition, the receiver-operating characteristic (ROC) is determined to analyze the effect of correlation on the system performance and optimal arrangement. For demonstration of the effect of cross-correlation on damage detection a numerical analysis is carried out using ABAQUS standard. Finally, it is concluded that by increasing the correlation among actuator-sensor pairs, the performance of the SHM system decreases.