Sensor Network Configuration Effect on Detection Sensitivity of an Acousto-Ultrasound-Based Active SHM System

Abstract

An investigation was performed to assess the sensitivity of an acousto-ultrasound SHM system to changes in the configuration of the sensor network for the detection of fatigue cracks. Two different sensor networks were selected for the study on a riveted aluminum plate subjected to cyclic loads. The detection sensitivity of the system was quantified by the probability of detection (POD) approach that is used in traditional NDE but in this case modified for active SHM systems. The objective of the present work is to propose a POD analysis framework for reliable quantification of SHM systems. The proposed framework takes into account environmental and operating uncertainties, including varying temperature and loading conditions. The main concept is outlined and its operational applicability is addressed for hotspot damage detection monitoring. The issues of proper data collection, both experimental and model-based using numerical models, are discussed, while a model-based temperature compensation method is employed to treat the environmental uncertainties. In addition, the active sensing diagnostics based on ultrasonic guided waves and signal processing techniques are presented. Finally, the overview of the statistical POD analysis methodology is outlined and discussed. The effectiveness of the proposed framework is experimentally assessed via its application to an aluminum coupon, with two distinct sensor network configurations attached, under varying temperatures and loading conditions. The POD results with respect to increasing crack length are extracted and the SHM system’s reliability for damage quantification for both configurations are compared and discussed.