Structural health management of aerospace hotspots under fatigue loading

Sustainability and life-cycle assessments of aerospace systems, such as aircraft structures and propulsion systems, represent growing challenges in engineering. Hence, there has been an increasing demand in using structural health monitoring (SHM) techniques for continuous monitoring of these systems in an effort to improve safety and reduce maintenance costs. The current research is part of an ongoing multidisciplinary effort to develop a robust SHM framework resulting in improved models for damage-state awareness and life prediction, and enhancing capability of future aircraft systems. Lug joints, a typical structural hotspot, were chosen as the test article for the current study.;The thesis focuses on integrated SHM techniques for damage detection and characterization in lug joints. Piezoelectric wafer sensors (PZTs) are used to generate guided Lamb waves as they can be easily used for onboard applications. Sensor placement in certain regions of a structural component is not feasible due to the inaccessibility of the area to be monitored. Therefore, a virtual sensing concept is introduced to acquire sensor data from finite element (FE) models. A full three dimensional FE analysis of lug joints with piezoelectric transducers, accounting for piezoelectrical-mechanical coupling, was performed in Abaqus and the sensor signals were simulated. These modeled sensors are called virtual sensors. A combination of real data from PZTs and virtual sensing data from FE analysis is used to monitor and detect fatigue damage in aluminum lug joints. Experiments were conducted on lug joints under fatigue loads and sensor signals collected were used to validate the simulated sensor response. An optimal sensor placement methodology for lug joints is developed based on a detection theory framework to maximize the detection rate and minimize the false alarm rate. The placement technique is such that the sensor features can be directly correlated to damage. The technique accounts for a number of factors, such as actuation frequency and strength, minimum damage size, damage detection scheme, material damping, signal to noise ratio and sensing radius. Advanced information processing methodologies are discussed for damage diagnosis. A new, instantaneous approach for damage detection, localization and quantification is proposed for applications to practical problems associated with changes in reference states under different environmental and operational conditions. Such an approach improves feature extraction for state awareness, resulting in robust life prediction capabilities.