UC San Diego

In many structural applications the use of composite material systems in both retrofit and new design modes has expanded greatly. The performance benefits from composites such as weight reduction with increased strength, corrosion resistance, and improved thermal and acoustic properties, are balanced by a host of failure modes whose genesis and progression are not yet well understood. As such, structural health monitoring (SHM) plays a key role for in-situ assessment for the purposes of performance/operations optimization, maintenance planning, and overall life cycle cost reduction. In this work, arrays of fiber Bragg grating optical strain sensors are attached to glass-epoxy solid laminate composite specimens that were subsequently subjected to specific levels of fully reversed cyclic loading. The fatigue loading was designed to impose strain levels in the panel that would induce damage to the laminate at varying numbers of cycles. The objectives of this test series were to assess the ability of the fiber Bragg grating sensors to detect fatigue damage (using previously developed SHM algorithms) and to establish a dataset for the development of a prognostic model to be applied to a random magnitude of fully reversed strain loading. The prognostic approach is rooted in the Failure Forecast Method, whereby the periodic feature rate-of-change was regressed against time to arrive at a failure estimate. An uncertainty model for the predictor was built so that a probability density function could be computed around the time-of-failure estimate, from which mean, median, and mode predictors were compared for robustness.