James E. Warner

IWSHM 2015: Probabilistic fatigue damage prognosis using surrogate models trained via three-dimensional finite element analysis

Utilizing inverse uncertainty quantification techniques, structural health monitoring (SHM) can be integrated with damage progression models to form a probabilistic prediction of a structure’s remaining useful life (RUL). However, damage evolution in realistic structures is physically complex. Accurately representing this behavior requires high-fidelity models which are typically computationally prohibitive. In this paper, high-fidelity fatigue crack growth simulation times are reduced by three orders of magnitude using a model based on a set of surrogate models trained via three-dimensional finite element analysis. The developed crack growth modeling approach is experimentally validated using SHM-based damage diagnosis data. A probabilistic prediction of RUL is formed for a metallic, single-edge notch tension specimen with a fatigue crack growing under mixed-mode conditions.

A Diagnosis-Prognosis Feedback Loop for Improved Performance Under Uncertainties

The feed-forward relationship between diagnosis and prognosis is the foundation of both aircraft structural health management and the digital twin concept. Measurements of structural response are obtained either in-situ with mounted sensor networks or offline using more traditional techniques (e.g., nondestructive evaluation). Diagnosis algorithms process this information to detect and quantify damage and then feed this data forward to a prognostic framework. A prognosis of the structures future operational readiness (e.g., remaining useful life or residual strength) is then made and is used to inform mission- critical decision-making. Years of research have been devoted to improving the elements of this process, but the process itself has not changed significantly. Here, a new approach is proposed in which prognosis information is not only fed forward for decision-making, but it is also fed back to the forthcoming diagnosis. In this way, diagnosis algorithms can take advantage of a priori information about the expected state of health, rather than operating in an uninformed condition. As a feasibility test, a diagnosis-prognosis feedback loop of this manner is demonstrated. The approach is applied to a numerical example in which fatigue crack growth is simulated in a simple aluminum alloy test specimen. A prognosis was derived from a set of diagnoses which provided feedback to a subsequent set of diagnoses. Improvements in accuracy and a reduction in uncertainty in the prognosis- informed diagnoses were observed when compared with an uninformed diagnostic approach.

Probabilistic Fatigue Damage Prognosis Using a Surrogate Model Trained Via 3D Finite Element Analysis

Utilizing inverse uncertainty quantification techniques, structural health monitoring data can be integrated with damage progression models to form probabilistic predictions of a structure’s remaining useful life. However, damage evolution in realistic structures is physically complex. Accurately representing this behavior requires high-fidelity models which are typically computationally prohibitive. In this paper, high-fidelity fatigue crack growth simulation times are significantly reduced using a surrogate model trained via finite element analysis. The new approach is applied to experimental damage diagnosis data to form a probabilistic prediction of remaining useful life for a test specimen under mixed-mode conditions. doi: 10.12783/SHM2015/299