An important challenge for the aircraft industry consists to predict the currents on the fastening assemblies in order to avoid sparking, which can lead to accident, especially for fuel tank fasteners. In the literature, it has been demonstrated that the contact resistance plays a major role in the current path on fasteners. Nevertheless, these contact resistances cannot be well determined and vary greatly. As a result, the prediction of current must be done in a statistical way. Usually, it requires several aircraft simulations with several set of contact resistances, which represents a significant computational cost. This article proposes a machine learning model, which allows us to predict the currents in the fastening assemblies of an aircraft fuel tank in a few seconds. This model is built from a database of FDTD simulations of the aircraft fuel tank in the lightning frequency range 100 Hz to 1 MHz. The FDTD modeling is depicted in detail in this article based on previous work. From this database, several machine leaning approaches are explored (k-nearest neighbors, support vector regression, XGBoost, and a neural network). As a result of this study, XGBoost presents the best performances. Further investigations using XGBoost highlights the ability of the model to predict well the current for most fasteners and frequencies, even with a small amount of simulations as training data. Moreover, the proposed model allows us to perform a parametric analysis, which underline the ability of the model to provide results in agreement with the physical effects of the issue (current paths, resistive effects, inductive effects, etc.). The results presented are promising for the use of the proposed methodology in the aeronautical industry.
Handling out-of-distribution (OOD) samples has become a major stake in the realworld deployment of machine learning systems. This work explores the application of self-supervised contrastive learning to the simultaneous detection of two types of OOD samples: unseen classes and adversarial perturbations. Since in practice the distribution of such samples is not known in advance, we do not assume access to OOD examples. We show that similarity functions trained with contrastive learning can be leveraged with the maximum mean discrepancy (MMD) two-sample test to verify whether two independent sets of samples are drawn from the same distribution. Inspired by this approach, we introduce CADet (Contrastive Anomaly Detection), a method based on image augmentations to perform anomaly detection on single samples. CADet compares favorably to adversarial detection methods to detect adversarially perturbed samples on ImageNet. Simultaneously, it achieves comparable performance to unseen label detection methods on two challenging benchmarks: ImageNet-O and iNaturalist. CADet is fully self-supervised and requires neither labels for in-distribution samples nor access to OOD examples. IntroductionWhile modern machine learning systems have achieved countless successful real-world applications, handling out-of-distribution (OOD) inputs remains a tough challenge of significant importance. The problem is especially acute for high-dimensional problems like image classification. Models are typically trained in a close-world setting but inevitably faced with novel input classes when deployed in the real world. The impact can range from displeasing customer experience to dire consequences in the case of safety-critical applications such as autonomous driving (Kitt et al., 2010) Preprint. Under review.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.