Many physics-based and surrogate models used in structural health monitoring are affected by different sources of uncertainty such as model approximations and simplified assumptions. Optimal structural health monitoring and prognostics are only possible with uncertainty quantification that leads to an informed course of action. In this article, a Bayesian neural network using variational inference is applied to learn a damage feature from a high-fidelity finite element model. Bayesian neural networks can learn from small and noisy data sets and are more robust to overfitting than artificial neural networks, which make it very suitable for applications such as structural health monitoring. Also, uncertainty estimates obtained from a trained Bayesian neural network model are used to build a cost-informed decision-making process. To demonstrate the applicability of Bayesian neural networks, an example of this approach applied to miter gates is presented. In this example, a degradation model based on real inspection data is used to simulate the damage evolution.

Condition assessments and rating systems are frequently used by field engineers to assess inland navigation assets and components. The goal of these assessments is to initiate effective risk-informed budget plans for maintenance and repair/replace. Ideally, a degradation model of every component failure mode in the gate would facilitate maintenance decision-making. However, sometimes there is no clear physical understanding how a damage progresses in time; for example, it isn't clear how the bearing gaps change in time in the quoin blocks of a miter gate. Therefore, this is one motivation for the framework proposed in this paper, which integrates Structural Health Monitoring with a Markov transition matrix built from historical condition assessment.To show the applicability of this framework, two examples are presented of how to find the optimal time to plan for maintenance of components in miter gates i) static maintenance planning based on operational condition assessment (OCA) ratings only and ii) dynamic maintenance planning based on integration of damage diagnostics based on monitoring data and failure prognosis based on OCA ratings. In addition, this paper presents a new Bayesian approach to estimate the ratio of errors in the OCA ratings, which allows for improved accuracy in OCA rating-based prognosis.

No abstract

The global precipitation measurement (GPM) mission is an international satellite mission to obtain accurate observations of precipitation on a global scale every 3 h. Its (GPM) core satellite was launched on 27 February 2014 with two science instruments: the microwave imager and the dual-frequency precipitation radar. Ground validation is an integral part of the GPM mission where instruments are deployed to complement and correlate with spacecraft instruments. The dual-frequency, dual-polarization, Doppler radar (D3R) is a critical ground validation instrument that was developed for the GPM program. This paper describes the salient features of the D3R in the context of the GPM ground validation mission. The engineering and architectural overview of the radar is described, and observations from successful GPM ground validation field experiments are presented.