skip to main content

Search for: All records

Creators/Authors contains: "Bourgeois, D."

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. Dissimilar metal welds (DMWs) are routinely used in the oil and gas industries for structural joining of high-strength steels to eliminate the need for post-weld heat treatment (PWHT) in field welding. Hydrogen-assisted cracking (HAC) can occur in DMWs during subsea service under cathodic protection. DMWs of two material combinations, 8630 steel/FM 625 and F22 steel/FM 625, produced with two welding procedures, non-temper bead (BS1) and temper bead (BS3), in the as-welded and PWHT conditions were investigated in this study. These DMWs were subjected to metallurgical characterization and testing with the delayed hydrogen cracking test (DHCT) to identify the effects of base metal composition, welding and PWHT procedures on their HAC susceptibility. The HAC susceptibility was ranked using the time to failure in the DHCT at loads equivalent to 90% of the base metal yield strength (YS) and the apparent stress threshold for HAC. A criterion for resistance to HAC in the testing conditions of DHCT was also established. The results of this study showed that 8630/FM 625 DMWs were more susceptible to HAC than the F22/FM 625 DMWs. PWHT did not sufficiently reduce the HAC susceptibility of the 8630/FM 625 and F22/FM 625 BS1 welds. DMWs produced using BS3 performed better than BS1 DMWs. The post-weld heat-treated F22/FM 625 BS3 DMW passed the HAC resistance criterion. 
    more » « less
  2. Dissimilar metal welds (DMWs) are commonly used when a high strength steel is overlaid with a corrosion resistant alloy (CRA) for petrochemical applications. There have been reported failures of these DMWs during subsea service while under cathodic protection (CP). These failures are caused by local hydrogen embrittlement of susceptible microstructures that form at the weld fusion boundary. Hydrogen-assisted cracking (HAC) occurs as a result of the local embrittlement and is influenced by base/filler metal combinations, and welding and post-weld heat treatment (PWHT) procedures. A delayed hydrogen cracking test was used to simulate tensile load and hydrogen charging on 8630-FM 625 weld. The failure of this sample was recorded using a high-speed camera to capture the crack initiation and propagation during failure. Fractography was performed using a scanning electron microscope (SEM) along with energy dispersive spectroscopy (EDS). The fracture surfaces, EDS measurement and video timestamps revealed brittle fracture nucleation in the planar growth and CGHAZ regions of the weld. The cracking continued to propagate through the same regions of the weld leading to final ductile failure (microvoid coalescence) in the cellular dendritic region of the weld. 
    more » « less
  3. null (Ed.)
  4. Graph Neural Networks (GNNs) are a powerful tool for machine learning on graphs. GNNs combine node feature information with the graph structure by recursively passing neural messages along edges of the input graph. However, incorporating both graph structure and feature information leads to complex models and explaining predictions made by GNNs remains unsolved. Here we propose GNNEXPLAINER, the first general, model-agnostic approach for providing interpretable explanations for predictions of any GNN-based model on any graph-based machine learning task. Given an instance, GNNEXPLAINER identifies a compact subgraph structure and a small subset of node features that have a crucial role in GNN’s prediction. Further, GNNEXPLAINER can generate consistent and concise explanations for an entire class of instances. We formulate GNNEXPLAINER as an optimization task that maximizes the mutual information between a GNN’s prediction and distribution of possible subgraph structures. Experiments on synthetic and real-world graphs show that our approach can identify important graph structures as well as node features, and outperforms alternative baseline approaches by up to 43.0% in explanation accuracy. GNNEXPLAINER provides a variety of benefits, from the ability to visualize semantically relevant structures to interpretability, to giving insights into errors of faulty GNNs. 
    more » « less