In this report we discuss training a deep learning seismic signal detection model on 3-component stations from the International Monitoring System (IMS) using the PhaseNet architecture. Using 14 years of associated signals from the International Data Centre’s (IDC) Late Event Bulletin (LEB), we auto-curated training data consisting of signal windows containing associated arrivals, and noise windows that contain no LEB-associated signals. We trained several models using different waveform window durations (30 seconds and 100 seconds), with and without bandpass filtering. We evaluated the effectiveness of our models using associated signals from the Unconstrained Global Event Bulletin (UGEB) and found that several of our models outperformed the signal detections from the IDC’s Selected Event List 3 (SEL3) arrival table. The SEL3 bulletin evaluated on the UGEB dataset with 100-second waveform windows registered a precision and recall of .15 and .48, respectively, versus .19 and .59 for our filtered-data model. For the 30-second waveform window dataset, the SEL3 bulletin achieved a precision and recall of .31 and .47, respectively, versus .32 and .60 for our filtered-data model. Finally, our models detected signals from all source-to-receiver distances, suggesting it is feasible to use a single PhaseNet model for the IMS network.
This report documents the results of a long-term (5.79 year) exposure of 4-point bend corrosion test samples in the inlet and outlet vents of four spent nuclear fuel dry storage systems at the Maine Yankee Independent Spent Fuel Storage Installation. The goal of the test was to evaluate the corrosiveness of salt aerosols in a realistic near-marine environment, providing a data set for improved understanding of stress corrosion cracking of spent nuclear fuel dry storage canisters. Examination of the samples after extraction showed minor corrosion was present, mostly on rough-ground surfaces. However, dye penetrant testing showed that no SCC cracks were present. Dust collected on coupons co-located with the corrosion specimens was analyzed by scanning electron microscopy and leached to determine the soluble salts present. The dust was mostly organic material (pollen and stellate trichomes), with lesser detrital mineral grains. Salts present were a mix of sea-salts and continental salts, with chloride dominating the anions, but significant amounts of nitrate were also present. Both corrosion samples and dust samples showed evidence of wetting, indicating entry of water into the vents. The results of this field test suggest that the environment at Maine Yankee is not highly aggressive, although extrapolation from the periodically wetted vent samples to the hot, dry, canister surface may be difficult. No stress corrosion cracks were observed, but minor corrosion was present despite high nitrate concentrations in the salts. These observations may help address the ongoing question of the importance of nitrate in suppressing corrosion and SCC.
We can improve network pruning by leveraging the loss-topography extraction techniques used by projective integral updates for variational inference. Low-variance Hessians facilitate more aggressive pruning by providing better loss approximations when a parameter is removed.
A dry etching process to transfer the pattern of a photonic integrated circuit design for high-speed laser communications is described. The laser stack under consideration is a 3.2-µm-thick InGaAs/InAlAs/InAlGaAs epitaxial structure grown by molecular beam epitaxy. The etching was performed using Cl2-based inductively-coupled-plasma and reactive-ion-etching (ICP-RIE) reactors. Four different recipes are presented in two similar ICP-RIE reactors, with special attention paid to the etched features formed with various hard mask compositions, in-situ passivations, and process temperatures. The results indicate that it is possible to produce high-aspect-ratio features with sub-micron separation on this multilayer structure. Additionally, the results of the etching highlight the tradeoffs involved with the corresponding recipes.
The final quality of any AI/ML system is directly related to the quality of the input data used to train the system. In this case, we are trying to build a reliable image classifier that can correctly identify electrical components in x-ray images. The classification confidence is directly related to the quality of the labels in the training data, which are used in developing the AI/ML classifier. Incorrect or incomplete labels can substantially hinder the performance of the system during the training process, as it tries to compensate for variations that should not exist. Image labels are entered by subject matter experts, and in general can be assumed to be correct. However, this is not a guarantee, so developing ways to measure label quality and help identify or reject bad labels is important, especially as the database continues to grow. Given the current size of the database, a full manual review of each component is not feasible. This report will highlight the current state of the “RECON” x-ray image database and summarize several recent developments to try to help ensure high quality labeling both now and in the future. Questions that we hope to answer with this development include: 1) Are there any components with incorrect labels? 2) Can we suggest labels for components that are marked “Unknown”? 3) What kind of overall confidence do we have in the quality of the existing labels? 4) What systems or procedures can we put in place to maximize label quality?
Sandia National Laboratories (SNL) performed a high-altitude nuclear electromagnetic pulse (HEMP) critical generation station component vulnerability test campaign with a focus on high-frequency, conducted early-time (E1) HEMP for the Department of Energy (DOE) Office of Cybersecurity, Energy Security, and Emergency Response (CESER). This report provides vulnerability test results to investigate component response and/or damage thresholds to reasonable HEMP threat levels that will help to inform site vulnerability assessments, mitigation planning, and modeling calibrations. This work details testing of North American Electric (NAE) magnetic motor starters to determine the effects of conducted HEMP environments. Motor starters are the control elements that provide power to motors throughout a generating plant; a starter going offline would cause loss of power to critical pumps and compressors, which could lead to component damage or unplanned plant outages. Additionally, failed starters would be unable to support plant startup. Six industrial motor starters were tested: two 2 horsepower (HP) starters with breaker disconnects and typical protection equipment, two 20 HP starters with breaker disconnects, and two 20 HP starters with fused disconnects. Each starter was placed in a circuit with a generator and inductive motor matching the starter rating. The conducted EMP insult was injected on the power cables passing through the motor starter, with separate tests for the generator and motor sides of the starter.
Derivative computation is a key component of optimization, sensitivity analysis, uncertainty quantification, and the solving of nonlinear problems. Automatic differentiation (AD) is a powerful technique for evaluating such derivatives, and in recent years, has been integrated into programming environments such as Jax, PyTorch, and TensorFlow to support derivative computations needed for training of machine learning models, facilitating wide-spread use of these technologies. The C++ language has become the de facto standard for scientific computing due to numerous factors, yet language complexity has made the wide-spread adoption of AD technologies for C++ difficult, hampering the incorporation of powerful differentiable programming approaches into C++ scientific simulations. This is exacerbated by the increasing emergence of architectures, such as GPUs, with limited memory capabilities and requiring massive thread-level concurrency. C++ AD tools must effectively use these environments to bring novel scientific simulations to next-generation DOE experimental and observational facilities. In this project, we investigated source transformation-based automatic differentiation using LLVM compiler infrastructure to automatically generate portable and efficient gradient computations of Kokkos-based code. We have demonstrated that our proposed strategy is feasible by investigating the usage of a prototype LLVM-based source transformation tool to generate gradients of simple functions made of sequences of simple Kokkos parallel regions. Speedups of up to 500x compared to Sacado were observed on NVIDIA V100 GPU.