Publications

Abstract not provided.

Causal discovery algorithms construct hypothesized causal graphs that depict causal dependencies among variables in observational data. While powerful, the accuracy of these algorithms is highly sensitive to the underlying dynamics of the system in ways that have not been fully characterized in the literature. In this report, we benchmark the PCMCI causal discovery algorithm in its application to gridded spatiotemporal systems. Effectively computing grid-level causal graphs on large grids will enable analysis of the causal impacts of transient and mobile spatial phenomena in large systems, such as the Earth’s climate. We evaluate the performance of PCMCI with a set of structural causal models, using simulated spatial vector autoregressive processes in one- and two-dimensions. We develop computational and analytical tools for characterizing these processes and their associated causal graphs. Our findings suggest that direct application of PCMCI is not suitable for the analysis of dynamical spatiotemporal gridded systems, such as climatological data, without significant preprocessing and downscaling of the data. PCMCI requires unrealistic sample sizes to achieve acceptable performance on even modestly sized problems and suffers from a notable curse of dimensionality. This work suggests that, even under generous structural assumptions, significant additional algorithmic improvements are needed before causal discovery algorithms can be reliably applied to grid-level outputs of earth system models.

The increasing use of machine learning (ML) models to support high-consequence decision making drives a need to increase the rigor of ML-based decision making. Critical problems ranging from climate change to nonproliferation monitoring rely on machine learning for aspects of their analyses. Likewise, future technologies, such as incorporation of data-driven methods into the stockpile surveillance and predictive failure analysis for weapons components, will all rely on decision-making that incorporates the output of machine learning models. In this project, our main focus was the development of decision scientific methods that combine uncertainty estimates for machine learning predictions, with a domain-specific model of error costs. Other focus areas include uncertainty measurement in ML predictions, designing decision rules using multiobjecive optimization, the value of uncertainty reduction, and decision-tailored uncertainty quantification for probability estimates. By laying foundations for rigorous decision making based on the predictions of machine learning models, these approaches are directly relevant to every national security mission that applies, or will apply, machine learning to data, most of which entail some decision context.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

This report contains the written footprint of a Sandia-hosted workshop held in Albuquerque, New Mexico, June 22-23, 2016 on “Complex Systems Models and Their Applications: Towards a New Science of Verification, Validation and Uncertainty Quantification,” as well as of pre-work that fed into the workshop. The workshop’s intent was to explore and begin articulating research opportunities at the intersection between two important Sandia communities: the complex systems (CS) modeling community, and the verification, validation and uncertainty quantification (VVUQ) community The overarching research opportunity (and challenge) that we ultimately hope to address is: how can we quantify the credibility of knowledge gained from complex systems models, knowledge that is often incomplete and interim, but will nonetheless be used, sometimes in real-time, by decision makers?

Abstract not provided.

Genetic algorithms provide attractive options for performing nonlinear multi-objective combinatorial design optimization, and they have proven very useful for optimizing individual systems. However, conventional genetic algorithms fall short when performing holistic portfolio optimizations in which the decision variables also include the integer counts of multiple system types over multiple time periods. When objective functions are formulated as analytic functions, we can formally differentiate with respect to system counts and use the resulting gradient information to generate favorable mutations in the count variables. We apply several variations on this basic idea to an idealized hanging chain example to obtain >> 1000x speedups over conventional genetic algorithms in both single - and multi-objective cases. We develop a more complex example of a notional military portfolio that includes combinatorial design variables and dependency constraints between the design choices. In this case, our initial results are mixed, but many variations are still open to further research.

Abstract not provided.

In 2012, Hurricane Sandy devastated much of the U.S. northeast coastal areas. Among those hardest hit was the small community of Hoboken, New Jersey, located on the banks of the Hudson River across from Manhattan. This report describes a city-wide electrical infrastructure design that uses microgrids and other infrastructure to ensure the city retains functionality should such an event occur in the future. The designs ensure that up to 55 critical buildings will retain power during blackout or flooded conditions and include analysis for microgrid architectures, performance parameters, system control, renewable energy integration, and financial opportunities (while grid connected). The results presented here are not binding and are subject to change based on input from the Hoboken stakeholders, the integrator selected to manage and implement the microgrid, or other subject matter experts during the detailed (final) phase of the design effort.

Emerging semiconductor switches based on the wide-bandgap semiconductor GaN have the potential to significantly improve the efficiency of portable power applications such as transportable energy storage. Such applications are likely to become more widespread as renewables such as wind and solar continue to come on-line. However, the long-term reliability of GaN-based power devices is relatively unexplored. In this paper, we describe joint work between Sandia National Laboratories and MIT on highvoltage AlGaN/GaN high electron mobility transistors. It is observed that the nature of current collapse is a strong function of bias conditions as well as device design, where factors such as Al composition in the barrier layer and surface passivation play a large role. Thermal and optical recovery experiments are performed to ascertain the nature of charge trapping in the device. Additionally, Kelvin-force microscopy measurements are used to evaluate the surface potential within the device. © The Electrochemical Society.

The complexity of net-centric system of systems (SoS) being fielded today has the military leadership increasingly dependent on modeling and simulation (M&S) tools for evaluating performance. Several types of M&S tools are required to model different aspects of military systems, yet these tools often have different computational fidelities in terms of time and scale. Current approaches using direct information transfer between M&S tools, such as High Level Architecture (HLA) and MATREX, do not provide the mechanisms for disparate tools to make direct use of each other's information [1], [2]. Thus, many military SoS analyses assume perfect communications, an unrealistic assumption that leaves a gap for conducting more comprehensive analyses for large-scale, net-centric SoS problems. This research addresses this gap by developing general purpose methodologies to bridge the gap between diverse M&S tools resulting in a capability that enables military decision makers to evaluate comms system performance effects at a SoS level [3]. This paper discusses the methodology, including parameter selection, data generation, surrogate modeling and SoS simulation results. © 2012 IEEE.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Carrier generation characteristics in n-type substrate SiC MOS capacitors induced by sub-bandgap energy light are reported. The generation rate is high enough to create an inversion layer in ∼20 minutes with monochromatic light (front side illumination) of energy 2.1 eV (intensity ∼5×10 16 cm-2s-1) in 4H-SiC for electric fields smaller than 1 MV/cm. Generation and recovery results strongly indicate involvement of a metastable defect whose efficiency as a generation center increases under hole-rich and decreases under electron-rich conditions. The generation dependence on bias history and light energy shows the defect to have properties consistent with the metastable silicon vacancy / carbon vacancy-antisite complex (VSi/Vc-CSi). © (2012) Trans Tech Publications.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Flexible AC Transmission Systems (FACTS) use advanced power electronics to minimize reactive power loss on the grid. Power devices used in FACTS systems must be capable of switching several thousand amps at voltages of 1-10 kV. Traditionally, these systems have relied on silicon thyristors, but recently have also began to incorporate insulated gate bipolar transistors. FACTS systems present an opportunity for emerging SiC and GaN power transistors, which offer major efficiency gains. However, for these advanced materials to be considered for use in high consequence grid level systems like FACTS controllers, excellent reliability must be demonstrated. ©The Electrochemical Society.

Power conversion systems for energy storage and other distributed energy resource applications are among the drivers of the important role that power electronics plays in providing reliable electricity. Wide band gap semiconductors such as silicon carbide (SiC) and gallium nitride (GaN) will help increase the performance and efficiency of power electronic equipment while condition monitoring (CM) and prognostics and health management (PHM) will increase the operational availability of the equipment and thereby make it more cost effective. Voltage and/or temperature stress testing were performed on a number of SiC devices in order to accelerate failure modes and to identify measureable shifts in electrical characteristics which may provide early indication of those failures. Those shifts can be interpreted and modeled to provide prognostic signatures for use in CM and/or PHM. Such experiments will also lead to a deeper understanding of basic device physics and the degradation mechanisms behind failure.

The project's goals are: (1) use experiments and modeling to investigate and characterize stress-related failure modes of post-silicon power electronic (PE) devices such as silicon carbide (SiC) and gallium nitride (GaN) switches; and (2) seek opportunities for condition monitoring (CM) and prognostics and health management (PHM) to further enhance the reliability of power electronics devices and equipment. CM - detect anomalies and diagnose problems that require maintenance. PHM - track damage growth, predict time to failure, and manage subsequent maintenance and operations in such a way to optimize overall system utility against cost. The benefits of CM/PHM are: (1) operate power conversion systems in ways that will preclude predicted failures; (2) reduce unscheduled downtime and thereby reduce costs; and (3) pioneering reliability in SiC and GaN.

Abstract not provided.

This report provides the DOE and industry with a general process for analyzing power electronics reliability. The analysis can help with understanding the main causes of failures, downtime, and cost and how to reduce them. One approach is to collect field maintenance data and use it directly to calculate reliability metrics related to each cause. Another approach is to model the functional structure of the equipment using a fault tree to derive system reliability from component reliability. Analysis of a fictitious device demonstrates the latter process. Optimization can use the resulting baseline model to decide how to improve reliability and/or lower costs. It is recommended that both electric utilities and equipment manufacturers make provisions to collect and share data in order to lay the groundwork for improving reliability into the future. Reliability analysis helps guide reliability improvements in hardware and software technology including condition monitoring and prognostics and health management.

Abstract not provided.