Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “probabilistic methods”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 793 records · Page 44
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Continued Discussion of Failure Mode Modeling and Overall Component Reliability: Is the Data Missing or Censored?

This paper is a continuation of a paper presented at the 13th Probabilistic Safety Assessment and Management Conference, in which a discussion of modeling failure modes of complex components was presented; see Paulos and Smith (2016). The paper demonstrated how the typical method of treating failure modes as being exponential in nature may yield optimistic predictions when estimating how improvements to components will perform in the real world. Instead of relying on textbook methods, it is more accurate to model the failure modes as a race in time; unfortunately, this does not give a closed-form solution. This work uses a simulation to solve for the model of the world, and is compared to the standard methodology of treating the failure modes as exponential random failures. The standard method is shown to have optimistic predictions, which will lead to prediction errors when failure modes were removed or “fixed.” The failure mode methodology presented in the first paper treated the data as being censored when the test stopped. In this paper, we will compare the results from treating the data as both censored and missing, and comparisons will be shown showing which method is more accurate and under what conditions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Fast and efficient identification of anomalous galaxy spectra with neural density estimation

ABSTRACT Current large-scale astrophysical experiments produce unprecedented amounts of rich and diverse data. This creates a growing need for fast and flexible automated data inspection methods. Deep learning algorithms can capture and pick up subtle variations in rich data sets and are fast to apply once trained. Here, we study the applicability of an unsupervised and probabilistic deep learning framework, the probabilistic auto-encoder, to the detection of peculiar objects in galaxy spectra from the SDSS survey. Different to supervised algorithms, this algorithm is not trained to detect a specific feature or type of anomaly, instead it learns the complex and diverse distribution of galaxy spectra from training data and identifies outliers with respect to the learned distribution. We find that the algorithm assigns consistently lower probabilities (higher anomaly score) to spectra that exhibit unusual features. For example, the majority of outliers among quiescent galaxies are E+A galaxies, whose spectra combine features from old and young stellar population. Other identified outliers include LINERs, supernovae, and overlapping objects. Conditional modelling further allows us to incorporate additional information. Namely, we evaluate the probability of an object being anomalous given a certain spectral class, but other information such as metrics of data quality or estimated redshift could be incorporated as well. We make our code publicly available.

Böhm, Vanessa↗
DOE OSTI2021

Validation of Fault Displacements from Dynamic Rupture Simulations against the Observations from the 1992 Landers Earthquake

ABSTRACT Coseismic fault displacements in large earthquakes have caused significant damage to structures and lifelines on and near fault lines. Coseismic displacements represent a real threat, especially to distributed infrastructure systems. For infrastructure systems that can not avoid active faults, engineering displacement demands are defined using probabilistic fault-displacement hazard analyses (PFDHA). However, PFDHA models are sparse and poorly constrained partly due to the scarcity of detailed fault-displacement observations. Advancements in dynamic rupture simulation methods make them an attractive approach to address this important issue. Because fault displacements can be simulated for various geologic conditions as constrained by current knowledge about earthquake processes, they can be used to supplement the observation datasets. In addition to providing on-fault displacements, when used with appropriate constitutive models for the bulk medium, they can capture off-fault distributed inelastic deformations as well. For viable extrapolation, simulations must first be validated against data. In this article, we summarize the calibration and validation of the dynamic rupture model against the observations of the well-documented 1992 Landers earthquake. We defined a preferred model that reproduces several first-order fault-displacement metrics such as the on-fault partition of the total displacement, the mean fault-zone width, and the location of the peak displacement. Simulated ground motions consistent with the observations ensure that all physics important to modeling have been properly parameterized. For the extrapolation, we generated a suite of dynamic rupture models to quantify expected fault-displacement metrics, their intercorrelations, and magnitude dependencies, which are in part supported by the Landers and other recent earthquakes. Our validation and extrapolation exercise paves the way for using dynamic rupture modeling to quantitatively address fault-displacement hazard on a broader scale. The results are promising and are expected to be useful to inform PFDHA model development.

Geochemistry & Geophysics↗
NASA NTRS1973

A unified method for analyzing mission reliability for fault tolerant computer systems.

For fault-tolerant computer systems consisting of multiple classes of modules, a unified method for analyzing mission reliability is proposed and evaluated. The analysis proceeds by generalizing the notions of standby and N modular redundancy into a concept called hybrid-degraded redundancy. The probabilistic evaluation of the unified redundancy concept is then developed to yield, for a given modular class, the joint distribution of success and the number of nonfailed modules from that class, at special times. With this information, a Markov chain analysis gives the reliability of an entire sequence of phases (mission profile).

Bricker, J. L.↗
NASA NTRS2008

Assuring Life in Composite Systems

A computational simulation method is presented to assure life in composite systems by using dynamic buckling of smart composite shells as an example. The combined use of composite mechanics, finite element computer codes, and probabilistic analysis enable the effective assessment of the dynamic buckling load of smart composite shells. A universal plot is generated to estimate the dynamic buckling load of composite shells at various load rates and probabilities. The shell structure is also evaluated with smart fibers embedded in the plies right below the outer plies. The results show that, on the average, the use of smart fibers improved the shell buckling resistance by about 9% at different probabilities and delayed the buckling occurrence time. The probabilistic sensitivities results indicate that uncertainties in the fiber volume ratio and ply thickness have major effects on the buckling load. The uncertainties in the electric field strength and smart material volume fraction have moderate effects and thereby in the assured life of the shell.

Chamis, Christos c.↗
DOE OSTI2024

Assessing and benchmarking the fidelity of posterior inference methods for astrophysics data analysis

In this era of large and complex astronomical survey data, interpreting, validating, and comparing inference techniques becomes increasingly difficult. This is particularly critical for emerging inference methods like Simulation-Based Inference (SBI), which offer significant speedup potential and posterior modeling flexibility, especially when deep learning is incorporated. We present a study to assess and compare the performance and uncertainty prediction capability of Bayesian inference algorithms – from traditional MCMC sampling of analytic functions to deep learning-enabled SBI. We focus on testing the capacity of hierarchical inference modeling in those scenarios. Before we extend this study to cosmology, we first use astrophysical simulation data to ensure interpretability. We demonstrate a probabilistic programming implementation of hierarchical and non-hierarchical Bayesian inference using simulations derived from the DeepBench software library, a benchmarking tool developed by our group that generates simple and controllable astrophysical objects from first principles. This study will enable astronomers and physicists to harness the inference potential of these methods with confidence.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

Probabilistic Physics-Informed Graph Convolutional Network for Active Distribution System Voltage Prediction

Here this letter proposes a novel data-driven probabilistic physics-informed graph convolutional network (GCN) for active distribution system voltage prediction with PVs and EVs. It leverages both measurements and network topology to accurately and efficiently predict node voltages without the need for an accurate distribution system power flow model. The dropout-enabled Bayesian inference is developed to achieve uncertainty quantification of the voltage prediction. Thanks to the network model embedding, it also has robustness against topology changes, a key difference with existing machine learning-based approaches. Comparison results with other state-of-the-art machine learning methods on a realistic 759-node distribution system demonstrate that the proposed method can achieve better accuracy and robustness under different scenarios.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2020

Thermal Overloading Risk Mitigation With a Semi-Analytical Probabilistic Model on Branch Current

A semi-analytical formulation is presented in this paper for the probability computation of branch current in multiphase systems. The developed formula is derived based on the linear power flow model in rectangular coordinates. The system uncertainty injections can be renewable energy resources or loads and are modeled using a Gaussian mixture model (GMM). The developed formula can be used to compute the line current violation probability as well as integrate into optimal power flow problem as chance-constraint relaxation. Here, the proposed formula is first compared with the Matlab embedded numerical integration function to show its performance. Besides, the semi-analytical formula is validated and compared with the Monte Carlo simulation method using the IEEE 123-bus system, EPRI Ckt5, and Ckt7 systems.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2026

Active learning using hybrid surrogate tool life modeling for machining process optimization

Here, this paper describes an active learning approach for part-to-part iterative machining process optimization using a hybrid surrogate tool life model. A probabilistic interpolating tool life model is developed by combining the empirical Taylor-type tool life equation and the model fit error. The probabilistic tool life model is then used to calculate the machining cost per part distribution. The optimal machining parameters are selected using an expected improvement in machining cost per part criterion. The method is validated numerically using experimental results; the results show a median convergence error of 2.2% after three tests over 400 simulations. The method is validated experimentally on two industrial applications for Ti-6Al-4V roughing resulting in a cost per part reduction greater than 23% after two tests. The described method is a robust solution for rapid convergence to optimal machining parameters in an industrial production environment.

Active learning↗
DOE OSTI2024

Machine Learning‐Assisted Microearthquake Location Workflow for Monitoring the Newberry Enhanced Geothermal System

Abstract Enhanced geothermal systems (EGS) offer a sustainable energy source but face challenges in accurately locating microearthquakes induced during reservoir stimulation. Locating these microearthquakes provides reliable feedback on the stimulation progress. Current deep learning methods for locating earthquakes require extensive data sets for training, which is problematic as detected microearthquakes are often limited. To address the scarcity of training data, we propose a practical workflow using probabilistic multilayer perceptron (PMLP) which predicts microearthquake locations from cross‐correlation time lags in waveforms. Utilizing a 3D velocity model of Newberry site derived from ambient noise interferometry, we generate numerous synthetic microearthquakes and 3D acoustic waveforms for PMLP training. Accurate synthetic tests prompt us to apply the trained network to the 2012 and 2014 stimulation field waveforms. To enhance the accuracy of source localization, we carefully handpick the P‐arrival times. Predictions on the 2012 stimulation data set show major microseismic activity at depths of 0.5–1.2 km, correlating with a known casing leakage scenario. In the 2014 data set, the majority of predictions concentrate at 2.0–2.9 km depths, consistent with results obtained from conventional physics‐based inversion, and align with the presence of natural fractures from 2.0 to 2.7 km. We validate our findings by comparing the synthetic and field picks, demonstrating a satisfactory match for the first arrivals. By combining the benefits of quick inference speeds and accurate location predictions, we demonstrate the feasibility of using realistic synthetic data set to locate microseismicity for EGS monitoring.

15 GEOTHERMAL ENERGY↗
DOE OSTI2025

Using the optimal combined index weight ratio to improve the probability of anomaly detection in big area additive manufacturing

Big Area Additive Manufacturing (BAAM) of composites requires significant time, energy, and material, so it is critical to reduce production inefficiencies to make functional parts without multiple iterations. Statistical process control coupled with Principal Component Analysis (PCA) is a powerful technique that provides a quick, computationally inexpensive, and intuitive way for operators to detect defects that form in a manufacturing process without massive datasets. Recently, a combined index that is a weighted sum of the Hotelling's T 2 and squared residual error statistics has been proposed that can be monitored in one chart, improving interpretation accuracy and simplicity. However, the literature does not offer a formal method to optimise the weights. Here, we introduce two new approaches to the traditional weight selection approach using simulated and BAAM image data. Approach 1 uses a theoretically motivated optimum inspired by probabilistic principal component analysis. Approach 2 systematically varies the ratio of the weights to find the optimum. We show that approach 1 delivers optimal anomaly detection performance in select cases while approach 2 fares better in practice. Surprisingly, we also show that choosing a more complex PCA model has a minimal negative impact on anomaly detection performance compared to a more simplistic model.

3-dimensional printing↗
DOE OSTI2023

From Optimization to Sampling Through Gradient Flows

Optimization and sampling algorithms play a central role in science and engineering as they enable finding optimal predictions, policies, and recommendations, as well as expected and equilibrium states of complex systems. The notion of “optimality” is formalized by the choice of an objective function, while the notion of an “expected” state is specified by a probabilistic model for the distribution of states. Optimizing rugged objective functions and sampling multimodal distributions is computationally challenging, especially in high-dimensional problems. Here, for this reason, many optimization and sampling methods have been developed by researchers working in disparate fields such as Bayesian statistics, molecular dynamics, genetics, quantum chemistry, machine learning, weather forecasting, econometrics, and medical imaging.

Trillos, N. García↗
DOE OSTI2023

Load Shedding for Voltage Regulation With Probabilistic Agent Compliance

With the increased observability and controllability of distribution systems, the share of behind-the-meter systems is trending upwards rapidly. As a consequence, the impact of human behaviors on system performance can no longer be ignored and should be reflected in the energy management system models. In this paper, we discuss the problem of distribution system voltage control by active power curtailment where the agent compliance of the load curtailment signal is probabilistic. We discuss the modeling of the optimal voltage control problem with probabilistic agent compliance as a chance-constrained optimization problem, its tractable safe approximation using convex restriction, and a scenario-based mixed-integer reformulation as well as the associated solution method based on augmented Lagrangian method. The numerical simulation on IEEE test system validates the effectiveness of the proposed approach in obtaining high-quality feasible load curtailment signal with low computational cost, which makes it a viable tool for real time decision making.

augmented Lagrangian method↗
DOE OSTI2020

Probabilistic Solar Power Forecasting Using Bayesian Model Averaging

There is rising interest in probabilistic forecasting to mitigate risks from solar power uncertainty, but the numerical weather prediction (NWP) ensembles readily available to system operators are often biased and underdispersed. We propose a Bayesian model averaging (BMA) post-processing method suitable for forecasting power from utility-scale photovoltaic (PV) plants at multiple time horizons up to at least the day-ahead timescale. BMA is a kernel dressing technique for NWP ensembles in which the forecast is a weighted sum of member-specific probability density functions. We tailor BMA for utility-scale PV forecasting by modeling power clipping at the AC inverter rating and advance the theory of BMA with a new beta kernel parameterization that accommodates theoretical constraints not previously addressed. BMA is demonstrated for a case study of 11 utility-scale PV plants in Texas, forecasting at hourly resolution for the complete year 2018. BMA's mixture-model approach mitigates underdispersion of the raw ensemble to significantly improve forecast calibration, while consistently outperforming an ensemble model output statistics (EMOS) parametric approach from the literature. At 4-hour lead time, the BMA post-processing achieves continuous ranked probability skill scores of 2--36% over the raw ensemble, with consistent performance at multiple lead times suitable for power system operations.

14 SOLAR ENERGY↗
DOE OSTI2020

Probabilistic Forecasting of Generators Startups and Shutdowns in the MISO System Based on Random Forest

Solving security constrained unit commitment (SCUC) problems to plan an economical generation schedule for day-head electricity market has been an important research topic in recent years. Mixed integer programming method (MIP), the-state-of-art approach for solving SCUC problem, is known computationally hard when the number of binary status variables is large. In this paper, a machine learning-based algorithm - random forest (RF), was applied to forecast the startups (SU) and shutdowns (SD) hours of generators, based on historical hourly system condition observations in the Midcontinent Independent System Operator (MISO) system. The main purpose is to reduce the number of binary status variables, by fixing the SU/SD hours to a narrow range of high confidence. This would significantly reduce the size of the decision space, and therefore speed up SCUC solutions with reduced uncertainty.

Lin, Xinming↗
NASA NTRSDate not supplied

Sensorimotor Application of Proposed Methods to Combine the Effects of Multiple Countermeasures for PRisM

Risk associated with human systems is challenging to quantify but is critical for the mission planning and decision making required to enable future Lunar and Martian missions. To address this gap, the Crew Health and Performance-Probabilistic Risk Assessment (CHP-PRA) project is developing an integrated computational model for CHP mission risk. Much like how MEDPRAT is designed to allow medical resource trades informed by medical risk, CHP-PRA will enable analogous trades in human system risks across multiple CHP functions and capabilities. Human performance is one component of the risk intended to be captured by CHP-PRA through the Performance Risk Model (PRisM). The sensorimotor risk area provides a good frame of reference for investigating the structure of a performance model because most tasks that astronauts are expected to perform require input from the sensory system and/or movement/motor control. Additionally, sensorimotor countermeasures are an area of particular concern for NASA’s Human Research Program because of the increased sensorimotor risk associated with surface operations in Lunar and Martian missions. Thus, a tool that can quickly compare risk reductions of potential countermeasures would be beneficial in guiding research and development of effective countermeasures. In this proof of concept, we present a systematic way to combine multiple performance data sets for humans subjected to different countermeasures such that we can predict the countermeasure(s) that optimize astronaut performance on relevant tasks. PRisM assumes that both the tests that are used to measure countermeasure effectiveness (input data) and the tasks we use to represent astronaut performance, can be broken down and represented as a function/vector of the human systems required to perform that test/task. Through mathematical combination, test data are used to predict performance on astronaut tasks that use similar systems. We propose that when combining countermeasures evaluated using the same test, that only one value should be used to represent their combined effectiveness. We start our analysis with the assumption that two countermeasures together will perform better than each countermeasure individually. Our initial implementation of this framework compares various space motion sickness countermeasures and the most up to date analysis will be demonstrated at the IWS.

Caroline R Austin↗
DOE OSTI2024

Denoising diffusion probabilistic models for generative alloy design

Inverse material design is an extremely challenging optimization task made difficult by, in part, the highly nonlinear relationship linking performance with composition. Quantitative approaches have improved significantly owing to advances in high throughput experimentation and computational thermodynamics. However, existing physics-based tools are mostly forward models; input a chemistry and obtain a prediction. More recently the materials community has leveraged advances in the machine learning community to establish novel inverse design frameworks. Very recently denoising diffusion probabilistic models have been shown to be extremely powerful generators producing synthetic data of various modalities e.g. images, text, audio, tables, etc.. In this work a novel framework for alloy design and optimization is proposed leveraging these class of models. Five key generative tasks are demonstrated (1) unconditional generation (2) composition conditioned generation (3) property conditioned generation (4) multi-feedstock conditioned generation and (5) generative optimization. These methods were tested on three case studies: high entropy alloy design, superalloy binder jet additive manufacturing, and in-situ dual-feedstock wire-arc additive manufacturing. Results indicate that the established models are extremely flexible, expressive, and robust. The architecture’s flexibility and training procedure empower the model to learn complex intra-compositional and composition-property relationships. Furthermore, the probabilistic nature of these models makes them well suited for addressing solution non-uniqueness and tackling uncertainty quantification tasks. While the fidelity and quantity of the underlying training data is paramount, we envision that future alloy design frameworks will make extensive use of these kinds of machine learning models as “search” tools bolstering the utility of experimental and computational approaches.

36 MATERIALS SCIENCE↗
NASA NTRSDate not supplied

Blind Validation Study of PRICE TruePlanning and SEER-H

Two of the primary parametric costing tools used to estimate the development and production cost of future spacecraft hardware are PRICE TruePlanning by PRICE Systems, and System Estimation and Evaluation of Resources-Hardware (SEER-H) by Galorath. These are standard tools used by NASA and industry to estimate the cost of new aerospace hardware. This study is an independent verification of the accuracy of these tools and was originally published in 2018. The purpose of this presentation is to circulate the findings of that study among the NASA cost community. Both PRICE Systems and Galorath have completed internal validation studies of their parametric cost estimating tools; however, they only provided the results of the studies and did not detail the exact methods used to perform the validation. In the present study, cost estimators used PRICE TruePlanning and SEER-H to estimate the cost of twelve different past NASA science missions. The estimators were prevented from knowing the actual cost of the missions in an effort to minimize cognitive biases. In the present study, SEER-H had an average error of 23%, median error of -0.3%, with a standard deviation of 43%. PRICE had an average error of 52%, median error of 50%, and standard deviation of 45%. Nine of the twelve mission's actual costs fell within the 80% confidence interval of SEER's probabilistic estimates; however, none of the missions fell within PRICE’s confidence intervals. There were several factors independent of PRICE and SEER-H which may have affected the accuracy of the results in the present study including: uncertainty in the technical data used for the estimates, the methods used to estimate uncertainty in spacecraft component mass and numbers of prototypes, and the experience of the estimators.

SEER↗