Search NASA⌕ Search

SEARCH · Search NASA

Results for “statistical methods for engineering applications”

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 163 records · Page 9

Ceramics Analysis and Reliability Evaluation of Structures (CARES). Users and programmers manual

This manual describes how to use the Ceramics Analysis and Reliability Evaluation of Structures (CARES) computer program. The primary function of the code is to calculate the fast fracture reliability or failure probability of macroscopically isotropic ceramic components. These components may be subjected to complex thermomechanical loadings, such as those found in heat engine applications. The program uses results from MSC/NASTRAN or ANSYS finite element analysis programs to evaluate component reliability due to inherent surface and/or volume type flaws. CARES utilizes the Batdorf model and the two-parameter Weibull cumulative distribution function to describe the effect of multiaxial stress states on material strength. The principle of independent action (PIA) and the Weibull normal stress averaging models are also included. Weibull material strength parameters, the Batdorf crack density coefficient, and other related statistical quantities are estimated from four-point bend bar or unifrom uniaxial tensile specimen fracture strength data. Parameter estimation can be performed for single or multiple failure modes by using the least-square analysis or the maximum likelihood method. Kolmogorov-Smirnov and Anderson-Darling goodness-of-fit tests, ninety percent confidence intervals on the Weibull parameters, and Kanofsky-Srinivasan ninety percent confidence band values are also provided. The probabilistic fast-fracture theories used in CARES, along with the input and output for CARES, are described. Example problems to demonstrate various feature of the program are also included. This manual describes the MSC/NASTRAN version of the CARES program.

Nemeth, Noel N.↗

Statistical tools for prognostics and health management of complex systems

Prognostics and Health Management (PHM) is increasingly important for understanding and managing today's complex systems. These systems are typically mission- or safety-critical, expensive to replace, and operate in environments where reliability and cost-effectiveness are a priority. We present background on PHM and a suite of applicable statistical tools and methods. Our primary focus is on predicting future states of the system (e.g., the probability of being operational at a future time, or the expected remaining system life) using heterogeneous data from a variety of sources. We discuss component reliability models incorporating physical understanding, condition measurements from sensors, and environmental covariates; system reliability models that allow prediction of system failure time distributions from component failure models; and the use of Bayesian techniques to incorporate expert judgments into component and system models.

42 ENGINEERING↗

NASA Tech Briefs, January 2005

Topics covered include: Fiber-Optic Sensor Would Monitor Growth of Polymer Film; Sensors for Pointing Moving Instruments Toward Each Other; Pd/CeO2/SiC Chemical Sensors; Microparticle Flow Sensor; Scattering-Type Surface-Plasmon-Resonance Biosensors; Diode-Laser-Based Spectrometer for Sensing Gases; Improved Cathode Structure for a Direct Methanol Fuel Cell; X-Band, 17-Watt Solid-State Power Amplifier; Improved Anode for a Direct Methanol Fuel Cell; Tools for Designing and Analyzing Structures; Interactive Display of Scenes with Annotations; Solving Common Mathematical Problems; Tools for Basic Statistical Analysis; Program Calculates Forces in Bolted Structural Joints; Integrated Structural Analysis and Test Program; Molybdate Coatings for Protecting Aluminum Against Corrosion; Synthesizing Diamond from Liquid Feedstock; Modifying Silicates for Better Dispersion in Nanocomposites; Powder-Collection System for Ultrasonic/Sonic Drill/Corer; Semiautomated, Reproducible Batch Processing of Soy; Hydrogen Peroxide Enhances Removal of NOx from Flue Gases; Subsurface Ice Probe; Real-Time Simulation of Aeroheating of the Hyper-X Airplane; Using Laser-Induced Incandescence To Measure Soot in Exhaust; Method of Real-Time Principal-Component Analysis; Insect-Inspired Flight Control for Unmanned Aerial Vehicles; Domain Compilation for Embedded Real-Time Planning; Semantic Metrics for Analysis of Software; Simulation of Laser Cooling and Trapping in Engineering Applications; Large Fluvial Fans and Exploration for Hydrocarbons; Doping-Induced Interband Gain in InAs/AlSb Quantum Wells; Development of Software for a Lidar-Altimeter Processor; Upgrading the Space Shuttle Caution and Warning System; and Fractal Reference Signals in Pulse-Width Modulation.

Source record↗

Numerical and Experimental Study of Mixing Processes Associated with Hydrogen and High Hydrogen Content Fuels

As simulation capability improves exponentially with increasingly more cost effective CPUs and hardware, it can be used “routinely” for engineering applications. Many commercial products are available and they are marketed as increasingly powerful and easy to use. The question remains as to the overall accuracy of results obtained. To support the validation of the CFD, a hierarchical experiment was established in which the type of fuel injection (radial, axial) as well as level of swirl (non-swirling, swirling) could be systematically varied. The effort was limited to time efficient approaches (i.e., generally RANS approaches) although limited assessment of time resolved methods (i.e., unsteady RANS and LES) were considered. Careful measurements of the flowfield velocity and fuel concentration were made using both intrusive and non-intrusive methods. This database was then used as the basis for the assessment of the CFD approach. The numerical studies were carried out with a statistically based matrix. As a result, the effect of turbulence model, fuel type, axial plane, turbulent Schmidt number, and injection type could be studied using analysis of variance. The results for the non-swirling cases could be analyzed as planned, and demonstrate that turbulence model selection, turbulence Schmidt number, and the type of injection will strongly influence the agreement with measured values. Interestingly, the type of fuel used (either hydrogen or methane) has no influence on the accuracy of the simulations. For axial injection, the selection of proper turbulence Schmidt number is important, whereas for radial injection, the results are relatively insensitive to this parameter. In general, it was found that the nature of the flowfield influences the performance of the predictions. This result implies that it is difficult to establish a priori the “best” simulation approach to use. However, the insights from the relative orientation of the jet and flow do offer some guidance for which approach to take. Overall, the results underscore the importance of model “anchoring” (i.e., “tuning” the model to provide “reasonable” agreement with a well characterized geometry/flow). Finally, the results obtained have been carefully compiled into a standalone database following a standard format that is contained in an Appendix. This database is thus available for use by others for CFD modeling evaluations.

08 HYDROGEN↗

Antifreeze Hydrogels from Amphiphilic Statistical Copolymers

Prevention of ice formation is a critical issue for many applications, but routes to overcome the large thermodynamic driving force for crystallization of water at significant supercooling are limited. Here, we demonstrate that supramolecular hydrogels formed from statistical copolymers of 2-hydroxyethyl acrylate (HEA) and 2-( N -ethylperfluorooctane sulfonamido)ethyl methacrylate (FOSM) exhibit a degree of ice formation suppression unprecedented in a synthetic material. The mechanisms of ice prevention by these hydrogels mimic two methods used by nature: (1) hydrogen bonding of water to highly hydrophilic macromolecular chains and (2) nanoconfinement of water between hydrophobic moieties. From systematic variation in the copolymer composition to control the nanoscale (<4 nm) separation of the self-assembled hydrophobic nanodomains, the main mechanism by which these supramolecular hydrogels inhibit large amounts of water from freezing appears to be soft nanoconfinement. Nearly complete ice inhibition was achieved in hydrogels when the nanodomain separation was <3 nm (i.e., confinement volume ~15 nm 3 ) where <290 water molecules are present. Dielectric spectroscopy is consistent with two primary populations of water: a population of water with a bulk-like dynamics as well as T g (136 K) and a minority population of water with suppressed dynamics and an enhanced T g near 151 K that is attributed to interfacial water. The nanostructured design of these supramolecular hydrogels provides a blueprint concept for controlling and manipulating ice formation in concentrated soft matter using the length scale between hydrophobic domains and the hydrophilicity of the network water-soluble component. Furthermore, these insights have the potential to provide solutions to challenges with ice in engineering applications where confinement of water to nanoscale dimensions is possible.

36 MATERIALS SCIENCE↗

Identifying Hydrometeorological Factors Influencing Reservoir Releases Using Machine Learning Methods

Simulation of reservoir releases plays a critical role in social-economic functioning and our nation's security. How-ever, it is challenging to predict the reservoir release accurately because of many influential factors from natural environments and engineering controls such as the reservoir inflow and storage. Moreover, climate change and hydrological intensification causing the extreme precipitation and temperature make the accurate prediction of reservoir releases even more challenging. Machine learning (ML) methods have shown some successful applications in simulating reservoir releases. However, previous studies mainly used inflow and storage data as inputs and only considered their short-term influences (e.g, previous one or two days). In this work, we use long short-term memory (LSTM) networks for reservoir release prediction based on four input variables including inflow, storage, precipitation, and temperature and consider their long-term influences. We apply the LSTM model to 30 reservoirs in Upper Colorado River Basin, United States. We analyze the prediction performance using six statistical metrics. More importantly, we investigate the influence of the input hydrometeorological factors, as well as their temporal effects on reservoir release decisions. Results indicate that inflow and storage are the most influential factors but the inclusion of precipitation and temperature can further improve the prediction of release especially in low flows. Additionally, the inflow and storage have a relatively long-term effect on the release. These findings can help optimize the water resources management in the reservoirs.

Fan, Ming↗

Quantifying uncertainty from material inhomogeneity.

Most engineering materials are inherently inhomogeneous in their processing, internal structure, properties, and performance. Their properties are therefore statistical rather than deterministic. These inhomogeneities manifest across multiple length and time scales, leading to variabilities, i.e. statistical distributions, that are necessary to accurately describe each stage in the process-structure-properties hierarchy, and are ultimately the primary source of uncertainty in performance of the material and component. When localized events are responsible for component failure, or when component dimensions are on the order of microstructural features, this uncertainty is particularly important. For ultra-high reliability applications, the uncertainty is compounded by a lack of data describing the extremely rare events. Hands-on testing alone cannot supply sufficient data for this purpose. To date, there is no robust or coherent method to quantify this uncertainty so that it can be used in a predictive manner at the component length scale. The research presented in this report begins to address this lack of capability through a systematic study of the effects of microstructure on the strain concentration at a hole. To achieve the strain concentration, small circular holes (approximately 100 {micro}m in diameter) were machined into brass tensile specimens using a femto-second laser. The brass was annealed at 450 C, 600 C, and 800 C to produce three hole-to-grain size ratios of approximately 7, 1, and 1/7. Electron backscatter diffraction experiments were used to guide the construction of digital microstructures for finite element simulations of uniaxial tension. Digital image correlation experiments were used to qualitatively validate the numerical simulations. The simulations were performed iteratively to generate statistics describing the distribution of plastic strain at the hole in varying microstructural environments. In both the experiments and simulations, the deformation behavior was found to depend strongly on the character of the nearby microstructure.

36 MATERIALS SCIENCE↗

Overview of the SAE G-11 RMSL (Reliability, Maintainability, Supportability, and Logistics) Division Activities and Technical Projects

The SAE G-11 RMSL (Reliability, Maintainability, Supportability, and Logistics) Division activities include identification and fulfillment of joint industry, government, and academia needs for development and implementation of RMSL technologies. Four Projects in the Probabilistic Methods area and two in the area of RMSL have been identified. These are: (1) Evaluation of Probabilistic Technology - progress has been made toward the selection of probabilistic application cases. Future effort will focus on assessment of multiple probabilistic softwares in solving selected engineering problems using probabilistic methods. Relevance to Industry & Government - Case studies of typical problems encountering uncertainties, results of solutions to these problems run by different codes, and recommendations on which code is applicable for what problems; (2) Probabilistic Input Preparation - progress has been made in identifying problem cases such as those with no data, little data and sufficient data. Future effort will focus on developing guidelines for preparing input for probabilistic analysis, especially with no or little data. Relevance to Industry & Government - Too often, we get bogged down thinking we need a lot of data before we can quantify uncertainties. Not True. There are ways to do credible probabilistic analysis with little data; (3) Probabilistic Reliability - probabilistic reliability literature search has been completed along with what differentiates it from statistical reliability. Work on computation of reliability based on quantification of uncertainties in primitive variables is in progress. Relevance to Industry & Government - Correct reliability computations both at the component and system level are needed so one can design an item based on its expected usage and life span; (4) Real World Applications of Probabilistic Methods (PM) - A draft of volume 1 comprising aerospace applications has been released. Volume 2, a compilation of real world applications of probabilistic methods with essential information demonstrating application type and timehost savings by the use of probabilistic methods for generic applications is in progress. Relevance to Industry & Government - Too often, we say, 'The Proof is in the Pudding'. With help from many contributors, we hope to produce such a document. Problem is - not too many people are coming forward due to proprietary nature. So, we are asking to document only minimum information including problem description, what method used, did it result in any savings, and how much?; (5) Software Reliability - software reliability concept, program, implementation, guidelines, and standards are being documented. Relevance to Industry & Government - software reliability is a complex issue that must be understood & addressed in all facets of business in industry, government, and other institutions. We address issues, concepts, ways to implement solutions, and guidelines for maximizing software reliability; (6) Maintainability Standards - maintainability/serviceability industry standard/guidelines and industry best practices and methodologies used in performing maintainability/ serviceability tasks are being documented. Relevance to Industry & Government - Any industry or government process, project, and/or tool must be maintained and serviced to realize the life and performance it was designed for. We address issues and develop guidelines for optimum performance & life.

Singhal, Surendra N.↗

Multilevel Monte Carlo Estimators For Derivative-Free Optimization Under Uncertainty

Optimization is a key tool for scientific and engineering applications; however, in the presence of models affected by uncertainty, the optimization formulation needs to be extended to consider statistics of the quantity of interest. Optimization under uncertainty (OUU) deals with this endeavor and requires uncertainty quantification analyses at several design locations; i.e., its overall computational cost is proportional to the cost of performing a forward uncertainty analysis at each design location. An OUU workflow has two main components: an inner loop strategy for the computation of statistics of the quantity of interest, and an outer loop optimization strategy tasked with finding the optimal design, given a merit function based on the inner loop statistics. Here, in this work, we propose to alleviate the cost of the inner loop uncertainty analysis by leveraging the so-called multilevel Monte Carlo (MLMC) method, which is able to allocate resources over multiple models with varying accuracy and cost. The resource allocation problem in MLMC is formulated by minimizing the computational cost given a target variance for the estimator. We consider MLMC estimators for statistics usually employed in OUU workflows and solve the corresponding allocation problem. For the outer loop, we consider a derivative-free optimization strategy implemented in the SNOWPAC library; our novel strategy is implemented and released in the Dakota software toolkit. We discuss several numerical test cases to showcase the features and performance of our approach with respect to its Monte Carlo single fidelity counterpart.

97 MATHEMATICS AND COMPUTING↗

Improved silicon nitride for advanced heat engines

Silicon nitride is a high temperature material currently under consideration for heat engine and other applications. The objective is to improve the net shape fabrication technology of Si3N4 by injection molding. This is to be accomplished by optimizing the process through a series of statistically designed matrix experiments. To provide input to the matrix experiments, a wide range of alternate materials and processing parameters was investigated throughout the whole program. The improvement in the processing is to be demonstrated by a 20 percent increase in strength and a 100 percent increase in the Weibull modulus over that of the baseline material. A full characterization of the baseline process was completed. Material properties were found to be highly dependent on each step of the process. Several important parameters identified thus far are the starting raw materials, sinter/hot isostatic pressing cycle, powder bed, mixing methods, and sintering aid levels.

Yeh, H. C.↗

A Frequency Domain Methodology for Quantitative Evaluation of Diffuse Wavefield With Applications to Seismic Imaging

Abstract Ambient Noise Imaging (ANI) of subsurface structures relies on seismic interferometry of diffuse seismic wavefields. However, the lack of effective methods to quantify and identify highly diffuse waves hampers applications of ANI, particularly in evaluating seismic attenuation and monitoring structural changes with high temporal resolution. Conventional ANI approaches require data normalization, which effectively suppresses the non‐diffuse component with large amplitude but also results in significant loss of amplitude and phase information in the continuous seismic records. In this study, we propose a frequency domain method to quantitatively evaluate the degree of diffuseness of seismic wavefields by analyzing their statistical characteristics of modal amplitudes for stationarity and randomness. Tests on synthetic waveform and field nodal records show that the proposed method can effectively distinguish between diffuse and non‐diffuse waveforms for either single‐ or three‐component data. As an application, we identify a 60‐s‐long diffuse coda of a local M 2.2 earthquake recorded by a dense nodal array on the San Jacinto Fault Zone, and successfully extract high‐quality dispersion curve andQ‐value without performing data normalization. These results are consistent with those obtained by conventional methods that assess the correlation between coherency and the Green's function, and by modeling ballistic waves generated by road traffic. Our proposed method can advance the imaging of subsurface velocity and attenuation structures as well as monitoring temporal changes for scientific studies and engineering applications.

Geochemistry & Geophysics↗

Computing Risk of Pyrotechnic Devices Using Lot Acceptance Testing

This paper presents an application of statistical engineering to solve the problem of determining the effects of reducing test sample sizes and environment levels in lot acceptance testing (LAT) of ordnance devices used by NASA and DoD space systems. Unlike environmental testing of dedicated qualification hardware, ordnance devices have a test-like-you-fly exception and use lot acceptance sampling or LAT. LAT is performed by randomly selecting a predefined number of devices from an individual lot and exposing them to a more severe dynamic shock and vibration testing environments than flight (sometimes as high as 6 decibels over flight environments) and then performing a destructive test on the device. Recent trends have been suggested to reduce LAT costs by reducing environment test levels or the number of devices per LAT, or both. A method is given here to compute flight risk, so the proposed methods can be compared by subject matter experts to the baseline LAT methods. In addition, this method can be used to determine the best methods to include in future LAT standards for pyrotechnic devices.

Computing Risk↗

A Factorial Data Rate and Dwell Time Experiment in the National Transonic Facility

This report is an introductory tutorial on the application of formal experiment design methods to wind tunnel testing, for the benefit of aeronautical engineers with little formal experiment design training. It also describes the results of a Study to determine whether increases in the sample rate and dwell time of the National Transonic Facility data system Would result in significant changes in force and moment data. Increases in sample rate from 10 samples per second to 50 samples per second were examined, as were changes in dwell time from one second per data point to two seconds. These changes were examined for a representative aircraft model in a range of tunnel operating conditions defined by angles of attack from 0 deg to 3.8 degrees, total pressure from 15.0 psi to 24.1 psi, and Mach numbers from 0.52 to 0.82. No statistically significant effect was associated with the change in sample rate. The change in dwell time from one second to two seconds affected axial force measurements, and to a lesser degree normal force measurements. This dwell effect comprises a "rectification error" caused by incomplete cancellation of the positive and negative elements of certain low frequency dynamic components that are not rejected by the one-Hz low-pass filters of the data system. These low frequency effects may be due to tunnel circuit phenomena and other sources. The magnitude of the dwell effect depends on dynamic pressure, with angle of attack and Mach number influencing the strength of this dependence. An analysis is presented which suggests that the magnitude of the rectification error depends on the ratio of measurement dwell time to the period of the low-frequency dynamics, as well as the amplitude of the dynamics The essential conclusion of this analysis is that extending the dwell time (or, equivalently, replicating short-dwell data points) reduces the rectification error.

DeLoach, R.↗

Statistical Analysis of Abnormal Electric Power Grid Behavior

Pacific Northwest National Laboratory is developing a technique to analyze Phasor Measurement Unit data to identify typical patterns, atypical events and precursors to a blackout or other undesirable event. The approach combines a data-driven multivariate analysis with an engineering-model approach. The method identifies atypical events, provides a plane English description of the event, and the capability to use drill-down graphics for detailed investigations. The tool can be applied to the entire grid, individual organizations (e.g. TVA, BPA), or specific substations (e.g., TVA_CUMB). The tool is envisioned for (1) event investigations, (2) overnight processing to generate a Morning Report that characterizes the previous days activity with respect to previous activity over the previous 10-30 days, and (3) potentially near-real-time operation to support the grid operators. This paper presents the current status of the tool and illustrations of its application to real world PMU data collected in three 10-day periods in 2007.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Near Infrared Spectroscopy Calibration for Wood Chemistry: Which Chemometric Technique Is Best for Prediction and Interpretation?

This paper addresses the precision in factor loadings during partial least squares (PLS) and principal components regression (PCR) of wood chemistry content from near infrared reflectance (NIR) spectra. The precision of the loadings is considered important because these estimates are often utilized to interpret chemometric models or selection of meaningful wavenumbers. Standard laboratory chemistry methods were employed on a mixed genus/species hardwood sample set. PLS and PCR, before and after 1st derivative pretreatment, was utilized for model building and loadings investigation. As demonstrated by others, PLS was found to provide better predictive diagnostics. However, PCR exhibited a more precise estimate of loading peaks which makes PCR better for interpretation. Application of the 1st derivative appeared to assist in improving both PCR and PLS loading precision, but due to the small sample size, the two chemometric methods could not be compared statistically. This work is important because to date most research works have committed to PLS because it yields better predictive performance. But this research suggests there is a tradeoff between better prediction and model interpretation. Future work is needed to compare PLS and PCR for a suite of spectral pretreatment techniques.

09 BIOMASS FUELS↗

LES on unstructured deforming meshes: Towards reciprocating IC engines

A variable explicit/implicit characteristics-based advection scheme that is second-order accurate in space and time has been developed recently for unstructured deforming meshes (O'Rourke & Sahota 1996a). To explore the suitability of this methodology for Large-Eddy Simulation (LES), three subgrid-scale turbulence models have been implemented in the CHAD CFD code (O'Rourke & Sahota 1996b): a constant-coefficient Smagorinsky model, a dynamic Smagorinsky model for flows having one or more directions of statistical homogeneity, and a Lagrangian dynamic Smagorinsky model for flows having no spatial or temporal homogeneity (Meneveau et al. 1996). Computations have been made for three canonical flows, progressing towards the intended application of in-cylinder flow in a reciprocating engine. Grid sizes were selected to be comparable to the coarsest meshes used in earlier spectral LES studies. Quantitative results are reported for decaying homogeneous isotropic turbulence, and for a planar channel flow. Computations are compared to experimental measurements, to Direct-Numerical Simulation (DNS) data, and to Rapid-Distortion Theory (RDT) where appropriate. Generally satisfactory evolution of first and second moments is found on these coarse meshes; deviations are attributed to insufficient mesh resolution. Issues include mesh resolution and computational requirements for a specified level of accuracy, analytic characterization of the filtering implied by the numerical method, wall treatment, and inflow boundary conditions. To resolve these issues, finer-mesh simulations and computations of a simplified axisymmetric reciprocating piston-cylinder assembly are in progress.

Haworth, D. C.↗

Data Collection and Analysis Challenges and Mitigation Strategies for Quantitative Human Factors Research Studies in Nuclear Power Plant Modernization

The United States (U.S.) Department of Energy (DOE) Light Water Reactor Sustainability (LWRS) program Plant Modernization Pathway is conducting targeted research and development (R&D) to address aging and reliability concerns with the legacy instrumentation and control and related information systems of the U.S. LWR fleet. In this effort, the application of human factors engineering (HFE) provides an important role in ensuring new digital plant technologies enable broad innovation and business improvement with continued operational safety. Evaluation is a key activity in HFE, which often occurs iteratively through the system design lifecycle. While qualitative methods are important in collecting information of how users perform tasks through observations, quantitative methods are equally important in assessing system design based on performance. In collecting and analyzing this quantitative data, there are notable challenges in the nuclear HFE domain that may threaten the validity and reliability of the inferences made in these studies. Notable challenges include small sample size and limited resources, large error variance and small effect size, an “adding test features to losing degrees of freedom” dilemma, non-normal distribution, and heterogeneity of variance. The results in control room usability studies are often statistically non-significant, which makes it hard to interpret. This work discusses these challenges across different scientific viewpoints and provides real-world examples of these challenges in practice. Collectively, the objective of this work is to position these challenges to the larger data science community as a means of identifying future opportunities to address these issues.

99 GENERAL AND MISCELLANEOUS↗