Engineering Papers⌕ Search

Engineering topics

Wang, Liping

Publications and source records attributed to Wang, Liping.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Data-Efficient Dimensionality Reduction and Surrogate Modeling of High-Dimensional Stress Fields

Tensor datatypes representing field variables like stress, displacement, velocity, etc., have increasingly become a common occurrence in data-driven modeling and analysis of simulations. Numerous methods [such as convolutional neural networks (CNNs)] exist to address the meta-modeling of field data from simulations. As the complexity of the simulation increases, so does the cost of acquisition, leading to limited data scenarios. Modeling of tensor datatypes under limited data scenarios remains a hindrance for engineering applications. Here, in this article, we introduce a direct image-to-image modeling framework of convolutional autoencoders enhanced by information bottleneck loss function to tackle the tensor data types with limited data. The information bottleneck method penalizes the nuisance information in the latent space while maximizing relevant information making it robust for limited data scenarios. The entire neural network framework is further combined with robust hyperparameter optimization. We perform numerical studies to compare the predictive performance of the proposed method with a dimensionality reduction-based surrogate modeling framework on a representative linear elastic ellipsoidal void problem with uniaxial loading. The data structure focuses on the low-data regime (fewer than 100 data points) and includes the parameterized geometry of the ellipsoidal void as the input and the predicted stress field as the output. The results of the numerical studies show that the information bottleneck approach yields improved overall accuracy and more precise prediction of the extremes of the stress field. Additionally, an in-depth analysis is carried out to elucidate the information compression behavior of the proposed framework.

artificial intelligence↗
DOE OSTI2024

Efficient mapping between void shapes and stress fields using Deep Convolutional Neural Networks with sparse data

Establishing fast and accurate structure-to-property relationships is an important component in the design and discovery of advanced materials. Physics-based simulation models like the finite element method (FEM) are often used to predict deformation, stress, and strain fields as a function of material microstructure in material and structural systems. Such models may be computationally expensive and time intensive if the underlying physics of the system is complex. This limits their application to solve inverse design problems and identify structures that maximize performance. In such scenarios, surrogate models are employed to make the forward mapping computationally efficient to evaluate. However, the high dimensionality of the input microstructure and the output field of interest often renders such surrogate models inefficient, especially when dealing with sparse data. Deep convolutional neural network (CNN) based surrogate models have shown great promise in handling such high-dimensional problems. In this paper, a single ellipsoidal void structure under a uniaxial tensile load represented by a linear elastic, high-dimensional and expensive-to-query, FEM model. We consider two deep CNN architectures, a modified convolutional autoencoder framework with a fully connected bottleneck and a UNet CNN, and compare their accuracy in predicting the von Mises stress field for any given input void shape in the FEM model. Additionally, a sensitivity analysis study is performed using the two approaches, where the variation in the prediction accuracy on unseen test data is studied through numerical experiments by varying the number of training samples from 20 to 100.

surrogate modeling; convolutional neural networks;↗
DOE OSTI2024

Enhanced Hydrogen Evolution Catalysis of Pentlandite due to the Increases in Coordination Number and Sulfur Vacancy during Cubic‐Hexagonal Phase Transition

Abstract The search for new phases is an important direction in materials science. The phase transition of sulfides results in significant changes in catalytic performance, such as MoS 2 and WS 2 . Cubic pentlandite [cPn, (Fe, Ni) 9 S 8 ] can be a functional material in batteries, solar cells, and catalytic fields. However, no report about the material properties of other phases of pentlandite exists. In this study, the unit‐cell parameters of a new phase of pentlandite, sulfur‐vacancy enriched hexagonal pentlandite (hPn), and the phase boundary between cPn and hPn are determined for the first time. Compared to cPn, the hPn shows a high coordination number, more sulfur vacancies, and high conductivity, which result in significantly higher hydrogen evolution performance of hPn than that of cPn and make the non‐nano rock catalyst hPn superior to other most known nanosulfide catalysts. The increase of sulfur vacancies during phase transition provides a new approach to designing functional materials.

Chemistry↗
DOE OSTI2023

The interactive indoor-outdoor building energy modeling for enhancing the predictions of urban microclimates and building energy demands

There is a lack of an urban building energy modeling framework that considers the influence of surrounding buildings and local urban climate on building thermal performance. This can lead to inaccurate results since the thermal performance of individual buildings is heavily influenced by their surrounding built and climatic environment. This study establishes an interactive indoor-outdoor building energy modeling method to enhance the predictions of urban microclimates and building energy demands by coupling an urban physics model with a physics-based building energy model. Validation of the interactive coupling scheme uses field measurement datasets. Parametric simulation and analysis are conducted to understand the influence of the roof-to-canyon width ratio, canyon orientation, and ground vegetation fraction on canyon temperature, building energy consumption, and energy demand. Furthermore, the impacts of building energy model complexity (e.g., detailed vs. simplified building models) and coupling approaches on canyon temperature and building energy profiles are demonstrated using two case study buildings. In comparison with the one-way coupling approach, cooling energy consumption predicted with the dynamic two-way coupling approach varies by 3.5% and 0.5% for the detailed medium office building model and high-rise building model, respectively, and peak cooling demand varies by 8.4% and 7.0% for the detailed medium office building model and high-rise building model, respectively. Here this study also suggests that adopting a complex two-way coupling approach with environmental data exchange at various elevations is necessary for modeling tall buildings at the urban scale.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

A chloroplast diacylglycerol lipase modulates glycerolipid pathway balance in Arabidopsis

SUMMARY Two parallel pathways compartmentalized in the chloroplast and the endoplasmic reticulum contribute to thylakoid lipid synthesis in plants, but how these two pathways are coordinated during thylakoid biogenesis and remodeling remains unknown. We report here the molecular characterization of a homologous ADIPOSE TRIGLYCERIDE LIPASE‐LIKE gene, previously referred to as ATGLL . The ATGLL gene is ubiquitously expressed throughout development and rapidly upregulated in response to a wide range of environmental cues. We show that ATGLL is a chloroplast non‐regioselective lipase with a hydrolytic activity preferentially towards 16:0 of diacylglycerol (DAG). Comprehensive lipid profiling and radiotracer labeling studies revealed a negative correlation of ATGLL expression and the relative contribution of the chloroplast lipid pathway to thylakoid lipid biosynthesis. Additionally, we show that genetic manipulation of ATGLL expression resulted in changes in triacylglycerol levels in leaves. We propose that ATGLL, through affecting the level of prokaryotic DAG in the chloroplast, plays important roles in balancing the two glycerolipid pathways and in maintaining lipid homeostasis in plants.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Concurrent Pressure-Induced Spin-State Transitions and Jahn–Teller Distortions in MnTe

Cooperative Jahn-Teller (JT) distrotion accompanied by electromagnetic-lattice coupling plays a crucial role in numerous important phenomena in materials science. The JT distortion occurring in high-spin 3d 4 and 3d 9 configurations in octahedral complexes is common, but similarly large distortion due to lifting the degeneracy of t 2g orbits has so far seldomly observed. Here we report the discovery of pressure-induced cooperative JT effect coupled with a large volume-collapse of ~12.8% and a magnetic moment collapse from S = 5/2 to S = 1/2 at ~20.9 GPa in MnTe. The first-principles calculation indicates that weakened p-d hybridization and enhanced d states localization result in lifting t 2g orbital degeneracy in Mn 2+ (d 5 ) which induced a cooperative JT effect-triggered electromagnetic-lattice coupling. Furthermore, these findings suggest a new mechanism for the cooperative JT distortion via pressure induced ligand distortion in antiferromagnetic materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Bayesian learning of orthogonal embeddings for multi-fidelity Gaussian Processes

Uncertainty propagation in complex engineering systems often poses significant computational challenges related to modeling and quantifying probability distributions of model outputs, as those emerge as the result of various sources of uncertainty that are inherent in the system under investigation. Gaussian Processes regression (GPs) is a robust meta-modeling technique that allows for fast model prediction and exploration of response surfaces. Multi-fidelity variations of GPs further leverage information from cheap and low fidelity model simulations in order to improve their predictive performance on the high fidelity model. In order to cope with the high volume of data required to train GPs in high dimensional design spaces, a common practice is to introduce latent design variables that are typically projections of the original input space to a lower dimensional subspace, and therefore substitute the problem of learning the initial high dimensional mapping, with that of training a GP on a low dimensional space. Here in this paper, we present a Bayesian approach to identify optimal transformations that map the input points to low dimensional latent variables. The \projection" mapping consists of an orthonormal matrix that is considered a priori unknown and needs to be inferred jointly with the GP parameters, conditioned on the available training data. The proposed Bayesian inference scheme relies on a two-step iterative algorithm that samples from the marginal posteriors of the GP parameters and the projection matrix respectively, both using Markov Chain Monte Carlo (MCMC) sampling. In order to take into account the orthogonality constraints imposed on the orthonormal projection matrix, a Geodesic Monte Carlo sampling algorithm is employed, that is suitable for exploiting probability measures on manifolds. We extend the proposed framework to multi-fidelity models using GPs including the scenarios of training multiple outputs together. We validate our framework on three synthetic problems with a known lower-dimensional subspace. The benefits of our proposed framework, are illustrated on the computationally challenging aerodynamic optimization of a last-stage blade for an industrial gas turbine, where we study the effect of an 85-dimensional shape parameterization of a three-dimensional airfoil on two output quantities of interest, specifically on the aerodynamic efficiency and the degree of reaction

42 ENGINEERING↗
DOE OSTI2021

Inverse Aerodynamic Design of Gas Turbine Blades using Probabilistic Machine Learning

Abstract One of the critical components in Industrial Gas Turbines (IGT) is the turbine blade. Design of turbine blades needs to consider multiple aspects like aerodynamic efficiency, durability, safety and manufacturing, which make the design process sequential and iterative. The sequential nature of these iterations forces a long design cycle time, ranging from several months to years. Due to the reactionary nature of these iterations, little effort has been made to accumulate data in a manner that allows for deep exploration and understanding of the total design space. This is exemplified in the process of designing the individual components of the IGT resulting in a potential unrealized efficiency. To overcome the aforementioned challenges, we demonstrate a probabilistic inverse design machine learning framework, namely PMI (PMI), to carry out an explicit inverse design. PMI calculates the design explicitly without costly iteration and overcomes the challenges associated with ill-posed inverse problems. In this work the framework will be demonstrated on inverse aerodynamic design of three-dimensional turbine blades.

Engineering↗
DOE OSTI2020

Structural disorder, sublattice melting, and thermo-elastic properties of anti-perovskite Li 3 OBr under high pressure and temperature

Here, in situ synchrotron x-ray diffraction experiments were conducted in the system Li-O-Br at pressures up to 6.5 GPa and temperatures up to the melting points. The thermal equation of state determined from the P–V–T measurements indicates that crystalline anti-perovskite Li 3 OBr exhibits similar thermo-elastic properties to common salts. At temperatures that are 50–75 °C below the crystal melting, a pre-melting behavior was inferred based on weakened Bragg intensity and a substantial increase in the diffuse scattering. Concurrent with the onset of pre-melting, an anomalous increase in the lattice volume was observed and is interpreted as anion sublattice disorder. These findings support previous ab initio molecular dynamics simulations based on a simple anion interchange mechanism, indicating that structural disorder below crystal melting temperature can be a main driving force for Li-sublattice melting and superionic transition in the Li 3 OBr anti-perovskite.

36 MATERIALS SCIENCE↗
NASA NTRS1999

Structural Reliability Analysis and Optimization: Use of Approximations

This report is intended for the demonstration of function approximation concepts and their applicability in reliability analysis and design. Particularly, approximations in the calculation of the safety index, failure probability and structural optimization (modification of design variables) are developed. With this scope in mind, extensive details on probability theory are avoided. Definitions relevant to the stated objectives have been taken from standard text books. The idea of function approximations is to minimize the repetitive use of computationally intensive calculations by replacing them with simpler closed-form equations, which could be nonlinear. Typically, the approximations provide good accuracy around the points where they are constructed, and they need to be periodically updated to extend their utility. There are approximations in calculating the failure probability of a limit state function. The first one, which is most commonly discussed, is how the limit state is approximated at the design point. Most of the time this could be a first-order Taylor series expansion, also known as the First Order Reliability Method (FORM), or a second-order Taylor series expansion (paraboloid), also known as the Second Order Reliability Method (SORM). From the computational procedure point of view, this step comes after the design point identification; however, the order of approximation for the probability of failure calculation is discussed first, and it is denoted by either FORM or SORM. The other approximation of interest is how the design point, or the most probable failure point (MPP), is identified. For iteratively finding this point, again the limit state is approximated. The accuracy and efficiency of the approximations make the search process quite practical for analysis intensive approaches such as the finite element methods; therefore, the crux of this research is to develop excellent approximations for MPP identification and also different approximations including the higher-order reliability methods (HORM) for representing the failure surface. This report is divided into several parts to emphasize different segments of the structural reliability analysis and design. Broadly, it consists of mathematical foundations, methods and applications. Chapter I discusses the fundamental definitions of the probability theory, which are mostly available in standard text books. Probability density function descriptions relevant to this work are addressed. In Chapter 2, the concept and utility of function approximation are discussed for a general application in engineering analysis. Various forms of function representations and the latest developments in nonlinear adaptive approximations are presented with comparison studies. Research work accomplished in reliability analysis is presented in Chapter 3. First, the definition of safety index and most probable point of failure are introduced. Efficient ways of computing the safety index with a fewer number of iterations is emphasized. In chapter 4, the probability of failure prediction is presented using first-order, second-order and higher-order methods. System reliability methods are discussed in chapter 5. Chapter 6 presents optimization techniques for the modification and redistribution of structural sizes for improving the structural reliability. The report also contains several appendices on probability parameters.

Grandhi, Ramana V.↗
NASA NTRS1997

Can the Topex/Poseidon Altimetry Data be Used to Estimate Air Sea Heat Flux in the North Atlantic?

We propose a novel approach to directly invert large-scale anomalous annual net air-sea heat flux from the Topex/Poseidon altimetry data in the North Atlantic. The major advantage of this new approach over the conventional bulk formula approach is that it does not need those climate parameters used in the conventional bulk formula calculation in estimating the large-scale anomalous annual net air-sea heat flux. Comparison with expendable bathythermograph data demonstrates that it is a feasible approach.

Wang, Liping↗