The optimal design of a structural system poses significant challenges, such as the high dimensionality of random variable space and the cost of computational simulations repeated during optimization. In this new paper, co-authored by Dr. Jungho Kim at UC Berkeley (Kim Jungho; a former Ph.D. student of my research group at Seoul National University) and Prof. Ziqi Wang (Ziqi Wang) at UC Berkeley, we proposed to adaptively train a heteroscedastic Gaussian process based surrogate model described in a low dimensional active subspace for accurate and efficient optimization of structures. The proposed method was successfully demonstrated and tested by three examples: (1) nonlinear mathematical functions of up to 100 random variables, (2) a space truss structure described by 57 random variables, and (3) a steel lattice transmission tower subjected to seismic excitations described by 100 random variables. Congratulations, Dr. Kim!
Junho Song’s Post
More Relevant Posts
-
In 2007, I began my PhD with some trepidation at Columbia University. I wanted to solve hard problems and I always loved probability and statistics, but this was a different path for me and I wasn't sure where it would lead. I chose to work with a wonderful adviser, incredible scholar, and even better person: George Deodatis. During my time in Columbia University Civil Engineering & Engineering Mechanics, I studied stochastic processes and their applications to civil engineering and broader mechanics applications. I did some really interesting work on simulation of non-Gaussian and non-stationary stochastic processes specifically that propelled my career forward in stochastic modeling and uncertainty quantification. I owe my current position in the Johns Hopkins Department of Civil and Systems Engineering and the journey that has led me to this point to that initial work with George, and I cannot imagine how it could have turned out better. Today, I am very excited to share that George and I have published a new review paper that covers the Spectral Representation Method, the stochastic simulation framework first pioneered by George's adviser, the great Masanobu Shinozuka, then by George, and to which I have had the good fortune to contribute over the past decade. This brings my PhD work full-circle, reminds me of the legacy of scholars I unknowingly stepped behind some 17 years ago, and (hopefully) marks a continuation of that legacy. Here's the paper!!! 😀 👇 https://lnkd.in/ezjmQTTC
The Spectral Representation Method: A framework for simulation of stochastic processes, fields, and waves
sciencedirect.com
To view or add a comment, sign in
-
I am pleased to share that our research on "Domain decomposition with local time discretization for the nonlinear Stokes–Biot system" has been published in the Journal of Computational and Applied Mathematics. A big thanks to Dr. Hyesuk Lee, my PhD advisor, from Clemson University for her valuable mentorship and collaboration. Abstract: This work presents a #domain_decomposition method for the #fluid-#poroelastic structure interaction (FPSI) system, which utilizes local time integration for subproblems. To derive the domain decomposition scheme, we introduce a Lagrange multiplier and define time-dependent Steklov– Poincaré-type operators based on the interface conditions. These operators are employed to transform the coupled system into an evolutionary #nonlinear interface problem, which is then solved using an iterative algorithm. This approach provides the flexibility to use different time discretization schemes and step sizes in subdomains, making it an efficient method for simulating #multiphysics systems. We present numerical tests for both non-physical and physical problems to demonstrate the accuracy and efficiency of this method.
Domain decomposition with local time discretization for the nonlinear Stokes-Biot system
sciencedirect.com
To view or add a comment, sign in
-
While driving through the mountains of Chile, George and I were suddenly confronted with a giant Andean Condor, seemingly bigger than our car, flying straight at our windshield! At the last minute, the magnificent bird swooped up, leaving us breathless and wishing we had been fast enough to grab out iPhones to make a video. The experience reminded us of a project module that we wrote on “Taking Flight: Modeling Condor Populations.” The module, which is free to download at https://lnkd.in/eswSEF_w, briefly supplies the necessary scientific background and projects appropriate for students to develop using system dynamics or age-structure modeling or agent-based simulations. The module is part of a collection of project modules (https://lnkd.in/gSTZ98FP) associated with our textbook, “Introduction to Computational Science: Modeling and Simulation for the Sciences” (2nd Ed., Princeton U. Press, https://lnkd.in/dpUNnRg). Video by DanielPeckham, CC BY 3.0 <https://lnkd.in/eQF9Zjv4>, via Wikimedia Commons, segments from https://lnkd.in/esMtupz9
To view or add a comment, sign in
-
🛠️ Proud to share my latest publication! M.P. Rezaei, G. Kudra, M. Ghodsi, J. Awrejcewicz, "Overcoming stretching and shortening assumptions in Euler–Bernoulli theory using nonlinear Hencky’s beam models: Applicable to partly-shortened and partly-stretched beams," Journal of Sound and Vibration, 118807, ISSN 0022-460X, 2024. (DOI: 10.1016/j.jsv.2024.118807). This study proposes nonlinear Hencky’s beam models to address limitations in the Euler–Bernoulli beam theory, offering a fresh perspective on beams with complex boundary conditions. It was a challenging but deeply rewarding research effort. Special thanks to the Polish National Science Center for supporting this work under the PRELUDIUM 22 grant No. 2023/49/N/ST8/00823. 🙏 Check it out here: https://lnkd.in/dHwWMeg9 #Research #Engineering #StructuralDynamics #NonlinearModeling #AcademicPublishing #NSCGrantSupport
Overcoming stretching and shortening assumptions in Euler–Bernoulli theory using nonlinear Hencky’s beam models: Applicable to partly-shortened and partly-stretched beams
sciencedirect.com
To view or add a comment, sign in
-
These 17 equations have revolutionized science, technology, and our understanding of the natural world (by Ian Stewart, 2013): 1. Pythagoras's Theorem (Pythagoras, 530 BC): Architecture, construction, navigation, computer graphics. 2. Logarithms (John Napier, 1610): Financial analysis, population growth modeling, signal processing. 3. Calculus (Newton, 1668): Physics, engineering, economics, computer graphics, optimization. 4. Law of Gravity (Newton, 1687): Celestial mechanics, satellite navigation, space exploration. 5. Wave Equation (d'Almbert, 1747) Acoustics, optics, telecommunications, seismology. 6. The square root of -1 (Euler, 1750): Mechanics, engineering, motion control systems. 7. Euler's Polyhedra Formula (Euler, 1751): Engineering, topological invariant, modern physics. 8. Normal distribution (Gauss, 1810): Physics, biology, social sciences 9. Fourier Transform (Fourier, 1822): Signal processing, image compression, data analysis. 10. Navier-Stokes Equations (Navier, Stokes, 1845): Fluid mechanics, aerodynamics, weather prediction. 11. Maxwell's Equations (Maxwell, 1865): Electromagnetism, electrical engineering, telecommunications. 12. The Second Law of Thermodynamics (Boltzmann, 1874): Heat engines, energy conversion, entropy. 13. Relativity (Einstein, 1905): Nuclear physics, nuclear energy, atomic bombs. Astrophysics, cosmology, GPS technology. 14. Schrödinger's Equation (Schrödinger, 1927): Quantum mechanics, atomic and molecular physics, material science. 15. Information Theory (Shannon, 1949): Data compression, communication systems, cryptography. 16. Chaos Theory (R. May, 1975): Weather prediction, complex systems, cryptography. 17. Black-Scholes Equation (Black, Scholes, 1990): Option pricing, quantitative finance, financial mathematics.
To view or add a comment, sign in
-
While driving through the mountains of Chile, Angela and I were suddenly confronted with a giant Andean Condor, seemingly bigger than our car, flying straight at our windshield! At the last minute, the magnificent bird swooped up, leaving us breathless and wishing we had been fast enough to grab out iPhones to make a video. The experience reminded us of a project module that we wrote on “Taking Flight: Modeling Condor Populations.” The module, which is free to download at https://lnkd.in/eGSDiKvQ, briefly supplies the necessary scientific background and projects appropriate for students to develop using system dynamics or age-structure modeling or agent-based simulations. The module is part of a collection of project modules (https://lnkd.in/eJS5yQse) associated with our textbook, “Introduction to Computational Science: Modeling and Simulation for the Sciences” (2ndEd., Princeton U. Press, https://lnkd.in/eKw4SNYg). Video by DanielPeckham, CC BY 3.0 <https://lnkd.in/eGSRcv6M>, via Wikimedia Commons, segments from https://lnkd.in/ejmVJtYw
To view or add a comment, sign in
-
📃Scientific paper: Surface parameterisation and spectral synthesis of rapidly rotating stars. Vega as a testbed Abstract: Spectral synthesis is a powerful tool with which to find the fundamental parameters of stars. Models are usually restricted to single values of temperature and gravity, and assume spherical symmetry. This approximation breaks down for rapidly rotating stars.This paper presents a joint formalism to allow a computation of the stellar structure, namely, the photospheric radius, the effective temperature, and gravity, as a function of the colatitude for rapid rotators with radiative envelopes, and a subsequent method to build the corresponding synthetic spectrum. The structure of the star is computed using a semi-analytical approach, which is easy to implement from a computational point of view and which reproduces very accurately the results of much more complex codes. Once R, T, and g are computed, the suite of code atlas and synthe, by R. Kurucz are used to synthesise spectra for a mesh of cells in which the star is divided. The appropriate limb-darkening coefficients are also computed, and the final output spectrum is built for a given inclination of the rotation axis with respect to the line of sight. All the geometrical transformations required are described in detail. The combined formalism has been applied to Vega, a rapidly rotating star almost seen pole-on, as a testbed. The structure reproduces the results from interferometric studies and the synthetic spectrum matches the peculiar shape of the spectral lines well. Contexts where this formalism can be applie... Continued on ES/IODE ➡️ https://etcse.fr/y1R ------- If you find this interesting, feel free to follow, comment and share. We need your help to enhance our visibility, so that our platform continues to serve you.
Surface parameterisation and spectral synthesis of rapidly rotating stars. Vega as a testbed
ethicseido.com
To view or add a comment, sign in
-
I am delighted to share that my book has just been published by Springer Nature! 📕 This book represents an enlarged and extensively reworked version of my doctoral thesis, including new results and additional commentary. In it, I focus on presenting the computational modelling approach we have developed for studying how atoms preferentially arrange themselves in multicomponent alloys, before applying the approach to a number of canonical alloy systems, including high-entropy alloys. It has been a real pleasure working with Zach and the rest of the team at Springer on this project! 🚀 You can find the book here: https://lnkd.in/gxwteGv3 #Physics #MaterialsScience #Alloys #MaterialsModelling
🚀 Great news for all you computational metallurgists out there! 🚀 I'm delighted to share that Dr. Christopher Woodgate's book, "Modelling Atomic Arrangements in Multicomponent Alloys—A Perturbative, First-Principles-Based Approach," has just been published! Based on his outstanding doctoral thesis, this work is now part of our esteemed Springer Series in Materials Science — and I couldn't think of a better home for it! The book offers a fresh perspective on phase stability in high-entropy alloys, presenting a method that captures the essential physics with impressive computational efficiency. Chris also explores the paramagnetic state of these systems using the disordered local moment (DLM) picture, delivering remarkable insights into these fascinating and complex materials. Chris has excelled in making this complex topic accessible and invaluable to anyone in materials science, physics, or chemistry. Join me in giving a heartfelt congratulations to Chris, who will also soon be continuing his research at the University of Bristol on an independent research fellowship. This move is well-deserved and no doubt bolstered by this exceptional work. 👏 👏👏 Thank you, Chris, for the opportunity to help bring this important project to life—it’s been a pleasure! https://lnkd.in/g_-WJH29 #HighEntropyAlloys #ComputationalMetallurgy #MaterialsScience #MulticomponentAlloys #AtomicModelling #Springer
Modelling Atomic Arrangements in Multicomponent Alloys
link.springer.com
To view or add a comment, sign in
-
📃Scientific paper: Surface parameterisation and spectral synthesis of rapidly rotating stars. Vega as a testbed Abstract: Spectral synthesis is a powerful tool with which to find the fundamental parameters of stars. Models are usually restricted to single values of temperature and gravity, and assume spherical symmetry. This approximation breaks down for rapidly rotating stars.This paper presents a joint formalism to allow a computation of the stellar structure, namely, the photospheric radius, the effective temperature, and gravity, as a function of the colatitude for rapid rotators with radiative envelopes, and a subsequent method to build the corresponding synthetic spectrum. The structure of the star is computed using a semi-analytical approach, which is easy to implement from a computational point of view and which reproduces very accurately the results of much more complex codes. Once R, T, and g are computed, the suite of code atlas and synthe, by R. Kurucz are used to synthesise spectra for a mesh of cells in which the star is divided. The appropriate limb-darkening coefficients are also computed, and the final output spectrum is built for a given inclination of the rotation axis with respect to the line of sight. All the geometrical transformations required are described in detail. The combined formalism has been applied to Vega, a rapidly rotating star almost seen pole-on, as a testbed. The structure reproduces the results from interferometric studies and the synthetic spectrum matches the peculiar shape of the spectral lines well. Contexts where this formalism can be applie... Continued on ES/IODE ➡️ https://etcse.fr/y1R ------- If you find this interesting, feel free to follow, comment and share. We need your help to enhance our visibility, so that our platform continues to serve you.
Surface parameterisation and spectral synthesis of rapidly rotating stars. Vega as a testbed
ethicseido.com
To view or add a comment, sign in
Structural Engineer at Swanton Consulting Ltd
9moCongratulations professor! Is the time difference between 50 to 100 variables very high for this work? Took my attention because computational cost is one of the most important factors when we have many variables to obtain. I usually optimise around 10 variables of hysteretic models using MLE, and it can take and hour for a multilinear model, or a day for a Bouc Wen class model