Department of Mechanics: Seminar: Abstract Liu 2019: Difference between revisions

From Wiki @ Department of mechanics
Jump to navigation Jump to search
(Created page with "=== Mechanistic Machine Learning Methods for Mechanical Science and Design/Optimization of Lightweight Material Systems === ==== Wing Kam Liu, Northwestern University, Evans...")
 
Line 7: Line 7:
'''Friday, 6 September 2019, 11am'''  
'''Friday, 6 September 2019, 11am'''  


Masonry material, generally constituted by regular or random assemblage of blocks and
As in all everyday applications, in engineering problems, the volume of data has increased substantially compared to even a decade ago but analyzing big data is expensive and time-consuming. Data-driven methods, which have been enabled in the past decade by the availability of sensors, data storage, and computational resources, are taking center stage across many disciplines (physical and information) of science. We now have highly scalable solutions for problems in object detection and recognition, machine translation, text-to-speech conversion, recommender systems, and information retrieval. All of these solutions attain state-of-the-art performance when trained with large amounts of data. However, purely data-driven approaches for machine learning present difficulties when the data is '''scarce''' and of ''variable fidelity'' relative to the complexity of the system.  
mortar joints, is characterized by a strong nonlinear constitutive response. Similarly to other
heterogeneous materials, the macroscopic or structural behavior depends on the kinematic and
static phenomena occurring at the mesoscopic level, i.e. at the constituents’ observation level. In
particular, anisotropy, stiffness degradation and irreversible displacement observed at the macro-
scale are the results of the opening-closing, sliding and dilatancy occurring at the joints. It follows
that the development of reliable stress analyses still represents a demanded challenge task.
In the last few decades, the development of multi-scale computational homogenization (CH)
techniques has been increasing. These techniques are characterized by the fact that the macroscopic
medium is considered homogeneous and its response is obtained from the solution of a mesoscopic
Boundary Value Problem (BVP) formulated for a representative volume element or unit cell (UC).
For masonry material characterized by a regular texture, a first-order homogenization scheme based
on a discontinuous-continuous approach is presented. At the mesoscopic level the formation and
propagation of fracture is modeled employing a UC consisting of an elastic unit surrounded by
elasto-plastic zero-thickness interfaces [1], characterized by a discontinuous displacement field. The
choice of adopting an elasto-plastic response of mortar represents a good compromise between ease
of applicability and effective representation of the decohesion process occurring at the joint level.
At the macroscopic level, instead, the model maintains the continuity of the displacement field. The
inelastic effects are enclosed in a smeared way, introducing a strain localization band established on
the basis of a spectral analysis of the UC acoustic tensor. Another key-point is the numerical
solution of the UC BVP, which is obtained by means of a more cost-effectiveness mesh-free model,
for this reason the strategy has been named FE·Meshless. Both linear and periodic boundary
conditions have been applied to the UC. It will be show that the meshless UC response strongly
reduce the number of degrees of freedom with respect to a standard FE discretization under
uniform displacements, in addition to the this aspect, meshless UC periodic response allows to
obtain the anti-periodicity of tractions using a considerably reduced number of degrees of freedom
with respect to the FE response. The FE·Meshless strategy has been validated through numerical
applications presented in [2-3] . 



'''References''' <br>
An open problem in data-driven methods for mechanical science is the efficient and accurate description of heterogeneous material behavior that strongly depends on complex microstructure. To explore the future development and the adaptation of data-driven methods, new mathematical and computational paradigms and broad flexible frameworks are needed, which can lead to probabilistic predictions using the minimum amount of information that can be processed expeditiously and be sufficiently accurate for decision making under uncertainty. Integrating multi-fidelity data into large-scale simulations is necessary to speed up the computation but also to deal with the “hidden physics” not captured by the lack of resolution or the lack of proper constitutive laws or boundary conditions. A number of applications will be presented. A number of material systems applications will be presented.
[1] Giambanco, G., Rizzo, S. and Spallino R. (2001). Numerical analysis of masonry structures via
 
interface models. Computer Methods in Applied Mechanics and Engineering, 190(49-50),
''Professor Wing Kam Liu is the Walter P. Murphy Professor of Northwestern University, Director of Global Center on Advanced Material Systems and Simulation, President and Past President of the International Association for Computational Mechanics (IACM) (President (2014-2018) Past President (2018-2024)), Past Chair (2017-2018) (Chair 2015-2016) of the US National Committee on TAM and Member of Board of International Scientific Organizations, both within the US National Academies. Selected synergistic activities includes the development of ICME multiscale theories, methods, and software with experimental validations for the analysis, design and  manufacture of engineering material systems, materials design, advanced and additive manufacturing; and technology transfer. He has over 38 years of engineering and manufacturing consulting, including a broad array of companies and industries, small businesses, and international corporations. Liu’s selected honors include Japan Society of Computational Engineering Sciences Grand Prize; Computational Mechanics Award from Japanese Society of Mechanical Engineers; Honorary Professorship from Dalian University of Technology, IACM Gauss-Newton Medal (highest honor) and Computational Mechanics Award; ASME Dedicated Service Award, ASME Robert Henry Thurston Lecture Award, ASME Gustus L. Larson Memorial Award, ASME Pi Tau Sigma Gold Medal and ASME Melville Medal; John von Neumann Medal (highest honor) and Computational Structural Mechanics Award from the US Association of Computational Mechanics (USACM). He was the founding Director of the NSF Summer Institute on Nano Mechanics and Materials and Founding Chair of the ASME NanoEngineering Council. He is the editor of two International Journals and honorary editor of two journals and has been a consultant for more than 20 organizations. Liu has written four books; and he is a Fellow of ASME, ASCE, USACM, AAM, and IACM.''
6493-6511.<br>
[2] Giambanco, G., La Malfa Ribolla, E. and Spada, A. (2018). Meshless meso-modeling of masonry
in the computational homogenization framework. Meccanica, (53)7, 1673-1697.<br>
[3] Spada, A., Giambanco, G. and La Malfa Ribolla, E. A FE-Meshless multiscale approach for
masonry materials. Procedia Engineering, 109, 364-371, Favignana, June 22-24, 2015. 


Revision as of 14:15, 29 August 2019

Mechanistic Machine Learning Methods for Mechanical Science and Design/Optimization of Lightweight Material Systems

Wing Kam Liu, Northwestern University, Evanston, Illinois, USA

Room B-366, Faculty of Civil Engineering, CTU in Prague

Friday, 6 September 2019, 11am

As in all everyday applications, in engineering problems, the volume of data has increased substantially compared to even a decade ago but analyzing big data is expensive and time-consuming. Data-driven methods, which have been enabled in the past decade by the availability of sensors, data storage, and computational resources, are taking center stage across many disciplines (physical and information) of science. We now have highly scalable solutions for problems in object detection and recognition, machine translation, text-to-speech conversion, recommender systems, and information retrieval. All of these solutions attain state-of-the-art performance when trained with large amounts of data. However, purely data-driven approaches for machine learning present difficulties when the data is scarce and of variable fidelity relative to the complexity of the system.

An open problem in data-driven methods for mechanical science is the efficient and accurate description of heterogeneous material behavior that strongly depends on complex microstructure. To explore the future development and the adaptation of data-driven methods, new mathematical and computational paradigms and broad flexible frameworks are needed, which can lead to probabilistic predictions using the minimum amount of information that can be processed expeditiously and be sufficiently accurate for decision making under uncertainty. Integrating multi-fidelity data into large-scale simulations is necessary to speed up the computation but also to deal with the “hidden physics” not captured by the lack of resolution or the lack of proper constitutive laws or boundary conditions. A number of applications will be presented. A number of material systems applications will be presented.

Professor Wing Kam Liu is the Walter P. Murphy Professor of Northwestern University, Director of Global Center on Advanced Material Systems and Simulation, President and Past President of the International Association for Computational Mechanics (IACM) (President (2014-2018) Past President (2018-2024)), Past Chair (2017-2018) (Chair 2015-2016) of the US National Committee on TAM and Member of Board of International Scientific Organizations, both within the US National Academies. Selected synergistic activities includes the development of ICME multiscale theories, methods, and software with experimental validations for the analysis, design and manufacture of engineering material systems, materials design, advanced and additive manufacturing; and technology transfer. He has over 38 years of engineering and manufacturing consulting, including a broad array of companies and industries, small businesses, and international corporations. Liu’s selected honors include Japan Society of Computational Engineering Sciences Grand Prize; Computational Mechanics Award from Japanese Society of Mechanical Engineers; Honorary Professorship from Dalian University of Technology, IACM Gauss-Newton Medal (highest honor) and Computational Mechanics Award; ASME Dedicated Service Award, ASME Robert Henry Thurston Lecture Award, ASME Gustus L. Larson Memorial Award, ASME Pi Tau Sigma Gold Medal and ASME Melville Medal; John von Neumann Medal (highest honor) and Computational Structural Mechanics Award from the US Association of Computational Mechanics (USACM). He was the founding Director of the NSF Summer Institute on Nano Mechanics and Materials and Founding Chair of the ASME NanoEngineering Council. He is the editor of two International Journals and honorary editor of two journals and has been a consultant for more than 20 organizations. Liu has written four books; and he is a Fellow of ASME, ASCE, USACM, AAM, and IACM.

Retrieved from "https://mech.fsv.cvut.cz/wiki/index.php?title=Department_of_Mechanics:_Seminar:_Abstract_Liu_2019&oldid=4691"