Special lssues
Table of Content

Data-driven Computational Modeling and Simulations

Submission Deadline: 01 August 2018 (closed)

Guest Editors

Gregory J Wagner (gregory.wagner@northwestern.edu)
Greg Wagner received his Ph.D. in Mechanical Engineering from Northwestern University in 2001. He spent over 12 years as a staff member and later manager in the Thermal/Fluid Science and Engineering department at Sandia National Laboratories in Livermore, CA, where his work included multiscale and multiphysics computational methods, multiphase and particulate flow simulation, extended timescale methods for atomistic simulation, and large-scale engineering code development. In January 2015 he joined the faculty of the Mechanical Engineering department at Northwestern. His current research focuses on applying novel simulation methods and high performance computing to multiscale and multiphysics problems, including additive manufacturing in metals, environmental transport, and multiphase flows.

WaiChing Sun (wsun@columbia.edu)
WaiChing Sun works in theoretical and computational mechanics for porous and geological materials. He obtained his B.S. from UC Davis (2005); M.S. (geomechanics) from Stanford (2007); M.A. degree from Princeton (2008); and Ph.D. in theoretical and applied mechanics from Northwestern (2011). Prior to joining Columbia, he was a senior member of technical staff in the mechanics of materials department at Sandia National Laboratories (Livermore, CA). He is the recipient of the Zienkiewicz Numerical Methods in Engineering Prize in 2016, US Air Force Young Investigator Program Award in 2017, Dresden Fellowship in 2016, US Army Young Investigator Program Award in 2015, and the Caterpillar Best Paper Prize in 2013, among others.

Miguel Bessa (M.A.Bessa@tudelft.nl)
Miguel Bessa's research involves understanding and modeling materials at every scale in a unique experimentally-validated and self-consistent computational framework. He envisions a new era of data-driven design of materials and structures based on Physics-informed machine learning, reduced order models and genetic optimization. Miguel started as an Assistant Professor in Materials Science at the Delft University of Technology on August 2017. He received his PhD in Mechanical Engineering from Northwestern University in September 2016 (Fulbright scholar; 4.0 GPA), and was a postdoctoral scholar in Aerospace at the California Institute of Technology until August 2017.


In recent years, the amount of digital data collectively generated by humans has doubled every two years or less. In the field of computational mechanics, not only is the amount of data increasing, but the form of that data has become increasingly diverse, due in part to technology advancements in sensors, micro-CT imaging, and high-speed cameras. However, making use of these increasingly complex and rich data to enhance scientific and engineering predictions remains a challenging problem. The objective of this special issue is to provide a platform to exchange ideas and advance knowledge on data-driven computational modeling and simulations. We are particularly interested in new ideas that integrate big data analytics and machine learning with existing human knowledge to create, calibrate, verify, and validate forward prediction models as well as inverse problems. Potential topics may include, but are not limited to:

• Reduced-order models and other methods to accelerate data generation and collection

• Data clustering, fusion, mining and feature extraction for computational mechanics

• Uncertainty quantification in the context of Big Data

• Applications of machine/deep learning for constitutive laws and model discovery

• Applications of machine/deep learning to design materials or structures

• Solutions to inverse problems

• Data-driven closure models for sub-grid scales, such as turbulence

• Issues related to the training, verification and validation of data-driven models


Machine learning, deep learning, uncertainty quantification, reduced-order modeling, coherent structures, classification, forward prediction, verification and validation

Published Papers

  • Open Access


    A Deep Learning-Based Computational Algorithm for Identifying Damage Load Condition: An Artificial Intelligence Inverse Problem Solution for Failure Analysis

    Shaofei Ren, Guorong Chen , Tiange Li , Qijun Chen, Shaofan Li
    CMES-Computer Modeling in Engineering & Sciences, Vol.117, No.3, pp. 287-307, 2018, DOI:10.31614/cmes.2018.04697
    (This article belongs to this Special Issue: Data-driven Computational Modeling and Simulations)
    Abstract In this work, we have developed a novel machine (deep) learning computational framework to determine and identify damage loading parameters (conditions) for structures and materials based on the permanent or residual plastic deformation distribution or damage state of the structure. We have shown that the developed machine learning algorithm can accurately and (practically) uniquely identify both prior static as well as impact loading conditions in an inverse manner, based on the residual plastic strain and plastic deformation as forensic signatures. The paper presents the detailed machine learning algorithm, data acquisition and learning processes, and validation/verification examples. This development may have… More >

  • Open Access


    Data-Driven Prediction of Mechanical Properties in Support of Rapid Certification of Additively Manufactured Alloys

    Fuyao Yan, Yu hin Chan, Abhinav Saboo, Jiten Shah, Gregory B. Olson, Wei Chen
    CMES-Computer Modeling in Engineering & Sciences, Vol.117, No.3, pp. 343-366, 2018, DOI:10.31614/cmes.2018.04452
    (This article belongs to this Special Issue: Data-driven Computational Modeling and Simulations)
    Abstract Predicting the mechanical properties of additively manufactured parts is often a tedious process, requiring the integration of multiple stand-alone and expensive simulations. Furthermore, as properties are highly location-dependent due to repeated heating and cooling cycles, the properties prediction models must be run for multiple locations before the part-level performance can be analyzed for certification, compounding the computational expense. This work has proposed a rapid prediction framework that replaces the physics-based mechanistic models with Gaussian process metamodels, a type of machine learning model for statistical inference with limited data. The metamodels can predict the varying properties within an entire part in… More >

  • Open Access


    Data-Driven Upscaling of Orientation Kinematics in Suspensions of Rigid Fibres

    Adrien Scheuer, Amine Ammar, Emmanuelle Abisset-Chavanne, Elias Cueto, Francisco Chinesta, Roland Keunings, Suresh G. Advani
    CMES-Computer Modeling in Engineering & Sciences, Vol.117, No.3, pp. 367-386, 2018, DOI:10.31614/cmes.2018.04278
    (This article belongs to this Special Issue: Data-driven Computational Modeling and Simulations)
    Abstract Describing the orientation state of the particles is often critical in fibre suspen-sion applications. Macroscopic descriptors, the so-called second-order orientation tensor (or moment) leading the way, are often preferred due to their low computational cost. Clo-sure problems however arise when evolution equations for the moments are derived from the orientation distribution functions and the impact of the chosen closure is often unpre-dictable. In this work, our aim is to provide macroscopic simulations of orientation that are cheap, accurate and closure-free. To this end, we propose an innovative data-based approach to the upscaling of orientation kinematics in the context of fibre… More >

  • Open Access


    Machine Learning Models of Plastic Flow Based on Representation Theory

    R. E. Jones, J. A. Templeton, C. M. Sanders, J. T. Ostien
    CMES-Computer Modeling in Engineering & Sciences, Vol.117, No.3, pp. 309-342, 2018, DOI:10.31614/cmes.2018.04285
    (This article belongs to this Special Issue: Data-driven Computational Modeling and Simulations)
    Abstract We use machine learning (ML) to infer stress and plastic flow rules using data from representative polycrystalline simulations. In particular, we use so-called deep (multilayer) neural networks (NN) to represent the two response functions. The ML process does not choose appropriate inputs or outputs, rather it is trained on selected inputs and output. Likewise, its discrimination of features is crucially connected to the chosen inputoutput map. Hence, we draw upon classical constitutive modeling to select inputs and enforce well-accepted symmetries and other properties. In the context of the results of numerous simulations, we discuss the design, stability and accuracy of… More >

  • Open Access


    Data Mining and Machine Learning Methods Applied to 3 A Numerical Clinching Model

    Marco Götz, Ferenc Leichsenring, Thomas Kropp, Peter Müller, Tobias Falk, Wolfgang Graf, Michael Kaliske, Welf-Guntram Drossel
    CMES-Computer Modeling in Engineering & Sciences, Vol.117, No.3, pp. 387-423, 2018, DOI:10.31614/cmes.2018.04112
    (This article belongs to this Special Issue: Data-driven Computational Modeling and Simulations)
    Abstract Numerical mechanical models used for design of structures and processes are very complex and high-dimensionally parametrised. The understanding of the model characteristics is of interest for engineering tasks and subsequently for an efficient design. Multiple analysis methods are known and available to gain insight into existing models. In this contribution, selected methods from various fields are applied to a real world mechanical engineering example of a currently developed clinching process. The selection of introduced methods comprises techniques of machine learning and data mining, in which the utilization is aiming at a decreased numerical effort. The methods of choice are basically… More >

Share Link