Erik Faust
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Erik Faust has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Schlüter
Institute for Mechanics, Technical University of Darmstadt, Darmstadt, Germany
2 shared publications - Henning Müller
Institut für Mechanik, Technische Universität Darmstadt, Darmstadt, Germany
2 shared publications Lisa Scheunemann
Institute of Applied Mechanics RWTH Aachen University Aachen Germany
2 shared publications- Lisa Scheunemann
Institute of Applied Mechanics RWTH Aachen University Aachen Germany
2 shared publications - Lisa Scheunemann
Chair of Applied Mechanics Department of Mechanical and Process Engineering RPTU Kaiserslautern-Landau
2 shared publications - Dong Zhao
Institute of Applied Mechanics (IFAM) RWTH Aachen University Aachen Germany
1 shared publication
Latest Publications
- Open Access
- 21 September 2026
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 26 (4)
Strain‐Space Hyperreduction Using Empirically Corrected Cluster Cubature With Linear and Quadratic Approximation Spaces
Strain‐space reduced cubature techniques such as the Empirically Corrected Cluster Cubature (E3C) considerably reduce the integration effort incurred in computational homogenisation problems. In this work, we apply E3C in tandem with projection‐based model reduction techniques utilising both linear and quadratic approximation spaces. For simulations on an example porous hyperelastic representative volume element (RVE), E3C leads to considerably improved performance, whether with linear or quadratic approximation spaces, over alternative hyperreduction approaches explored in our previous work. As expected, a quadratic ansatz allows for the construction of smaller approximation spaces than a linear one, with equal accuracy. However, in the example explored here, this does not translate into an improved tradeoff between accuracy and runtime.
- 21 September 2026
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 26 (4)
Open AccessStrain‐Space Hyperreduction Using Empirically Corrected Cluster Cubature With Linear and Quadratic Approximation Spaces
Strain‐space reduced cubature techniques such as the Empirically Corrected Cluster Cubature (E3C) considerably reduce the integration effort incurred in computational homogenisation problems. In this work, we apply E3C in tandem with projection‐based model reduction techniques utilising both linear and quadratic approximation spaces. For simulations on an example porous hyperelastic representative volume element (RVE), E3C leads to considerably improved performance, whether with linear or quadratic approximation spaces, over alternative hyperreduction approaches explored in our previous work. As expected, a quadratic ansatz allows for the construction of smaller approximation spaces than a linear one, with equal accuracy. However, in the example explored here, this does not translate into an improved tradeoff between accuracy and runtime.
- Open Access
- 2 April 2026
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 26 (2)
Nonlinear Model Order Reduction on Polynomial Manifolds for Computational Homogenisation Problems
Model order reduction (MOR) techniques utilising nonlinear approximation spaces can search for solutions to computational homogenisation problems on low‐dimensional approximation spaces. In combination with hyperreduction techniques, this allows for computations on representative volume elements (RVEs) to be accelerated by multiple orders of magnitude. In this work, we apply a polynomial manifold approach to efficiently compute solutions to a hyperelastic quasi‐static RVE problem.
- 2 April 2026
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 26 (2)
Open AccessNonlinear Model Order Reduction on Polynomial Manifolds for Computational Homogenisation Problems
Model order reduction (MOR) techniques utilising nonlinear approximation spaces can search for solutions to computational homogenisation problems on low‐dimensional approximation spaces. In combination with hyperreduction techniques, this allows for computations on representative volume elements (RVEs) to be accelerated by multiple orders of magnitude. In this work, we apply a polynomial manifold approach to efficiently compute solutions to a hyperelastic quasi‐static RVE problem.
- Open Access
- 9 December 2025
- Published by Springer Nature in Computational Mechanics
- Vol. 77 (4) , 983-1023
A manifold learning approach to nonlinear model order reduction of quasi-static problems in solid mechanics: proof of concept and parameter study
The proper orthogonal decomposition (POD)—a popular model order reduction (MOR) method—may require significant model dimensionalities to successfully capture a nonlinear solution manifold resulting from a parameterised quasi-static solid-mechanical problem. The local basis method by Amsallem et al. (Int J Numer Meth Eng 92(10):891–916, 2012) addresses this deficiency by introducing a locally, rather than globally, linear approximation of the solution manifold. However, this generally successful approach comes with some limitations, especially in the data-poor setting. In this investigation, we instead propose a graph-based manifold learning approach to nonlinear projection-based MOR which uses a global, continuously nonlinear approximation of the solution manifold. Approximations of local tangents to the solution manifold, which are necessary for a Galerkin scheme, are computed in the online phase. As an example application for the resulting nonlinear MOR algorithms, we consider simple representative volume element computations. On this example, the manifold learning approach Pareto-dominates the POD and local basis method in terms of the error achieved using a range of model dimensionalities. This contribution discusses the broad framework of the presented manifold learning approach in view of an application within MOR schemes; hyperreduction is beyond the scope of this article.
- 9 December 2025
- Published by Springer Nature in Computational Mechanics
- Vol. 77 (4) , 983-1023
Open AccessA manifold learning approach to nonlinear model order reduction of quasi-static problems in solid mechanics: proof of concept and parameter study
The proper orthogonal decomposition (POD)—a popular model order reduction (MOR) method—may require significant model dimensionalities to successfully capture a nonlinear solution manifold resulting from a parameterised quasi-static solid-mechanical problem. The local basis method by Amsallem et al. (Int J Numer Meth Eng 92(10):891–916, 2012) addresses this deficiency by introducing a locally, rather than globally, linear approximation of the solution manifold. However, this generally successful approach comes with some limitations, especially in the data-poor setting. In this investigation, we instead propose a graph-based manifold learning approach to nonlinear projection-based MOR which uses a global, continuously nonlinear approximation of the solution manifold. Approximations of local tangents to the solution manifold, which are necessary for a Galerkin scheme, are computed in the online phase. As an example application for the resulting nonlinear MOR algorithms, we consider simple representative volume element computations. On this example, the manifold learning approach Pareto-dominates the POD and local basis method in terms of the error achieved using a range of model dimensionalities. This contribution discusses the broad framework of the presented manifold learning approach in view of an application within MOR schemes; hyperreduction is beyond the scope of this article.
- Open Access
A hyperreduced manifold learning approach to nonlinear model order reduction for the homogenisation of hyperelastic RVEs
In a recent work, we proposed a graph-based manifold learning scheme for the nonlinear Galerkin-reduction of quasi-static solid mechanical problems [1]. The resulting nonlinear approximation spaces can closely and flexibly represent nonlinear solution manifolds. The present work discusses how this nonlinear model order reduction (MOR) approach can be employed to reduce online computational costs by multiple orders of magnitude while retaining high levels of accuracy. We integrate two popular hyperreduction methods into the nonlinear MOR framework and discuss how we achieve an algorithmic complexity which is independent from the original system size. Furthermore, improvements are made to the local online linearisation scheme for the sake of performance and robustness. On an example RVE problem, the MOR scheme accelerates computations by more than two orders of magnitude with little training data and negligible loss of accuracy. Additionally, the algorithm Pareto-dominates alternative approaches in the trade-off between accuracy and runtime on the considered example.
Open AccessA hyperreduced manifold learning approach to nonlinear model order reduction for the homogenisation of hyperelastic RVEs
In a recent work, we proposed a graph-based manifold learning scheme for the nonlinear Galerkin-reduction of quasi-static solid mechanical problems [1]. The resulting nonlinear approximation spaces can closely and flexibly represent nonlinear solution manifolds. The present work discusses how this nonlinear model order reduction (MOR) approach can be employed to reduce online computational costs by multiple orders of magnitude while retaining high levels of accuracy. We integrate two popular hyperreduction methods into the nonlinear MOR framework and discuss how we achieve an algorithmic complexity which is independent from the original system size. Furthermore, improvements are made to the local online linearisation scheme for the sake of performance and robustness. On an example RVE problem, the MOR scheme accelerates computations by more than two orders of magnitude with little training data and negligible loss of accuracy. Additionally, the algorithm Pareto-dominates alternative approaches in the trade-off between accuracy and runtime on the considered example.
- Open Access
- 23 May 2025
- Published by Elsevier in Computer Methods in Applied Mechanics and Engineering
- Vol. 443
Extending the Lattice Boltzmann Method to non-linear elastodynamics
- 23 May 2025
- Published by Elsevier in Computer Methods in Applied Mechanics and Engineering
- Vol. 443
Open AccessExtending the Lattice Boltzmann Method to non-linear elastodynamics
- Open Access
A manifold learning approach to nonlinear model order reduction of quasi-static problems in solid mechanics
The proper orthogonal decomposition (POD) -- a popular projection-based model order reduction (MOR) method -- may require significant model dimensionalities to successfully capture a nonlinear solution manifold resulting from a parameterised quasi-static solid-mechanical problem. The local basis method by Amsallem et al. [1] addresses this deficiency by introducing a locally, rather than globally, linear approximation of the solution manifold. However, this generally successful approach comes with some limitations, especially in the data-poor setting. In this proof-of-concept investigation, we instead propose a graph-based manifold learning approach to nonlinear projection-based MOR which uses a global, continuously nonlinear approximation of the solution manifold. Approximations of local tangents to the solution manifold, which are necessary for a Galerkin scheme, are computed in the online phase. As an example application for the resulting nonlinear MOR algorithms, we consider simple representative volume element computations. On this example, the manifold learning approach Pareto-dominates the POD and local basis method in terms of the error and runtime achieved using a range of model dimensionalities.
Open AccessA manifold learning approach to nonlinear model order reduction of quasi-static problems in solid mechanics
The proper orthogonal decomposition (POD) -- a popular projection-based model order reduction (MOR) method -- may require significant model dimensionalities to successfully capture a nonlinear solution manifold resulting from a parameterised quasi-static solid-mechanical problem. The local basis method by Amsallem et al. [1] addresses this deficiency by introducing a locally, rather than globally, linear approximation of the solution manifold. However, this generally successful approach comes with some limitations, especially in the data-poor setting. In this proof-of-concept investigation, we instead propose a graph-based manifold learning approach to nonlinear projection-based MOR which uses a global, continuously nonlinear approximation of the solution manifold. Approximations of local tangents to the solution manifold, which are necessary for a Galerkin scheme, are computed in the online phase. As an example application for the resulting nonlinear MOR algorithms, we consider simple representative volume element computations. On this example, the manifold learning approach Pareto-dominates the POD and local basis method in terms of the error and runtime achieved using a range of model dimensionalities.
- Open Access
- 1 August 2024
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 24 (2)
Nonlinear model order reduction using local basis methods on RVEs: A parameter study and comparison of different variants
All possible solutions to a static representative volume element (RVE) problem lie on a solution manifold, which is often quite low‐dimensional. Solutions of a hyperelastic 3D RVE problem, for example, are determined by the macroscopic deformation gradient, meaning that its solution manifold is only six‐dimensional. Established projection‐based model order reduction (MOR) methods such as the Proper Orthogonal Decomposition (POD), however, often require much higher‐dimensional approximation spaces to successfully capture such a nonlinear solution manifold on account of their linearity. One approach to modelling the solution manifold more closely is the local basis method by Amsallem et al.. It utilises a locally, rather than globally, linear approximation of the solution manifold. In this work, we investigate several variants of this algorithm with respect to their performance on an example RVE problem.
- 1 August 2024
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 24 (2)
Open AccessNonlinear model order reduction using local basis methods on RVEs: A parameter study and comparison of different variants
All possible solutions to a static representative volume element (RVE) problem lie on a solution manifold, which is often quite low‐dimensional. Solutions of a hyperelastic 3D RVE problem, for example, are determined by the macroscopic deformation gradient, meaning that its solution manifold is only six‐dimensional. Established projection‐based model order reduction (MOR) methods such as the Proper Orthogonal Decomposition (POD), however, often require much higher‐dimensional approximation spaces to successfully capture such a nonlinear solution manifold on account of their linearity. One approach to modelling the solution manifold more closely is the local basis method by Amsallem et al.. It utilises a locally, rather than globally, linear approximation of the solution manifold. In this work, we investigate several variants of this algorithm with respect to their performance on an example RVE problem.
- Open Access
- 26 July 2023
- Published by Springer Nature in Computational Mechanics
- Vol. 73 (2) , 317-339
Dirichlet and Neumann boundary conditions in a lattice Boltzmann method for elastodynamics
Recently, Murthy et al. (Commun Comput Phys 2:23, 2017. http://dx.doi.org/10.4208/cicp.OA-2016-0259 ) and Escande et al. (Lattice Boltzmann method for wave propagation in elastic solids with a regular lattice: theoretical analysis and validation, 2020. arXiv.doi:1048550/ARXIV.2009.06404. arXiv:2009.06404) adopted the Lattice Boltzmann Method (LBM) to model the linear elastodynamic behaviour of isotropic solids. The LBM is attractive as an elastodynamic solver because it can be parallelised readily and lends itself to finely discretised simulations of dynamic effects in continua, allowing transient phenomena such as wave propagation to be modeled efficiently. This work proposes simple local boundary rules which approximate the behaviour of Dirichlet and Neumann boundary conditions with an LBM for elastic solids. The boundary rules are shown to be consistent with the target boundary values in the first order. An empirical convergence study is performed for the transient tension loading of a rectangular plate, with a Finite Element (FE) simulation being used as a reference. Additionally, we compare results produced by the LBM for the sudden loading of a stationary crack with an analytical solution from Freund (Dynamic fracture mechanics. Cambridge Monographs on Mechanics. Cambridge University Press, Cambridge, 1990. https://doi.org/10.1017/CBO9780511546761).
- 26 July 2023
- Published by Springer Nature in Computational Mechanics
- Vol. 73 (2) , 317-339
Open AccessDirichlet and Neumann boundary conditions in a lattice Boltzmann method for elastodynamics
Recently, Murthy et al. (Commun Comput Phys 2:23, 2017. http://dx.doi.org/10.4208/cicp.OA-2016-0259 ) and Escande et al. (Lattice Boltzmann method for wave propagation in elastic solids with a regular lattice: theoretical analysis and validation, 2020. arXiv.doi:1048550/ARXIV.2009.06404. arXiv:2009.06404) adopted the Lattice Boltzmann Method (LBM) to model the linear elastodynamic behaviour of isotropic solids. The LBM is attractive as an elastodynamic solver because it can be parallelised readily and lends itself to finely discretised simulations of dynamic effects in continua, allowing transient phenomena such as wave propagation to be modeled efficiently. This work proposes simple local boundary rules which approximate the behaviour of Dirichlet and Neumann boundary conditions with an LBM for elastic solids. The boundary rules are shown to be consistent with the target boundary values in the first order. An empirical convergence study is performed for the transient tension loading of a rectangular plate, with a Finite Element (FE) simulation being used as a reference. Additionally, we compare results produced by the LBM for the sudden loading of a stationary crack with an analytical solution from Freund (Dynamic fracture mechanics. Cambridge Monographs on Mechanics. Cambridge University Press, Cambridge, 1990. https://doi.org/10.1017/CBO9780511546761).
- Open Access
- 1 June 2015
- Published by SAGE Publications in Justice Research and Policy
- Vol. 16 (1) , 81-98
Sex Offenders in the Federal Correctional System
Driven largely by changes in federal legislation, which increased resources for law enforcement and enhanced mandatory minimum sentences, the number of sex offenders handled by the federal criminal justice system has increased over the last 5 years at rates second only to illegal immigration offenses. The heightened attention to sex offending by federal law enforcement coupled with a greater certainty of punishment imposed has resulted in a record number of sex offenders in federal prison and on supervision in the community. Child pornography offenders, for example, share a similar likelihood of receiving a prison sentence (given being charged and convicted) as drug, weapon, and violent offenders. Sex offenders are also serving longer terms in federal prison and on supervision—terms that are expected to increase further in the next decade. This article considers changes in the federal response to sex offenders and growth in prison and supervision populations. Implications for the management of sex offenders in prison—including prison-based residential and nonresidential treatment programs and civil commitment for individuals adjudicated as sexually dangerous—and offenders under supervision in the community are discussed.
- 1 June 2015
- Published by SAGE Publications in Justice Research and Policy
- Vol. 16 (1) , 81-98
Open AccessSex Offenders in the Federal Correctional System
Driven largely by changes in federal legislation, which increased resources for law enforcement and enhanced mandatory minimum sentences, the number of sex offenders handled by the federal criminal justice system has increased over the last 5 years at rates second only to illegal immigration offenses. The heightened attention to sex offending by federal law enforcement coupled with a greater certainty of punishment imposed has resulted in a record number of sex offenders in federal prison and on supervision in the community. Child pornography offenders, for example, share a similar likelihood of receiving a prison sentence (given being charged and convicted) as drug, weapon, and violent offenders. Sex offenders are also serving longer terms in federal prison and on supervision—terms that are expected to increase further in the next decade. This article considers changes in the federal response to sex offenders and growth in prison and supervision populations. Implications for the management of sex offenders in prison—including prison-based residential and nonresidential treatment programs and civil commitment for individuals adjudicated as sexually dangerous—and offenders under supervision in the community are discussed.