Showing 1–36 of 36 results for author: Exl, L

Physics-informed machine learning and stray field computation with application to micromagnetic energy minimization

Authors: Sebastian Schaffer, Thomas Schrefl, Harald Oezelt, Alexander Kovacs, Leoni Breth, Norbert J. Mauser, Dieter Suess, Lukas Exl

Abstract: We study the full 3d static micromagnetic equations via a physics-informed neural network (PINN) ansatz for the continuous magnetization configuration. PINNs are inherently mesh-free and unsupervised learning models. In our approach we can learn to minimize the total Gibbs free energy with additional conditional parameters, such as the exchange length, by a single low-parametric neural network mod… ▽ More We study the full 3d static micromagnetic equations via a physics-informed neural network (PINN) ansatz for the continuous magnetization configuration. PINNs are inherently mesh-free and unsupervised learning models. In our approach we can learn to minimize the total Gibbs free energy with additional conditional parameters, such as the exchange length, by a single low-parametric neural network model. In contrast, traditional numerical methods would require the computation and storage of a large number of solutions to interpolate the continuous spectrum of quasi-optimal magnetization configurations. We also consider the important and computationally expensive stray field problem via PINNs, where we use a basically linear learning ansatz, called extreme learning machines (ELM) within a splitting method for the scalar potential. This reduces the stray field training to a linear least squares problem with precomputable solution operator. We validate the stray field method by means of numerical example comparisons from literature and illustrate the full micromagnetic approach via the NIST $μ$MAG standard problem $\#3$. △ Less

Submitted 31 January, 2023; originally announced January 2023.

Comments: 17 pages, 10 figures

MSC Class: 62P35; 68T05; 65Z05

Bridging Fidelities to Predict Nanoindentation Tip Radii Using Interpretable Deep Learning Models

Authors: Claus O. W. Trost, Stanislav Zak, Sebastian Schaffer, Christian Saringer, Lukas Exl, Megan J. Cordill

Abstract: As the need for miniaturized structural and functional materials has increased,the need for precise materials characterizaton has also expanded. Nanoindentation is a popular method that can be used to measure material mechanical behavior which enables high-throughput experiments and, in some cases, can also provide images of the indented area through scanning. Both indenting and scanning can cause… ▽ More As the need for miniaturized structural and functional materials has increased,the need for precise materials characterizaton has also expanded. Nanoindentation is a popular method that can be used to measure material mechanical behavior which enables high-throughput experiments and, in some cases, can also provide images of the indented area through scanning. Both indenting and scanning can cause tip wear that can influence the measurements. Therefore, precise characterization of tip radii is needed to improve data evaluation. A data fusion method is introduced which uses finite element simulations and experimental data to estimate the tip radius in situ in a meaningful way using an interpretable multi-fidelity deep learning approach. By interpreting the machine learning models, it is shown that the approaches are able to accurately capture physical indentation phenomena. △ Less

Submitted 1 July, 2022; originally announced July 2022.

Comments: 11 pages, 8 figures, for used codes and updated figure 5 see https://github.com/materialsguy/Predict_Nanoindentation_Tip_Wear

MSC Class: 68T07; 74S05; 91A12

Description of collective magnetization processes with machine learning models

Authors: Alexander Kornell, Lukas Exl, Leoni Breth, Johann Fischbacher, Alexander Kovacs, Harald Oezelt, Markus Gusenbauer, Masao Yano, Noritsugu Sakuma, Akihito Kinoshita, Tetsuya Shoji, Akira Kato, Norbert J. Mauser, Thomas Schrefl

Abstract: This work introduces a latent space method to calculate the demagnetization reversal process of multigrain permanent magnets. The algorithm consists of two deep learning models based on neural networks. The embedded Stoner-Wohlfarth method is used as a reduced order model for computing samples to train and test the machine learning approach. The work is a proof of concept of the method since the u… ▽ More This work introduces a latent space method to calculate the demagnetization reversal process of multigrain permanent magnets. The algorithm consists of two deep learning models based on neural networks. The embedded Stoner-Wohlfarth method is used as a reduced order model for computing samples to train and test the machine learning approach. The work is a proof of concept of the method since the used microstructures are simple and the only varying parameters are the magnetic anisotropy axes of the hard magnetic grains. A convolutional autoencoder is used for nonlinear dimensionality reduction, in order to reduce the required amount of training samples for predicting the hysteresis loops. We enriched the loss function with physical information about the underlying problem which increases the accuracy of the machine learning approach. A deep learning regressor is operating in the latent space of the autoencoder. The predictor takes a series of previously encoded magnetization states and outputs the next magnetization state along the hysteresis curve. To predict the complete demagnetization loop, we apply a recursive learning scheme. △ Less

Submitted 7 May, 2022; originally announced May 2022.

Exploring the hysteresis properties of nanocrystalline permanent magnets using deep learning

Authors: Alexander Kovacs, Lukas Exl, Alexander Kornell, Johann Fischbacher, Markus Hovorka, Markus Gusenbauer, Leoni Breth, Harald Oezelt, Masao Yano, Noritsugu Sakuma, Akihito Kinoshita, Tetsuya Shoji, Akira Kato, Thomas Schrefl

Abstract: We demonstrate the use of model order reduction and neural networks for estimating the hysteresis properties of nanocrystalline permanent magnets from microstructure. With a data-driven approach, we learn the demagnetization curve from data-sets created by grain growth and micromagnetic simulations. We show that the granular structure of a magnet can be encoded within a low-dimensional latent spac… ▽ More We demonstrate the use of model order reduction and neural networks for estimating the hysteresis properties of nanocrystalline permanent magnets from microstructure. With a data-driven approach, we learn the demagnetization curve from data-sets created by grain growth and micromagnetic simulations. We show that the granular structure of a magnet can be encoded within a low-dimensional latent space. Latent codes are constructed using a variational autoencoder. The map** of structure code to hysteresis properties is a multi-target regression problem. We apply deep neural network and use parameter sharing, in order to predict anchor points along the demagnetization curves from the magnet's structure code. The method is applied to study the magnetic properties of nanocrystalline permanent magnets. We show how new grain structures can be generated by interpolation between two points in the latent space and how the magnetic properties of the resulting magnets can be predicted. △ Less

Submitted 30 March, 2022; originally announced March 2022.

Comments: 17 pages, 8 figures

MSC Class: 68T07 ACM Class: I.2.m

Magnetostatics and micromagnetics with physics informed neural networks

Authors: Alexander Kovacs, Lukas Exl, Alexander Kornell, Johann Fischbacher, Markus Hovorka, Markus Gusenbauer, Leoni Breth, Harald Oezelt, Dirk Praetorius, Dieter Suess, Thomas Schrefl

Abstract: Partial differential equations and variational problems can be solved with physics informed neural networks (PINNs). The unknown field is approximated with neural networks. Minimizing the residuals of the static Maxwell equation at collocation points or the magnetostatic energy, the weights of the neural network are adjusted so that the neural network solution approximates the magnetic vector pote… ▽ More Partial differential equations and variational problems can be solved with physics informed neural networks (PINNs). The unknown field is approximated with neural networks. Minimizing the residuals of the static Maxwell equation at collocation points or the magnetostatic energy, the weights of the neural network are adjusted so that the neural network solution approximates the magnetic vector potential. This way, the magnetic flux density for a given magnetization distribution can be estimated. With the magnetization as an additional unknown, inverse magnetostatic problems can be solved. Augmenting the magnetostatic energy with additional energy terms, micromagnetic problems can be solved. We demonstrate the use of physics informed neural networks for solving magnetostatic problems, computing the magnetization for inverse problems, and calculating the demagnetization curves for two-dimensional geometries. △ Less

Submitted 7 June, 2021; originally announced June 2021.

Comments: 22 pages, 9 figures

MSC Class: I.6.3; I.6.5; I.2.6

Conditional physics informed neural networks

Authors: Alexander Kovacs, Lukas Exl, Alexander Kornell, Johann Fischbacher, Markus Hovorka, Markus Gusenbauer, Leoni Breth, Harald Oezelt, Masao Yano, Noritsugu Sakuma, Akihito Kinoshita, Tetsuya Shoji, Akira Kato, Thomas Schrefl

Abstract: We introduce conditional PINNs (physics informed neural networks) for estimating the solution of classes of eigenvalue problems. The concept of PINNs is expanded to learn not only the solution of one particular differential equation but the solutions to a class of problems. We demonstrate this idea by estimating the coercive field of permanent magnets which depends on the width and strength of loc… ▽ More We introduce conditional PINNs (physics informed neural networks) for estimating the solution of classes of eigenvalue problems. The concept of PINNs is expanded to learn not only the solution of one particular differential equation but the solutions to a class of problems. We demonstrate this idea by estimating the coercive field of permanent magnets which depends on the width and strength of local defects. When the neural network incorporates the physics of magnetization reversal, training can be achieved in an unsupervised way. There is no need to generate labeled training data. The presented test cases have been rigorously studied in the past. Thus, a detailed and easy comparison with analytical solutions is made. We show that a single deep neural network can learn the solution of partial differential equations for an entire class of problems. △ Less

Submitted 6 April, 2021; originally announced April 2021.

Comments: 18 pages, 11 figures

MSC Class: I.2.6

Machine learning methods for the prediction of micromagnetic magnetization dynamics

Authors: Sebastian Schaffer, Norbert J. Mauser, Thomas Schrefl, Dieter Suess, Lukas Exl

Abstract: Machine learning (ML) entered the field of computational micromagnetics only recently. The main objective of these new approaches is the automatization of solutions of parameter-dependent problems in micromagnetism such as fast response curve estimation modeled by the Landau-Lifschitz-Gilbert (LLG) equation. Data-driven models for the solution of time- and parameter-dependent partial differential… ▽ More Machine learning (ML) entered the field of computational micromagnetics only recently. The main objective of these new approaches is the automatization of solutions of parameter-dependent problems in micromagnetism such as fast response curve estimation modeled by the Landau-Lifschitz-Gilbert (LLG) equation. Data-driven models for the solution of time- and parameter-dependent partial differential equations require high dimensional training data-structures. ML in this case is by no means a straight-forward trivial task, it needs algorithmic and mathematical innovation. Our work introduces theoretical and computational conceptions of certain kernel and neural network based dimensionality reduction approaches for efficient prediction of solutions via the notion of low-dimensional feature space integration. We introduce efficient treatment of kernel ridge regression and kernel principal component analysis via low-rank approximation. A second line follows neural network (NN) autoencoders as nonlinear data-dependent dimensional reduction for the training data with focus on accurate latent space variable description suitable for a feature space integration scheme. We verify and compare numerically by means of a NIST standard problem. The low-rank kernel method approach is fast and surprisingly accurate, while the NN scheme can even exceed this level of accuracy at the expense of significantly higher costs. △ Less

Submitted 4 July, 2021; v1 submitted 16 March, 2021; originally announced March 2021.

Prediction of magnetization dynamics in a reduced dimensional feature space setting utilizing a low-rank kernel method

Authors: Lukas Exl, Norbert J. Mauser, Sebastian Schaffer, Thomas Schrefl, Dieter Suess

Abstract: We establish a machine learning model for the prediction of the magnetization dynamics as function of the external field described by the Landau-Lifschitz-Gilbert equation, the partial differential equation of motion in micromagnetism. The model allows for fast and accurate determination of the response to an external field which is illustrated by a thin-film standard problem. The data-driven meth… ▽ More We establish a machine learning model for the prediction of the magnetization dynamics as function of the external field described by the Landau-Lifschitz-Gilbert equation, the partial differential equation of motion in micromagnetism. The model allows for fast and accurate determination of the response to an external field which is illustrated by a thin-film standard problem. The data-driven method internally reduces the dimensionality of the problem by means of nonlinear model reduction for unsupervised learning. This not only makes accurate prediction of the time steps possible, but also decisively reduces complexity in the learning process where magnetization states from simulated micromagnetic dynamics associated with different external fields are used as input data. We use a truncated representation of kernel principal components to describe the states between time predictions. The method is capable of handling large training sample sets owing to a low-rank approximation of the kernel matrix and an associated low-rank extension of kernel principal component analysis and kernel ridge regression. The approach entirely shifts computations into a reduced dimensional setting breaking down the problem dimension from the thousands to the tens. △ Less

Submitted 19 July, 2021; v1 submitted 13 August, 2020; originally announced August 2020.

MSC Class: 62P35; 68T05; 65Z05

Optimal control of the self-bound dipolar droplet formation process

Authors: Jan-Frederik Mennemann, Tim Langen, Lukas Exl, Norbert J. Mauser

Abstract: Dipolar Bose-Einstein condensates have recently attracted much attention in the world of quantum many body experiments. While the theoretical principles behind these experiments are typically supported by numerical simulations, the application of optimal control algorithms could potentially open up entirely new possibilities. As a proof of concept, we demonstrate that the formation process of a si… ▽ More Dipolar Bose-Einstein condensates have recently attracted much attention in the world of quantum many body experiments. While the theoretical principles behind these experiments are typically supported by numerical simulations, the application of optimal control algorithms could potentially open up entirely new possibilities. As a proof of concept, we demonstrate that the formation process of a single dipolar droplet state could be dramatically accelerated using advanced concepts of optimal control. More specifically, our optimization is based on a multilevel B-spline method reducing the number of required cost function evaluations and hence significantly reducing the numerical effort. Moreover, our strategy allows to consider box constraints on the control inputs in a concise and efficient way. To further improve the overall efficiency, we show how to evaluate the dipolar interaction potential in the generalized Gross-Pitaevskii equation without sacrificing the spectral convergence rate of the underlying time-splitting spectral method. △ Less

Submitted 29 May, 2019; originally announced May 2019.

Learning time-step** by nonlinear dimensionality reduction to predict magnetization dynamics

Authors: Lukas Exl, Norbert J. Mauser, Thomas Schrefl, Dieter Suess

Abstract: We establish a time-step** learning algorithm and apply it to predict the solution of the partial differential equation of motion in micromagnetism as a dynamical system depending on the external field as parameter. The data-driven approach is based on nonlinear model order reduction by use of kernel methods for unsupervised learning, yielding a predictor for the magnetization dynamics without a… ▽ More We establish a time-step** learning algorithm and apply it to predict the solution of the partial differential equation of motion in micromagnetism as a dynamical system depending on the external field as parameter. The data-driven approach is based on nonlinear model order reduction by use of kernel methods for unsupervised learning, yielding a predictor for the magnetization dynamics without any need for field evaluations after a data generation and training phase as precomputation. Magnetization states from simulated micromagnetic dynamics associated with different external fields are used as training data to learn a low-dimensional representation in so-called feature space and a map that predicts the time-evolution in reduced space. Remarkably, only two degrees of freedom in feature space were enough to describe the nonlinear dynamics of a thin-film element. The approach has no restrictions on the spatial discretization and might be useful for fast determination of the response to an external field. △ Less

Submitted 29 January, 2020; v1 submitted 8 April, 2019; originally announced April 2019.

MSC Class: 37M05; 62P35; 65Z05

Journal ref: Commun Nonlinear Sci Numer Simul 84 (2020), 105205

Micromagnetics of rare-earth efficient permanent magnets

Authors: Johann Fischbacher, Alexander Kovacs, Markus Gusenbauer, Harald Oezelt, Lukas Exl, Simon Bance, Thomas Schrefl

Abstract: The development of permanent magnets containing less or no rare-earth elements is linked to profound knowledge of the coercivity mechanism. Prerequisites for a promising permanent magnet material are a high spontaneous magnetization and a sufficiently high magnetic anisotropy. In addition to the intrinsic magnetic properties the microstructure of the magnet plays a significant role in establishing… ▽ More The development of permanent magnets containing less or no rare-earth elements is linked to profound knowledge of the coercivity mechanism. Prerequisites for a promising permanent magnet material are a high spontaneous magnetization and a sufficiently high magnetic anisotropy. In addition to the intrinsic magnetic properties the microstructure of the magnet plays a significant role in establishing coercivity. The influence of the microstructure on coercivity, remanence, and energy density product can be understood by {using} micromagnetic simulations. With advances in computer hardware and numerical methods, hysteresis curves of magnets can be computed quickly so that the simulations can readily provide guidance for the development of permanent magnets. The potential of rare-earth reduced and free permanent magnets is investigated using micromagnetic simulations. The results show excellent hard magnetic properties can be achieved in grain boundary engineered NdFeB, rare-earth magnets with a ThMn12 structure, Co-based nano-wires, and L10-FeNi provided that the magnet's microstructure is optimized. △ Less

Submitted 28 March, 2019; originally announced March 2019.

Journal ref: J. Phys. D: Appl. Phys. 51 (2018) 193002

Learning magnetization dynamics

Authors: Alexander Kovacs, Johann Fischbacher, Harald Oezelt, Markus Gusenbauer, Lukas Exl, Florian Bruckner, Dieter Suess, Thomas Schrefl

Abstract: Deep neural networks are used to model the magnetization dynamics in magnetic thin film elements. The magnetic states of a thin film element can be represented in a low dimensional space. With convolutional autoencoders a compression ratio of 1024:1 was achieved. Time integration can be performed in the latent space with a second network which was trained by solutions of the Landau-Lifshitz-Gilber… ▽ More Deep neural networks are used to model the magnetization dynamics in magnetic thin film elements. The magnetic states of a thin film element can be represented in a low dimensional space. With convolutional autoencoders a compression ratio of 1024:1 was achieved. Time integration can be performed in the latent space with a second network which was trained by solutions of the Landau-Lifshitz-Gilbert equation. Thus the magnetic response to an external field can be computed quickly. △ Less

Submitted 22 March, 2019; originally announced March 2019.

Computational micromagnetics with Commics

Authors: Carl-Martin Pfeiler, Michele Ruggeri, Bernhard Stiftner, Lukas Exl, Matthias Hochsteger, Gino Hrkac, Joachim Schöberl, Norbert J. Mauser, Dirk Praetorius

Abstract: We present our open-source Python module Commics for the study of the magnetization dynamics in ferromagnetic materials via micromagnetic simulations. It implements state-of-the-art unconditionally convergent finite element methods for the numerical integration of the Landau-Lifshitz-Gilbert equation. The implementation is based on the multiphysics finite element software Netgen/NGSolve. The simul… ▽ More We present our open-source Python module Commics for the study of the magnetization dynamics in ferromagnetic materials via micromagnetic simulations. It implements state-of-the-art unconditionally convergent finite element methods for the numerical integration of the Landau-Lifshitz-Gilbert equation. The implementation is based on the multiphysics finite element software Netgen/NGSolve. The simulation scripts are written in Python, which leads to very readable code and direct access to extensive post-processing. Together with documentation and example scripts, the code is freely available on GitLab. △ Less

Submitted 14 December, 2018; originally announced December 2018.

Journal ref: Computer Physics Communications, 248 (2020), 106965

Magnetic microstructure machine learning analysis

Authors: Lukas Exl, Johann Fischbacher, Alexander Kovacs, Harald Oezelt, Markus Gusenbauer, Kazuya Yokota, Tetsuya Shoji, Gino Hrkac, Thomas Schrefl

Abstract: We use a machine learning approach to identify the importance of microstructure characteristics in causing magnetization reversal in ideally structured large-grained Nd$_2$Fe$_{14}$B permanent magnets. The embedded Stoner-Wohlfarth method is used as a reduced order model for determining local switching field maps which guide the data-driven learning procedure. The predictor model is a random fores… ▽ More We use a machine learning approach to identify the importance of microstructure characteristics in causing magnetization reversal in ideally structured large-grained Nd$_2$Fe$_{14}$B permanent magnets. The embedded Stoner-Wohlfarth method is used as a reduced order model for determining local switching field maps which guide the data-driven learning procedure. The predictor model is a random forest classifier which we validate by comparing with full micromagnetic simulations in the case of small granular test structures. In the course of the machine learning microstructure analysis the most important features explaining magnetization reversal were found to be the misorientation and the position of the grain within the magnet. The lowest switching fields occur near the top and bottom edges of the magnet. While the dependence of the local switching field on the grain orientation is known from theory, the influence of the position of the grain on the local coercive field strength is less obvious. As a direct result of our findings of the machine learning analysis we show that edge hardening via Dy-diffusion leads to higher coercive fields. △ Less

Submitted 11 August, 2018; originally announced August 2018.

Comments: 17 pages, 15 figures

Journal ref: J. Phys. Mater. 2 (2019) 014001

Many-body physics in two-component Bose-Einstein condensates in a cavity: fragmented superradiance and polarization

Authors: Axel U. J. Lode, Fritz S. Diorico, Rugway Wu, Paolo Molignini, Luca Papariello, Rui Lin, Camille Lévêque, Lukas Exl, Marios C. Tsatsos, R. Chitra, Norbert J. Mauser

Abstract: We consider laser-pumped one-dimensional two-component bosons in a parabolic trap embedded in a high-finesse optical cavity. Above a threshold pump power, the photons that populate the cavity modify the effective atom trap and mediate a coupling between the two components of the Bose-Einstein condensate. We calculate the ground state of the laser-pumped system and find different stages of self-org… ▽ More We consider laser-pumped one-dimensional two-component bosons in a parabolic trap embedded in a high-finesse optical cavity. Above a threshold pump power, the photons that populate the cavity modify the effective atom trap and mediate a coupling between the two components of the Bose-Einstein condensate. We calculate the ground state of the laser-pumped system and find different stages of self-organization depending on the power of the laser. The modified potential and the laser-mediated coupling between the atomic components give rise to rich many-body physics: an increase of the pump power triggers a self-organization of the atoms while an even larger pump power causes correlations between the self-organized atoms -- the BEC becomes fragmented and the reduced density matrix acquires multiple macroscopic eigenvalues. In this fragmented superradiant state, the atoms can no longer be described as two-level systems and the map** of the system to the Dicke model breaks down. △ Less

Submitted 29 January, 2018; originally announced January 2018.

Comments: 8 pages, 3 figures, software available at http://ultracold.org

Preconditioned nonlinear conjugate gradient method for micromagnetic energy minimization

Authors: Lukas Exl, Johann Fischbacher, Alexander Kovacs, Harald Oezelt, Markus Gusenbauer, Thomas Schrefl

Abstract: Fast computation of demagnetization curves is essential for the computational design of soft magnetic sensors or permanent magnet materials. We show that a sparse preconditioner for a nonlinear conjugate gradient energy minimizer can lead to a speed up by a factor of 3 and 7 for computing hysteresis in soft magnetic and hard magnetic materials, respectively. As a preconditioner an approximation of… ▽ More Fast computation of demagnetization curves is essential for the computational design of soft magnetic sensors or permanent magnet materials. We show that a sparse preconditioner for a nonlinear conjugate gradient energy minimizer can lead to a speed up by a factor of 3 and 7 for computing hysteresis in soft magnetic and hard magnetic materials, respectively. As a preconditioner an approximation of the Hessian of the Lagrangian is used, which only takes local field terms into account. Preconditioning requires a few additional sparse matrix vector multiplications per iteration of the nonlinear conjugate gradient method, which is used for minimizing the energy for a given external field. The time to solution for computing the demagnetization curve scales almost linearly with problem size. △ Less

Submitted 11 January, 2018; originally announced January 2018.

Comments: 18 pages, 3 figures

An optimization approach for dynamical Tucker tensor approximation

Authors: Lukas Exl

Abstract: An optimization-based approach for the Tucker tensor approximation of parameter-dependent data tensors and solutions of tensor differential equations with low Tucker rank is presented. The problem of updating the tensor decomposition is reformulated as fitting problem subject to the tangent space without relying on an orthogonality gauge condition. A discrete Euler scheme is established in an alte… ▽ More An optimization-based approach for the Tucker tensor approximation of parameter-dependent data tensors and solutions of tensor differential equations with low Tucker rank is presented. The problem of updating the tensor decomposition is reformulated as fitting problem subject to the tangent space without relying on an orthogonality gauge condition. A discrete Euler scheme is established in an alternating least squares framework, where the quadratic subproblems reduce to trace optimization problems, that are shown to be explicitly solvable and accessible using SVD of small size. In the presence of small singular values, instability for larger ranks is reduced, since the method does not need the (pseudo) inverse of matricizations of the core tensor. Regularization of Tikhonov type can be used to compensate for the lack of uniqueness in the tangent space. The method is validated numerically and shown to be stable also for larger ranks in the case of small singular values of the core unfoldings. Higher order explicit integrators of Runge-Kutta type can be composed. △ Less

Submitted 7 February, 2018; v1 submitted 28 September, 2017; originally announced September 2017.

Comments: 18 pages, 10 figures

MSC Class: 90C30; 15A69; 15A18

On the limits of coercivity in permanent magnets

Authors: J. Fischbacher, A. Kovacs, H. Oezelt, M. Gusenbauer, T. Schrefl, L. Exl, D. Givord, N. M. Dempsey, G. Zimanyi, M. Winklhofer, G. Hrkac, R. Chantrell, N. Sakuma, M. Yano, A. Kato, T. Shoji, A. Manabe

Abstract: The maximum coercivity that can be achieved for a given hard magnetic alloy is estimated by computing the energy barrier for the nucleation of a reversed domain in an idealized microstructure without any structural defects and without any soft magnetic secondary phases. For Sm$_{1-z}$Zr$_z$(Fe$_{1-y}$Co$_y$)$_{12-x}$Ti$_x$ based alloys, which are considered an alternative to Nd$_2$Fe$_{14}$B magne… ▽ More The maximum coercivity that can be achieved for a given hard magnetic alloy is estimated by computing the energy barrier for the nucleation of a reversed domain in an idealized microstructure without any structural defects and without any soft magnetic secondary phases. For Sm$_{1-z}$Zr$_z$(Fe$_{1-y}$Co$_y$)$_{12-x}$Ti$_x$ based alloys, which are considered an alternative to Nd$_2$Fe$_{14}$B magnets with lower rare-earth content, the coercive field of a small magnetic cube is reduced to 60 percent of the anisotropy field at room temperature and to 50 percent of the anisotropy field at elevated temperature (473K). This decrease of the coercive field is caused by misorientation, demagnetizing fields and thermal fluctuations. △ Less

Submitted 6 August, 2017; originally announced August 2017.

Comments: 5 pages, 4 figues

MSC Class: 82D40

Journal ref: Appl. Phys. Lett. 111, 072404 (2017)

The extrapolated explicit midpoint scheme for variable order and step size controlled integration of the Landau-Lifschitz-Gilbert equation

Authors: Lukas Exl, Norbert J. Mauser, Thomas Schrefl, Dieter Suess

Abstract: A practical and efficient scheme for the higher order integration of the Landau-Lifschitz-Gilbert (LLG) equation is presented. The method is based on extrapolation of the two-step explicit midpoint rule and incorporates adaptive time step and order selection. We make use of a piecewise time-linear stray field approximation to reduce the necessary work per time step. The approximation to the interp… ▽ More A practical and efficient scheme for the higher order integration of the Landau-Lifschitz-Gilbert (LLG) equation is presented. The method is based on extrapolation of the two-step explicit midpoint rule and incorporates adaptive time step and order selection. We make use of a piecewise time-linear stray field approximation to reduce the necessary work per time step. The approximation to the interpolated operator is embedded into the extrapolation process to keep in step with the hierarchic order structure of the scheme. We verify the approach by means of numerical experiments on a standardized NIST problem and compare with a higher order embedded Runge-Kutta formula. The efficiency of the presented approach increases when the stray field computation takes a larger portion of the costs for the effective field evaluation. △ Less

Submitted 7 March, 2017; originally announced March 2017.

Journal ref: J. Comput. Phys. 346 (2017) 14-24

Conjugate gradient methods in micromagnetics

Authors: J. Fischbacher, A. Kovacs, H. Oezelt, T. Schrefl, L. Exl, J. Fidler, D. Suess, N. Sakuma, M. Yano, A. Kato, T. Shoji, A. Manabe

Abstract: Conjugate gradient methods for energy minimization in micromagnetics are compared. When the step length in the line search is controlled, conjugate gradient techniques are a fast and reliable way to compute the hysteresis properties of permanent magnets. The method is applied to investigate demagnetizing effects in NdFe12 based permanent magnets. The reduction of the coercive field by demagnetizin… ▽ More Conjugate gradient methods for energy minimization in micromagnetics are compared. When the step length in the line search is controlled, conjugate gradient techniques are a fast and reliable way to compute the hysteresis properties of permanent magnets. The method is applied to investigate demagnetizing effects in NdFe12 based permanent magnets. The reduction of the coercive field by demagnetizing effects is 1.4 T at 450 K. △ Less

Submitted 20 January, 2017; originally announced January 2017.

A GPU accelerated and error-controlled solver for the unbounded Poisson equation in three dimensions

Authors: Lukas Exl

Abstract: An efficient solver for the three dimensional free-space Poisson equation is presented. The underlying numerical method is based on finite Fourier series approximation. While the error of all involved approximations can be fully controlled, the overall computation error is driven by the convergence of the finite Fourier series of the density. For smooth and fast-decaying densities the proposed met… ▽ More An efficient solver for the three dimensional free-space Poisson equation is presented. The underlying numerical method is based on finite Fourier series approximation. While the error of all involved approximations can be fully controlled, the overall computation error is driven by the convergence of the finite Fourier series of the density. For smooth and fast-decaying densities the proposed method will be spectral accurate. The method scales with $\mathcal{O}(N\log N)$ operations, where $N$ is the total number of discretization points in the Cartesian grid. The majority of the computational costs come from fast Fourier transforms (FFT), which makes it ideal for GPU computation. Several numerical computations on CPU and GPU validate the method and show efficiency and convergence behavior. Tests are performed using the Vienna Scientific Cluster 3 (VSC3). A free MATLAB implementation for CPU and GPU is provided to the interested community. △ Less

Submitted 6 December, 2016; originally announced December 2016.

Journal ref: Comput Phys Commun 221 (2017) 352-357

A magnetostatic energy formula arising from the $L^2$-orthogonal decomposition of the stray field

Authors: Lukas Exl

Abstract: A formula for the magnetostatic energy of a finite magnet is proven. In contrast to common approaches, the new energy identity does not rely on evaluation of a nonlocal boundary integral inside the magnet or the solution of an equivalent Dirichlet problem. The formula is therefore computationally efficient, which is also shown numerically. Algorithms for the simulation of magnetic materials could… ▽ More A formula for the magnetostatic energy of a finite magnet is proven. In contrast to common approaches, the new energy identity does not rely on evaluation of a nonlocal boundary integral inside the magnet or the solution of an equivalent Dirichlet problem. The formula is therefore computationally efficient, which is also shown numerically. Algorithms for the simulation of magnetic materials could benefit from incorporating the presented representation of the energy. In addition, a natural analogue for the energy via the magnetic induction is given. Proofs are carried out within a setting which is suitable for common discretizations in computational micromagnetics. △ Less

Submitted 23 July, 2018; v1 submitted 7 November, 2016; originally announced November 2016.

MSC Class: 65N99; 82D40

Journal ref: J. Math. Anal. Appl. 467 (2018) 230-237

Highly Parallel Demagnetization Field Calculation Using the Fast Multipole Method on Tetrahedral Meshes with Continuous Sources

Authors: Pietro Palmesi, Lukas Exl, Florian Bruckner, Claas Abert, Dieter Suess

Abstract: The long-range magnetic field is the most time-consuming part in micromagnetic simulations. Improvements both on a numerical and computational basis can relief problems related to this bottleneck. This work presents an efficient implementation of the Fast Multipole Method [FMM] for the magnetic scalar potential as used in micromagnetics. We assume linearly magnetized tetrahedral sources, treat the… ▽ More The long-range magnetic field is the most time-consuming part in micromagnetic simulations. Improvements both on a numerical and computational basis can relief problems related to this bottleneck. This work presents an efficient implementation of the Fast Multipole Method [FMM] for the magnetic scalar potential as used in micromagnetics. We assume linearly magnetized tetrahedral sources, treat the near field directly and use analytical integration on the multipole expansion in the far field. This approach tackles important issues like the vectorial and continuous nature of the magnetic field. By using FMM the calculations scale linearly in time and memory. △ Less

Submitted 1 June, 2016; originally announced June 2016.

Grain-size dependent demagnetizing factors in permanent magnets

Authors: Simon Bance, Bernhard Seebacher, Thomas Schrefl, Lukas Exl, Michael Winklhofer, Gino Hrkac, Gergely Zimanyi, Tetsuya Shoji, Masao Yano, Noritsugu Sakuma, M Ito, Akira Kato, Akira Manabe

Abstract: The coercive field of permanent magnets decreases with increasing grain size. The grain size dependence of coercivity is explained by a size dependent demagnetizing factor. In Dy free Nd$_2$Fe$_{14}$B magnets the size dependent demagnetizing factor ranges from 0.2 for a grain size of 55 nm to 1.22 for a grain size of 8300 nm. The comparison of experimental data with micromagnetic simulations sugge… ▽ More The coercive field of permanent magnets decreases with increasing grain size. The grain size dependence of coercivity is explained by a size dependent demagnetizing factor. In Dy free Nd$_2$Fe$_{14}$B magnets the size dependent demagnetizing factor ranges from 0.2 for a grain size of 55 nm to 1.22 for a grain size of 8300 nm. The comparison of experimental data with micromagnetic simulations suggests that the grain size dependence of the coercive field in hard magnets is due to the non-uniform magnetostatic field in polyhedral grains. △ Less

Submitted 27 March, 2016; originally announced March 2016.

MSC Class: 82D40

Journal ref: J. App. Phys. vol. 116 no. 23 pp. 233903 (2014)

Numerical optimization of writer geometries for bit patterned magnetic recording

Authors: A. Kovacs, H. Oezelt, S. Bance, J. Fischbacher, M. Gusenbauer, F. Reichel, L. Exl, T. Schrefl, M. E. Schabes

Abstract: A fully-automated pole-tip shape optimization tool, involving write head geometry construction, meshing, micromagnetic simulation and evaluation, is presented. Optimizations have been performed for three different writing schemes (centered, staggered and shingled) for an underlying bit patterned media with an areal density of 2.12 Tdots/in$^2$ . Optimizations were performed for a single-phase medi… ▽ More A fully-automated pole-tip shape optimization tool, involving write head geometry construction, meshing, micromagnetic simulation and evaluation, is presented. Optimizations have been performed for three different writing schemes (centered, staggered and shingled) for an underlying bit patterned media with an areal density of 2.12 Tdots/in$^2$ . Optimizations were performed for a single-phase media with 10 nm thickness and a mag spacing of 8 nm. From the computed write field and its gradient and the minimum energy barrier during writing for islands on the adjacent track, the overall write error rate is computed. The overall write errors are 0.7, 0.08, and 2.8 x 10$^{-5}$ for centered writing, staggered writing, and shingled writing. △ Less

Submitted 14 April, 2015; originally announced April 2015.

Journal ref: J. Appl. Phys. 115, 17B704 (2014)

Accurate and efficient computation of nonlocal potentials based on Gaussian-sum approximation

Authors: Lukas Exl, Norbert J. Mauser, Yong Zhang

Abstract: We introduce an accurate and efficient method for a class of nonlocal potential evaluations with free boundary condition, including the 3D/2D Coulomb, 2D Poisson and 3D dipolar potentials. Our method is based on a Gaussian-sum approximation of the singular convolution kernel and Taylor expansion of the density. Starting from the convolution formulation, for smooth and fast decaying densities, we m… ▽ More We introduce an accurate and efficient method for a class of nonlocal potential evaluations with free boundary condition, including the 3D/2D Coulomb, 2D Poisson and 3D dipolar potentials. Our method is based on a Gaussian-sum approximation of the singular convolution kernel and Taylor expansion of the density. Starting from the convolution formulation, for smooth and fast decaying densities, we make a full use of the Fourier pseudospectral (plane wave) approximation of the density and a separable Gaussian-sum approximation of the kernel in an interval where the singularity (the origin) is excluded. Hence, the potential is separated into a regular integral and a near-field singular correction integral, where the first integral is computed with the Fourier pseudospectral method and the latter singular one can be well resolved utilizing a low-order Taylor expansion of the density. Both evaluations can be accelerated by fast Fourier transforms (FFT). The new method is accurate (14-16 digits), efficient ($O(N \log N)$ complexity), low in storage, easily adaptable to other different kernels, applicable for anisotropic densities and highly parallelable. △ Less

Submitted 19 January, 2015; originally announced January 2015.

MSC Class: 33F05; 44A35; 65E05; 65R10; 65T50

Journal ref: J. Comput. Phys. 327 (2016) 629-642

LaBonte's method revisited: An effective steepest descent method for micromagnetic energy minimization

Authors: Lukas Exl, Simon Bance, Franz Reichel, Thomas Schrefl, Hans Peter Stimming, Norbert J. Mauser

Abstract: We present a steepest descent energy minimization scheme for micromagnetics. The method searches on a curve that lies on the sphere which keeps the magnitude of the magnetization vector constant. The step size is selected according to a modified Barzilai-Borwein method. Standard linear tetrahedral finite elements are used for space discretization. For the computation of static hysteresis loops the… ▽ More We present a steepest descent energy minimization scheme for micromagnetics. The method searches on a curve that lies on the sphere which keeps the magnitude of the magnetization vector constant. The step size is selected according to a modified Barzilai-Borwein method. Standard linear tetrahedral finite elements are used for space discretization. For the computation of static hysteresis loops the steepest descent minimizer is faster than a Landau-Lifshitz micromagnetic solver by more than a factor of two. The speed up on a graphic processor is 4.8 as compared to the fastest single-core CPU implementation. △ Less

Submitted 23 September, 2013; originally announced September 2013.

Journal ref: J. Appl. Phys. 115, 17D118 (2014)

Guided self-assembly of magnetic beads for biomedical applications

Authors: Markus Gusenbauer, Ha Nguyen, Franz Reichel, Lukas Exl, Simon Bance, Johann Fischbacher, Harald Özelt, Alexander Kovacs, Martin Brandl, Thomas Schrefl

Abstract: Micromagnetic beads are widely used in biomedical applications for cell separation, drug delivery, and hypothermia cancer treatment. Here we propose to use self-organized magnetic bead structures which accumulate on fixed magnetic seeding points to isolate circulating tumor cells. The analysis of circulating tumor cells is an emerging tool for cancer biology research and clinical cancer management… ▽ More Micromagnetic beads are widely used in biomedical applications for cell separation, drug delivery, and hypothermia cancer treatment. Here we propose to use self-organized magnetic bead structures which accumulate on fixed magnetic seeding points to isolate circulating tumor cells. The analysis of circulating tumor cells is an emerging tool for cancer biology research and clinical cancer management including the detection, diagnosis and monitoring of cancer. Microfluidic chips for isolating circulating tumor cells use either affinity, size or density capturing methods. We combine multiphysics simulation techniques to understand the microscopic behavior of magnetic beads interacting with Nickel accumulation points used in lab-on-chip technologies. Our proposed chip technology offers the possibility to combine affinity and size capturing with special antibody-coated bead arrangements using a magnetic gradient field created by Neodymium Iron Boron permanent magnets. The multiscale simulation environment combines magnetic field computation, fluid dynamics and discrete particle dynamics. △ Less

Submitted 30 May, 2013; originally announced May 2013.

Simulation of magnetic active polymers for versatile microfluidic devices

Authors: Markus Gusenbauer, Harald Özelt, Johann Fischbacher, Franz Reichel, Lukas Exl, Simon Bance, Nadezhda Kataeva, Claudia Binder, Hubert Brückl, Thomas Schrefl

Abstract: We propose to use a compound of magnetic nanoparticles (20-100 nm) embedded in a flexible polymer (Polydimethylsiloxane PDMS) to filter circulating tumor cells (CTCs). The analysis of CTCs is an emerging tool for cancer biology research and clinical cancer management including the detection, diagnosis and monitoring of cancer. The combination of experiments and simulations lead to a versatile micr… ▽ More We propose to use a compound of magnetic nanoparticles (20-100 nm) embedded in a flexible polymer (Polydimethylsiloxane PDMS) to filter circulating tumor cells (CTCs). The analysis of CTCs is an emerging tool for cancer biology research and clinical cancer management including the detection, diagnosis and monitoring of cancer. The combination of experiments and simulations lead to a versatile microfluidic lab-on-chip device. Simulations are essential to understand the influence of the embedded nanoparticles in the elastic PDMS when applying a magnetic gradient field. It combines finite element calculations of the polymer, magnetic simulations of the embedded nanoparticles and the fluid dynamic calculations of blood plasma and blood cells. With the use of magnetic active polymers a wide range of tunable microfluidic structures can be created. The method can help to increase the yield of needed isolated CTCs. △ Less

Submitted 30 May, 2013; originally announced May 2013.

Journal ref: EPJ Web of Conferences. Vol. 40. EDP Sciences, 2013

Non-uniform FFT for the finite element computation of the micromagnetic scalar potential

Authors: Lukas Exl, Thomas Schrefl

Abstract: We present a quasi-linearly scaling, first order polynomial finite element method for the solution of the magnetostatic open boundary problem by splitting the magnetic scalar potential. The potential is determined by solving a Dirichlet problem and evaluation of the single layer potential by a fast approximation technique based on Fourier approximation of the kernel function. The latter approximat… ▽ More We present a quasi-linearly scaling, first order polynomial finite element method for the solution of the magnetostatic open boundary problem by splitting the magnetic scalar potential. The potential is determined by solving a Dirichlet problem and evaluation of the single layer potential by a fast approximation technique based on Fourier approximation of the kernel function. The latter approximation leads to a generalization of the well-known convolution theorem used in finite difference methods. We address it by a non-uniform FFT approach. Overall, our method scales O(M + N + N log N) for N nodes and M surface triangles. We confirm our approach by several numerical tests. △ Less

Submitted 3 March, 2014; v1 submitted 14 May, 2013; originally announced May 2013.

Journal ref: J. Comput. Phys. 270C (2014) 490-505

magnum.fe: A micromagnetic finite-element simulation code based on FEniCS

Authors: Claas Abert, Lukas Exl, Florian Bruckner, André Drews, Dieter Suess

Abstract: We have developed a finite-element micromagnetic simulation code based on the FEniCS package called magnum.fe. Here we describe the numerical methods that are applied as well as their implementation with FEniCS. We apply a transformation method for the solution of the demagnetization-field problem. A semi-implicit weak formulation is used for the integration of the Landau-Lifshitz-Gilbert equation… ▽ More We have developed a finite-element micromagnetic simulation code based on the FEniCS package called magnum.fe. Here we describe the numerical methods that are applied as well as their implementation with FEniCS. We apply a transformation method for the solution of the demagnetization-field problem. A semi-implicit weak formulation is used for the integration of the Landau-Lifshitz-Gilbert equation. Numerical experiments show the validity of simulation results. magnum.fe is open source and well documented. The broad feature range of the FEniCS package makes magnum.fe a good choice for the implementation of novel micromagnetic finite-element algorithms. △ Less

Submitted 15 January, 2013; v1 submitted 14 January, 2013; originally announced January 2013.

FFT-based Kronecker product approximation to micromagnetic long-range interactions

Authors: Lukas Exl, Claas Abert, Norbert J. Mauser, Thomas Schrefl, Hans Peter Stimming, Dieter Suess

Abstract: We derive a Kronecker product approximation for the micromagnetic long range interactions in a collocation framework by means of separable sinc quadrature. Evaluation of this operator for structured tensors (Canonical format, Tucker format, Tensor Trains) scales below linear in the volume size. Based on efficient usage of FFT for structured tensors, we are able to accelerate computations to quasi… ▽ More We derive a Kronecker product approximation for the micromagnetic long range interactions in a collocation framework by means of separable sinc quadrature. Evaluation of this operator for structured tensors (Canonical format, Tucker format, Tensor Trains) scales below linear in the volume size. Based on efficient usage of FFT for structured tensors, we are able to accelerate computations to quasi linear complexity in the number of collocation points used in one dimension. Quadratic convergence of the underlying collocation scheme as well as exponential convergence in the separation rank of the approximations is proved. Numerical experiments on accuracy and complexity confirm the theoretical results. △ Less

Submitted 14 December, 2012; originally announced December 2012.

Comments: 4 figures

MSC Class: 65N99; 82D40

Journal ref: Math. Models Methods Appl. Sci. 24 (2014) 1877-1901

Numerical Methods for the Stray-Field Calculation: A Comparison of recently developed Algorithms

Authors: Claas Abert, Lukas Exl, Gunnar Selke, André Drews, Thomas Schrefl

Abstract: Different numerical approaches for the stray-field calculation in the context of micromagnetic simulations are investigated. We compare finite difference based fast Fourier transform methods, tensor grid methods and the finite-element method with shell transformation in terms of computational complexity, storage requirements and accuracy tested on several benchmark problems. These methods can be s… ▽ More Different numerical approaches for the stray-field calculation in the context of micromagnetic simulations are investigated. We compare finite difference based fast Fourier transform methods, tensor grid methods and the finite-element method with shell transformation in terms of computational complexity, storage requirements and accuracy tested on several benchmark problems. These methods can be subdivided into integral methods (fast Fourier transform methods, tensor-grid method) which solve the stray field directly and in differential equation methods (finite-element method), which compute the stray field as the solution of a partial differential equation. It turns out that for cuboid structures the integral methods, which work on cuboid grids (fast Fourier transform methods and tensor grid methods) outperform the finite-element method in terms of the ratio of computational effort to accuracy. Among these three methods the tensor grid method is the fastest. However, the use of the tensor grid method in the context of full micromagnetic codes is not well investigated yet. The finite-element method performs best for computations on curved structures. △ Less

Submitted 19 April, 2012; originally announced April 2012.

A tunable cancer cell filter using magnetic beads: cellular and fluid dynamic simulations

Authors: Markus Gusenbauer, Ivan Cimrak, Simon Bance, Lukas Exl, Franz Reichel, Harald Oezelt, Thomas Schrefl

Abstract: In the field of biomedicine magnetic beads are used for drug delivery and to treat hyperthermia. Here we propose to use self-organized bead structures to isolate circulating tumor cells using lab-on-chip technologies. Typically blood flows past microposts functionalized with antibodies for circulating tumor cells. Creating these microposts with interacting magnetic beads makes it possible to tune… ▽ More In the field of biomedicine magnetic beads are used for drug delivery and to treat hyperthermia. Here we propose to use self-organized bead structures to isolate circulating tumor cells using lab-on-chip technologies. Typically blood flows past microposts functionalized with antibodies for circulating tumor cells. Creating these microposts with interacting magnetic beads makes it possible to tune the geometry in size, position and shape. We develop a simulation tool that combines micromagnetics, discrete particle dynamics and fluid dynamics, in order to design micropost arrays made of interacting beads. For the simulation of blood flow we use the Lattice-Boltzmann method with immersed elastic blood cell models. Parallelization distributes large fluid and particle dynamic simulations over available resources to reduce overall calculation time. △ Less

Submitted 5 October, 2011; originally announced October 2011.

Self-organizing magnetic beads for biomedical applications

Authors: Markus Gusenbauer, Alexander Kovacs, Franz Reichel, Lukas Exl, Simon Bance, Harald Ozelt, Thomas Schrefl

Abstract: In the field of biomedicine magnetic beads are used for drug delivery and to treat hyperthermia. Here we propose to use self-organized bead structures to isolate circulating tumor cells using lab-on-chip technologies. Typically blood flows past microposts functionalized with antibodies for circulating tumor cells. Creating these microposts with interacting magnetic beads makes it possible to tune… ▽ More In the field of biomedicine magnetic beads are used for drug delivery and to treat hyperthermia. Here we propose to use self-organized bead structures to isolate circulating tumor cells using lab-on-chip technologies. Typically blood flows past microposts functionalized with antibodies for circulating tumor cells. Creating these microposts with interacting magnetic beads makes it possible to tune the geometry in size, position and shape. We developed a simulation tool that combines micromagnetics and discrete particle dynamics, in order to design micropost arrays made of interacting beads. The simulation takes into account the viscous drag of the blood flow, magnetostatic interactions between the magnetic beads and gradient forces from external aligned magnets. We developed a particle-particle particle-mesh method for effective computation of the magnetic force and torque acting on the particles. △ Less

Submitted 5 October, 2011; originally announced October 2011.

Journal ref: Journal of Magnetism and Magnetic Materials 324.6 (2012): 977-982

Fast stray field computation on tensor grids

Authors: Lukas Exl, Winfried Auzinger, Simon Bance, Markus Gusenbauer, Franz Reichel, Thomas Schrefl

Abstract: A direct integration algorithm is described to compute the magnetostatic field and energy for given magnetization distributions on not necessarily uniform tensor grids. We use an analytically-based tensor approximation approach for function-related tensors, which reduces calculations to multilinear algebra operations. The algorithm scales with N^(4/3) for N computational cells used and with N^(2/3… ▽ More A direct integration algorithm is described to compute the magnetostatic field and energy for given magnetization distributions on not necessarily uniform tensor grids. We use an analytically-based tensor approximation approach for function-related tensors, which reduces calculations to multilinear algebra operations. The algorithm scales with N^(4/3) for N computational cells used and with N^(2/3) (sublinear) when magnetization is given in canonical tensor format. In the final section we confirm our theoretical results concerning computing times and accuracy by means of numerical examples. △ Less

Submitted 30 September, 2011; v1 submitted 27 September, 2011; originally announced September 2011.

Comments: 16 pages, 1 figure, submitted to the Journal of Computational Physics

Journal ref: J. Comput. Phys. 231 (2012) 2840-2850