arXiv subjects

Stefan Kollmannsberger

Publications and source records attributed to Stefan Kollmannsberger.

At least 19 recordsLinked to original sources

arXiv2026

Machine Learning Inference Limits of Routine Cement Characterization for CEM I Performance: Evidence From a Multi-Producer Dataset

Routine cement performance characterization provides continuous quality control data, but its information content for performance inference and transferability across independent producers remains uncertain. This study analyzes 476 cement records from 23 European producers, collected in one laboratory over 27 years, to determine what can be inferred from routine measurements. The analysis focuses on CEM I and combines oxide chemistry, Blaine fineness, particle-size distribution descriptors, physical properties, and derived Bogue and equivalent-alkali descriptors with machine learning attribution and producer-transfer tests. For CEM I, fineness is the strongest descriptor family for strength class and water demand, but oxide chemistry contributes a comparable signal when evaluated jointly. Blaine and compact particle-size distribution representations are largely interchangeable within the descriptor space, indicating that the dominant recoverable fineness information is captured by routine measurements. Equivalent alkali shows a consistent negative association with 28-day strength, through K$_2$O in this dataset. Strength class and water demand can be recovered from routine cement characterization data. The early-strength designation is recovered only as a population-level tendency, not a physically separable class, because early-strength development can arise from combinations of fineness, sulfate--alkali chemistry, phase assemblage, and plant practice. Producer-holdout tests show that absolute prediction errors remain comparable across the held-out producers in this dataset, whereas recovery of within-producer strength variation is producer-dependent. Routine CEM I characterization therefore supports useful performance inference across producers, while exposing producer-specific variation whose recovery may require additional speciation.

cond-mat.mtrl-sci↗

arXiv2026

Lightweight return-mapping surrogates for multiscale plasticity: a practical guide

This paper presents a practical guide to building lightweight neural-network surrogates for the plastic return-mapping process in concurrent multiscale (FE2) simulations. Rather than proposing a new architecture, we show how a deliberately simple feed-forward network, structured to mirror the classical return-mapping update, can replace the prohibitively expensive nested fine-scale solves that dominate the cost of conventional FE2 schemes based on FFT homogenization at the meso-scale. We walk through the full workflow: generating training data from incremental homogenization analyses, constructing a compact yet sufficient dataset, embedding material symmetries directly into the mapping, and deploying the trained network as a user-defined material subroutine (UMAT) in a standard finite-element solver -- enabling widespread use. A sensitivity study examines the model's robustness to data density, increment size, and mesh refinement, and we characterize the regimes in which the surrogate holds and where it breaks down. For the macroscopically isotropic, two-dimensional plane-stress setting considered here, the surrogate reproduces the reference response while reducing the per-analysis cost from hours to seconds with speed-ups up to 30,000 over standard FE2. The approach extends naturally to three dimensions and to weaker symmetry assumptions, given an appropriate sampling strategy and dataset.

math.NA↗

arXiv2026

Reconstructing effective ultrasound transducer models via distributed source inversion

Accurate modeling of ultrasound wave propagation is essential for high-fidelity simulation and imaging in ultrasonic testing. A primary challenge lies in characterizing the excitation source, particularly for transducers with large apertures relative to the acoustic wavelengths. In such cases, non-uniform excitation and spatial interference significantly affect the resulting radiation patterns. This paper proposes a distributed source inversion strategy to reconstruct an effective spatio-temporal transducer model that reproduces experimentally measured wavefields. The reconstructed source model captures aperture-dependent phase and amplitude variations without the need for detailed knowledge of the transducer structure. The approach is validated using directivity measurements on an aluminum half-cylinder, where simulations incorporating the reconstructed source model show close agreement with experimental directivity patterns and waveform shapes. Finally, synthetic studies on reverse time migration and full-waveform inversion demonstrate that accurate transducer modeling is critical for the success of simulation-based imaging and inversion workflows and significantly improves reconstruction quality.

physics.med-ph↗

arXiv2025

A Memory Efficient Adjoint Method to Enable Billion Parameter Optimization on a Single GPU in Dynamic Problems

Dynamic optimization is currently limited by sensitivity computations that require information from full forward and adjoint wave fields. Since the forward and adjoint solutions are computed in opposing time directions, the forward solution must be stored. This requires a substantial amount of memory for large-scale problems even when using check pointing or data compression techniques. As a result, the problem size is memory bound rather than bound by wall clock time, when working with modern GPU-based implementations that have limited memory capacity. To overcome this limitation, we introduce a new approach for approximate sensitivity computation based on the adjoint method (for self-adjoint problems) that relies on the principle of superposition. The approximation allows an iterative computation of the sensitivity, reducing the memory burden to that of the solution at a small number of time steps, i.e., to the number of degrees of freedom. This enables sensitivity computations for problems with billions of degrees of freedom on current GPUs, such as the A100 from NVIDIA (from 2020). We demonstrate the approach on full waveform inversion and transient acoustic topology optimization problems, relying on a highly efficient finite difference forward solver implemented in CUDA. Phenomena such as damping cannot be considered, as the approximation technique is limited to self-adjoint problems.

cs.CE↗

arXiv2025

Generalized eigenvalue stabilization for immersed explicit dynamics

Explicit time integration for immersed finite element discretizations severely suffers from the influence of poorly cut elements. In this contribution, we propose a generalized eigenvalue stabilization (GEVS) strategy for the element mass matrices of cut elements to cure their adverse impact on the critical time step size of the global system. We use spectral basis functions, specifically $C^0$ continuous Lagrangian interpolation polynomials defined on Gauss-Lobatto-Legendre (GLL) points, which, in combination with its associated GLL quadrature rule, yield high-order convergent diagonal mass matrices for uncut elements. Moreover, considering cut elements, we combine the proposed GEVS approach with the finite cell method (FCM) to guarantee definiteness of the system matrices. However, the proposed GEVS stabilization can directly be applied to other immersed boundary finite element methods. Numerical experiments demonstrate that the stabilization strategy achieves optimal convergence rates and recovers critical time step sizes of equivalent boundary-conforming discretizations. This also holds in the presence of weakly enforced Dirichlet boundary conditions using either Nitsche's method or penalty formulations.

cs.CE↗

arXiv2025

On the efficiency of explicit and semi-explicit immersed boundary finite element methods for wave propagation problems

Immersed boundary methods have attracted substantial interest in the last decades due to their potential for computations involving complex geometries. Often these cannot be efficiently discretized using boundary-fitted finite elements. Immersed boundary methods provide a simple and fully automatic discretization based on Cartesian grids and tailored quadrature schemes that account for the geometric model. It can thus be described independently of the grid, e.g., by image data obtained from computed tomography scans. The drawback of such a discretization lies in the potentially small overlap between certain elements in the grid and the geometry. These badly cut elements with small physical support pose a particular challenge for nonlinear and/or dynamic simulations. In this work, we focus on problems in structural dynamics and acoustics and concentrate on solving them with explicit time-marching schemes. In this context, badly cut elements can lead to unfeasibly small critical time step sizes. We investigate the performance of implicit-explicit time marching schemes and two stabilization methods developed in previous works as potential remedies. While these have been studied before with regard to their effectiveness in increasing the critical time step size, their numerical efficiency has only been considered in terms of accuracy per degree of freedom. In this paper, we evaluate the computation time required for a given accuracy, which depends not only on the number of degrees of freedom but also on the selected spatial discretization, the sparsity patterns of the system matrices, and the employed time-marching scheme.

cs.CE↗

arXiv2025

A plastic damage model with mixed isotropic-kinematic hardening for low-cycle fatigue in 7020 aluminum

The paper at hand presents an in-depth investigation into the fatigue behavior of the high-strength aluminum alloy EN AW-7020 T6 using both experimental and numerical approaches. Two types of specimens are investigated: a dog-bone specimen subjected to cyclic loading in a symmetric strain-controlled regime, and a compact tension specimen subjected to repeated loading and unloading, which leads to damage growth from the notch tip. Experimental data from these tests are used to identify the different phases of fatigue. Subsequently, a plastic-damage model is developed, incorporating J2 plasticity with Chaboche-type mixed isotropic-kinematic hardening. A detailed investigation reveals that the Chaboche model must be blended with a suitable isotropic hardening and combined with a proper damage growth model to accurately describe cyclic fatigue including large plastic strains up to failure. Multiple back-stress components with independent properties are superimposed, and exponential isotropic hardening with saturation effects is introduced to improve alignment with experimental results. For damage, different stress splits are tested, with the deviatoric/volumetric split proving successful in reproducing the desired degradation in peak stress and stiffness. A nonlinear activation function is introduced to ensure smooth transitions between tension and compression. Two damage indices, one for the deviatoric part and one for the volumetric part, are defined, each of which is governed by a distinct trilinear damage growth function. The governing differential equation of the problem is regularized by higher-order gradient terms to address the ill-posedness induced by softening. Finally, the plasticity model is calibrated using finite element simulations of the dog-bone test and subsequently applied to the cyclic loading of the compact tension specimen.

math.NA↗

arXiv2025

Full-waveform inversion via the scaled boundary finite element method

We begin by addressing the time-domain full-waveform inversion using the adjoint method. Next, we derive the scaled boundary semi-weak form of the scalar wave equation in heterogeneous media through the Galerkin method. Unlike conventional formulations, the resulting system incorporates variable density and two additional terms involving its spatial derivative. As a result, the coefficient matrices are no longer constant and depend on the radial coordinate, rendering the common solution methods inapplicable. Thus, we introduce a radial discretization scheme within the framework of the scaled boundary finite element method. We employ finite difference approximation, yet the choice underlying our ansatz is made for demonstration purposes and remains flexible. Next, we introduce an algorithmic condensation procedure to compute the dynamic stiffness matrices on the fly. Therefore, we maneuver around the need to introduce auxiliary unknowns. As a result, the optimization problem is structured in a two-level hierarchy. We obtain the Fréchet kernel by computing the zero-lag cross-correlations of the forward and adjoint wavefields, and solve the minimization problems iteratively by moving downhill on the cost function hypersurface through the limited-memory BFGS algorithm. The numerical results demonstrate the effectiveness and robustness of the new formulation and show that using the simplified differential equation along with the conventional formulation is highly inferior to applying the complete form of the differential equation. This approach effectively decomposes the computational load into independent local problems and a single coupled global system, making the solution method highly parallelizable. We demonstrate that, with a simple OpenMP implementation using 12 threads on a personal laptop, the new formulation outperforms the existing approach in terms of computation time.

math.NA↗

arXiv2024

Quantitative Comparison of the Total Focusing Method, Reverse Time Migration, and Full Waveform Inversion for Ultrasonic Imaging

Phased array ultrasound is a widely used technique in non-destructive testing. Using piezoelectric elements as both sources and receivers provides a significant gain in information and enables more accurate defect detection. When all source-receiver combinations are used, the process is called full matrix capture. The total focusing method~(TFM), which exploits such datasets, relies on a delay and sum algorithm to sum up the signals on a pixel grid. However, TFM only uses the first arriving p-waves, making it challenging to size complex-shaped defects. By contrast, more advanced methods such as reverse time migration~(RTM) and full waveform inversion~(FWI) use full waveforms to reconstruct defects. Both methods compare measured signals with ultrasound simulations. While RTM identifies defects by convolving forward and backward wavefields once, FWI iteratively updates material models to reconstruct the actual distribution of material properties. This study compares TFM, RTM, and FWI for six specimens featuring circular defects or Y-shaped notches. The reconstructed results are first evaluated qualitatively using different thresholds and then quantitatively using metrics such as AUPRC, AUROC, and F1-score. The results show that FWI performs best in most cases, both qualitatively and quantitatively.

cs.CE↗

arXiv2024

Accelerating Full Waveform Inversion By Transfer Learning

Full waveform inversion (FWI) is a powerful tool for reconstructing material fields based on sparsely measured data obtained by wave propagation. For specific problems, discretizing the material field with a neural network (NN) improves the robustness and reconstruction quality of the corresponding optimization problem. We call this method NN-based FWI. Starting from an initial guess, the weights of the NN are iteratively updated to fit the simulated wave signals to the sparsely measured data set. For gradient-based optimization, a suitable choice of the initial guess, i.e., a suitable NN weight initialization, is crucial for fast and robust convergence. In this paper, we introduce a novel transfer learning approach to further improve NN-based FWI. This approach leverages supervised pretraining to provide a better NN weight initialization, leading to faster convergence of the subsequent optimization problem. Moreover, the inversions yield physically more meaningful local minima. The network is pretrained to predict the unknown material field using the gradient information from the first iteration of conventional FWI. In our computational experiments on two-dimensional domains, the training data set consists of reference simulations with arbitrarily positioned elliptical voids of different shapes and orientations. We compare the performance of the proposed transfer learning NN-based FWI with three other methods: conventional FWI, NN-based FWI without pretraining and conventional FWI with an initial guess predicted from the pretrained NN. Our results show that transfer learning NN-based FWI outperforms the other methods in terms of convergence speed and reconstruction quality.

cs.LG↗

arXiv2024

Neural Networks for Generating Better Local Optima in Topology Optimization

Neural networks have recently been employed as material discretizations within adjoint optimization frameworks for inverse problems and topology optimization. While advantageous regularization effects and better optima have been found for some inverse problems, the benefit for topology optimization has been limited -- where the focus of investigations has been the compliance problem. We demonstrate how neural network material discretizations can, under certain conditions, find better local optima in more challenging optimization problems, where we here specifically consider acoustic topology optimization. The chances of identifying a better optimum can significantly be improved by running multiple partial optimizations with different neural network initializations. Furthermore, we show that the neural network material discretization's advantage comes from the interplay with the Adam optimizer and emphasize its current limitations when competing with constrained and higher-order optimization techniques. At the moment, this discretization has only been shown to be beneficial for unconstrained first-order optimization.

cs.LG↗

arXiv2023

The Discontinuous Strain Method: accurately representing fatigue and failure

Fatigue simulation requires accurate modeling of unloading and reloading. However, classical ductile damage models treat deformations after complete failure as irrecoverable -- which leads to unphysical behavior during unloading. This unphysical behavior stems from the continued accumulation of plastic strains after failure, resulting in an incorrect stress state at crack closure. As a remedy, we introduce a discontinuity strain in the additive elasto-plastic strain decomposition, which absorbs the excess strain after failure. This allows representing pre- and post-cracking regimes in a fully continuous setting, wherein the transition from the elasto-plastic response to cracking can be triggered at any arbitrary stage in a completely smooth manner. Moreover, the presented methodology does not exhibit the spurious energy release observed in hybrid approaches. In addition, our approach guarantees mesh-independent results by relying on a characteristic length scale -- based on the discretization's resolution. We name this new methodology the discontinuous strain method. The proposed approach requires only minor modifications of conventional plastic-damage routines. To convey the method in a didactic manner, the algorithmic modifications are first discussed for one- and subsequently for two-/three-dimensional implementations. Using a simple ductile constitutive model, the discontinuous strain method is validated against established two-dimensional benchmarks. The method is, however, independent of the employed constitutive model. Elastic, plastic, and damage models may thus be chosen arbitrarily. Furthermore, computational efforts associated with the method are minimal, rendering it advantageous for accurately representing low-cycle fatigue but potentially also for other scenarios requiring a discontinuity representation within a plastic-damage framework.

cs.CE↗

arXiv2023

From ductile damage to unilateral contact via point-wise implicit discontinuity at the infinitesimal element level

Ductile damage models and cohesive laws incorporate the material plasticity entailing the growth of irrecoverable deformations even after complete failure. This unrealistic growth remains concealed until the unilateral effects arising from the crack closure emerge. We address this issue by proposing a new strategy to cope with the entire process of failure, from the very inception in the form of diffuse damage to the final stage, i.e. the emergence of sharp cracks. To this end, we introduce a new strain field, termed discontinuity strain, to the conventional additive strain decomposition to account for discontinuities in a continuous sense so that the standard principle of virtual work applies. We treat this strain field similar to a strong discontinuity, yet without introducing new kinematic variables and nonlinear boundary conditions. In this paper, we demonstrate the effectiveness of this new strategy at a simple ductile damage constitutive model. The model uses a scalar damage index to control the degradation process. The discontinuity strain field is injected into the strain decomposition if this damage index exceeds a certain threshold. The threshold corresponds to the limit at which the induced imperfections merge and form a discrete crack. With three-point bending tests under pure mode I and mixed-mode conditions, we demonstrate that this augmentation does not show the early crack closure artifact which is wrongly predicted by plastic damage formulations at load reversal. We also use the concrete damaged plasticity model provided in Abaqus commercial finite element program for our comparison. Lastly, a high-intensity low-cycle fatigue test demonstrates the unilateral effects resulting from the complete closure of the induced crack.

cs.CE↗

arXiv2023

Transfer Learning Enhanced Full Waveform Inversion

We propose a way to favorably employ neural networks in the field of non-destructive testing using Full Waveform Inversion (FWI). The presented methodology discretizes the unknown material distribution in the domain with a neural network within an adjoint optimization. To further increase efficiency of the FWI, pretrained neural networks are used to provide a good starting point for the inversion. This reduces the number of iterations in the Full Waveform Inversion for specific, yet generalizable settings.

cs.LG↗

arXiv2023

On the radial discretization of finite element spaces in the scaled boundary finite element method

The scaled boundary finite element method is known for its capability in reproducing highly-detailed solution fields. This, however, is only attainable in those cases where analytical solutions exist. Many others invoke the use of numerical methods that only provide the response of boundaries. Hence, no information on the inner-subdomain solution fields can be recovered. As a remedy, we propose a new solution scheme by which the interior fields of subdomains can be recovered.

math.NA↗

arXiv2023

Implicit-Explicit Time Integration for the Immersed Wave Equation

Immersed boundary methods simplify mesh generation by embedding the domain of interest into an extended domain that is easy to mesh, introducing the challenge of dealing with cells that intersect the domain boundary. Combined with explicit time integration schemes, the finite cell method introduces a lower bound for the critical time step size. Explicit transient analyses commonly use the spectral element method due to its natural way of obtaining diagonal mass matrices through nodal lumping. Its combination with the finite cell method is called the spectral cell method. Unfortunately, a direct application of nodal lumping in the spectral cell method is impossible due to the special quadrature necessary to treat the discontinuous integrand inside the cut cells. We analyze an implicit-explicit (IMEX) time integration method to exploit the advantages of the nodal lumping scheme for uncut cells on one side and the unconditional stability of implicit time integration schemes for cut cells on the other. In this hybrid, immersed Newmark IMEX approach, we use explicit second-order central differences to integrate the uncut degrees of freedom that lead to a diagonal block in the mass matrix and an implicit trapezoidal Newmark method to integrate the remaining degrees of freedom (those supported by at least one cut cell). The immersed Newmark IMEX approach preserves the high-order convergence rates and the geometric flexibility of the finite cell method. We analyze a simple system of spring-coupled masses to highlight some of the essential characteristics of Newmark IMEX time integration. We then solve the scalar wave equation on two- and three-dimensional examples with significant geometric complexity to show that our approach is more efficient than state-of-the-art time integration schemes when comparing accuracy and runtime.

cs.CE↗

arXiv2023

Deep Learning in Deterministic Computational Mechanics

The rapid growth of deep learning research, including within the field of computational mechanics, has resulted in an extensive and diverse body of literature. To help researchers identify key concepts and promising methodologies within this field, we provide an overview of deep learning in deterministic computational mechanics. Five main categories are identified and explored: simulation substitution, simulation enhancement, discretizations as neural networks, generative approaches, and deep reinforcement learning. This review focuses on deep learning methods rather than applications for computational mechanics, thereby enabling researchers to explore this field more effectively. As such, the review is not necessarily aimed at researchers with extensive knowledge of deep learning -- instead, the primary audience is researchers at the verge of entering this field or those who attempt to gain an overview of deep learning in computational mechanics. The discussed concepts are, therefore, explained as simple as possible.

cs.LG↗

arXiv2023

Isogeometric Multi-Resolution Full Waveform Inversion based on the Finite Cell Method

Full waveform inversion (FWI) is an iterative identification process that serves to minimize the misfit of model-based simulated and experimentally measured wave field data, with the goal of identifying a field of parameters for a given physical object. The inverse optimization process of FWI is based on forward and backward solutions of the (elastic or acoustic) eave equation. In a previous paper [1], we explored opportunities of using the finite cell method (FCM) as the wave field solver to incorporate highly complex geometric models. Furthermore, we demonstrated that the identification of the model's density outperforms that of the velocity -- particularly in cases where unknown voids characterized by homogeneous Neumann boundary conditions need to be detected. The paper at hand extends this previous study: The isogeometric finite cell analysis (IGA-FCM) -- a combination of isogeometric analysis (IGA) and FCM -- is applied for the wave field solver, with the advantage that the polynomial degree and subsequently also the sampling frequency of the wave field can be increased quite easily. Since the inversion efficiency strongly depends on the accuracy of the forward and backward wave field solution and of the gradient of the functional, consistent and lumped mass matrix discretization are compared. The resolution of the grid describing the unknown material density is the decouple from the knot span grid. Finally, we propose an adaptive multi-resolution algorithm that refines the material grid only locally using an image processing-based refinement indicator. The developed inversion framework allows fast and memory-efficient wave simulation and object identification. While we study the general behavior of the proposed approach on 2D benchmark problems, a final 3D problem shows that it can also be used to identify voids in geometrically complex spatial structures.

cs.CE↗