arXiv subjects

Stefan Hiemer

Publications and source records attributed to Stefan Hiemer.

5 recordsLinked to original sources

arXiv2025

X-ray particle tracking velocimetry for steady-state rheological characterization: Case study of a complex polymer melt flow in material extrusion additive manufacturing

We introduce X-ray Particle Tracking Velocimetry (XPTV) as a promising method to quantitatively resolve the velocity field and associated rheological information of polymer melt flow within the nozzle of a fused filament fabrication (FFF) printer. Employing tungsten powder as tracer particles embedded within a polymer filament, we investigate melt flow dynamics through an aluminum nozzle in a custom setup comparable to commercial printers. The velocity profiles obtained via XPTV reveal significant deviations from classical Newtonian flow, highlighting complex heterogeneous and non-isothermal behavior within the melt. From these measurements, we determine the local infinitesimal strain rate tensor and correlate flow-induced non-Newtonian effects to spatially varying temperature distributions, reflecting incomplete thermal homogenization within the nozzle. We complement the experiments with computational fluid dynamics simulations of the flow inside the printing nozzle, incorporating filament melting through an enthalpy-porosity formulation and treating the air-polymer melt interface using a two-phase approach. The simulated velocity profiles agree closely with the XPTV measurements across the investigated operating conditions, supporting the experimental interpretation. Our findings demonstrate the capability of XPTV to quantify both velocity fields and rheological properties, underscoring its potential as a tool for investigating opaque polymer melt flows in additive manufacturing, industrial processing, and rheology. To our knowledge, this is the first application of XPTV to polymer melt rheology. It enables measurements that are inaccessible to conventional optical methods.

cond-mat.soft↗

arXiv2025

Fiber bundle model of thermally activated creep failure

An equal load sharing fiber bundle model for thermally activated breakdown is developed using transition state theory to describe the rate of elementary failures. The lifetime distribution, average, variance and their asymptotic limits for uniform fiber failure thresholds are derived and found to be in excellent agreement with simulations. The asymptotic scaling with regards to the number of fibers matches analytical approximations in the low temperature limit derived by Roux and co-workers for a model of thermal breakdown by stationary Gaussian noise. For the case of randomly distributed fiber failure strengths, the lifetime distribution is derived as a multidimensional integral with no closed form solution. Simulations with different fiber strength distributions indicate that, in the limit of large fiber numbers, the statistics of bundle lifetimes shows a similar asymptotic scaling for distributed and for uniform thresholds. Fiber breakage by thermal activation occurs in avalanches triggered by individual thermally activated failure events, and the asymptotic avalanche size distribution obtained from the simulations matches earlier theoretical results derived for quasistatic loading.

cond-mat.stat-mech↗

arXiv2022

Predicting Creep Failure by Machine Learning -- Which Features Matter?

Spatial and temporal features are studied with respect to their predictive value for failure time prediction in subcritical failure with machine learning (ML). Data are generated from simulations of a novel, brittle random fuse model (RFM), as well as elasto-plastic finite element simulations (FEM) of a stochastic plasticity model with damage, both models considering stochastic thermally activated damage/failure processes in disordered materials. Fuse networks are generated with hierarchical and nonhierarchical architectures. Random forests - a specific ML algorithm - allow us to measure the feature importance through a feature's average error reduction. RFM simulation data are found to become more predictable with increasing system size and temperature. Increasing the load or the scatter in local materials properties has the opposite effect. Damage accumulation in these models proceeds in stochastic avalanches, and statistical signatures such as avalanche rate or magnitude have been discussed in the literature as predictors of incipient failure. However, in the present study such features proved of no measurable use to the ML models, which mostly rely on global or local strain for prediction. This suggests the strain as viable quantity to monitor in future experimental studies as it is accessible via digital image correlation.

cond-mat.mtrl-sci↗

arXiv2022

Relating plasticity to dislocation properties by data analysis: scaling vs. machine learning approaches

Plasticity modelling has long been based on phenomenological models based on ad-hoc assuption of constitutive relations, which are then fitted to limited data. Other work is based on the consideration of physical mechanisms which seek to establish a physical foundation of the observed plastic deformation behavior through identification of isolated defect processes ('mechanisms') which are observed either experimentally or in simulations and then serve to formulate so-called physically based models. Neither of these approaches is adequate to capture the complexity of plastic deformation which belongs into the realm of emergent collective phenomena, and to understand the complex interplay of multiple deformation pathways which is at the core of modern high performance structural materials. Data based approaches offer alternative pathways towards plasticity modelling whose strengths and limitations we explore here for a simple example, namely the interplay between rate and dislocation density dependent strengthening mechanisms in fcc metals.

cond-mat.mtrl-sci↗

arXiv2022

Predicting the failure of two-dimensional silica glasses

Being able to predict the failure of materials based on structural information is a fundamental issue with enormous practical and industrial relevance for the monitoring of devices and components. Thanks to recent advances in deep learning, accurate failure predictions are becoming possible even for strongly disordered solids, but the sheer number of parameters used in the process renders a physical interpretation of the results impossible. Here we address this issue and use machine learning methods to predict the failure of simulated two dimensional silica glasses from their initial undeformed structure. We then exploit Gradient-weighted Class Activation Mapping (Grad-CAM) to build attention maps associated with the predictions, and we demonstrate that these maps are amenable to physical interpretation in terms of topological defects and local potential energies. We show that our predictions can be transferred to samples with different shape or size than those used in training, as well as to experimental images. Our strategy illustrates how artificial neural networks trained with numerical simulation results can provide interpretable predictions of the behavior of experimentally measured structures.

cond-mat.mtrl-sci↗