@unpublished{55159, abstract = {{We introduce a method based on Gaussian process regression to identify discrete variational principles from observed solutions of a field theory. The method is based on the data-based identification of a discrete Lagrangian density. It is a geometric machine learning technique in the sense that the variational structure of the true field theory is reflected in the data-driven model by design. We provide a rigorous convergence statement of the method. The proof circumvents challenges posed by the ambiguity of discrete Lagrangian densities in the inverse problem of variational calculus. Moreover, our method can be used to quantify model uncertainty in the equations of motions and any linear observable of the discrete field theory. This is illustrated on the example of the discrete wave equation and Schrödinger equation. The article constitutes an extension of our previous article arXiv:2404.19626 for the data-driven identification of (discrete) Lagrangians for variational dynamics from an ode setting to the setting of discrete pdes.}}, author = {{Offen, Christian}}, keywords = {{System identification, inverse problem of variational calculus, Gaussian process, Lagrangian learning, physics informed machine learning, geometry aware learning}}, pages = {{28}}, title = {{{Machine learning of discrete field theories with guaranteed convergence and uncertainty quantification}}}, year = {{2024}}, } @article{58491, abstract = {{Similar to bulk metal forming, clinch joining is characterised by large plastic deformations and a variety of different 3D stress states, including severe compression. However, inherent to plastic forming is the nucleation and growth of defects, whose detrimental effects on the material behaviour can be described by continuum damage models and eventually lead to material failure. As the damage evolution strongly depends on the stress state, a stress-state-dependent model is utilised to correctly track the accumulation. To formulate and parameterise this model, besides classical experiments, so-called modified punch tests are also integrated herein to enhance the calibration of the failure model by capturing a larger range of stress states and metal-forming-specific loading conditions. Moreover, when highly ductile materials are considered, such as the dual-phase steel HCT590X and the aluminium alloy EN AW-6014 T4 investigated here, strong necking and localisation might occur prior to fracture. This can alter the stress state and affect the actual strain at failure. This influence is captured by coupling plasticity and damage to incorporate the damage-induced softening effect. Its relative importance is shown by conducting inverse parameter identifications to determine damage and failure parameters for both mentioned ductile metals based on up to 12 different experiments.}}, author = {{Friedlein, Johannes and Böhnke, Max and Schlichter, Malte and Bobbert, Mathias and Meschut, Gerson and Mergheim, Julia and Steinmann, Paul}}, issn = {{2504-4494}}, journal = {{Journal of Manufacturing and Materials Processing}}, keywords = {{ductile damage, stress-state dependency, failure, parameter identification, punch test, clinching}}, number = {{4}}, publisher = {{MDPI AG}}, title = {{{Material Parameter Identification for a Stress-State-Dependent Ductile Damage and Failure Model Applied to Clinch Joining}}}, doi = {{10.3390/jmmp8040157}}, volume = {{8}}, year = {{2024}}, } @inproceedings{62078, abstract = {{Fiber reinforced plastics (FRP) exhibit strongly non-linear deformation behavior. To capture this in simulations, intricate models with a variety of parameters are typically used. The identification of values for such parameters is highly challenging and requires in depth understanding of the model itself. Machine learning (ML) is a promising approach for alleviating this challenge by directly predicting parameters based on experimental results. So far, this works mostly for purely artificial data. In this work, two approaches to generalize to experimental data are investigated: a sequential approach, leveraging understanding of the constitutive model and a direct, purely data driven approach. This is exemplary carried out for a highly non-linear strain rate dependent constitutive model for the shear behavior of FRP.The sequential model is found to work better on both artificial and experimental data. It is capable of extracting well suited parameters from the artificial data under realistic conditions. For the experimental data, the model performance depends on the composition of the experimental curves, varying between excellently suiting and reasonable predictions. Taking the expert knowledge into account for ML-model training led to far better results than the purely data driven approach. Robustifying the model predictions on experimental data promises further improvement. }}, author = {{Gerritzen, Johannes and Hornig, Andreas and Winkler, Peter and Gude, Maik}}, booktitle = {{ECCM21 - Proceedings of the 21st European Conference on Composite Materials}}, isbn = {{978-2-912985-01-9}}, keywords = {{Direct parameter identification, Machine learning, Convolutional neural networks, Strain rate dependency, Fiber reinforced plastics, woven composites, segmentation, synthetic training data, x-ray computed tomography}}, pages = {{1252–1259}}, publisher = {{European Society for Composite Materials (ESCM)}}, title = {{{Direct parameter identification for highly nonlinear strain rate dependent constitutive models using machine learning}}}, doi = {{10.60691/yj56-np80}}, volume = {{3}}, year = {{2024}}, } @inproceedings{42163, abstract = {{The article shows how to learn models of dynamical systems from data which are governed by an unknown variational PDE. Rather than employing reduction techniques, we learn a discrete field theory governed by a discrete Lagrangian density $L_d$ that is modelled as a neural network. Careful regularisation of the loss function for training $L_d$ is necessary to obtain a field theory that is suitable for numerical computations: we derive a regularisation term which optimises the solvability of the discrete Euler--Lagrange equations. Secondly, we develop a method to find solutions to machine learned discrete field theories which constitute travelling waves of the underlying continuous PDE.}}, author = {{Offen, Christian and Ober-Blöbaum, Sina}}, booktitle = {{Geometric Science of Information}}, editor = {{Nielsen, F and Barbaresco, F}}, keywords = {{System identification, discrete Lagrangians, travelling waves}}, location = {{Saint-Malo, Palais du Grand Large, France}}, pages = {{569--579}}, publisher = {{Springer, Cham.}}, title = {{{Learning discrete Lagrangians for variational PDEs from data and detection of travelling waves}}}, doi = {{10.1007/978-3-031-38271-0_57}}, volume = {{14071}}, year = {{2023}}, } @inbook{34211, abstract = {{Nowadays, clinching is a widely used joining technique, where sheets are joined by pure deformation to create an interlock without the need for auxiliary parts. This leads to advantages such as reduced joining time and manufacturing costs. On the other hand, the joint strength solely relies on directed material deformation, which renders an accurate material modelling essential to reliably predict the joint forming. The formation of the joint locally involves large plastic strains and possibly complex non-proportional loading paths, as typical of many metal forming applications. Consequently, a finite plasticity formulation is utilised incorporating a Chaboche–Rousselier kinematic hardening law to capture the Bauschinger effect. Material parameters are identified from tension–compression tests on miniature spec- imens for the dual-phase steel HCT590X. The resulting material model is implemented in LS-Dyna to study the locally diverse loading paths and give a quantitative statement on the importance of kinematic hardening for clinching. It turns out that the Bauschinger effect mainly affects the springback of the sheets and has a smaller effect on the joint forming itself.}}, author = {{Friedlein, Johannes and Mergheim, Julia and Steinmann, Paul}}, booktitle = {{The Minerals, Metals & Materials Series}}, isbn = {{9783031062117}}, issn = {{2367-1181}}, keywords = {{Clinching, Material modelling, Kinematic hardening, Parameter identification, Bauschinger effect}}, publisher = {{Springer International Publishing}}, title = {{{Influence of Kinematic Hardening on Clinch Joining of Dual-Phase Steel HCT590X Sheet Metal}}}, doi = {{10.1007/978-3-031-06212-4_31}}, year = {{2022}}, } @inproceedings{36339, abstract = {{Al-Li-based alloys are an attractive material for aircraft and aerospace applications. Preparation of these alloys by twin-roll casting (TRC), which combines rapid metal solidification and subsequent plastic reduction in a single processing step, could improve the properties of the alloys compared to materials prepared by conventional direct-chill casting. A commonly used approach for identifying primary phases is a chemical analysis by energy dispersive spectroscopy (EDS). More accurate results can be achieved by combining the method with diffraction analysis. This process can be considerably simplified in microscopes equipped with automated crystal orientation and phase mapping (ACOM-TEM). Al-Cu-Li-Mg-Zr alloy was prepared by twin-roll casting. A combination of TEM and STEM images with chemical analysis by EDS and ACOM-TEM was used to obtain complex information about phases of boundary primary particles. The efficiency of the individual methods for the phase identification in TRC Al-Li-based alloys is discussed.}}, author = {{BAJTOŠOVÁ, Lucia and Grydin, Olexandr and STOLBCHENKO, Mykhailo and Schaper, Mirko and KŘIVSKÁ, Barbora and KRÁLÍK, Rostislav and ŠLAPÁKOVÁ, Michaela and CIESLAR, Miroslav}}, booktitle = {{METAL 2022 Conference Proeedings}}, issn = {{2694-9296}}, keywords = {{Al-Cu-Li-M-Zr-Fe alloy, twin-roll casting, phase identification, ACOM-TEM}}, location = {{Brno}}, publisher = {{TANGER Ltd.}}, title = {{{Phase identification in twin-roll cast Al-Li alloys}}}, doi = {{10.37904/metal.2022.4437}}, year = {{2022}}, } @inproceedings{26539, abstract = {{In control design most control strategies are model-based and require accurate models to be applied successfully. Due to simplifications and the model-reality-gap physics-derived models frequently exhibit deviations from real-world-systems. Likewise, purely data-driven methods often do not generalise well enough and may violate physical laws. Recently Physics-Guided Neural Networks (PGNN) and physics-inspired loss functions separately have shown promising results to conquer these drawbacks. In this contribution we extend existing methods towards the identification of non-autonomous systems and propose a combined approach PGNN-L, which uses a PGNN and a physics-inspired loss term (-L) to successfully identify the system's dynamics, while maintaining the consistency with physical laws. The proposed method is demonstrated on two real-world nonlinear systems and outperforms existing techniques regarding complexity and reliability.}}, author = {{Götte, Ricarda-Samantha and Timmermann, Julia}}, booktitle = {{2022 3rd International Conference on Artificial Intelligence, Robotics and Control (AIRC)}}, keywords = {{data-driven, physics-based, physics-informed, neural networks, system identification, hybrid modelling}}, location = {{Cairo, Egypt}}, pages = {{67--76}}, title = {{{Composed Physics- and Data-driven System Identification for Non-autonomous Systems in Control Engineering}}}, doi = {{10.1109/AIRC56195.2022.9836982}}, year = {{2022}}, } @inproceedings{31066, abstract = {{While trade-offs between modeling effort and model accuracy remain a major concern with system identification, resorting to data-driven methods often leads to a complete disregard for physical plausibility. To address this issue, we propose a physics-guided hybrid approach for modeling non-autonomous systems under control. Starting from a traditional physics-based model, this is extended by a recurrent neural network and trained using a sophisticated multi-objective strategy yielding physically plausible models. While purely data-driven methods fail to produce satisfying results, experiments conducted on real data reveal substantial accuracy improvements by our approach compared to a physics-based model. }}, author = {{Schön, Oliver and Götte, Ricarda-Samantha and Timmermann, Julia}}, booktitle = {{14th IFAC Workshop on Adaptive and Learning Control Systems (ALCOS 2022)}}, keywords = {{neural networks, physics-guided, data-driven, multi-objective optimization, system identification, machine learning, dynamical systems}}, location = {{Casablanca, Morocco}}, number = {{12}}, pages = {{19--24}}, title = {{{Multi-Objective Physics-Guided Recurrent Neural Networks for Identifying Non-Autonomous Dynamical Systems}}}, doi = {{https://doi.org/10.1016/j.ifacol.2022.07.282}}, volume = {{55}}, year = {{2022}}, } @inproceedings{9992, abstract = {{State-of-the-art industrial compact high power electronic packages require copper-copper interconnections with larger cross sections made by ultrasonic bonding. In comparison to aluminium-copper, copper-copper interconnections require increased normal forces and ultrasonic power, which might lead to substrate damage due to increased mechanical stresses. One option to raise friction energy without increasing vibration amplitude between wire and substrate or bonding force is the use of two-dimensional vibration. The first part of this contribution reports on the development of a novel bonding system that executes two-dimensional vibrations of a tool-tip to bond a nail- like pin onto a copper substrate. Since intermetallic bonds only form properly when surfaces are clean, oxide free and activated, the geometries of tool-tip and pin were optimised using finite element analysis. To maximize the area of the bonded annulus the distribution of normal pressure was optimized by varying the convexity of the bottom side of the pin. Second, a statistical model obtained from an experimental parameter study shows the influence of different bonding parameters on the bond result. To find bonding parameters with the minimum number of tests, the experiments have been planned using a D-optimal experimental design approach.}}, author = {{Dymel, Collin and Eichwald, Paul and Schemmel, Reinhard and Hemsel, Tobias and Brökelmann, Michael and Hunstig, Matthias and Sextro, Walter}}, booktitle = {{(Proceedings of 7th Electronics System-Integration Technology Conference, Dresden, Germany)}}, keywords = {{ultrasonic wire-bonding, bond-tool design, parameter identification, statistical engineering}}, pages = {{1--6}}, title = {{{Numerical and statistical investigation of weld formation in a novel two-dimensional copper-copper bonding process}}}, year = {{2018}}, } @inproceedings{13222, abstract = {{When performing measurements, the effects of the measurement system itself on the measured data generally must be eliminated. Consequently, those effects, i.e. the system’s dynamic behavior, need to be known. For the piezo-composite transducers in an ultrasonic transmission line, a model based approach is used to describe their dynamic behavior and take into account its dependence on the environment temperature and the acoustic impedance of the target medium. Temperature-dependent model parameters are presented, which are obtained by performing a multiplepart identification process on the transducer model, based on electrical impedance measurements [1]. The identification process uses an inverse approach for optimizing a subset of the model parameters. Additionally, algorithmic differentiation methods are used to determine accurate derivatives. In a final optimization step, impedance measurements taken at different temperatures are used to determine the temperature dependencies of the model parameters. These can then be used to assess the plausibility of the identification results. Additionally, the parameters can be expressed as polynomials in the temperature to take different operating conditions into account.}}, author = {{Webersen, Manuel and Bause, Fabian and Rautenberg, Jens and Henning, Bernd}}, booktitle = {{AMA Conferences 2015}}, keywords = {{piezo-composite, transducer, temperature dependency, identification, plausibility}}, location = {{Nürnberg}}, pages = {{195--200}}, title = {{{Identification of temperature-dependent model parameters of ultrasonic piezo-composite transducers}}}, year = {{2015}}, } @article{9876, abstract = {{Piezoelectric inertia motors use the inertia of a body to drive it by means of a friction contact in a series of small steps. It has been shown previously in theoretical investigations that higher velocities and smoother movements can be obtained if these steps do not contain phases of stiction (''stick-slip`` operation), but use sliding friction only (''slip-slip`` operation). One very promising driving option for such motors is the superposition of multiple sinusoidal signals or harmonics. In this contribution, the theoretical results are validated experimentally. In this context, a quick and reliable identification process for parameters describing the friction contact is proposed. Additionally, the force generation potential of inertia motors is investigated theoretically and experimentally. The experimental results confirm the theoretical result that for a given maximum frequency, a signal with a high fundamental frequency and consisting of two superposed sine waves leads to the highest velocity and the smoothest motion, while the maximum motor force is obtained with signals containing more harmonics. These results are of fundamental importance for the further development of high-velocity piezoelectric inertia motors.}}, author = {{Hunstig, Matthias and Hemsel, Tobias and Sextro, Walter}}, issn = {{0939-1533}}, journal = {{Archive of Applied Mechanics}}, keywords = {{Inertia motor, High velocity, Stick-slip motor, Slip-slip operation, Friction parameter identification}}, pages = {{1--9}}, publisher = {{Springer Berlin Heidelberg}}, title = {{{High-velocity operation of piezoelectric inertia motors: experimental validation}}}, doi = {{10.1007/s00419-014-0940-0}}, year = {{2014}}, } @inproceedings{11816, abstract = {{In this paper, we consider the Maximum Likelihood (ML) estimation of the parameters of a GAUSSIAN in the presence of censored, i.e., clipped data. We show that the resulting Expectation Maximization (EM) algorithm delivers virtually biasfree and efficient estimates, and we discuss its convergence properties. We also discuss optimal classification in the presence of censored data. Censored data are frequently encountered in wireless LAN positioning systems based on the fingerprinting method employing signal strength measurements, due to the limited sensitivity of the portable devices. Experiments both on simulated and real-world data demonstrate the effectiveness of the proposed algorithms.}}, author = {{Hoang, Manh Kha and Haeb-Umbach, Reinhold}}, booktitle = {{38th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2013)}}, issn = {{1520-6149}}, keywords = {{Gaussian processes, Global Positioning System, convergence, expectation-maximisation algorithm, fingerprint identification, indoor radio, signal classification, wireless LAN, EM algorithm, ML estimation, WiFi indoor positioning, censored Gaussian data classification, clipped data, convergence properties, expectation maximization algorithm, fingerprinting method, maximum likelihood estimation, optimal classification, parameters estimation, portable devices sensitivity, signal strength measurements, wireless LAN positioning systems, Convergence, IEEE 802.11 Standards, Maximum likelihood estimation, Parameter estimation, Position measurement, Training, Indoor positioning, censored data, expectation maximization, signal strength, wireless LAN}}, pages = {{3721--3725}}, title = {{{Parameter estimation and classification of censored Gaussian data with application to WiFi indoor positioning}}}, doi = {{10.1109/ICASSP.2013.6638353}}, year = {{2013}}, } @article{11892, abstract = {{For an environment to be perceived as being smart, contextual information has to be gathered to adapt the system's behavior and its interface towards the user. Being a rich source of context information speech can be acquired unobtrusively by microphone arrays and then processed to extract information about the user and his environment. In this paper, a system for joint temporal segmentation, speaker localization, and identification is presented, which is supported by face identification from video data obtained from a steerable camera. Special attention is paid to latency aspects and online processing capabilities, as they are important for the application under investigation, namely ambient communication. It describes the vision of terminal-less, session-less and multi-modal telecommunication with remote partners, where the user can move freely within his home while the communication follows him. The speaker diarization serves as a context source, which has been integrated in a service-oriented middleware architecture and provided to the application to select the most appropriate I/O device and to steer the camera towards the speaker during ambient communication.}}, author = {{Schmalenstroeer, Joerg and Haeb-Umbach, Reinhold}}, journal = {{IEEE Journal of Selected Topics in Signal Processing}}, keywords = {{audio streaming, audio visual data streaming, context information speech, face identification, face recognition, image segmentation, middleware, multimodal telecommunication, online diarization, service oriented middleware architecture, sessionless telecommunication, software architecture, speaker identification, speaker localization, speaker recognition, steerable camera, telecommunication computing, temporal segmentation, terminal-less telecommunication, video streaming}}, number = {{5}}, pages = {{845--856}}, title = {{{Online Diarization of Streaming Audio-Visual Data for Smart Environments}}}, doi = {{10.1109/JSTSP.2010.2050519}}, volume = {{4}}, year = {{2010}}, }