Search

Abstract This survey discusses a posteriori error estimation for model order reduction of parametric systems, including linear and nonlinear, time-dependent and steady systems. We focus on introducing the error estimators we have proposed in the past few years and comparing them with the most related error estimators from the literature. For a clearer comparison, we have translated some existing error bounds proposed in function spaces into the vector space $${\mathbb {C}}^n$$ C n and provide the corresponding proofs in $$\mathbb C^n$$ C n . Some new insights into our proposed error estimators are explored. Moreover, we review our newly proposed error estimator for nonlinear time-evolution systems, which is applicable to reduced-order models solved by arbitrary time-integration solvers. Our recent work on multi-fidelity error estimation is also briefly discussed. Finally, we derive a new inf-sup-constant-free output error estimator for nonlinear time-evolution systems. Numerical results for three examples show the robustness of the new error estimator.

Abstract We investigate the existence of solutions of weakly nonlinear periodic boundary value problems for systems of ordinary differential equations with switchings and the construction of these solutions. We consider the critical case where the equation for the generating constants of a weakly nonlinear periodic boundary-value problem with switchings does not turn into an identity. We improve the classification of critical and noncritical cases and construct an iterative algorithm for finding solutions of weakly nonlinear periodic boundary value problems with switchings in the critical case. As examples of application of the constructed iterative scheme, we obtain approximations to the solutions of a periodic boundary value problem for the mathematical model of nonisothermal chemical reactions. To check the accuracy of the proposed approximations, we evaluate discrepancies in the original equation.

Abstract The purpose of this work is the development of a trained artificial neural network for surrogate modeling of the mechanical response of elasto-viscoplastic grain microstructures. To this end, a U-Net-based convolutional neural network (CNN) is trained using results for the von Mises stress field from the numerical solution of initial-boundary-value problems (IBVPs) for mechanical equilibrium in such microstructures subject to quasi-static uniaxial extension. The resulting trained CNN (tCNN) accurately reproduces the von Mises stress field about 500 times faster than numerical solutions of the corresponding IBVP based on spectral methods. Application of the tCNN to test cases based on microstructure morphologies and boundary conditions not contained in the training dataset is also investigated and discussed.

Abstract We propose an adaptive moment-matching framework for model order reduction of quadratic-bilinear systems. In this framework, an important issue is the selection of those shift frequencies where moment-matching is to be achieved. So far, the choice often has been random or linked to the linear part of the nonlinear system. In this paper, we extend the use of an existing a posteriori error bound for general linear time invariant systems to quadratic-bilinear systems and develop a greedy-type framework to select a good choice of interpolation points for the construction of the projection matrices. The results are compared with standard quadratic-bilinear projection methods and we observe that the approximations obtained by the proposed method yield high accuracy.

Measurement noise is an integral part of collecting data of a physical process. Thus, noise removal is necessary to draw conclusions from these data, and it often becomes essential to construct dynamical models using these data. We discuss a methodology to learn differential equation(s) using noisy and irregularly sampled measurements. In our methodology, the main innovation can be seen in the integration of deep neural networks with the neural ordinary differential equations (ODEs) approach. Precisely, we aim at learning a neural network that provides (approximately) an implicit representation of the data and an additional neural network that models the vector fields of the dependent variables. We combine these two networks by constraints using neural ODEs. The proposed framework to learn a model describing the vector field is highly effective under noisy measurements. The approach can handle scenarios where dependent variables are unavailable at the same temporal grid. Moreover, a particular structure, e.g. second order with respect to time, can easily be incorporated. We demonstrate the effectiveness of the proposed method for learning models using data obtained from various differential equations and present a comparison with the neural ODE method that does not make any special treatment to noise. Additionally, we discuss an ensemble approach to improve the performance of the proposed approach further.

Abstract To counter the volatile nature of renewable energy sources, gas networks take a vital role. But, to ensure fulfillment of contracts under these circumstances, a vast number of possible scenarios, incorporating uncertain supply and demand, has to be simulated ahead of time. This many-query gas network simulation task can be accelerated by model reduction, yet, large-scale, nonlinear, parametric, hyperbolic partial differential(-algebraic) equation systems, modeling natural gas transport, are a challenging application for model order reduction algorithms. For this industrial application, we bring together the scientific computing topics of: mathematical modeling of gas transport networks, numerical simulation of hyperbolic partial differential equation, and parametric model reduction for nonlinear systems. This research resulted in the morgen (Model Order Reduction for Gas and Energy Networks) software platform, which enables modular testing of various combinations of models, solvers, and model reduction methods. In this work we present the theoretical background on systemic modeling and structured, data-driven, system-theoretic model reduction for gas networks, as well as the implementation of morgen and associated numerical experiments testing model reduction adapted to gas network models.

The control of general nonlinear systems is a challenging task in particular for large-scale models as they occur in the semi-discretization of partial differential equations (PDEs) of, say, fluid flow. In order to employ powerful methods from linear numerical algebra and linear control theory, one may embed the nonlinear system in the class of linear parameter varying (LPV) systems. In this work, we show how convolutional neural networks can be used to design LPV approximations of incompressible Navier-Stokes equations. In view of a possibly low-dimensional approximation of the parametrization, we discuss the use of deep neural networks (DNNs) in a semi-discrete PDE context and compare their performance to an approach based on proper orthogonal decomposition (POD). For a streamlined training of DNNs directed to the PDEs in a Finite Element (FEM) framework, we also discuss algorithmical details of implementing the proper norms in general loss functions.

Mathematical models are essential to analyze and understand the dynamics of complex systems. Recently, data-driven methodologies have gotten a lot of attention which is leveraged by advancements in sensor technology. However, the quality of obtained data plays a vital role in learning a good and reliable model. Therefore, in this paper, we propose an efficient heuristic methodology to collect data both in the frequency domain and the time domain, aiming at having more information gained from limited experimental data than equidistant points. In the frequency domain, the interpolation points are restricted to the imaginary axis as the transfer function can be estimated easily on the imaginary axis. The efficiency of the proposed methodology is illustrated by means of several examples, and its robustness in the presence of noisy data is shown.

Die Katalyse ist ein sehr komplexes und interdisziplinäres wissenschaftliches Gebiet, das die effiziente Herstellung einer Vielzahl von Produkten für verschiedene Branchen und in unterschiedlichen Produktionsgrößen ermöglicht. Somit ist die Katalyse eine der Schlüsseltechnologien zur Bewältigung wesentlicher Herausforderungen wie einer nachhaltigen Energieversorgung oder des Klimawandels. Das Datenmanagement in der Katalyse ist derzeit überwiegend auf institutioneller oder Arbeitsgruppenebene organisiert und basiert auf lokalen Konventionen. Übergreifende Repositorien und Standards, wie Daten inklusive Metadaten gespeichert werden sollen, existieren nur in rudimentärer Form. Um den höchstmöglichen Mehrwert aus der Katalyseforschung zu generieren, ist ein grundlegender Wandel in der Katalyseforschung und den katalyseverwandten Wissenschaften wie Chemieingenieurwesen und Verfahrenstechnik erforderlich. Die verschiedenen Disziplinen der Katalyse im Hinblick auf das Datenmanagement zusammenzuführen, ist die zentrale Herausforderung. Aus all diesen Notwendigkeiten heraus wurde im Rahmen der Initiative zur Nationalen Forschungsdateninfrastruktur (NFDI) das NFDI4Cat-Konsortium gegründet. Als eines von derzeit neun geförderten Konsortien konzentriert sich NFDI4Cat auf die Wissenschaften rund um die Katalyse. Dieser Artikel beschreibt die Motivation der Gründungsmitglieder für dieses im Oktober 2020 gestartete Konsortium und wie die Ziele in Bezug auf die Fachcommunity, die technische Infrastruktur und insbesondere das Forschungsdatenmanagement erreicht werden sollen. Darüber hinaus eröffnet die dynamische Entwicklung moderner Technologien zur Erfassung, Bereitstellung und zur Analyse großer Datenmengen neue Möglichkeiten für wissenschaftlichen Fortschritt in den Einzeldisziplinen und insbesondere an deren Nahtstellen.

Abstract This paper studies parametric reduced-order modeling via the interpolation of linear multiple-input multiple-output reduced-order, or, more general, surrogate models in the frequency domain. It shows that realization plays a central role and two methods based on different realizations are proposed. Interpolation of reduced-order models in the Loewner representation is equivalent to interpolating the corresponding frequency response functions. Interpolation of reduced-order models in the (real) pole-residue representation is equivalent to interpolating the positions and residues of the poles. The latter pole-matching approach proves to be more natural in approximating system dynamics. Numerical results demonstrate the high efficiency and wide applicability of the pole-matching method. It is shown to be efficient in interpolating surrogate models built by several different methods, including the balanced truncation method, the Krylov method, the Loewner framework, and a system identification method. It is even capable of interpolating a reduced-order model built by a projection-based method and a reduced-order model built by a data-driven method. Its other merits include low computational cost, small size of the parametric reduced-order model, relative insensitivity to the dimension of the parameter space, and capability of dealing with complicated parameter dependence.

We address the numerical solution of Lyapunov, algebraic and differential Riccati equations, via the matrix sign function, on platforms equipped with general-purpose multicore processors and, optionally, one or more graphics processing units (GPUs). In particular, we review the solvers for these equations, as well as the underlying methods, analyze their concurrency and scalability and provide details on their parallel implementation. Our experimental results show that this class of hardware provides sufficient computational power to tackle large-scale problems, which only a few years ago would have required a cluster of computers.

The high performance computing community has traditionally focused uniquely on the reduction of execution time, though in the last years, the optimization of energy consumption has become a main issue. A reduction of energy usage without a degradation of performance requires the adoption of energy-efficient hardware platforms accompanied by the development of energy-aware algorithms and computational kernels. The solution of linear systems is a key operation for many scientific and engineering problems. Its relevance has motivated an important amount of work, and consequently, it is possible to find high performance solvers for a wide variety of hardware platforms. In this work, we aim to develop a high performance and energy-efficient linear system solver. In particular, we develop two solvers for a low-power CPU-GPU platform, the NVIDIA Jetson TK1. These solvers implement the Gauss-Huard algorithm yielding an efficient usage of the target hardware as well as an efficient memory access. The experimental evaluation shows that the novel proposal reports important savings in both time and energy-consumption when compared with the state-of-the-art solvers of the platform.

For the modelling and the numerical approximation of problems with time-dependent Dirichlet boundary conditions one can call on several consistent and inconsistent approaches. We show that spatially discretized boundary control problems can be brought into a standard state space form accessible for standard optimization and model reduction techniques. We discuss several methods that base on standard finite-element discretizations, propose a newly developed problem formulation, and investigate their performance in numerical examples. We illustrate that penalty schemes require a wise choice of the penalization parameters in particular for iterative solves of the algebraic equations. Incidentally we confirm that standard finite element discretizations of higher order may not achieve the optimal order of convergence in the treatment of boundary forcing problems and that convergence estimates by the common method of manufactured solutions can be misleading.

We analyze the performance-power-energy balance of a conventional Intel Xeon mul- ticore processor and two low-power architectures –an Intel Atom processor and a sys- tem with a quad-core ARM Cortex A9+NVIDIA Quadro 1000M– using a high perfor- mance implementation of Gauss-Jordan elimination (GJE) for matrix inversion. The blocked version of this algorithm employed in the experimental evaluation mostly com- prises matrix-matrix products, so that the results from the evaluation carry beyond the simple matrix inversion and are representative for a wide variety of dense linear algebra operations/codes.

We study numerical methods for finding the maximal symmetric positive definite solution of the nonlinear matrix equation X=Q+LX−1LT, where Q is symmetric positive definite and L is nonsingular. Such equations arise for instance in the analysis of stationary Gaussian reciprocal processes over a finite interval. Its unique largest positive definite solution coincides with the unique positive definite solution of a related discrete-time algebraic Riccati equation (DARE). We discuss how to use the butterfly SZ algorithm to solve the DARE. This approach is compared to several fixed-point and doubling-type iterative methods suggested in the literature.