Advanced Modeling and Simulation in Engineering Sciences最新文献

We extend (EUCLID Efficient Unsupervised Constitutive Law Identification and Discovery)-a data-driven framework for automated material model discovery-to pressure-sensitive plasticity models, encompassing arbitrarily shaped yield surfaces with convexity constraints and non-associated flow rules. The method only requires full-field displacement and boundary force data from one single experiment and delivers constitutive laws as interpretable mathematical expressions. We construct a material model library for pressure-sensitive plasticity models with non-associated flow rules in four steps: (1) a Fourier series describes an arbitrary yield surface shape in the deviatoric stress plane; (2) a pressure-sensitive term in the yield function defines the shape of the shear failure surface and determines plastic deformation under tension; (3) a compression cap term determines plastic deformation under compression; (4) a non-associated flow rule may be adopted to avoid the excessive dilatancy induced by plastic deformations. In contrast to traditional parameter identification methods, EUCLID is equipped with a sparsity promoting regularization to restrain the number of model parameters (and thus modeling features) to the minimum needed to accurately interpret the data, thus achieving a compromise between model simplicity and accuracy. The convexity of the learned yield surface is guaranteed by a set of constraints in the inverse optimization problem. We demonstrate the proposed approach in multiple numerical experiments with noisy data, and show the ability of EUCLID to accurately select a suitable material model from the starting library.

This article provides a reduced-order modelling framework for turbulent compressible flows discretized by the use of finite volume approaches. The basic idea behind this work is the construction of a reduced-order model capable of providing closely accurate solutions with respect to the high fidelity flow fields. Full-order solutions are often obtained through the use of segregated solvers (solution variables are solved one after another), employing slightly modified conservation laws so that they can be decoupled and then solved one at a time. Classical reduction architectures, on the contrary, rely on the Galerkin projection of a complete Navier-Stokes system to be projected all at once, causing a mild discrepancy with the high order solutions. This article relies on segregated reduced-order algorithms for the resolution of turbulent and compressible flows in the context of physical and geometrical parameters. At the full-order level turbulence is modeled using an eddy viscosity approach. Since there is a variety of different turbulence models for the approximation of this supplementary viscosity, one of the aims of this work is to provide a reduced-order model which is independent on this selection. This goal is reached by the application of hybrid methods where Navier-Stokes equations are projected in a standard way while the viscosity field is approximated by the use of data-driven interpolation methods or by the evaluation of a properly trained neural network. By exploiting the aforementioned expedients it is possible to predict accurate solutions with respect to the full-order problems characterized by high Reynolds numbers and elevated Mach numbers.

The accurate modelling of structural dynamics is crucial across numerous engineering applications, such as Structural Health Monitoring (SHM), structural design optimisation, and vibration control. Often, these models originate from physics-based principles and can be derived from corresponding governing equations, often of differential equation form. However, complex system characteristics, such as nonlinearities and energy dissipation mechanisms, often imply that such models are approximative and often imprecise. This challenge is further compounded in SHM, where sensor data is often sparse, making it difficult to fully observe the system's states. Or in an additional context, in inverse modelling from noisy full-field data, modelling assumptions are compounded in to the observation uncertainty approximation. To address these issues, this paper explores the use of Physics-Informed Neural Networks (PINNs), a class of physics-enhanced machine learning (PEML) techniques, for the identification and estimation of dynamical systems. PINNs offer a unique advantage by embedding known physical laws directly into the neural network's loss function, allowing for simple embedding of complex phenomena, even in the presence of uncertainties. This study specifically investigates three key applications of PINNs: state estimation in systems with sparse sensing, joint state-parameter estimation, when both system response and parameters are unknown, and parameter estimation from full-field observation, within a Bayesian framework to quantify uncertainties. The results demonstrate that PINNs deliver an efficient tool across all aforementioned tasks, even in the presence of modelling errors. However, these errors tend to have a more significant impact on parameter estimation, as the optimization process must reconcile discrepancies between the prescribed model and the true system behavior. Despite these challenges, PINNs show promise in dynamical system modeling, offering a robust approach to handling model uncertainties.

This study proposes and compares two data-driven, non-intrusive reduced-order models (ROMs) for additive manufacturing (AM) processes: a combined proper orthogonal decomposition-artificial neural network (POD-ANN) and a convolutional autoencoder-multilayer perceptron (CAE-MLP). The POD-ANN model utilizes proper orthogonal decomposition to create a reduced-order model, which is then combined with an artificial neural network to establish a surrogate model linking the snapshot matrix to the input parameters. This approach effectively reduces the dimensionality of the high-fidelity snapshot matrix and constructs a regression framework for accurate predictions. Conversely, the CAE-MLP model employs a 1D convolutional autoencoder to reduce the spatial dimension of a high-fidelity snapshot matrix derived from numerical simulations. The compressed latent space is then projected onto the input variables using a multilayer perceptron (MLP) regression model. This method leverages deep learning techniques to handle the complexity of the data and improve prediction accuracy. The accuracy and efficiency of both models are evaluated through thermo-mechanical analysis of an AM-built part. The comparison of statistical moments from high-fidelity simulation results with ROM predictions reveals a strong correlation. Furthermore, the predictions are validated against experimental results at various locations. While both models demonstrate good agreement with experimental data, the CAE-MLP model outperforms the POD-ANN model in terms of prediction accuracy and performance. The findings highlight the potential of integrating reduced-order modeling techniques with machine learning algorithms to enhance the analysis of complex AM processes. The proposed models offer a robust framework for future research and applications in the field of additive manufacturing, providing high precision and efficiency.