Engineering with Computers最新文献

The Northern European Enclosure Dam (NEED) is a hypothetical project to prevent flooding in European countries following the rising ocean level due to melting arctic glaciers. This project involves the construction of two large dams between Scotland and Norway, as well as England and France. The anticipated cost of this project is 250 to 500 billion euros. In this paper, we present the simulation of the aftermath of flooding on the European coastline caused by a catastrophic break of this hypothetical dam. From our simulation results, we can observe a traveling wave after the accident, with a velocity of approximately 45 kms per hour, raising the sea level permanently inside the dammed region. This observation implies a need to construct additional dams or barriers protecting the Netherlands’ northern coastline and the Baltic Sea’s interior. Our simulations have been obtained using the following building blocks. First, a graph transformation model was applied to generate an adaptive mesh, refined towards the seabed and the seashore topography, approximating the topography of the Earth. We employ the composition graph grammar model to break the mesh’s triangular elements without generating hanging nodes. Second, the wave equation is formulated in a spherical latitude-longitude system of coordinates and solved by a high-order time integration scheme using the generalized (alpha) method. While our paper mainly focuses on the simulation of the NEED dam break, we also provide a stand-alone tool to generate an adaptive mesh of the whole Earth. We can use our software as a stand-alone package in FEniCS or other simulation software.

A machine learning (ML) model can provide a precise prediction very quickly, if it is well trained with a massive amount of reliable training data. A finite element method (FEM) is often employed to generate substantial training data. However, such a training process can be computationally burdensome especially for a geometrically complex structure. More critically, a specific size and/or configuration of a training model may confine the applicability of the trained model to the same kind only. In this study, we present a scalable ML approach with an efficient training strategy for micromechanical analysis of fiber-reinforced composite materials. Here, a scalable data-driven micromechanics model (SDMM) is proposed for predicting stresses in unidirectional composites with random fiber arrays. The training data for SDMM is defined in the unit of a fiber pair. A single dataset is composed of a stress value between a fiber pair and an image highlighting the pair with nearby fibers affecting the stress. Therefore, the training microstructures can be considerably small, but the pairwise ML model can be applied to every pair of adjacent two fibers inside a much larger microstructure. The scalability of SDMM is demonstrated by predicting the maximum principal stress values acting between every fiber pair in a super-sized representative volume element. The accuracy of the prediction results is evaluated by finite element analysis results. It is shown that a certain number of nearby fibers is required in the training datasets for accurate prediction.

Air-structure interaction is a key aspect to account for during the design of Micro Air Vehicles. In this context, modelisation and numerical simulations represent a powerful tool to analyse aerodynamic performances. This work proposes an advanced fluid–structure interaction numerical technique for the simulation of dragonfly wings, considered one of the most interesting model due to their complex flapping kinematic. The fluid subproblem, described by incompressible Navier–Stokes equations, is solved in a Finite Element Arbitrary-Lagrangian-Eulerian framework, while the solid subproblem is addressed using structural Finite Element, such as membranes and beams. Moreover, a novel remeshing algorithm based on connectivity manipulation and refinement procedure has been implemented to reduce element distortion in fluid mesh, thus increasing the accuracy of the fluid solution. Firstly, the deformation of a single hindwing has been studied. Secondly, the dragonfly model is enriched by incorporating the forewing and a simplified thorax geometry. Preliminary results highlight the complex dynamic of the fluid around the body as well as the efficiency of the proposed mesh generation algorithm.

This paper presents volume-constrained stress minimization-based, topology optimization. The maximum entropy (maxent) basis functions-based meshless method for two-dimensional linear elastic structures is explored. This work focuses to test the effectiveness of the meshless method in handling the stress singularities during the topology optimization process. The commonly used moving least square basis functions are replaced with maximum entropy basis functions, as the latter possess weak Kronecker delta property which leads to the finite element method (FEM) like displacement boundary conditions imposition. The maxent basis functions are calculated once at the beginning of the simulation and then used in optimization at every iteration. Young’s modulus for each background cell is interpolated using the modified solid isotropic material with penalization approach. An open source pre-processor CUBIT is used. A comparison of the proposed approach with the FEM is carried out using a diverse set of problems with simple and complex geometries of structured and unstructured discretization, to establish that maxent-based meshless methods perform better in tackling the stress singularities due to its smooth stress field.

This paper introduces a methodology designed to augment the inverse design optimization process in scenarios constrained by limited compute, through the strategic synergy of multi-fidelity evaluations, machine learning models, and optimization algorithms. The proposed methodology is analyzed on two distinct engineering inverse design problems: airfoil inverse design and the scalar field reconstruction problem. It leverages a machine learning model trained with low-fidelity simulation data, in each optimization cycle, thereby proficiently predicting a target variable and discerning whether a high-fidelity simulation is necessitated, which notably conserves computational resources. Additionally, the machine learning model is strategically deployed prior to optimization to compress the design space boundaries, thereby further accelerating convergence toward the optimal solution. The methodology has been employed to enhance two optimization algorithms, namely Differential Evolution and Particle Swarm Optimization. Comparative analyses illustrate performance improvements across both algorithms. Notably, this method is adaptable across any inverse design application, facilitating a synergy between a representative low-fidelity ML model, and high-fidelity simulation, and can be seamlessly applied across any variety of population-based optimization algorithms.

We propose a method for reducing the spatial discretization error of coarse computational fluid dynamics (CFD) problems by enhancing the quality of low-resolution simulations using deep learning. We feed the model with fine-grid data after projecting it to the coarse-grid discretization. We substitute the default differencing scheme for the convection term by a feed-forward neural network that interpolates velocities from cell centers to face values to produce velocities that approximate the down-sampled fine-grid data well. The deep learning framework incorporates the open-source CFD code OpenFOAM, resulting in an end-to-end differentiable model. We automatically differentiate the CFD physics using a discrete adjoint code version. We present a fast communication method between TensorFlow (Python) and OpenFOAM (c++) that accelerates the training process. We applied the model to the flow past a square cylinder problem, reducing the error from 120% to 25% in the velocity for simulations inside the training distribution compared to the traditional solver using an x8 coarser mesh. For simulations outside the training distribution, the error reduction in the velocities was about 50%. The training is affordable in terms of time and data samples since the architecture exploits the local features of the physics.

Finite element (FE) analysis is one of the most accurate methods for predicting electromagnetic field scatter; however, it presents a significant computational overhead. In this study, we propose a data-driven parametric model-order reduction (pMOR) framework to predict the scattered electromagnetic field of FE analysis. The surface impedance of a coated component is selected as parameter of analysis. A physics-aware (PA) neural network incorporated within a least-squares hierarchical-variational autoencoder (LSH-VAE) is selected for the data-driven pMOR method. The proposed PA-LSH-VAE framework directly accesses the scattered electromagnetic field represented by a large number of degrees of freedom (DOFs). Furthermore, it captures the behavior along with the variation of the complex-valued multi-parameters. A parallel computing approach is used to generate the training data efficiently. The PA-LSH-VAE framework is designed to handle over 2 million DOFs, providing satisfactory accuracy and exhibiting a second-order speed-up factor.

The Volume-of-Fluid (VOF) method is commonly used for numerical simulations of phase change phenomena, such as nucleate boiling or droplet evaporation. A key issue with the standard VOF method is the averaging of the liquid and vapor properties in interface cells, which causes non-physical conjugate heat transfer with a solid wall. Therefore, we aim at a physical model for conjugate heat transfer between a solid and a multiphase fluid. The first measure for higher quality simulations is the splitting of the single temperature field in the fluid region into separate liquid and vapor temperature fields. The second measure is the development of a new, more physical temperature boundary condition for conjugate heat transfer between a solid region and a multiphase fluid, based on experimental results, theoretical models and theoretical considerations. In interface cells, the vapor phase is excluded from the conjugate heat transfer because only heat transfer to the liquid phase occurs resp. dominates. Additionally, the conjugate heat transfer between solid and liquid in the interface cells is performed with virtual subcells, depending on the respective volume fraction of the liquid phase. This new approach (we name it distinctive approach) is successfully validated for energy conservation, and stability issues are discussed for the first time. Significant differences to simulations with averaged properties are observed due to the (now) physically correct modeling of conjugate heat transfer. In our boiling cases, the more accurate numerical simulations lead to considerably larger bubble growth rates. Higher quality simulations are also expected for nearly all applications, where there is a three-phase contact line, be it vapor bubbles in nucleate boiling or droplets impacting on a heated surface.

The purpose of this paper is to point out that it is possible to evaluate the approximation solution of elliptic Partial differential equations (PDEs) on regular and irregular domains where no boundary conditions are defined on some part of the boundary domain. In the presence of trigonometric basis functions (TBFs), the backward substitution method (BSM) coupled with the radial basis functions neural network (RBFNN) is implemented very easily and works well. As a result, the approximation of the boundary conditions and the approximation of the PDE inside the solution domain is separated. The particular solution with an ungiven part of the inhomogeneous boundary condition is completely analyzed by the RBFNN method, and the efficiency and accuracy of the developed algorithms are discussed.