Colloquiam: Documents published in 2022

Comparing three calculation methods of load distribution in radial bearings

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:47:10 +0100

There are two basic methods for radial external load distribution calculation on rolling elements in a rolling element bearing: the discrete method and the integral method. Solving the discrete equilibrium equation using the Newton-Raphson scheme, more accurate results are derived than those based on the integral method, with small theoretical and computational efforts. The Sjövall's radial integral factors, as well as some approximations proposed in the literature, for lineand point-contacts, are given. Numerical approximations for the Sjövall's radial integrals are proposed. The approximations' errors with respect to the Sjövall's radial integral's numerical integration are shown.

Investigation of parameter-dependent material characteristics of additively manufactured specimens for data-driven part optimization

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:46:51 +0100

Direct Metal Laser Sintering (DMLS) is a complex production process including hosts of parameters and a multitude of physical phenomena, which make the simulation and modeling quite challenging. This work investigates the impact of modified printing parameters (e.g., hatch distance, laser power) on correlating material properties (e.g., Young's modulus, temperature gradient) of hardened aluminum specimens. The ultimate goal is to create a data model that enables data-driven and multi-physical optimization of mechanical components fabricated via DMLS.

On different discretisation strategies to solve the kinematical and equilibrium problem for masonry-like structures

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:46:38 +0100

Nowadays, there is a raising interest in the development of fast and robust tools to detect the consequences of settlements or loading changes in unreinforced masonry buildings, since they constitute a large part of world architectural heritage. Current tools, based on Finite Element Method or on Discrete Element Method are computationally cumbersome, from one side due to difficulties in dealing with unilateral materials, and on the other side, due to the need of formulating the problem as an explicit dynamics problem. The methods proposed here are based on the minimization problem of two different functionals, the Total Potential Energy, and the Total Complementary Energy, which allow to detect the stress and strain distribution developed under given load and given boundary settlements, through a minimization problem, which require a significantly lower computational cost and no material parameters, especially when rigidity assumption of the material is done. After illustrating the main characteristics of the two methods, they are applied to a case study, and the results are suitably described and discussed.

Robust discretizations for poroelastic problems: engineering and mathematical points of views in a presenation in pairs

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:46:23 +0100

This contribution is the proceeding of a presentation in pairs taking different viewpoints on the robustness of discretizations for poroelastic problems. These presentations are organised by the Young researcher committee to continue the tradition of fruitful interactions between applied mathematics and computational engineering. The engineering part of this contribution highlights key aspects of the theoretical framework and comments on robustness of common discretizations. Within the mathematical part of this contribution it is shown that the accurate approximation of the total stress tensor as well as the Darcy velocity are crucial to obtain reliability and robustness.

Transonic Buffet Simulation using a Partially-Averaged Navier-Stokes Approach

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:46:06 +0100

This work assesses the capability of the partially averaged Navier-Stokes (PANS) method to accurately reproduce self-sustained shock oscillations, also known as transonic buffet, occurring on supercritical aerofoils at high Reynolds numbers. Attention is paid to the comparison with unsteady Reynolds-averaged Navier Stokes (URANS) results to show the benefits of PANS, in resolving flow unsteadiness on affordable CFD grids. The role of the mesh metrics in the formulation of the PANS model is emphasized, as well as the relation of the mesh metrics with the spatiotemporal discretisation used for the numerical simulations. The aim is to extend the use of PANS to flow cases involving shock-wave boundary layer interactions to obtain accurate predictions without the need for very expensive computations.

Review of Micro and Mesoscale simulation methods for Laser Powder Bed Fusion

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:45:49 +0100

Additive manufacturing (AM) is an advanced method of manufacturing complex parts layer by layer until the required design is achieved. Laser powder bed fusion (L-PBF) is used to produce parts with high resolution because of low layer thickness. L-PBF is based on laser beam and material interaction where the powder material is melted and then solidified. This occurs in a short time frame of the order of 0.02 seconds and makes the whole process challenging to be studied in real time. Studies have shown the development of numerical methods and the use of simulation software to understand the laser beam and material interaction. This phenomenon is key to understanding the material behavior under melting and mechanical properties of the part produced by L-PBF process as it is directly linked with the solidification of the melted powder material. A detailed study of the laser beam and material interaction is needed on a microscale and mesoscale level as it provides a better understanding and helps in the development of the given material for the L-PBF process. This review provides a comprehensive understanding of the background for the use of simulation in AM and the different simulation scales of feature under interest. The main conclusion from this review is the need to develop a methodology to use simulation at micro and mesoscale level to understand the laser beam and material interaction and improve the efficiency of the L-PBF process using this data.

Strategic application of digital tools to enhance lifecycle cost: product design and optimization in metal based powder bed fusion

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:45:31 +0100

Additive manufacturing (AM) has undergone different phases of technological changes from being a mere manufacturing method for consumer goods, prototyping, and tooling to industrial series production of functional end-use parts. The seven AM sub-categories allow the creation of unprecedented designs that are otherwise impossible using conventional manufacturing (CM) methods. The layer-by-layer approach to manufacturing enables the creation of metal components with hollows and overhangs, often requiring sacrificial support structures which are removed prior to or during the post-processing phase. Factors such as poor part quality, high investment cost, low material efficiency, and long manufacturing time hindered the widespread adoption of AM in the past. The adoption of laser-based powder bed fusion for metals was particularly hindered due to reasons such as the need for support structures, demand for post-processing, the numerous affecting processing parameters and the lack of understanding of the interaction between laser beam and material. Technological advances in AM have helped users reduce or omit some of the limitations to adoption, such as optimized support structures for better material efficiency. Simulation-driven tool is one means offering ways to time-efficient product development and more superior structural components amidst the raw material and cost reductions. This study elucidates how such benefits are feasible via using simulation tools. Simulation-driven optimization of the product design, process, and manufacturing is revealed to change the design, support structures and postprocessing required to bring parts to the required reliability. Virtual manufacturing planning also gives a prior understanding of how processing parameters such as laser scan velocity, laser power, scanning strategy, hatch distance and others can be controlled; to achieve optimal interaction between laser beam and material for the required part quality. Simulation-driven design for additive manufacturing (DfAM) allows for agile design optimizing with design parameters and rules, boosting resource efficiency and productivity. This research proposes a life cycle cost (LCC)driven DfAM tool, which potentially improves service life and life cycle cost. The results provide insight into the simulation-driven DfAM of laser-based PBF and demonstrate the potential for LCC-based approaches to enhance the confidence in adopting PBF for metals.

Vectorised spectral/hp element matrix-free operator for anisotropic heat transport in tokamak edge plasma

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:45:10 +0100

A highly efficient matrix-free Helmholtz operator with single-instruction multipledata (SIMD) vectorisation is implemented in Nektar++ [1] and applied to the simulation of anisotropic heat transport in tokamak edge plasma. A tokamak is currently the leading candidate for a practical fusion reactor using the magnetic confinement approach to produce electricity through controlled thermonuclear fusion. Predicting the transport of heat in magnetized plasma is important to designing a safe tokamak design. Due to the ionized nature of plasma, the heat conduction of the magnetized plasma is highly anisotropic along the magnetic field lines. In this study, a variational form is proposed to simulate the anisotropic heat transport in magnetized plasma and the details of its mathematical derivation and implementation are presented. To accurately approximate the thermal load of plasma deposition on the wall of tokamak chamber, highly scalable and efficient algorithms are crucial. To achieve this, a matrix-free Helmholtz operator is implemented in the Nektar++ framework, utilising sum-factorisation to reduce the operation count and increase arithmetic intensity, and leveraging SIMD vectorisation to accelerate the computation on modern hardware. The performance of the implementation is assessed by measuring throughput and speed-up of the operators using deformed and regular quadrilateral and triangular elements.

The neural network multigrid solver for the Navier-Stokes equations and its application to 3D simulation

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:44:49 +0100

We investigate scaling and efficiency of the deep neural network multigrid method (DNN-MG), a novel neural network-based technique for the simulation of the Navier-Stokes equations that combines an adaptive geometric multigrid solver with a recurrent neural network with memory. The neural network replaces in DNN-MG one or multiple finest multigrid layers and provides a correction for the classical solve in the next time step. This leads to little degradation in the solution quality while substantially reducing the overall computational costs. At the same time, the use of the multigrid solver at the coarse scales allows for a compact network that is easy to train, generalizes well, and allows for the incorporation of physical constraints. In this work, we investigate how the network size affects training and solution quality and the overall runtime of the computations.

Evaluation of the performance portability layer of different linear solver packages with ALIEN, an open generic and extensible linear algebra framework

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:44:28 +0100

Applications to solve large and complex partial derivative equation systems often rely nowadays on frameworks like Arcane, Dune, Feel++. Linear solver packages like PETSc or Trilinos are used to manage linear systems and provide access to a wide range of algorithms. With the evolution of High-Performance Computing, the variety of the hardware features available in new architectures has considerably increased. ARM processors, AMD, Intel and Nvidia GP-GPUs, TPU and FPGA devices are now common. To handle the induced complexity, different strategies are adopted in each linear solver framework. One of them consists in introducing a new layer that provides abstractions to manage the performance portability and to enable several parallel programming models. In this paper, we evaluate the performance of linear solver packages that rely on tools like SYCL [16], Kokkos [8] or HARTS [11] to handle runtime systems like OpenMP, TBB, CUDA,. . . . A simulator to solve advection-diffusion problems has been developed with ALIEN, a C++ framework that provides a high level and unified API to handle large distributed matrices and vectors. We have benchmarked different solver algorithms, and have evaluated the efficiency of their implementations, and their capability to perform on different architectures, for instance, large number of cores, GP-GPU accelerators, or processors with large SIMD instructions.

A Factorization Algorithm for Sparse Matrix with Mixed Precision Arithmetic

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:44:10 +0100

A new hybrid algorithm for LDU -factorization for large sparse matrix combining iterative solver, which can keep the same accuracy as the classical factorization, is proposed. The last Schur complement will be generated by iterative solver for multiple right-hand sides using block GCR method with the factorization in lower precision as a preconditioner, which achieves mixed precision arithmetic, and then the Schur complement will be factorized in higher precision. In this algorithm, essential procedure is decomposition of the matrix into a union of moderate and hard parts, which is realized by LDU -factorization in lower precision with symmetric pivoting and threshold postponing technique.

Very fast FEM Poisson solvers on lower precision accelerator hardware

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:43:48 +0100

Graphics cards that are equipped with Tensor Core units designed for AI applications, for example the NVIDIA Ampere A100, promise very high peak rates concerning their computing power (156 TFLOP/s in single and 312 TFLOP/s in half precision in the case of the A100). This is only achieved when performing arithmetically intensive operations such as dense matrix multiplications in the aforementioned lower precision, which is an obstacle when trying to use this hardware for solving linear systems arising from PDEs discretized with the finite element method. In previous works, we delivered a proof of concept that the predecessor of the A100, the V100 and its Tensor Cores, can be exploited to a great extent when solving Poisson's equation on the unit square if a hardware-oriented direct solver based on prehandling via hierarchical finite elements and a Schur complement approach is used. In this work, using numerical results on an A100 graphics card, we show that the method also achieves a very high performance if Poisson's equation, which is discretized by linear finite elements, is solved on a more complex domain corresponding to a flow around a square configuration.

Material point method for large deformation seismic response analysis

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:43:30 +0100

Landslides triggered by earthquakes are one of the major seismic hazards and can cause large damages and fatalities. The material point method (MPM) has become a popular technique to model such large mass movements. A limitation of existing MPM implementations is the lack of appropriate boundary conditions to perform seismic response analysis of slopes. To bridge this gap, an extension to the basic MPM framework is presented for simulating the seismic triggering and subsequent collapse of slopes within a single analysis step. The concepts of a compliant base boundary and free-field columns are applied within the MPM framework enabling the direct application of input ground motions and accounting for the absorption of outgoing waves.

Penalized direct-dorcing method and power-law-based wall model for immersed-boundary numerical simulations of obstacles in turbulent flow

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:43:09 +0100

In this paper, academic and industrial test cases have been conducted in order to validate the approach of using a Penalized Direct Forcing method coupled with an immersed turbulent wall model. Good results are obtained compared to a body fitted mesh with the Werner & Wengle wall model. In a shortcoming second step, we can project the coupling between the immersed wall law and a K-epsilon model, as well as obstacle shape optimization during the flow computation.

A comparison between IBM with feedback forcing and a volume penalization method for compressible flows

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:42:48 +0100

The aim of this work is to model compressible flows involving shock waves past a solid obstacle using a non-conformal mesh. An Immersed Boundary Method (IBM) with feedback forcing and a volume penalization method are considered and compared. Both methods are validated on various test-cases. Accuracy and computational cost are discussed.

Simulation of the flow-acoustic-structural interaction in flow ducts using a partitioned approach in the time domain

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:42:28 +0100

Duct systems confining a subsonic air flow, such as ventilation ducts, often have a lightweight design. These lightweight constructions are easily excited by unsteady pressure fluctuations in the flow, causing structural vibrations and noise emissions. Designing effective solutions for this flow-acousticstructural problem requires a better understanding of the multi-physical interactions and efficient prediction tools. In this work, due to the confined configuration, the vibro-acoustic interaction is a strong two-way interaction and is modeled by coupling a flow-acoustic solver with a structural solver. The kinematic and dynamic continuity at the interface is ensured in this partitioned approach by a data exchange during runtime between the solvers. The data exchange is managed by the open-source coupling library preCICE [1]. The analysis of the flow-acoustic-structural interaction in a flexible flow duct with rectangular cross section was given in [2]. In this paper, the error resulting from the pressure mapping between both solvers is analyzed and an improved force mapping strategy is adopted.

Adaptive and flexible macro-micro coupling software

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:42:09 +0100

Many multiscale simulation problems require a many-to-one coupling between different scales. For such coupled problems, researchers oftentimes focus on the coupling methodology, but largely ignore software engineering and high-performance computing aspects. This can lead to inefficient use of hardware resources, on the one hand, but also inefficient use of human resources as solutions to typical technical coupling problems are constantly reinvented. This work proposes a flexible and application-agnostic software framework to couple independent simulation codes in a many-to-one fashion. To this end, we introduce a prototype of a new lightweight software component called Micro Manager, which allows us to reuse the coupling library preCICE for two-scale coupled problems. We demonstrate the applicability of the framework by a two-scale coupled heat conduction problem.

Data-driven models for shrinkage porosity prediction in aluminium casting

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:41:45 +0100

Casting is one of the most used processes to form metals like aluminium. A casting part can contain several defects that threaten its resistance. Shrinkage porosity is one of the major anomalies that designers try to avoid. For this purpose, rounds of numerical simulations should be performed with operating on a selection of parameters in order to minimize the presence of porosity in the casting part. In general, these approaches are time-cost with dependence on the complexity of the study case and the needed accuracy. In this paper, a methodology of data-driven porosity prediction for 3D parts is proposed in order to minimize the time-cost. A supervised learning algorithm is implemented to learn nodal porosity prediction using decision trees based method. A dataset is generated from a casting simulation software with operating on a selection of parameters. The training is realised on critical features vector extracted from nodal thermal history. Model order reduction method is used to interpolate thermal fields allover the parameter space. This interpolation is sufficiently accurate with minor errors. Promising results of shrinkage porosity prediction on a 3D study case are obtained. An evaluation of these results is performed with reference to the simulations results. This solution can contribute to open perspectives for more data-driven solutions that optimize the time-cost in the design stage.

Model order reduction of solidification problems

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:41:24 +0100

Advection driven problems are known to be difficult to model with a reduced basis because of a slow decay of the Kolmogorov N -width. This paper investigates how this challenge transfers to the context of solidification problems and tries to answer when and to what extend reduced order models (ROMs) work for solidification problems. In solidification problems, the challenge is not the advection per se, but rather a moving solidification front. This paper studies reduced spaces for 1D step functions that move in time, which can either be seen as advection of a quantity or as a moving solidification front. Furthermore, the reduced space of a 2D solidification test case is compared with the reduced space of an alloy solidification featuring a mushy zone. The results show that not only the PDE itself, but the smoothness of the solution is crucial for the decay of the singular values and thus the quality of a reduced space representation.

Data-driven enhanced model reduction for bifurcating models in computational fluid dynamics

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:41:08 +0100

We investigate various data-driven methods to enhance projection-based model reduction techniques with the aim of capturing bifurcating solutions. To show the effectiveness of the data-driven enhancements, we focus on the incompressible Navier-Stokes equations and different types of bifurcations. To recover solutions past a Hopf bifurcation, we propose an approach that combines proper orthogonal decomposition with Hankel dynamic mode decomposition. To approximate solutions close to a pitchfork bifurcation, we combine localized reduced models with artificial neural networks. Several numerical examples are shown to demonstrate the feasibility of the presented approaches.

Accelerated nonlinear PDE-constrained optimization by reduced order modelling

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:40:52 +0100

We present a framework to accelerate optimization of problems where the objective function is governed by a nonlinear partial differential equation (PDE) using projection-based reduced-order models (ROMs) and a trust-region (TR) method. To reduce the cost of objective function evaluations by several orders of magnitude, we replace the underlying full-order model (FOM) with a series of hyperreduced ROMs (HROMs) constructed on-the-fly. Each HROM is equipped with an online-efficient a posteriori error estimator, which is used to define a TR. Hyperreduction is performed following a goal-oriented empirical quadrature procedure, which guarantees first-order consistency of the HROM with the FOM at the TR center. This ensures the optimizer converges to a local minimum of the underlying FOM problem. We demonstrate the framework through optimization of a nonlinear thermal fin and pressure-matching inverse design of an airfoil under Euler flow and Reynolds-averaged Navier-Stokes flow.

On the matching of eigensolutions to parametric partial differential equations

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:40:34 +0100

In this paper a novel numerical approximation of parametric eigenvalue problems is presented. We motivate our study with the analysis of a POD reduced order model for a simple one dimensional example. In particular, we introduce a new algorithm capable to track the matching of eigenvalues when the parameters vary.

The impact of tube corrugation within the multi-disciplinary design optimization of a charge air cooler.

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:40:11 +0100

The following paper explores the impact of corrugated tubes within a charge air cooler (CAC) on overall cooler performance, cost and size, for the first time. Corrugated tubes have been demonstrated to perform better in terms of heat transfer, when compared to a smooth tube [2], however they have not been optimized in the context of a CAC. In this study, a CAC with corrugated tubes is compared against a similar system comprising of smooth tubes as a baseline design. Both CACs have common design parameters, such as number of tubes per rows, number of rows, number of passes, fins per meter, fin material, and tube material, while two additional design parameters exist i.e., groove depth, and pitch for the CAC with corrugated tubes, that characterizes the helical corrugation. These two systems are optimized to minimize manufacturing cost where cost is a function of cooler dimensions and material selection. Feasible designs are then obtained by satisfying dimension, pressure, weight, performance and vibrations based constraints. A vibration constraint introduced here is an addition to the current state of the art [3], making this approach, a multi-disciplinary one and the first of its kind. Finally, the optimum is compared which signifies the importance of a multi-disciplinary analysis for both cooler configurations.

Application of multiresolution analysis and deep learning to obtain failure pressure of corroded pipelines

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:32:57 +0100

The assessment of corroded pipelines is considered a very important task in the oil and gas industry. The present work aims to develop an efficient system to accurately predict the burst pressure of corroded pipelines with complex corrosion profiles through hybrid models using multiresolution analysis, numerical analysis, and metamodels. The corrosion profile is obtained from ultrasonic inspections and the data is provided as a river bottom profile. The real corrosion shapes are parametrized considering a discrete wavelet transform to reduce the amount of data that describes the defect. The coefficients obtained from the wavelet transform are used as inputs to feed a deep neural network system for quickly and accurately predict the burst pipeline pressure. Eight different steel materials are considered in the NN build process. Synthetic models that have similar statistics to real corrosion profiles are created and submitted to non-linear FEM analysis, for the different materials. The failure pressures obtained from the synthetic defects are used to train a neural network to predict the burst pressure of the pipelines. The results obtained with the deep neural networks are very accurate for all cases presented in this work.

A new numerical limit analysis-based strategy to retrofit masonry curved structures with FRCM systems

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:32:36 +0100

In most historic masonry structures, curved geometries, such as arches or vaults, are key structural components to the overall building stability. Therefore, it is crucial to assess their safety level with respect to changes in the boundary conditions (increased loads or settlements). If the safety level of the structure needs to be enhanced, a strategy to intervene and retrofit structural members is represented by the use of Fabric Reinforced Cementitious Matrix (FRCM) systems. These types of externally bonded composite materials, made of high-strength textiles embedded in inorganic matrices, are proven to be a particularly advantageous strengthening solution for curved masonry structures. Even though limit analysis approaches such as Thrust Network Analysis (TNA) have been widely used to assess structural stability, their use in a retrofitting framework is seldom explored. This paper proposes an automated procedure to design the FRCM reinforcement required in masonry structures based on an initial TNA assessment analysis. To perform these analyses, a nonlinear programming problem is implemented and solved to compute the minimum reinforcement required for stability. These quantities are then used to design the FRCM reinforcement according to existing regulations. Finally, the load-bearing capacity of the reinforced structure can be re-evaluated for different load cases ensuring that the structure is safe. The effectiveness of the proposed approach is benchmarked against laboratory tests and demonstrated on arched structures.

Design relationships for the strengthening of masonry walls with mortar-based composites

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:32:14 +0100

The paper presents a design method for the strengthening of masonry walls with fabric reinforced cementitious matrix (FRCM), steel reinforced grout (SRG) and composite reinforced mortar (CRM) systems. They have proved effective for the enhancement of structural capacity and are suitable for seismic retrofitting and for applications to architectural heritage. More recently, significant research efforts have been devoted to the development of testing/certification methods and of design guidelines. For this latter purpose, analytical relationships were developed, which are consistent with Eurocodes, are suitable for engineering practice, and have been incorporated in design guides. Both the bending strengthening under out-of-plane loads and the shear strengthening under in-plane loads are dealt with in the paper. The validation of the resisting models and the calibration of partial coefficients according to the design-by-testing approach are described. Assumptions, limitations and advantages are discussed, to promote the knowledge transfer from the academia to engineering practice and the proper use of FRCM, SRG and CRM for enhancing the safety level of the built environment.

Challenges of integrating adjoint simulations in industrial turbomachinery MDO

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:31:54 +0100

The design of turbomachinery creates a strong demand for the simultaneous optimization of multiple blade rows with regard to different disciplines including aerodynamics, aeroelasticity, and solid mechanics. Established gradient-free methods, typically surrogatebased methods, have been successfully applied to the optimization of single blade rows and pairs of adjacent rows, typically featuring in the order of 50 design variables per blade row. Gradient-free methods become inhibitively expensive through the increased number of design variables from simultaneous optimizations of many rows. Gradients obtained from adjoint simulations can help in transitioning to larger design spaces as they provide derivatives with respect to each design variable at a computational cost that only depends on the number of objectives. For the transition from gradient-free to gradientbased optimizations, a variety of challenges had to be solved, which will be outlined in this paper.

Analysis of high-order interpolation schemes for the finite-volume resolution of linear problems on unstructured meshes

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:31:13 +0100

Finite-volume strategies in fluid-structure interaction problems would be of crucial importance in many engineering applications such as in the analysis of reed valves in reciprocating compressors. The efficient implementation of this strategy passes from the formulation of reliable high-order schemes on 3D unstructured meshes. The development of high-order models is essential in bending-dominant problems, where the phenomenon of shear blocking appears. In order to solve this problem, it is possible to either increase the number of elements or increase the interpolation order of the main variable. Increasing the number of elements does not always yield good results and implies a very high computational cost that, in real problems, is inadmissible. Using unstructured meshes is also vital because they are necessary for real problems where the geometries are complex and depart from canonical rectangular or regular shapes. This work presents a series of tests to demonstrate the feasibility of a high-order model using finite volumes for linear elasticity on unstructured and structured meshes. The high-order interpolation will be performed using two different schemes such as the Moving Least Squares (MLS) and the Local Regression Estimators (LRE). The reliability of the method for solving 2D and 3D problems will be verified by solving some known test cases with an analytical solution such as a thin beam or problems where stress concentrations appear.

Aeroelastic Adjoint-Based Optimisation of Highly Flexible Aircraft Wing Configuration

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:30:56 +0100

This paper investigates benefits resulting from the use of coupled aeroelastic analysis for aerodynamic shape optimisation of a highly flexible wing. The study is carried out on the eXternal Research Forum model (XRF-1) specified by Airbus Commercial Aircraft, representative of a long-range aircraft configuration. Improvements delivered by considering aeroelastic effects for the evaluation of both the aerodynamic performance and the associated gradients are assessed with respect to the results obtained by freezing the wing flexibility in both primal and adjoint computations. An analysis of the impact on the different drag components is also illustrated based on the far-field drag breakdown. Results show that for induced drag, engaging flexibility only at the primal level still allows to capture first-order gain on the final performance. However, engaging coupled-adjoint sensitivities is key to completely master wave drag reduction on the considered highly flexible wing. Performance improvement obtained by increasing the number of design parameters is also investigated.

Approximating the full-field temperature evolution in 3D electronic systems from randomized “Minecraft” systems

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:30:38 +0100

Neural Networks as fast physics simulators have a large potential for many engineering design tasks. Prerequisites for a wide-spread application are an easy-to-use workflow for generating training datasets in a reasonable time, and the capability of the network to generalize to unseen systems. In contrast to most previous works where training systems are similar to the evaluation dataset, we propose to adapt the type of training system to the network architecture. Specifically, we apply a fully convolutional network and, thus, design 3D systems of randomly located voxels with randomly assigned physical properties. The idea is tested for the transient heat diffusion in electronic systems. Training only on random 'Minecraft' systems, we obtain good generalization to electronic systems four times as large as the training systems (one-step prediction error of 0.07 % vs 0.8 %).

In-plane shear capacity response of FRCM-strengthened masonry walls

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:30:21 +0100

This paper aims at investigating the in-plane shear response of FRCM-strengthened masonry walls. To this end, available results of experimental tests are collected, accounting for the masonry substrate made with bricks and mortar joints and several FRCM materials applied with different strengthening configurations. The contribution of the composite material to the masonry wall shear capacity is evaluated. The influence of some geometrical and mechanical parameters on the shear strength of the retrofitted walls is assessed. Available analytical design formulations are implemented to the database and commented.

Empirical usability fragility curves for unreinforced masonry buildings affected by the 2009 L’Aquila earthquake

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:30:02 +0100

Recent earthquakes occurred in Italy highlighted the great vulnerability of the Italian building stoke that registered significant economic losses. In this context, many vulnerability models were developed in the literature to obtain a reliable loss assessment. They often focused on damage fragility curves definitions, intending to estimate the damage suffered by the buildings after the seismic events. Nevertheless, in the last years, the attention of different research groups is moved toward the prediction of the building usability, i.e. the condition of a building being habitable or occupiable after a seismic event. In fact, recent researches highlighted that usability is stronger correlated with direct and indirect costs than structural damage. Consequently, the prediction of usability performance represents a valid indicator for the economic funding distribution after an earthquake. From this perspective, this paper aims to develop typological usability fragility curves for Italian unreinforced-masonry buildings to be used for seismic risk assessment on a large scale. The proposed empirical model was calibrated from the observed data collected after the 2009 L'Aquila earthquake, including more than 56 000 unreinforced-masonry buildings. The database was increased to estimate the effective number of usable buildings in the study area. Then, the structural parameters affecting the usability assessment were investigated, and three parameters (construction timespan, number of stories, and state of repair), available both on the post-earthquake database and Italian census, were selected to define different typological classes. The usability fragility curves were defined as a function of peak ground acceleration for two building usability states strongly correlated to repair and population assistance costs: partially unusable and unusable. The curves represent a sound tool to be used as part of a risk model for assessing earthquake impact in terms of both economic and societal losses.

Bayesian system identification and dynamic virtualization using incomplete noisy measurements

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:29:42 +0100

This study presents the application of Bayesian Expectation-Maximization (BEM) methodology to coupled state-input-parameter estimation in both linear and nonlinear structures. The BEM is built upon a Bayesian foundation, which utilizes the EM algorithm to deliver accurate estimates for latent states, model parameters, and input forces while updating noise characteristics effectively. This feature allows for quantifying associated uncertainties using response-only measurements. The proposed methodology is equipped with a recursive backward-forward Bayesian estimator that provides smoothed estimates of the state, input, and parameters during the Expectation step. Next, these estimates help calculate the most probable values of the noise parameters based on the observed data. This adaptive approach to the coupled estimation problem allows for real-time quantification of estimation uncertainties, whereby displacement, velocity, acceleration, strain, and stress states can be reconstructed for all degrees-of-freedom through virtual sensing. Through numerical examples, it is demonstrated that the BEM accurately estimates the unknown quantities based on the measured quantities, not only when a fusion of displacement and acceleration measurements is available but also in the presence of acceleration-only response measurements.

Seismic vulnerability of masonry structures through a mechanical-based approach

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:29:18 +0100

Enhancing the territorial resilience to natural events, such as earthquakes, is assuming a primary role in the current political debate. In the context of Disaster Risk Management, developing reliable vulnerability models for the seismic risk assessment at a territorial scale is an aspect of crucial importance. In this perspective, the paper presents a mechanical-based method for the evaluation of local-scale seismic fragility curves for unreinforced masonry buildings, based on the exposure data collected in the Italian CARTIS database. It uses a bidimensional finite element model and static nonlinear analyses to obtain the structural behaviour. Monte Carlo simulations are performed to propagate the uncertainties. Both local and global scale structural behaviour are considered to define the damage grade. A case-study regarding the city centre of Cosenza, in southern Italy, validates the proposal.

Bayesian Optimal Sensor Placement for Virtual Sensing and Strain Reconstruction

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:28:56 +0100

A Bayesian optimal sensor placement (OSP) framework is presented for virtual sensing in structures using output-only vibration measurements. Particularly, this probabilistic OSP scheme aims to enhance the reconstruction of dynamical responses (e.g., accelerations, displacements, strain, stresses) for updating structural reliability and safety, as well as fatigue lifetime prognosis. The OSP framework is formulated using information theory. The information gained from a sensor configuration is defined as the Kullback-Liebler divergence (KL-div) between the prior and posterior distributions of the response quantities of interest (QoI). The Gaussian nature of the response estimate for linear models of structures is employed, and the information gain is characterized in terms of the reconstruction error covariance matrix. A Kalman-based input-state estimation technique is integrated within an existing OSP strategy, aiming to obtain estimates of response QoI and their uncertainties. The design variables include the location, type and number of sensors. Heuristic algorithms are used to solve optimization problem and provide computationally efficient solutions. The effectiveness of the method is demonstrated using an example from structural dynamics.

Hierarchical Bayesian model for simulating the mechanical behavior of bare printed circuit boards with fixing

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:28:34 +0100

A method for probabilistic simulation of a bare printed circuit board fixed with bolted joints based on hierarchical Bayesian updating of a numerical model is presented in this paper. The objective is the determination of parameter uncertainties in a set of nominally identical boards and the propagation of these uncertainties to calculate probability distributions for the behavior of the mechanical system. The numerical model of the system is split into models for the circuit board, the bolts and a contact model that are updated separately.

LES-aided shape optimization of U-Bend channel

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:28:16 +0100

Reynolds-Averaged Navier-Stokes (RANS) simulations are inaccurate in predicting complex flow features (ex: separation regions), and therefore deriving an optimised shape using the RANS-adjoint framework does not yield a truly optimal geometry. With the purpose of obtaining accurate sensitivity to objective function of interest, we improve the RANS flowfield using the strategy of Singh et al. [1]. This involves multiplying a corrective factor to the production term in the Spalart-Allmaras (SA) turbulence model equation and solving the inverse problem to determine the appropriate field, which enables the RANS solution to match the high-fidelity data.The geometry of our interest is the U-Bend which is widely studied in literature in the context of gas turbine cooling, and which is known to be a challenging case for RANS simulations to reproduce. We use the mean flowfield from a large-eddy simulation of the U-Bend geometry as the high-fidelity data to which the RANS flowfield is fit using the strategy outlined above. We observe a clear improvement in the RANS flowfield by optimising for the field, the objective function to be minimized being L2-norm of the mean velocity difference between RANS and LES. We further show that adding an additional corrective factor () to the destruction term in the SA turbulence equation and simultaneously optimising for the field alongside the field results in a better match of the RANS flowfield with the corresponding LES flowfield. We also show that surface sensitivity map for the improved LES-aided flowfield varies significantly in comparison to the baseline SA-based flowfield for an objective function of interest, the total pressure loss in the U-Bend.

Using conservation laws to infer deep learning model accuracy of Richtmyer-Meshkov instabilities

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:27:53 +0100

Richtmyer-Meshkov Instability (RMI) is a complicated phenomenon that occurs when a shockwave passes through a perturbed interface. Over a thousand hydrodynamic simulations were performed to study the formation of RMI for a parameterized high velocity impact. Deep learning was used to learn the temporal mapping of initial geometric perturbations to the full-field hydrodynamic solutions of density and velocity. The continuity equation was used to include physical information into the loss function, however only resulted in very minor improvements at the cost of additional training complexity. Predictions from the deep learning model appear to accurately capture temporal RMI formations for a variety of geometric conditions within the domain. First principle physical laws were investigated to infer the accuracy of the model's predictive capability. While the continuity equation appeared to show no correlation with the accuracy of the model, conservation of mass and momentum were weakly correlated with accuracy. Since conservation laws can be quickly calculated from the deep learning model, they may be useful in applications where a relative accuracy measure is needed.

Vectorial limitation for multislope MUSCL schemes

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:27:32 +0100

In finite volume schemes with MUSCL interpolation of scalar variables at the faces of control volumes, a slope limiting function is used in order to prevent non-physical oscillations of the solution. More particularly, these functions are designed to ensure a certain monotonicity criterion at each face of the control volume, criterion which then ensures a stability property of the scheme. For vectorial variables, these slope limiting functions are generally applied componentwise, but this may result in a frame-dependance, as well as a loss of accuracy due to false detection of extrema. In this paper, a new vectorial interpolation method is introduced, which is frame-invariant, second-order accurate and stable in a sense that will be defined.

Remaining Useful Life prediction with a Deep Self-Supervised Learning Approach

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:27:12 +0100

With the increasing availability of data for Prognostics and Health Management (PHM), Deep Learning (DL) techniques are now the subject of considerable attention in Prognostics for Predictive Maintenance, achieving more accurate Remaining Useful Life (RUL) predictions. However, one of the major challenges for DL techniques resides in the difficulty of obtaining large amounts of labeled data on industrial systems. To overcome this lack of labeled data, an emerging learning technique is considered in this work : Self-Supervised Learning, a sub-category of unsupervised learning approaches. This paper aims to investigate whether pre-training DL models in a self-supervised way on unlabeled sensors data can be useful for downstream tasks in PHM (i.e. RUL estimation) with only limited amount of labelled data. A synthetic dataset composed of strain data is used. Results show that the self-supervised pretrained models significantly outperform the non pre-trained models in downstream Remaining Useful Life (RUL) prediction task, showing promising results in prognostic tasks when only limited labeled data is available.

Pros and cons of various equivalent frame models for nonlinear analysis of URM buildings

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:26:55 +0100

Brick masonry is considered as one of the old construction materials, and several cultural heritage assets are made of unreinforced masonry (URM), which is susceptible to earthquakes due to its brittle behavior. The equivalent frame method (EFM) is a nonlinear modeling method widely utilized for the seismic analysis of URM buildings with lower computational efforts than finite and discrete element methods. In this study, three macroelements, including the unified method (UM), composite spring method (CSM), and double modified multiple vertical line element model (DM-MVLEM), were utilized to model three case studies. The first case study is a full-scale two-story URM wall that was tested by applying the cyclic prescribed displacements, and two other case studies were developed by changing the configuration of openings. The second case study is with short piers, and weak spandrels exist in the third model. The efficiency of the methods in terms of the accuracy of the pushover results, prediction of damage patterns, and duration of the incremental dynamic analysis (IDA) are discussed. Finally, seismic fragility curves are provided to compare the IDA results.

Simulation of reacting flows using artificial neural networks: application to multi-regime combustion

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:26:39 +0100

The simulation of reactive flows is a major challenge in several industrial sectors, such as aeronautics or energy production. The coupling between fluid dynamics and chemistry comes however at a cost, as chemical processes involve a wide range of spatial and temporal scales. The resulting equations are stiff and require specific, and expensive, numerical methods. The use of machine learning to estimate the reaction rates has been recently proposed. In particular, Artificial Neural Networks (ANN) have the ability to perform interpolation on high-dimensional data and are thus particularly adapted to chemistry problems. A major issue is then to select an appropriate database on which to train the ANN. It must: (i) be representative of the targeted application; (ii) be sufficiently quick to generate. A promising strategy is to use 0-D stochastics reactors, which mimic reactive and mixing processes in systems while being cheap to compute. This methodology has been successfully applied to non-premixed combustion in the literature. In the present work, the aim is to investigate the ability of the 0D stochastic reactors to be used as a database for a wider range of combustion systems. More specifically, the focus will be on the ability to predict auto-ignition followed by premixed flame propagation. To that purpose, a 2-D turbulent case involving the auto-ignition of a hotspot in a hydrogen/air mixture and the subsequent propagation of a premixed flame is proposed. An ANN model based on stochastic reactors is then built and tested on (i) a 0-D auto-ignition case; (ii) a 1-D laminar premixed flame propagation; (ii) the full 2-D turbulent configuration. Using adequate data transformation at the input and output of the neural network, accurate results are obtained, highlighting the ability of the proposed strategy to deal with a large range of combustion applications.

Scalability analysis and performance modelling of layer-parallel training of deep residual networks using a non-linear multigrid-in-time algorithm

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:26:22 +0100

Replacing the traditional forward and backward passes in a residual network with a Multigrid-Reduction-in-Time (MGRIT) algorithm paves the way for exploiting parallelism across the layer dimension. In this paper, we evaluate the layer-parallel MGRIT algorithm with respect to convergence, scalability, and performance on regression problems. Specifically, we demonstrate that a few MGRIT iterations solve the systems of equations corresponding to the forward and backward passes in ResNets up to reasonable tolerances. We also demonstrate that the MGRIT algorithm breaks the scalability barrier created by the sequential propagation of data during the forward and backward passes. Moreover, we show that ResNet training using the layer-parallel algorithm significantly reduces the training time compared to the layer-serial algorithm on two non-linear regression tasks. We observe much more efficient training loss curves using layer-parallel ResNets as compared to the layer-serial ResNets on two regression tasks. We hypothesize that the error stemming from approximately solving the forward and backward pass systems using the MGRIT algorithm helps the optimization algorithm escape flat saddle-point-like plateaus or local minima on the optimization landscape. We validate this by illustrating that artificially injecting noise in a typical forward or backward propagation, allows the optimizer to escape a saddle-point-like plateau at network initialization.

Lessons Learnt from Chimera Method Application to a Deploying Krueger Device

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:18:58 +0100

The Chimera method [1] is an established method for simulation of overlapping grids. Meshing parts independently has made this method popular for complex geometries as well as moving bodies like propellers and rotors (e.g. [2]) or control surfaces (e.g. [3]). It is thus a promising method to simulate deflecting high-lift systems. The motion of the Krueger flap – as the most promising leading edge high-lift system device for laminar wing technology – is characterized by a relatively large movement (about 140 deg deflection) at a relatively high deflection speed (up to 200 deg/s) compared to classical leading edge devices. In terms of simulation, the grid properties of the overlapping mesh regions vary throughout the motion from a nearly sealed retracted position to a gapped flow in fully deflected position comparable to a slat device. This expects dynamic effects may get dominant and a valid simulation of this flow is needed for proper design and analysis. In the frame of the UHURA project1 , several partners applied their CFD capabilities based on Chimera in order to validate the method for this specific application in comparison to wind tunnel tests. The presentation outlines the different Chimera approaches ranging from structured/2D to hybrid/3D in steady and unsteady simulations for the different type of setups investigated, namely straight and swept wing with full-span and part-span Krueger flap. It summarizes common challenges and best practice for application of the Chimera approach for such a device.

Computational modelling of plasma electrolytic oxidation process induced damage in extruded Mg material

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:18:33 +0100

Magnesium (Mg) alloys are an attractive constructive material due to their light weight and high mechanical strength. Plasma electrolyte oxidation (PEO) treatment of Mg alloys creates a thin ceramic coating with protective effects against mechanical wear and corrosion. The coating properties like its porosity and thickness can be adjusted by PEO process parameters and at the same time affects the material behaviour under tensile strength. In this work, dedicated slow-strain rate experiments of differently PEO coated Mg alloy dog-bone shaped specimen were conducted and the coating porosity, thickness and crack spacing were analyzed in order to deduce a predictive Finite Element Method (FEM) damage model. The results indicate that the thicker, more porous coatings lead to material failure at smaller strains in plastic regions. The effect can be implemented via partial differential equation into the FEM model.

Parallel Finite Volume Code for Plasma with Unstructured Adaptive Mesh Refinement

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:18:14 +0100

The present paper describes a parallel unstructured-mesh Plasma simulation code based on Finite Volume method. The code dynamically refines and coarses mesh for accurate resolution of the different features regarding the electron density. Our purpose is to examine the performance of a new Parallel Adaptive Mesh Refinement (PAMR) procedure introduced on the ADAPT platform, which resolves of a relatively complicated system coupling the flow partial differential equations to the Poisson's equation. The implementation deals with the MUMPS parallel multi-frontal direct solver and mesh partitioning methods using METIS to improve the performance of the framework. The standard MPI is used to establish communication between processors. Performance analysis of the PAMR procedure shows the efficiency and the potential of the method for the propagation equations of ionization waves.

PREDICTION OF THE BIOMASS PARTICLES THROUGH THE PHYSICS-INFORMED NEURAL NETWORK

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:17:55 +0100

Woody biomass energy is a kind of renewable energy that contributes to the reduction of greenhouse gas emissions, the creation of healthier forests, and the reduction of wildfire danger. Generally speaking, simulations of the motion of biomass particles are a time-consuming process due to a large number of particles and required simulation time. We used a physicsinformed neural network (PINN) model to predict the motion of particles by including their equations of motion to reconstruct the velocity fields and reduce the processing effort. compare to the discrete element methods, the PINNs methods have the advantage of predicting the velocity fields without the knowledge of the simulation's boundary and initial conditions as well as geometry. It has shown that the proposed model has reliable prediction results with a mean percentage error in time less than 1 percent.

Development of a low-level, algebra-based library to provide platform portability on hybrid supercomputers

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:17:33 +0100

Continuous enhancement in hardware technologies enables scientific computing to advance incessantly and reach further aims. Since the start of the global race for exascale high-performance computing, massively-parallel devices of various architectures have been incorporated into the newest supercomputers, leading to an increasing hybridization of compute nodes. In this context of accelerated innovation, software portability and efficiency become crucial. Traditionally, scientific computing software development using mesh methods is based on calculations in iterative stencil loops over a discretized geometry--the mesh. Despite being intuitive and versatile, the interdependency between algorithms and their computational implementations in stencil applications usually results in a large number of subroutines and introduces an inevitable complexity when it comes to portability and sustainability. An alternative is to break the interdependency between the algorithm and its implementation, and then to cast the calculations into a minimalist set of kernels. Algebra-based implementations rely on a reduced set of basic linear algebra subroutines, which simplifies the deployment of software in hybrid computing systems. In this work, we tackle the development of a fully-portable, algebraic library that can be coupled beneath other high-level, algebra-oriented framework. Namely, this library provides platform portability in the simplest possible manner (i.e., the user develops applications in a purely sequential style). Internally, algebraic objects are distributed among computing devices using a multilevel decomposition approach. Data exchanges between computing units or between nodes are hidden by a multithreaded overlapping scheme.

Efficient strategies for solving the variable Poisson equation with large contrasts in the coefficients

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:17:09 +0100

Discrete versions of Poisson's equation with large contrasts in the coefficients result in very ill-conditioned systems. Thus, its iterative solution represents a major challenge, for instance, in porous media and multiphase flow simulations, where considerable permeability and density ratios are usually found. The existing strategies trying to remedy this are highly dependent on whether the coefficient matrix remains constant at each time iteration or not. In this regard, incompressible multiphase flows with high-density ratios are particularly demanding as their resulting Poisson equation varies along with the density field, making the reconstruction of complex preconditioners impractical. This work presents a strategy for solving such versions of the variable Poisson equation.Roughly, we first make it constant through an adequate approximation. Then, we block-diagonalise it through an inexpensive change of basis that takes advantage of mesh reflection symmetries, which are common in multiphase flows. Finally, we solve the resulting set of fully decoupled subsystems with virtually any solver. The numerical experiments conducted on a multiphase flow simulation prove the benefits of such an approach, resulting in up to 6.6x faster convergences.

TN, ROM, ML, PINNs – Four approaches for real-time temperature estimation in electric motors in comparison

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:16:47 +0100

ne of the major trends in electrical machines for automotive applications is towards higher power-densities and more integrated components. With that, accurate thermal management of the machine and capable cooling systems are of great significance to the safety and reliability of the traction system. Thermal simulations are an integral part in the design process of electrical drives. However, recently thermal models are also more frequently used in the context of machine control. The latter demanding for fast, yet accurate, temperature estimations.

Modelization of a molten salt thermal energy storage for concentrated solar power.

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:16:24 +0100

A numerical model for studying a storage tank for concentrated solar power is presented. The model consists of solving the heat equation for the solid part made from ceramic materials, a one-dimensional model for the molten salt circulating inside the solid, and a coupling between them. Then, some results are presented for a reference case with some typical parameters for the storage system.

Data-driven geometric modelling methods for digital twinning: manufacturing, geospatial and medical applications

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:16:01 +0100

In recent years there has been an explosion of interest in digital twinning in many disciplines, including the manufacturing, geospatial, and medical domains. A core topic of importance in modelling digital twins, is reconstruction of geometric models from raw data. Despite the diversity of requirements in the vast space of digital twin applications, methods for geometric reconstruction can often be transferred between disciplines with only minor modifications. In this paper we present some recent results related to how advances in machine learning over the last decade can be used for data-driven geometric reconstruction in the medical, geospatial and manufacturing domains.

A Parallel Solver for CFD based on the Alternating Anderson-Richardson Method

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:15:42 +0100

Recent advancement in the usage and deployment of large supercomputing resources require the need for algorithmic improvements to make use of the increased parallelism architecture. The Alternating Anderson-Richardson (AAR) method has been recently shown to exhibit good performance when solving problems in distributed parallel computers. This research will extend and investigate the performance of the AAR method to solve CFD problems using a modern compressible flow solver. This work will compare its performance and scalability against commonly used linear solvers, such as the Richardson method and the Generalised Minimal RESidual (GMRES), for solving large, sparse linear systems of equations arising from CFD applications. Results using a range of turbomachinery test cases demonstrate that the current AAR implementation offers significant performance improvement over the Richardson method. The speedup of AAR with respect to GMRES is less significant due to the load imbalance across partitions.

Next-generation HPC models for future Rotorcraft applications

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:15:18 +0100

Rotorcraft technologies pose great scientific and industrial challenges for numerical computing. As available computational resources approach the exascale, finer scales and therefore more accurate simulations of engineering test cases become accessible. However, shifting legacy workflows and optimizing parallel efficiency and scalability of existing software on new hardware is often demanding. This paper reports preliminary results in CFD and structural dynamics simulations using the T106A Low Pressure Turbine (LPT) blade geometry on Leonardo S.p.A.'s davinci-1 high-performance computing (HPC) facility. Time to solution and scalability are assessed for commercial packages Ansys Fluent, STAR-CCM+, and ABAQUS, and the open-source scientific computing framework PyFR. In direct numerical simulations of compressible fluid flow, normalized time to solution values obtained using PyFR are found to be up to 8 times smaller than those obtained using Fluent and STAR-CCM+. The findings extend to the incompressible case. All models offer weak and strong scaling in tests performed on up to 48 compute nodes, each with 4 Nvidia A100 GPUs. In linear elasticity simulations with ABAQUS, both the iterative solver and the direct solver provide speedup in preliminary scaling tests, with the iterative solver outperforming the direct solver in terms of time-to-solution and memory usage. The results provide a first indication of the potential of HPC architectures in scaling engineering applications towards certification by simulation, and the first step for the Company towards the use of cutting-edge HPC toolkits in the field of Rotorcraft technologies.

Wind Turbine Control using Machine Learning techniques

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:15:00 +0100

This article presents two model-free controllers for wind-turbine torque and pitch control. These controllers are based on reinforcement learning (RL) and Bayesian optimization (BO) and do not rely on any mathematical model of the wind-turbine dynamics, in contrast to classical approaches designed on linearized models. The model-free controllers were benchmarked against a proportional-integral-derivative (PID) regulator in a numerical environment using Blade Element Momentum theory for computing the aerodynamic torque and the blade loads. The results showed that the model-free approaches could increase power harvesting while reducing wind turbine loads.

Separation behaviour of small foreign objects in dry foods

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:14:45 +0100

It is difficult to detect small foreign objects such as hair and soft plastics in dry foods. In our laboratory, an electrostatic separator of small foreign objects in dry foods was developed. The separator consists of a grounded inclined rotating drum, a cylindrical electrode fixed at the centre of the drum, and a suction device. The principle of the separation is based on the difference in the charge per unit mass of the dry food and foreign objects. Although it was found that it is possible to obtain a high purity and a high recovery rate of foods using this separator, the understanding of the separation mechanism is still limited. In this study, we numerically investigated the separation behaviours of foreign objects and dry foods in the inclined rotating drum. The behaviours of the foreign objects and dry foods were calculated, considering the electrostatic force. The electric field strength was calculated by the finite difference method. The effect of the inclination angle of the rotating drum on the trajectory of the particles to be separated was investigated. To examine the validity of the calculation method, the experimental result was compared with the calculated result.

Geomiso ISA: a hybrid software for isogeometric analysis with plate elements and advanced spline techniques

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:14:26 +0100

In this paper, we propose the Geomiso ISA program (www.geomiso.com), a new hybrid software for applications on static isogeometric analysis with plate elements. It is based on the isogeometric analysis, the powerful generalization of the traditional finite element analysis, which, in combination with the plate theory, has attracted increasing attention in construction industry over the last decade, as it achieves efficient design-throughanalysis procedures and shows superior performance. This recently developed program is not just a plug-in, but a both on-premises and cloud-based software solution, applicable to thin (Kirchhoff-Love theory) and thick (Mindlin-Reissner theory) plates. It is used to simulate spline models of slabs and analyze their strength and behavior, while it has many features in common with both finite element software and design programs. This new software solution addresses the rising industrial need for seamless integration of computer-aided design and computer-aided analysis, while it appears to be more efficient to finite element software packages with major improvements, as it facilitates the geometry modeling within analysis, and achieves superior accuracy per degree-of-freedom with shortened computational cost. This is the first time ever such a both on-premises and cloud-based software package has been developed.

Isogeometric representations for digital twins subjected to dynamic excitations with GEOMISO DNL software

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:14:08 +0100

The new Geomiso DNL software is proposed to facilitate the use of isogeometric analysis for nonlinear inelastic dynamic applications. This hybrid software solution combines isogeometric analysis and 3D design with advanced spline techniques, such as NURBS and Tsplines. Its dual nature satisfies the rising industrial need for unification of the fields of computer-aided design (CAD) and computer-aided analysis (CAE), as it eliminates geometric errors by merging geometry design with mesh generation into a single procedure. This paper presents sample nonlinear applications in structural dynamics. Geomiso DNL is seen to handle these situations remarkably well, as the numerical examples exhibit significantly improved accuracy of the results, and reduced computational cost, when compared with finite element software packages. Geomiso DNL is not just a plug-in, but a both on-premises and cloud-based software, which enables engineers to simulate complex dynamic phenomena, whose impact on industrial products and structures in real-world environments can be more efficiently estimated. Taking advantage of the new horizons offered in the peak of the Industry 4.0 era, the physical twin feeds, via cloud technology, with real-time data its geometrically exact digital twin, while a dynamic analysis is performed and crucial results about structure safety and quality are obtained. It is argued that Geomiso DNL is a new, more efficient, alternative to FEA software. This is the first time ever such a cloud-based program has been developed.

Physics-driven digital twin for laser powder bed fusion on GPUs

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:13:44 +0100

Metal Additive Manufacturing (AM) processes such as Laser Powder Bed Fusion (LPBF) suffer from part distortion due to the localized melting and resolidification of the metal powder, which introduces stresses and strains. Despite becoming more and more important as a manufacturing process, options for simulating the printing process to predict the distortions are limited, especially because existing solutions often require very long computation times. In this work, we present the results of an implementation of the inherent strain method on graphics processing units (GPUs) that exploits the massive parallelism of the many GPU cores to speed up the simulations considerably compared to CPU-based implementations.

Immersed boundaries in hypersonic flows with considerations about high-fidelity and massive parallelism

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:13:23 +0100

This paper inscribes itself in the ongoing doctoral work of the first author which aims at adapting the immersed boundary conditions (IBC) technique to three-dimensional (3D) large eddy simulations (LES) of viscous hypersonic flows around complex vehicles. The work relies on a pre-existing in-house IBC code, HYPERION (HYPERsonic vehicle design using Immersed bOuNdaries), originally developed in two dimensions (2D) as a proof of concept that it is possible to use IBC in the presence of strongly shocked flows [4]. As a first step towards the optimization of the 3D HYPERION, we discuss in this paper a novel MPI/Open MP hybrid rasterization algorithm allowing for the detection of immersed cells in record time even for very large problems. We then consider the least-square-based reconstruction algorithm from HYPERION [4]. It was shown in the original paper that the number of neighbors used in the reconstruction is directly related to the condition number of the least-square matrix and an optimum can be found when the condition number reaches an asymptote. In 3D configurations it is found that the number of neighbors has to be very high to ensure the proper conditioning of the least-square matrix. If the computation is distributed on several MPI processes (as is always the case in 3D for realistic return times), gathering the information from that many neighbors can cause obvious communication issues it amounts to covering large stencils with unrealistically large MPI halos. We therefore introduce an algorithm designed for a hybrid MPI/OpenMP environment based on migratable tasks and the consensus algorithm developed by [9] to remedy the former shortcoming. Finally, we discuss the premise of the implementation of LES capabilities in HYPERION. The last milestone of the main author's doctoral work is indeed to study the feasibility of embedding wall laws in the IBC modeling and reconstruction algorithm to try and counteract the low accuracy of the near-wall phenomena caused by the lack of body-fitted mesh.

Vulnerability assessment of cultural heritage structures

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:13:04 +0100

Cultural heritage (CH) assets are the legacy of a society that are inherited from the past generations and can give us lessons for contemporary construction. Not only the formally recognized CH assets but also the non-CH structures and infrastructure, and the interconnection between them are crucial to be considered in a vulnerability assessment tool for the sustainable reconstruction of historic areas. Since most CH assets were not designed based on robust design codes to resist natural hazards such as earthquakes, vulnerability assessment and preservation are pivotal tasks for the authorities. For this aim, Hyperion, an H2020 project (Grant agreement No 821054), was formed in order to take advantage of existing tools and services together with novel technologies to deliver an integrated vulnerability assessment platform for improving the resiliency of historic areas. Geometric documentation is the first and most important step toward the generation of digital twins of CH assets that can be facilitated using 3D laser scanners or drone imaging. Afterward, the finite element method is an accurate approach for developing the simulation-based digital twins of cultural heritage assets. For calibration of the models, the result of the operational modal analysis from the ambient vibration testing using accelerometers can be utilized. Structural analysis for the prediction of the structural behavior or near real-time analysis can be carried out on the calibrated models. However, the full finite element analysis needs a lot of computational effort, and to tackle this limitation, equivalent frame methods can be utilized.

Digitization Principles for Application Scenarios towards Digital Twins of Organizations

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:12:47 +0100

In today's agile business ecosystems, digital twins (DTs) and especially digital twins of organizations (DTOs) allow for adaption through dynamically evolving models depicting organizational aspects such as production processes, data flows, human actors and interactions. A hybrid modelling approach is utilized, as the establishment of such DTOs either considered on their own or as part of a DT ecosystem is not trivial. Meta modelling and meta model merging patterns are applied to integrate heterogeneous perspectives and domain models. Two main research questions with respect to digitization towards digital twinning are discussed: First, which digitization principles/patterns are appropriate for DTOs? Patterns ranging from 'counting' to 'estimation' are introduced to fill digital models serving as a foundation for DTs with data. As a starting point, potential digitization principles for relevant characteristics of BPMN 'Modelling Method for Business Processes' and KPI 'Modelling Method for Key Performance Indicators' models are considered. Second, which principle/pattern is appropriate for which organizational structure? In order to ease the selection of suitable patterns for specific application scenarios, those will be associated with organizational structures like but not limited to construction processes, assembly processes or production processes each of them with domain-specific characteristics. A prototype consisting of three phases use case requirements collection, model design and digitization assistance builds upon (a) physical experimentations in the OMiLAB Innovation Corner using physical assets such as edge devices or sensors, (b) domain specific services considering software related aspects such as timeseries databases or simulation algorithms, and (c) modelling methods enabling the integration of physical and digital components. The paint production pilot from the European Change2Twin project serves as an application scenario evaluation use case. A notion of what the use case company intends to achieve by digital twinning and what is possible by introducing digital services is touched. The outlook presents how artificial intelligence may be introduced for the prototype to leverage the paint production use case and further application scenarios.

A Data-Driven Reduced Order Modeling Approach Applied in Context of Numerical Analysis and Optimization of Plastic Profile Extrusion

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 12:12:30 +0100

In course of this work, we examine the process of plastic profile extrusion, where a polymer melt is shaped inside the so-called extrusion die and fixed in its shape by solidification in the downstream calibration unit. More precise, we focus on the development of a data-driven reduced order model (ROM) for the purpose of predicting temperature distributions within the extruded profiles inside the calibration unit. Therein, the ROM functions as a first step to our overall goal of prediction based process control in order to avoid undesired warpage and damages of the final product.

A Validation Program for Dynamic High-Lift System Aerodynamics

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:49:39 +0100

The feasibility of laminar flow control technology for future wing is bound to the development of a leading edge high-lift system that complies with the requirements on smooth surfaces to enable maintaining the laminar boundary layer flow, such as a Krueger flap. Although in principle the aerodynamic performance of a Krueger flap is known, the unsteady behaviour of the flow during deployment and retraction is completely unknown. This is as even more important as during deployment the Krueger flap is exposed to highly unfavourable positions perpendicular to the flow. To mitigate the risk of unfavourable aircraft behaviour, it is therefore expected that a Krueger flap has to be deflected significantly fast and may trigger unsteady aerodynamic effects. The European H2020 project UHURA, running from September 2018 to August 2022, has been focusing on the unsteady flow behaviour around such high-lift system and will first time deliver a deeper understanding of critical flow features at this type of high-lift device during their deployment and retraction together with a validated numerical procedure for its simulation. UHURA performed detailed experimental measurements in several wind tunnels to obtain a unique data set for validation purposes of Computational Fluid Dynamics (CFD) software, including detailed flow measurements by Particle Image Velocimetry (PIV) and other optical measurement technologies.

Krueger High-Lift System Design Optimization

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:49:20 +0100

This work describes the cooperative/competitive design process that led to the definition of the Krueger flap to be used in the numerical and experimental tests of the European project UHURA. The project requirements are particularly challenging because it is necessary to develop a device with good aerodynamic high-lift characteristics, but it is necessary to consider many constraints of structural and kinematic nature. Indeed, the kinematics for its deployment is quite complex and imposes hard constraints on the Krueger shape, and the structural characteristics must allow it to withstand considerable structural stresses in the deployment phase which is studied in the wind tunnel.

Lattice Boltzmann simulation of a deploying Krueger device

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:49:05 +0100

In this paper, we describe the numerical simulations carried out within the H2020 UHURA project of the turbulent unsteady flow generated during the motion of a Krueger device for laminar wings using a commercial lattice Boltzmann solver based on a Wall-Modelled LES approach. The simulations are focused on reproducing one of the experimental test cases carried out in the ONERA-L1 wind tunnel during the UHURA project. The numerical method and the simulation setup are described. The simulation results are compared with the high-quality experimental data obtained in the ONERA-L1 wind tunnel in order to assess the accuracy of the predictions.

Numerical investigation of Hydrogen self-ignition and deflagration-to-detonation phenomena using automated meshing approach and detailed chemistry

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:48:46 +0100

Computational fluid dynamics (CFD) plays a critical role in designing safe storage and transport systems for hydrogen. Fine mesh resolution and detailed chemistry are essential for the accurate prediction of self-ignition and deflagration-to-detonation (DDT) in hydrogenair mixtures. However, simulating H2 venting and explosion in real-life scenarios (e.g., with complex obstacle shapes and a large computational domain) involves tedious meshing effort and several mesh iterations to capture flame and shock locations. This paper addresses these challenges by assessing the capability of a detailed-chemistry approach combined with automated meshing based on a cut-cell technique and Adaptive Mesh Refinement (AMR). Furthermore, three different turbulence-chemistry interaction modelling approaches are compared for self-ignition and DDT scenarios: a homogeneous reactor model, an eddy dissipation model, and a flame thickening approach.

Finite element model of kurpsai dam in kyrgyzstan based on actual response measured by extensive network of various sensors

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:48:26 +0100

The paper presents recent results of an ongoing collaborative research project focused on modelling the Kurpsai water dam in Kyrgyzstan. The research team includes scientists and engineers from the USA, Kyrgyzstan, and Uzbekistan. This water dam was selected for modelling because of the recent installation of an extensive network of various sensors aimed at monitoring its performance under seasonal changes, ambient vibration, and seismic excitation. The installed instrumentation network includes the following sensors: (1) a set of fiber-optic strainmeters and temperature meters, (2) a set of velocimeters for seismic monitoring, and (3) a set of GNSS receivers to measure absolute static displacements. A 3D model of the water dam was generated based on a utilization of the finite element approach. As a starting point the water dam’s concrete was assumed to be elastic material. The latter assumption is considered acceptable, because (as of today) only responses to relatively small excitations were measured by the sensors. The actual responses of the dam were compared to that of the finite element model to achieve a close correlation with each other. Resonant frequencies of the water dam and its vibrational modes were estimated from the model. In the next phase of the project, the research team is planning to update the geometry of the model based on laser scanning that will be conducted this year. Local anomalies (bulging areas, cracks and so on) of the water dam will be studied via an analysis of point clouds collected by the laser scanner. The fully developed model will be used in an extensive numerical study to predict the dam’s performance and its response to strong seismic events and other hazards.

Mathematical modeling and numerical results on the propagation of solitary waves on tensegrity lattices

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:48:04 +0100

This work discusses the mathematical properties of the interaction potential that characterizes tensegrity mass-spring chains, and its implications in terms of the propagation of compact compression waves in such systems when impacted by a striker. Numerical simulations show evidence of the dependence of the wave form on the speed of the propagating compression pulses, which change shape when passing from the sonic to the super-sonic wave propagation regime.

Wear modelling in elasto-plastic wheel-rail contact problems

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:47:50 +0100

The paper is concerned with the development of the numerical procedure to solve the wheel-rail contact problem and the computation of the distribution of surface flash temperatures, stresses as well as the wear evolution due to friction. The two-dimensional wheel-rail contact problem between a rigid wheel and an elasto-plastic rail lying on a rigid foundation is considered. The contact phenomenon includes Coulomb friction, frictional heat generation as well as the wear of the contacting surfaces. The displacement and stress of the rail in contact are governed by the coupled elasto-plastic and heat conductive equations. The wear depth function appears as an internal variable in the non-penetration condition updating the gap between the worn surfaces of the bodies. Moreover the dissipated energy due to friction is calculated to evaluate the loss of rail material and to determine the shape of the contacting surfaces during the wear evolution process. This contact problem is solved numerically using the finite element method as well as the operator splitting approach. The plastic flow and friction inequality conditions are reformulated as equality conditions using the nonlinear complementarity functions. The distribution of surface temperatures and stresses as well as the evolution of the shape of the contact surfaces and the wear depth are reported and discussed.

REDIM based model reduction of the decomposition of urea-water-solutions in films and droplets

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:47:31 +0100

Selective catalytic reduction (SCR) process with urea-water solution (UWS) is often used in automotive industry to decrease emissions of nitric oxides (NO x) in the exhaust gas. In this process the urea from UWS decomposes to isocyanic acid and ammonia, where the latter is needed to increase the efficiency of the NO xreduction on the catalyst surface. Along with the advantages of using UWS several drawbacks reduce the performance of a SCR system. Incomplete decomposition of urea leads to a formation of residuals affecting the efficiency of the exhaust gas systems. Therefore, the complete decomposition of urea and homogeneous distribution of the resulting ammonia in front of the SCR catalyst represent main challenges in improving the SCR technology. In order to investigate the process of the urea decomposition a detailed chemical kinetic mechanism in the liquid phase is employed. The results are compared with a commonly used approach to model urea decomposition as an evaporation with a following decomposition reaction in the gas phase. It is shown that by using such a mechanism, the decomposition of urea and the gas phase composition with the urea decomposition products can be described more accurately. However, implementing these mechanisms in computations (in CFD approaches) requires a large amount of computational (CPU) time and memory. The method of Reaction Diffusion Manifolds (REDIMs) is implemented for the reduction of the detailed chemical kinetics in the stage of urea decomposition such that the distribution of products of the urea decomposition can be captured accurately in the gas phase with only two reduced variables instead of the 7 gas phase species of the original model.

Interval field methods with local gradient control

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:47:13 +0100

This paper introduces a novel method to create an interval field based on measurement data. Such interval fields are typically used to describe a spatially distributed non-deterministic quantity, e.g., Young's modulus. The interval field is based on a number of measurement points, i.e., control points, expended throughout the domain by a set of basis functions. At the control point the non-deterministic quantity is known and bounded by an interval. However, at these measurement points information about the gradients might also be available. In addition, the non-deterministic quantity might be described better by estimating the gradients based on the other measurements. Hence, the proposed interval field method allows to incorporate this gradient information. The method is based on Inverse Distance Weighing (IDW) with an additional set of basis functions: one set of basis functions interpolates the value, and the second set of basis functions controls the gradient at the control points. The additional basis functions can be determined in two distinct ways: first, the gradients are available or can directly be measured at the control point, and second, a weighted average is taken with respect to all control points within the domain. In general, the proposed interval field provides a more versatile definition of an interval field compared to the standard implementation of inverse distance weighting. The application of the interval field is shown in a number of one-dimensional cases where a comparison with standard inverse distance weighting is made. In addition, a case study with a set of measurement data is used to illustrate the method and how different realisations are obtained.

A modified cohesive zone model for the simulation of mixed-mode fracture of co-consolidated thermoplastic laminates considering fiber bridging

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:46:57 +0100

Fiber bridging is a mechanism that may significantly alter the fracture behavior of composite laminates, adhesively bonded laminates, welded laminates, and co-consolidated laminates. It is therefore quite important for the finite element to take that mechanism into consideration. Such models have been developed for thermosetting laminates; however, this is not the case for thermoplastic laminates and thermoplastic joints. In the present work, a numerical model based on the cohesive zone modelling (CZM) approach has been developed to simulate mixed-mode fracture of co-consolidated thermoplastic laminates by considering fiber bridging. A modified traction separation law of tri-linear form has been developed by superimposing the bi-linear behaviors of the matrix and fibers. Initially, the data from mode I (DCB) and mode II (ENF) fracture toughness tests were used to construct the R-curves of the joints in the opening and sliding directions. The aforementioned curves were embedded into the numerical models through a user-defined material subroutine developed in the LS-Dyna FE code, in order to extract the fiber bridging law directly from the simulation results. The model was used to simulate fracture of a Single-Lap-Shear (SLS) specimen in which a considerable amount of fiber bridging was observed on the fracture area. The numerical results show that the developed model presented improved accuracy in comparison to the CZM employing the bilinear traction-separation law.

Phase field model for simulating fracture of ice

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:46:42 +0100

In the last decade, the phase field model has been established to simulate crack nucleation as well as crack propagation. In this variational approach the physically discontinuity of a crack is modeled by a continuous field variable that distinguishes between intact and broken material. The phase field model has been recently extended to viscoelastic materials in various ways, in which the rate dependent response of viscoelastic materials are taken into account. We propose a viscoelastic fracture phase field model and apply it to simulate the fracture in ice shelves. Thereby we consider the viscoelastic rheology of ice, which can be represented by a Maxwell model. The elastic response is often neglected in ice dynamic simulations but crucial for fracture mechanical studies. The numerical examples of this contribution are implemented and conducted in the finite element software FEniCS and data mimic typical situations in Antarctic and Greenland ice shelves.

Modelling of two-scale contact - Investigation of leakage in polymer seals at cryogenic temperatures

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:46:23 +0100

The prediction of leakage in polymer seals is still a particular challenge due to many dependencies: manufacturing inaccuracy, microparticles on the contact surface and surface asperity. Polymer seals, which are operated at cryogenic temperatures, undergo a material behaviour change at the so-called glass transition temperature. At this temperature, its behaviour changes from viscous/rubbery to glassy. There is a significant stiffening of the polymer material, which leads to a worse compensation of roughness in the contact surfaces. As a consequence the tightness of the valve may no longer be sufficiently given. The leakage through the valve is numerically investigated by a two-scale contact simulation, which is based on the concept of Representative Volume Elements, which are known in homogenization of microstructures. The deformations on the microstructure are prescribed by the macroscopic kinematics at the contact area. The mean microscopic friction coefficient is determined in Representative Contact Elements (RCE), which are node-wise linked to the macroscopic contact area. The RCEs surface texture is parameterized based on optical measurement data. As the polymer seal is operated over a wide temperature range, a fully coupled thermo-viscoelastic material model at finite strains is used to simulate the material behaviour at both scales. Due to the change from entropy to energy dominated behaviour over the glass transition temperature the model is extended to account for the transition from viscous/rubbery to glassy as the temperature is decreased. The surface asperity needs to be represented explicitly as the gap and volume between both contact surfaces and the fluid path through the seal are used to determine the leakage through the seal.

Design of piezoelectric lattice metamaterials

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:46:04 +0100

Piezoelectric lattice metamaterials are considered. A computationally-effective homogenisation method is developed based on the recent solution to the Saint-Venant problem for general anisotropic piezoelectric cylinders. A publicly available repository of unit cell topologies is used to identify piezoelectric metamaterials with optimal figures of merit.

Attenuation and localization of waves in taut cables with suspended masses

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:45:39 +0100

This work analyzes structural waves that propagate freely along taut cables, characterized by a discrete array of scatter elements. The outcomes underline the role played by the periodic distribution of such elements, whose presence alters the response of the system when subjected to propagating waves. Namely, when the domain is perfectly periodic, band gaps are found in the spectrum of the problem. It is also shown that the introduction of a defect of periodicity can lead to the appearance of eigenvalues inside band gaps, corresponding to a motion localized around the defect.

A thermo-coupled constitutive model for semi-crystalline polymers at finite strains: Application to varying degrees of crystallinity and temperatures

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:45:16 +0100

Thermoplastic materials are widely used for thermoforming and injection moulding processes, since their low density in combination with a high strength to mass ratio are interesting for various industrial applications. Semi-crystalline polymers make up a subcategory of thermoplastics, which partly crystallize after cool-down from the molten state. During the thermoforming process, residual stresses can arise, due to complex material behavior under different temperatures and strain rates. Therefore, computational models are needed to predict the material response and minimize production errors. This work presents a thermomechanically consistent phenomenological material formulation at finite strains, based on [1]. In order to account for the highly nonlinear material behavior, elasto-plastic and visco-elastic contributions are combined in the model formulation. To account for the crystalline regions, a hyperelastic-plastic framework is chosen, based on [2, 3]. Kinematic hardening of Arruda-Boyce form is incorporated in the formulation, as well as associated plastic flow. The material parameters depend on both, the temperature as well as the degree of crystallinity. A comparison to experiments with varying degrees of crystallinity and temperatures is presented, where a special blending technique ensures stable crystallinity conditions.

Finite element analyses of shear yielding and crazing in glassy polymers under cyclic mode I loading

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:44:56 +0100

A simple and yet physically motivated continuum-micromechanical model for crazing is developed, focussing on cyclic loading. The model features fibril drawing and fibril creep deformation, loose hanging fibrils upon unloading and the morphology change fibrils undergo between craze initiation up to a fully developed craze. The crazing model is implemented in a user material subroutine in the commercial finite element programme ABAQUS. The performance is investigated on a mode I crack growth boundary value problem under cyclic loading. Experimentally measured craze/crack opening profiles from the literature are reasonably-well captured by the model. The results exhibit further interesting model characteristics, such as a variation of the craze length in the course of a load cycle.

Immersed Boundary Analysis of Models with Internal State Variables: Applications to Hydrogels

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:44:39 +0100

Research on Soft Active Materials (SAMs) has flourished in recent years driven mainly by potential applications to actuation systems, tissue engineering, and soft robotics. These applications benefit from the unique properties of SAMs such as large deformations, a wide range of stimulants, and high motion complexities. Hydrogels are among the dominant members of SAMs. Their highly nonlinear chemo-mechanical transient behavior is described by equations that include rates of internal state variables representing the local swelling state of the gel. Hence, the simulation of hydrogels requires intricate numerical approaches with stabilization schemes. This paper presents an immersed boundary analysis technique to simulate models with internal state variables. A hydrogel model is used as an example to describe the components of the proposed technique. Level sets define the material layout on a fixed background mesh and a generalized version of the extended finite element method predicts the response. The influence of the internal state variables on the stability of the physical analysis is examined. While focusing on an XFEM approach for hydrogels, the presented theory can be extrapolated to similar applications using models with internal state variables (e.g., shape memory polymers) and other immersed boundary analysis technique (e.g., CutFEM).

Simple modeling for stiffness evaluation of bolted joints using interfacial element

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:44:22 +0100

This study has developed a novel finite element named interfacial element which simulates the contact between microscale asperities at contact surfaces of bolted joints. In this element, the contact is assumed to be the Hertzian contact of elastic asperities whose peak heights obey the Gaussian distribution. Based on this assumption, the stiffness of the interfacial element is derived from the compressive stress and the surface texture of the interfaces. On the other hand, it is necessary for large-scale simulations that target the entire vehicle body to reduce the number of nodes and elements in the finite element models. This study has further proposed simple modeling for stiffness evaluation of bolted joints using the interfacial element. Finite element simulations by simplified models in which heads, axes and holes of bolts were ignored were conducted and compared with detailed models and hammering tests. The results revealed that the mean value of the natural frequency of the simplified models had good agreement with that of the detailed models and the hammering tests though the calculation accuracy of the simplified models were lower than the detailed models. The bolt heads and the nuts could be ignored by increasing the density of the bolt axes to be equal to the total weight.

Development of a FE² Multiscale Model of Chloride Ions Transport in Recycled Aggregates Concrete

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:44:05 +0100

Decreasing CO2 emissions and preserving natural resources are necessary to the well-being of our civilisations. In the construction industry, recycling old concrete members could be part of the solution to reach theses objectives. Recycled Concrete Aggregates (RCA), obtained by crushing of demolished concrete structures, can substitute the Natural Aggregates (NA) inside the so-called Recycled Aggregates Concrete (RAC). The durability of RAC is not guaranteed in the current state of research. RCA are indeed composed of natural aggregates partially embedded in an adherent mortar paste, increasing the porosity and water absorption of RAC.

This research aims to better predict the influence of RCA on chloride ions ingress inside concrete. It started with an experimental phase where multiple experiments have been performed to determine the transfer properties and the chloride ions diffusion coefficients of a mortar paste and concretes made from NA or 100% RCA. In this context, the microstructure of the RCA influences deeply the permeability, water content distribution and chloride ingress. Therefore, these properties must be included into a numerical model that integrates the microstructural information in a proper way. A numerical homogenization technique, based on the Finite Element square (FE2 ) method [5, 13], is implemented into a coupled multiscale model of water flows and advection/diffusion of chlorides in saturated concrete, in order to model the complex flow behaviour encountered.

Probabilistic modeling of LCF failure times using a epidemiological crack percolation model

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 11:43:44 +0100

The analysis of standardized low cycle fatigue (LCF) experiments shows that the failure times widely scatter. Furthermore, mechanical components often fail before the deterministic failure time is reached. A possibility to overcome these problems is to consider probabilistic failure times. Our approach for probabilistic life prediction is based on the microstructure of the metal. Since we focus on nickel-base alloys we consider a coarse grained microstructure, with random oriented FCC grains. This leads to random distributed Schmid factors and different anisotropic stress in each grain. To gain crack initiation times, we use Coffin-MansonBasquin and Ramberg-Osgood equation on stresses corrected with probabilistic Schmid factors. Using these single grain crack initiation times, we have developed an epidemiological crack growth model over multiple grains. In this mesoscopic crack percolation model, cracked grains induce a stress increase in neighboring grains. This stress increase is realized using a machine learning model trained on data generated from finite element simulations. The resulting crack clusters are evaluated with a failure criterion based on a multimodal stress intensity factor. From the generated failure times, we calculate surface dependent hazard rates using a Monte Carlo framework. We compare the obtained failure time distributions to data from LCF experiments and find good coincidence of predicted and measured scatter bands.

Modeling imperfect interfaces in layered beams through multi- and single-variable zigzag kinematics

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:41:56 +0100

Multiscale structural models based on the coupling of a zigzag kinematics and a cohesive crack approach have been recently formulated to analyze the response of shear deformable layered structures with imperfect interfaces and describe progressive delamination fracture (Massabò, in Handbook of Damage Mechanics, Springer, 2022, pp.665-698). The zigzag kinematics accounts for zigzag effects associated to the elastic mismatch of the layers and displacement jumps due to interfacial imperfections, using a limited number of variables, which is independent of the number of layers. The effects of imperfect interfaces on the response of structures subjected to thermo-mechanical loading and on wave propagation and dispersion have been studied analytically and the advantages of this approach over discrete layer models and layerwise theories have been highlighted and discussed. In the presentation we review and discuss these models and present preliminary results on novel single-variable formulations, which have been inspired by a technique developed for homogeneous Timoshenko beams in (Kiendl et al.

Application of an adaptive stable GFEM for fracture propagation in plain concrete

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:41:40 +0100

Generalized finite element method (GFEM) has proven itself as a tool of choice over the conventional FEM in fracture analysis due to enhanced computational efficiency as well as allowing cracks to propagate independently of the domain mesh Thanks to the use of enrichments chosen based on the a priori knowledge of the solution behavior. With the many versions of the method's formulations in the literature, their stability issues, compared to the standard FEM, are often unresolved. This paper presents the use of an adaptive stable GFEM to plain concrete fracture propagation. Having verified the formulation's accuracy and stability based on the Linear Elastic Fracture Mechanics in previous studies and its two-scale (global-local) version on concrete fracture, the present work seeks to verify its capabilities in capturing the size effect behavior in concrete. A set of fracture simulations in geometrically similar experimental concrete beams, under a 3-point bending regime, is presented based on a bilinear cohesive model. In addition to the GFEM's agreement with the experimental load-displacement response and the effect of the initial notch-to-depth ratio, the simulation successfully captures the size effect behavior when presented on the popular Type II Size Effect plot the so-called Bazant's law.

Inverse Dynamics of Geometrically Exact Beams

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:41:25 +0100

This paper is concerned with the inverse dynamics of flexible mechanical systems whose motion is governed by quasi-linear hyperbolic partial differential equations. Problems that appear by applying classical solution strategies to the problem at hand, e.g. integrating the problem at hand sequentially in space and time will be adressed in this work. Motivated by the hyperbolic structure of the underlying initial boundary value problem, two methods that are based on a simultaneous space-time integration will be presented. Special emphasize will be given to the phenomena of wave propagation within geometrically exact beams and its relevance regarding the inverse dynamics problem.

Influence of Cooling Lubricants on the Interaction Between Indenter and Material Surface During Scratch Tests

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:41:09 +0100

Striving for the optimization and the increase of efficiency of various systems demands further developments of the classic manufacturing methods. Especially grinding processes, which are characterized by undefined cutting-edge geometries, reveal many fields where there still is a lack of understanding. In particular, the processes at and their effects on the individual abrasive grit are insufficiently researched and, therefore, do not allow sufficiently accurate behavior predictions. In order to optimize grinding processes and, ultimately, the resulting quality of the workpiece surface, it is necessary to look at the entire process in a holistic way. Due to the large number of influences to which the grinding process is subject, it is initially advisable to break down the process as far as possible into individual scratch tests and then gradually return to the overall process. One approach is the development and expansion of an FEM-based physical force model, which allows for the simulation and prediction of a scratch tests and, subsequently, also the entire grinding process with all relevant influencing factors. One of these influencing factors, which are essential but mostly unconsidered, are cooling lubricants, especially their tribologically favorable influence on the interaction between workpiece and indenter. Therefore, it is important to identify and investigate the different aspects, such as the friction phenomena of scratch tests that are influenced by the use of cooling lubricants. In addition to temperature and force characteristics, which have been found to differ with and without cooling lubricant, differences in the scratch geometry on the material surface have also been observed in recent tests. Based on these findings, this work examines the relationship between scratch geometry and cooling lubricant. It turned out that scratch tests conducted with cooling lubricants have an influence on the topography of the scratch on the workpiece surface in addition to the influence on the tangential and normal forces. The ratio of scratch width to scratch depth is used for evaluation. A reduction of this ratio is observerd in the scratches with cooling lubricants and is, therefore, interpreted as a reduction of the scratch width as a result of the use of cooling lubricants.

Simulation of phase transformations in polycrystalline shape memory alloys using fast Fourier transforms

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:40:50 +0100

The properties of shape memory alloys (SMA) are mainly influenced by phase transformations between austenite and martensite. The complex material behavior is described by a variational method which describes the evolution of the phase fractions. We combined the method with a microstructural analysis based on fast Fourier transformations. Such a highly resolved microstructural analysis comes along with a high computational effort.To reduce the later one, we propose a model order reduction technique that uses just a reduced set of Fourier modes, which is adapted to the underlying microstructure. The presentation of the theoretical background as well as of the implemented algorithm is followed by numerical results that underline the performance of our method.

Using a reduced set of Fourier modes in terms of a FFT-based microstructure simulation

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:40:32 +0100

Given a heterogeneous material, the mechanical behavior of its microstructure can be investigated by an algorithm that uses the Fourier representation of the Lippmann-Schwinger equation. Incorporating a model order reduction technique based on calculations with a reduced set of Fourier modes, the computational cost of this algorithm can be decreased. It was shown that the accuracy of this model order reduction technique strongly depends on the choice of Fourier modes by considering a geometrically adapted rather than a fixed sampling pattern to define the reduced set of Fourier modes. Since it is difficult to define a geometrically adapted sampling pattern for complex microstructures, additionally a strain-based sampling pattern was introduced. The accuracy and adaptability of this strain-based reduced set of Fourier modes is shown by incorporating a polycrystalline microstructure.

Fast and simple creation of powder beds for selective laser melting

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:40:14 +0100

In selective laser melting, components are produced by layer-by-layer melting of a powder bed. To investigate the interaction between the powder bed and the laser energy in the process, it is necessary to generate different powder bed configurations with a defined particle size distribution. For this purpose, based on different particle contact approaches, a simple algorithm for planar particle bed configurations was developed in the Julia programming language. Based on the so called 0and 1-particle-contact approaches, a monodisperse sphere packing with a filling ratio of up to 64%, and with a normally distributed particle size with a filling ratio of up to 67% were generated. With the 0-particle-contact approach, the individual powder beds could be generated more quickly, but showed an insufficient degree of filling. In contrast, the 1-particle-contact approach can produce powder beds realistically. An extension for spatial problems, as well as variations in the contact approaches, is given by the simple algorithm design and shall be implemented and further investigated in simulations of selective laser melting.

Efficient assessment of noise transmission through highly flexible slender structures

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:39:55 +0100

In this work, we propose an efficient methodology for the assessment of noise transmission through cables and hoses. An interactive simulation with a geometrically exact Cosserat rod enables simple and fast modelling of various configurations. Subsequently, we linearise the equations of motion at the static equilibrium for given boundary conditions and, using the resulting system matrices, compute the mechanical impedance matrix. The computation result, i.e. the impedance matrix, is available within seconds. The impedance matrix either can be used to compute reaction forces for given excitation or, if the excitation is unknown, allows to analyse the transmission of noise by looking at single matrix elements. The latter is especially useful in early, purely virtual development phases.

Homogenization of the constitutive properties of composite beam cross-sections

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:39:36 +0100

When modelling slender bodies made of composite materials as beams, homogenized stiffness coefficients must be obtained. In [2, 3], analytic expressions for these are obtained by comparing the solutions of some Saint-Venant extension, bending and torsion 3D linear elasticity problems with their corresponding beam theory counterparts. In [2], the authors provide general expressions for the determination of these coefficients for multilayered beams. The present work consists in the study of a homogenization procedure of the stiffness coefficients for circular cross-sections with two layers. This will help in the study of the constitutive behavior of unloaded shafts of endoscopes since their cross-section could be studied as a simplified model of a three-layers hollow circular cross-section. In preparation of this geometry, results of an experimental campaign carried out at KARL STORZ GmbH & Co. KG (Tallinn, Estonia) are presented in a second part of this paper. The purpose of the testing was the experimental characterization of the torsional stiffness of such devices.

Modeling the Effective Inelastic Behavior of Multi-Wire Cables Under Mechanical Load Using Finite Elements

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:39:12 +0100

In the development and manufacturing process of modern cars, cables and hoses are important system components. In automotive industry, virtual assembly planning and digital validation of system layouts require a fast and physically correct simulation of the mechanical behavior of cables and hoses. The mechanical response of cable systems and hoses under load is typically non-linear and inelastic due to their multi-component structure. However, those effects can hardly be observed and investigated separately in experiments. Thus, the authors recently presented simplified cable models using the commercial finite element tool ANSYS which take wire interactions into account. As cables in automotive applications are often subject to large deformations, finite beam elements with quadratic shape functions were used to discretize the single helix wires. The comparison of simulation results obtained for helix wire strands under bending with analytical results based on wire rope theory showed good agreement for the case of frictionless interactions. Furthermore, the modeling approach serves as a versatile toolbox for the investigation of material and structural inelastic effects which commonly occur when cables are deformed. A significant influence of structural parameters, such as the helix angle of the wires or the choice of friction model parameters, on the mechanical response could be found. In this work, this modeling approach is applied to the simulation of multi-wire strands consisting of parallel elastic wires under bending and torsion. The results of these mesoscopic simulations will be compared to experimental results.

Estimating stress fluctuations in polycrystals with an improved maximum entropy method

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:38:53 +0100

The prediction of local field statistics from effective properties is an open problem in the field of micromechanics. Partial information on the local field statistics is accessible from homogenization assumptions. In particular, exact phase-wise second moments of stresses can be calculated analytically from the effective strain energy density. In recent years, full-field calculations have become efficient enough to sample large ensembles of microstructures in the plastic regime (e.g. Gehrig et. al [4]). In the present work, the maximum entropy method known from statistical thermodynamics is used to estimate first and second moments of local stresses from known eigenstrain distributions. The simple and refined formulations of the maximum entropy method proposed by Kreher and Pompe [9] are considered. While the simple method yields satisfactory results for a large amount of material classes (cf. Krause and Böhlke [7]), we prove that it does not respect the linearity of the eigenstrain problem. We further show that neither method corresponds to the exact second moments of stresses known from the effective strain energy density. By incorporating additional information, we find an improved maximum entropy method. As an example, we analyze stress fluctuations in polycristalline titanium.For the exact analytical solution and the maximum entropy methods, we use the singular approximation and the Hashin-Shtrikman bounds. For comparison, we numerically approximate full-field statistics using an FFT approach. In all methods, the stress fluctuations caused by the anisotropy of the single crystal strongly influence the elastic-plastic transition.

Quantifying errors due to the Hertzian contact model in multi-sphere Discrete Element Modelling simulations

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:38:35 +0100

In Discrete Element Modelling (DEM), non-spherical particles are often emulated using clusters of rigidly connected spheres that can be either overlapping or not. Within this multi-spherical approach, two sources of error directly affecting the normal contact forces present can be identified. One is due to the difference between the true particle shape and the multispherical approximation; the other arises from the contact model used in the DEM simulations. The potential for inaccuracy in multi-sphere DEM simulations is well known. However, for a DEM simulator it remains unclear what error might be expected when multi-sphere particles are adopted. This contribution focuses on the role of the contact model as a source of error for the special case of two-sphere particles. Considering a single multi-spherical rod consisting of two identical spheres and quasi-statically compressing it until a total of 5% strain has been applied, the force response obtained through Finite Element Analysis (FEA) was compared against the Hertzian contact model. The process was repeated for spheres with varying degrees of overlap, ranging from 0 to 100%, and the relative errors of the FEA models against the Hertzian contact model were calculated. Up to a 60% sphere overlap, Hertz underpredicts the normal contact forces beyond 0.5% strain, where the disparity between the FEA model and Hertz forces is increasing monotonically with strain. However, beyond an overlap of 70%, Hertz overpredicts the normal contact forces with the sphere overlap being the main driver of this deviation. Future research will involve comparing these errors with the `shape' source of error by compressing perfect spherocylinders, considering rods composed of more spheres, and investigating how the error is affected by the relative orientation of two contacting particles.

Simulating Nonlinear Elastic Behaviour of Cables Using an Iterative Method

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:38:16 +0100

This contribution introduces a novel approach to simulate the nonlinear elastic bending behaviour of cables. Bending tests on real cables with complex structures clearly show the existence of nonlinear constitutive bending behaviour. In our current framework, only constant effective stiffness parameters are used. In order to enable nonlinear bending behaviour within the current framework, we propose an iterative method where, at each step, constant stiffness parameters are used and are updated according to the current cable state. The presented method is demonstrated by means of numerical experiments. Moreover, the inverse problem, i.e. the determination of state-dependent bending stiffness characteristic, is considered.

Director-based IGA beam elements for sliding contact problems

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:37:59 +0100

The geometrically exact beam theory is one of the most prominent non-linear beam models. It can be used to simulate aerial runways or pantograph-catenaries, where a sliding contact condition between two or more beams is used. A smooth discretization of at least C1continuity is needed to not introduce any unphysical kinks. This can be achieved using the isogeometric analysis, which we apply to a director-based formulation of the geometrically exact beam. For a stable time integration scheme we use an energy-momentum conserving scheme. Using the notion of the discrete gradient, an energy-momentum conserving algorithm is constructed, including the case of sliding contact between beams.

Nonlinear seismic analysis of reinforced concrete structures using POD reduced order method

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:37:42 +0100

This paper presents an extension of the Proper Orthogonal Decomposition method (POD) to nonlinear dynamic analysis of reinforced concrete multistory frame structure where the material nonlinearity is modeled by the multi-fiber section. To test the effectiveness of this approach, we first perform a nonlinear dynamic analysis under a seismic excitation using a direct implicit time integration scheme. Then, based on structural response observations, POD modes were extracted and used to reduce the structural system subjected to different earthquakes. A comparison was made between full model and reduced model analysis in order to assess the effectiveness of this technique.

Simulation of non-round particles in tribological three-body systems

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:37:25 +0100

Particles between the contact interfaces of two components in relative motion are present in many technical applications and can strongly influence the system behavior. In this context, the focus is often on the investigation of wear and damage. In addition to such undesirable phenomena, however, there is also the targeted use of hard particles, for example in the lapping process. In lapping, hard particles are intentionally inserted between a lapping disc and the workpiece surface to be processed in order to cause material removal with the help of the particles and to improve the morphology of the workpiece surface for certain applications. Many simulations of such tribological systems are based on the assumption of spherical particles. However, both, size and shape of the particles have an essential effect on the system behavior. Here, an approach is presented in which hard, arbitrarily shaped particles in tribological contacts can be studied a priori using the finite element method by performing indentation simulations for various particle orientations. Based on the results, an orientation-dependent particle model is created for simulations of the overall system, which includes particles in narrow gaps. This modular design allows direct control in the implementation of phenomenological effects and new insights into the behavior of such systems, as well as the estimation of the resulting surface topography.

Molecular dynamics investigation of the effect of interlayer cavities of the structure of calcium silicate hydrate at the atomistic scale

Jesús Sánchez Pinedo — Wed, 23 Nov 2022 10:37:09 +0100

Calcium silicate hydrate (CSH) gel, as the most important component of hydration products, has the most significant effect on the properties of hardened cement paste. One of the most critical factors affecting the mechanical properties of CSH is the interlayer cavities in the gel. In this study, the effect of these cavities on Young's modulus of CSH has been investigated. For this purpose, first, the atomic structure of CSH is created, and then interlayer cavities with different dimensions are created inside the structure. For modelling, first, a super cell with dimensions of 3 × 6 × 1 times the unit cell of Tobermorite is prepared, and then each of these layers are placed on both sides of the new cell, and a space is created between these two layers. This distance is basically the cavity between the layers.