Search

In this paper, a finite element procedure is used to study the dynamic behaviour of a paper web in a free span between two rollers, including effects of transport velocity and surrounding air. The paper web is modelled as a three-dimensional orthotropic structure. The influence of air is accounted for by utilizing fluid–solid interaction analyses based on acoustic theory. The contribution of transport velocity is included through gyroscopic matrices and forces. The structural response on harmonic excitations has been studied using linear and non-linear models. Results show that air significantly reduces eigenfrequencies of the web. So called “edge-flutter” is nothing but the result of skew tension profile. Excessive web vibration can be eliminated by adjusting the web tension. [Copyright &y& Elsevier]

A new shell element has been proposed for geometrically non-linear, finite element analyses of axially moving paper web. Web instability problems of paper travelling in a printing system, such as wrinkling and fluttering, pose special challenges in the numerical analyses. For examples, these are finite bending stiffness (as opposed to membrane), orthotropy, and non-uniform distributions of mechanical and hygro-properties (in different length scales). The new formulation has accounted for paper transport velocity and paper bending stiffness based on a mixed Lagrangian–Eulerian description of the motion. A natural coordinate system has been employed in the analytical and finite element formulations. Benchmarking with different finite elements in various tests showed that the proposed element is indeed more stable and reliable for the chosen application than existing elements. The usability of the shell element has been successfully demonstrated by two example problems: wrinkling of stretched isotropic and orthotropic membranes, and the vibration of a web showing a speed-tensioning effect above a critical web transport velocity. [Copyright &y& Elsevier]

• A novel surrogate-based crack identification method is presented. • Structure frequency shift characteristics is analyzed. • The crack identification is converted to an optimization problem. • Numerical examples and experimental validation prove satisfactory precision. A novel crack identification method is presented in this paper for cantilever beam-type structures within the scope of vibration-based damage identification. Natural frequencies are more easily to be measured yet are weakly sensitive to damage in comparison to other dynamic parameters such as mode shape and damping ratio. Therefore, this paper develops a surrogate-based crack identification method in order to overcome the low sensitivity of frequency to damage. The surrogate model was constructed using radial point interpolation method (RPIM) and subsequently trained with the frequency shift database established from parametric finite element analysis (FEA) modeling of cracked/ non-cracked cantilever beams. On this basis, the crack identification problem had been converted to a standard optimization model with crack length and location as design variables. Subsequently, genetic algorithm (GA) was employed for solving the optimization model to achieve crack parameters. In the end, numerical examples and experimental test have been conducted to validate the developed method, revealing maximum relative errors within 0.5% and 2.5%, respectively. [ABSTRACT FROM AUTHOR]

In this paper, the statistical finite element method is developed further to synthesize distributed rail response data with nonlinear finite element model predictions within and outside the measured loading range. In the data-generating model of the statistical finite element method, the distributed sensing data is decomposed into a finite element model component, a model-reality mismatch component, and a noise component. Each component is considered uncertain and is represented as a Gaussian random vector with a corresponding prior density. The finite element prior density is updated using the Bayesian statistical framework in light of the distributed fiber optic sensing data. The calculated posterior density enables one to infer the true structural response. The required finite element prior density is determined by solving a conventional stochastic forward problem using a polynomial chaos expansion of random fields and a non-intrusive pseudo-spectral projection approach. In this paper, a lab test involving a section of a rail (i.e., a beam-column structural member) instrumented with distributed fiber optic sensors and subjected to axial load causing progressive lateral buckling is considered to demonstrate the use of the statistical finite element method to improve rail response prediction. The results show improved prediction of true rail strain in linear and nonlinear regimes compared with a pure finite element model-based prediction. • Statistical finite element method enables integration of nonlinear model and sensor data. • Model selection/misspecification & data uncertainty incorporated using Bayesian formalism. • A polynomial chaos expansion is utilized to derive the necessary finite element prior. • The proposed approach is validated with progressive buckling tests and strain data. [ABSTRACT FROM AUTHOR]

• A methodology for analysis of large-scale masonry structures is presented which incorporates an inelastic macroscopic framework coupled with a series of artificial neural networks. • The neural networks serve the purpose of identification of material functions embedded in anisotropic formulation as well as the specification of localization plane. • The data required for training of neural networks is generated using 'virtual experiments' that involve a mesoscale finite element analysis of masonry wallets. • Finite element analysis of a large masonry wall with multiple openings is carried out. The results are compared with those based on a detailed mesoscale model. This paper presents a methodology for analysis of large-scale masonry structures. The approach involves development of a series of artificial neural networks which enable the identification of main variables employed in the macroscopic formulation that incorporates an inelastic constitutive law with embedded discontinuity. The data required for training of neural networks is generated using 'virtual experiments', whereby the 'equivalent' anisotropic response of masonry is obtained through a mesoscale finite element analysis of masonry wallets. The paper outlines the procedure for identification of approximation coefficients describing the orientation-dependency of strength, and other relevant parameters. A numerical example is provided involving analysis of a large masonry wall with multiple openings. The results of macroscale approach are compared with those based on a detailed mesoscale model for the same geometry and boundary conditions. [ABSTRACT FROM AUTHOR]

• A new methodology for multi-parameter identification of concrete dams is proposed. • For the first time, polynomial chaos expansion and slime mould algorithm are integrated. • A very high computational efficiency is achieved using the proposed methodology. This paper presents a novel methodology that combines polynomial chaos expansion and slime mould algorithm for multi-parameter identification of concrete dams. This methodology not only incorporates the merits of low computational cost in the polynomial chaos expansion and fast convergence of slime mould algorithm, but also considers the priori uncertainty in the input parameters by introducing statistical probability theory. By considering two examples with different complexity, this paper verifies the effectiveness of the proposed method with a univariate simply supported beam model, followed by a complex multivariate dam model to demonstrate its practicability in real engineering problems. In addition, parameter sensitivity analysis of the dam model is conducted at an extremely low cost by polynomial chaos expansion based on Sobol' indices. Furthermore, the conventional parameter identification methods based on optimization methods directly combined with the finite element model are employed for comparison, highlighting two distinct advantages of the proposed method: (i) the proposed method improves the computational efficiency by nearly 52 times while ensuring a high accuracy, and (ii) the classical non-population optimization algorithm, Bayesian optimization, is used for comparison, revealing the outstanding performance of slime mould algorithm in terms of convergence speed and robustness. The application of the proposed algorithm is not only limited to dams, but also it can be extended to any structure. [ABSTRACT FROM AUTHOR]

• This paper presents an improved algorithm for the Moore-Penrose continuation method. • Its ability to compute solution curves with critical points is greatly improved. • Difficulties observed using the standard approach are not encountered. • Easy implementation. Using typical solution strategies to compute the solution curve of challenging problems often leads to the break down of the algorithm. To improve the solution process, numerical continuation methods have proved to be a very efficient tool. However, these methods can still lead to undesired results. In particular, near severe limit points and cusps, the solution process frequently encounters one of the following situations: divergence of the algorithm, a change in direction which makes the algorithm backtrack on a part of the solution curve that has already been obtained and omitting important regions of the solution curve by converging to a point that is much farther than the one anticipated. Detecting these situations is not an easy task when solving practical problems since the shape of the solution curve is not known in advance. This paper will therefore present a modified Moore-Penrose continuation method that will include two key aspects to solve challenging problems: detection of problematic regions during the solution process and additional steps to deal with them. The proposed approach can either be used as a basic continuation method or simply activated when difficulties occur. Numerical examples will be presented to show the efficiency of the new approach. [ABSTRACT FROM AUTHOR]

A new approach is proposed to account for the impact of sprain variables for computational simulation of fracture in quasi-brittle materials. For an isotropic solid, these variables are equal to the Laplacians of the components of the displacement field, and they were introduced in a recent study by Zhang and Bazant to resolve spurious mesh effects for cases involving softening material response. This is possible by establishing localization-resisting forces which are conjugate to the displacements with respect to a sprain energy functional, defined as a function of the sprain variables. To enable the evaluation of the sprain variables in a conventional, C 0 continuum finite element approximation, this paper introduces a set of scalar fields, each being weakly equal to the Laplacian of one of the components of the displacement vector field. This leads to a straightforward implementation with standard continuum (solid) elements, eliminating the need for approximations satisfying higher continuity requirements or for ad-hoc, finite-difference algorithms which were adopted in the original pertinent work for the calculation of the sprain variables. • Use of localization-resisting forces in continuum simulations of strain softening, to eliminate spurious mesh-size effects. • These forces are governed by a set of scalar fields, weakly equivalent to the Laplacians of the displacement components. • Weak equivalence allows to obtain the Laplacians of the displacement field with conventional, C0 finite element approximations. • The method is found to eliminate spurious effects of mesh size and of mesh orientation with respect to crack propagation path. [ABSTRACT FROM AUTHOR]

Tangled (non-convex) elements, i.e. elements with negative Jacobian determinant, can lead to erroneous results in the standard finite element method (FEM). Constructing tangle-free, well-structured meshes for complex geometries is often impossible. Hence there is a need to explore analysis methods that can directly handle such tangled meshes. In this paper, we propose the isoparametric tangled finite element method (i-TFEM) for free and forced vibration problems over tangled meshes. By employing piece-wise invertible mapping, a variational formulation is derived, leading to a simple modification of the standard FEM stiffness and mass matrices with the incorporation of additional compatibility constraints. Moreover, i-TFEM reduces to standard FEM for non-tangled (regular) meshes. The proposed method is implemented for three types of elements: 4-node quadrilateral, 9-node quadrilateral, and 8-node hexahedral elements. The numerical results demonstrate that i-TFEM is able to consistently handle general tangled (non-convex) elements, enabling convenient meshing for complex geometries. • Isoparametric tangled finite element method (i-TFEM) is proposed for free and forced vibration problems over tangled meshes. • To handle concave elements using i-TFEM, simple modifications to the standard FEM stiffness and mass matrices are needed. • I-TFEM is implemented for three element types: 4-node quadrilateral, 9-node quadrilateral, and 8-node hexahedral elements. • The numerical results demonstrate that i-TFEM consistently handles general non-convex elements. [ABSTRACT FROM AUTHOR]

Highlights • A new and robust numerical method based on use of isoparametric elements is proposed for solving general thermal and mechanical problems. • Only one element is used for each collocation node and the element can be freely formed by itself and surround nodes. • The bandwidth of the formed final system coefficient matrix is extremely narrow. • A new 21-node quadratic element for 3D problems is constructed in this paper for the first time. Abstract In this paper, a new numerical method, named as the Free Element Collocation Method (FECM), is proposed for solving general engineering problems governed by the second order partial differential equations (PDEs). The method belongs to the group of the collocation method, but the spatial partial derivatives of physical quantities are computed based on the isoparametric elements as used in FEM. The key point of the method is that the isoparametric elements used can be freely formed by the nodes around the collocation node. To achieve a narrow bandwidth of the final system of equations, elements with a central node are recommended. For this purpose, a new 21-node quadratic element for 3D problems is constructed for the first time. Attributed to the use of isoparametric elements, FECM can result in more stable results than the traditional collocation method. In addition, the elements can be freely formed by local nodes, FECM has the advantage of mesh-free methods to fit complicated geometries of engineering problems. A number of numerical examples of 2D and 3D thermal and mechanical problems are given to demonstrate the correctness and efficiency of the proposed method. [ABSTRACT FROM AUTHOR]

• Higher-order layer-wise plate models for composite variable angle tow plate analysis. • Passive damping using viscoelastic inter-layers with frequency-dependent properties. • Mixed interpolation of tensorial components for more efficient and accurate solutions. In this paper, the damped free-vibration and frequency response analysis of variable-angle-tow composite plates embedding viscoelastic layers with frequency-dependent properties are performed. The governing equations are derived from the Principle of Virtual Displacements and higher-order Layer-Wise models are used for the unknown variables description in the thickness direction, furthermore a nine-node finite plate element is employed to solve them. The plate formulation presented in this paper is more efficient with respect to three-dimensional approaches, typically used for this kind of structures, and it is more accurate than existing reduced order methods thanks to the use of the mixed interpolation of tensorial components technique, employed to solve the problem of the shear and membrane locking phenomena. Different multilayered structures have been considered, laminated with unidirectional cross-ply layers or with curvilinear fibers path ones. The use of viscoelastic soft sheets permits to have a passive damping of the structural vibration. The frequency dependent properties of the viscoelastic materials are described through the use of a Kelvin-Voigt model. Numerical solutions are presented for the analysis of variable-angle-tow composites with different boundary-conditions and various lamination schemes. [ABSTRACT FROM AUTHOR]

• The collective performance of Masonry Aggregates affects the individual units. • A novel BIM to FEM framework for Masonry Aggregates is outlined. • The Level of Information Need for solid FEA of Masonry Aggregates is defined. • An automated process from IFC to tetrahedron finite elements is developed. • The developed methods follow Open BIM logic. Masonry building aggregates are complex historical structures comprising interconnected buildings. The safety of each building depends not only on its structural integrity but also on the interactions with neighbouring structures and their collective performance. The structural assessment of such complex constructions can benefit from BIM methodology. The latter eases modelling complicated solid geometries and assigning mechanical properties information essential for developing a reliable finite element model. Thus far, BIM to finite element analysis procedures applied to historic constructions have remained laborious, semi-automatic, and based on proprietary software. This paper introduces an open-source automated solid finite element analysis method using OpenBIM data (IFC). The tool implemented automatically creates tetrahedron mesh and assigns mechanical properties, self-weights, and fixities at the structure's base. The methodology is applied to an actual masonry building aggregate to validate its capability of working with complex geometries and scalability. [ABSTRACT FROM AUTHOR]

Exploring the wave propagation in nano-waveguides enables the development of on-chip nano-resonators, leading innovative capabilities that enhance acoustics and micro-mechanics. This paper conducts a numerical investigation into the dynamic properties of periodic nano-waveguides. The Second Strain Gradient (SSG) elasticity, which captures the size effects, is used in the determination of the constitutive relations. The weak form including the element matrices is deduced by the Hamilton's principle. Furthermore, a finite element method, based on the periodic structure theory, is introduced to explore the free wave propagation by solving the eigenvalue problems. The stiffness hardening phenomenon manifested in the dispersion curves, band structures, and energy flow vector fields is discussed. The sensitivity of higher order parameters existing in the SSG theory is analyzed. The numerical investigation for wave propagation in the periodic nano-waveguides through the SSG theory is an innovative work, highlighting the significance of the proposed methodology in explaining the special dynamic characteristics of complex nano-waveguides. • Modeling of 3D nano-waveguides by second strain gradient theory. • Eigenvalue problems within 2D wave finite element method framework. • 2D wave motion properties in two different nano-waveguides. • Sensitivity of higher order parameters existing in second strain gradient elasticity. [ABSTRACT FROM AUTHOR]

• FEM analyzes of two representative historic railroad arch masonry bridges are shown. • Careful selection of the masonry properties is crucial and always required. • Modern traffic loads may contribute to the damage of masonry railroad bridges. • Cohesive zone modeling provides efficient description of masonry cracking. • Broad analysis of numerous masonry bridges require effective approaches as shown. The problems encountered during the analyzes of structural response of historic masonry railroad arch bridges are described in this paper. The attention is mainly focused on the stiffness of the masonry arches, their strengths and appropriate estimation of railroad load intensity. Issues related to computational modelling of two, existing, almost 130 years old masonry arch railroad bridges are presented in this context. The main properties and results of detailed inspection of the structures are shown. Computational models that were created in the finite element method environment in order to efficiently describe the responses of the bridges under typical loading conditions and estimate their load carrying abilities are presented. The outcomes of several nonlinear static analyses that were conducted for this purpose are discussed. What is more the results of finite element analyses are reviewed against the inspected bridge condition and final conclusions are formulated on that basis. All the analyses allowed to find the possible causes of the deterioration of the bridges condition. [ABSTRACT FROM AUTHOR]

This paper presents a numerical model of coupled hydro-mechanical behaviour of partially saturated soils. The model is based on Bishop's effective stress concept and the Cam-Clay model, with the compression stiffness influenced by the effect of partial saturation. The mechanical model is accompanied by the simplified water transport model where the water pressure is the only variable and van Genuchten moisture retention curve was adopted. The governing equations are discretized in the framework of the finite element method, and the resulting model is implemented in the in-house open source software package SIFEL. To improve convergence of the Newton-Raphson method used at the global level of the equilibrium finding, the algorithmic stiffness matrix was derived for the used stress return algorithm. A numerical examples of foundation problem illustrates the effective application of the model and proved that the model can capture wetting collapse. • Coupled hydro-mechanical model for partially saturated soils. • Improvement of the model performance by the algorithmic stiffness matrix for the cutting plane stress algorithm. • The effect of the wetting collapse can be captured by the model, as demonstrated in the example. • Comparison of model performance on the partially and fully saturated cases of the footing strip problems. [ABSTRACT FROM AUTHOR]

This paper introduces a novel hybrid formulation of a peridynamics-based machine learning model for thermal diffusion analysis in one-dimensional and two-dimensional problems with evolving discontinuities. The hybrid model employs a multivariate linear regression approach to establish the relationship between the temperature values of material points, their neighboring points, and the applied external heat flux. The thermal modal analysis method uses the finite element method to generate training and testing data. An efficient numerical procedure is also developed to couple the peridynamics model and the peridynamics-based machine learning model. The model is analyzed under multiple configurations of micro-thermal conductivity functions for one-dimensional thermal bar problems, under both steady-state and transient loading conditions, to identify the configuration that exhibits superior convergence behavior towards the local solution. Furthermore, benchmark problems demonstrate the high accuracy of the formulated hybrid model, including the analysis of thermal plates with discontinuities, such as a plate with a hole, a plate with a pre-existing insulated crack, and plate with a bi-material interface with stationary and evolving discontinuances. The hybrid model effectively captures intricate discontinuities and boundaries while being computationally efficient, indicating its potential for thermal diffusion analysis in one- and two-dimensional problems with stationary and evolving discontinuities. • The peridynamics-based machine learning model is proposed for diffusion problems. • An efficient scheme is proposed to couple peridynamics and machine learning models. • Machine learning data is generated using modal analysis for different thermal conditions. • The efficiency and robustness of the model is demonstrated through numerical examples. • The model is straightforward to extend to diffusion problems associated with pitting corrosion. [ABSTRACT FROM AUTHOR]

• A multi-scale beam model is derived using FE2 framework and absolute nodal coordinate formulation. • It has both the efficiency of absolute nodal coordinate formulation and multi-scale framework. • Performance of the proposed method is demonstrated through numerical examples. In this paper, a multi-scale modeling approach for the dynamic response of flexible multibody system made of composite material is suggested using the absolute nodal coordinate formulation (ANCF) – multilevel finite element (FE2) method. On many occasions, the analysis of composite structures requires a multi-scale modeling approach. FE2 method is one of the famous multi-scale modeling methods. And it is a versatile computational homogenization approach that is commonly applicable to many different materials. Because the present FE2 is capable of computing the mechanical response at two different scales simultaneously, ANCF element will be used at the macroscopic scale. And the microscopic RVE at each macroscopic integration point will be obtained by the finite element method. Since ANCF element uses only the absolute nodes and gradients, the mass matrix will be constant. Due to such characteristics, when combined with FE2, the linearization process will become easier compared against the existing formulation. And Coriolis or centrifugal force will not need to be considered in the dynamic formulation. Therefore, computational cost will be reduced, and the relevant algebraic manipulation will become simplified. Several numerical examples are presented to demonstrate the improved accuracy and diminished computational cost of the present suggestion. [ABSTRACT FROM AUTHOR]

• Propose a smart component model replacement (SCMR) approach to simulate large-scale nonlinear structure. • The SCMR approach can finely simulate the local damage and failure mechanism in structures. • Present "replacement" strategies for smoothly replacing between different models. • The efficiency and accuracy of SCMR approach are verified by three application examples. This paper proposes a novel smart component model replacement (SCMR) approach that is not only efficient for large-scale nonlinear structural analysis, but also accurate in the refined simulation of components of interest. The method creates a FE model with a relatively coarse mesh to simulate the entire structure. During analysis, when a component is identified to predefined criteria defined by the users, its model will be replaced by another more refined "substructure" model that can simulate the damage and failure more accurately. A few "replacement" strategies are presented for smoothly transiting between different models: (1) A gradually transiting strategy is presented to avoid the possible jumps of responses due to the model differences to guarantee numerical stability; (2) The historical boundary conditions are applied to the substructure model to address the historical dependence of responses; (3) The boundary connection between main and substructure is considered by using a flexible client/server technology, to simplify the replacement. The SCMR approach is verified by three application examples. The results demonstrate that the proposed modeling approach can efficiently perform large-scale nonlinear structural analysis, while capturing the failure mechanism in RC structures well. [ABSTRACT FROM AUTHOR]

This paper presents a novel derivation of the direct parametrisation method for invariant manifolds able to build simulation-free reduced-order models for nonlinear piezoelectric structures, with a particular emphasis on applications to Micro Electro Mechanical Systems. The constitutive model adopted accounts for the hysteretic and electrostrictive response of the piezoelectric material by resorting to the Landau-Devonshire theory of ferroelectrics. Results are validated with full-order simulations operated with a harmonic balance finite element method to highlight the reliability of the proposed reduction procedure. Numerical results show a remarkable gain in terms of computing time as a result of the dimensionality reduction process over low dimensional invariant sets. Results are also compared with experimental data to highlight the remarkable benefits of the proposed model order reduction technique. • Model order reduction for structures with geometric and piezoelectric nonlinearity. • Direct computation of the invariant manifold from finite element problem. • Efficient technique accounting for electrostrictive response of the piezoelectric. • High accuracy of the results compared to full-order simulation. • Application to micro-electro mechanical systems with complex geometry. [ABSTRACT FROM AUTHOR]

• A modified stain energy release rate method is proposed to enable the method for solid element. • An algorithm is proposed to map the crack information obtained by vision method in the FE model. • The effectiveness and generality of the proposed framework are validated using three cracking FE configurations. Vision inspection information of damages in civil engineering structures is hard to be transformed into an analysis-available resistance model to conduct in-service condition assessment. This paper proposes an updating crack-embedded finite element modeling framework using periodical vision inspection information, which offers a convenient crack embedding approach for safety assessment of cracked reinforced concrete (RC) frame structures. For this purpose, a modified stain energy release rate (SERR) method is proposed to enable the method for solid element with high accuracy. Consequently, an algorithm is proposed to map the crack information obtained by vision method in the FE model. The combination of modified SERR method and proposed mapping algorithm could achieve flexibility matrix updating of cracked element. In addition, the effectiveness and generality of the proposed framework are validated using three cracking configurations. This research can provide a feasible framework for the rapid condition assessment of in-service civil engineering structures, showing a great prospect of intelligent disaster prevention. [ABSTRACT FROM AUTHOR]

This publication presents a novel numerical avenue for the assessment of masonry structures. To circumvent the labor-intensive finite element method modeling and pre-processing step, the isogeometric analysis is enhanced within this contribution to cope with the necessities for masonry structures. This incorporates the study of apparent kinematic shell formulations, including their respective properties. Additionally, the connection towards non-linear material laws is outlined, and an empirically evaluated approach for the element size regularization is proposed. This generic approach would be inherently valid for other size dependent constitutive material models, as e.g. concrete models. The current paper is explored in the context of a number of experimental problems. These validation examples are used to provide structural assessments that extend beyond the scope of the experiment. Finally, a potential example that provides an outlook towards anticipated applications is presented. • Advancing the isogeometric analysis to cope with shell-based masonry structures. • Proposing a damage model in the context of the isogeometric analysis. • Newly proposed element length regularization technique for NURBS. • Enhancement of parametric and CAD-integrated workflow for masonry analysis. • Comprehensive mechanical assessments of masonry structures. [ABSTRACT FROM AUTHOR]

We present novel overlapping finite elements for a new paradigm of solution proposed in our previous papers, see Bathe (2016) and Bathe and Zhang (2017). We give the formulation of the new overlapping elements and the solutions of basic numerical examples to investigate the robustness and efficiency of the new finite elements. The results show that the new overlapping elements are quite distortion insensitive and the numerical integration of the element matrices is efficient. The computational effort to integrate the matrices is much less than in meshfree methods. Finally, we illustrate the complete solution scheme of the new paradigm using the overlapping finite elements in the analysis of the bracket problem already considered in Bathe and Zhang (2017). While the paper proposes and studies new overlapping elements, we conclude that further research is needed to fully harvest the potential of the new analysis approach. [ABSTRACT FROM AUTHOR]

Predicting the mechanical response of components requires simplifications and idealizations that affect the fidelity of the results and introduce errors. Some errors correspond to the limited knowledge of intrinsic physical attributes while others are introduced by the modeling framework and mathematical approximations. This paper studies the dependence of the force-displacement response of threaded fasteners on modeling attributes such as geometry, material, and friction resistance using finite element simulations. A systematic comparison of 1 D , 2.5 D or 3 D computational models demonstrates the influence of model properties and the limitations of the methodologies. Finally, the paper discusses the sources of model inputs and model form errors for threaded fasteners. [ABSTRACT FROM AUTHOR]

• Strength-based structural/material layout optimization for solid reinforced concrete. • The design cost is minimized based on the individual price of concrete and reinforcement. • The optimization is strength-based, i.e., strength scaling parameters on the concrete and reinforcement ratios are used. • The material layout is defined based on the concrete strength scaling parameters and reinforcement ratios. Whether monetary, environmental, or otherwise, the cost of the structure is always an important aspect when designing reinforced concrete. A structure with a low cost is desirable, but other considerations must be made. First and foremost, the structures should be safe. The two conditions define the structural optimization problem that this paper seeks to solve as finding the structure with the lowest cost, which is also safe. Validation of reinforced concrete in the ultimate limit state is often performed using plasticity and strength-based methods. This paper combines Finite Element Limit Analysis (FELA) with a yield criterion that includes material layout variables to create a strength-based structural optimization framework for solid reinforced concrete structures. The yield criterion combines the Modified Mohr–Coulomb criterion, where variables scale the compressive strength, with the inclusion of smeared reinforcement, where variables scale the reinforcement ratios. The framework is presented and shown in examples using free optimization and examples where material layout variables in parts of the structure are coupled and optimized as a group. [ABSTRACT FROM AUTHOR]

• The use of the Embedded Finite Element Method (E-FEM) to simulate local material heterogeneities is analyzed in detail. • A new fully consistent weak discontinuity enhancement function is proposed. • Basic simulations are used to compare the classical approach, commonly used in literature, and the new fully consistent approach proposed in the paper. • A discussion evaluates both model performance and ease of implementation. This paper analyses in detail the use of the Embedded Finite Element Method (E-FEM) to simulate local material heterogeneities. The work starts by a short review on the evolution of weak discontinuity models within the E-FEM framework to discuss how they account for the presence of multiple materials within a single element structure. A theoretical basis is introduced through some mathematical weak discontinuity definitions and the Hu-Washizu variational principle, for then establishing a set of requirements for retaining variational and kinematic consistency for any weak discontinuity enhancement proposal. From a general definition of a displacement enhancement field, two particular enhancement functions are derived by considering different consistency requirements: one which has been typically used in previous works and other which truly possesses variational consistency. A discussion is held on enhancement stability properties and the impact to global finite element solution processes. In the end, numerical simulations are made to assess the performance of each of these enhancements on the task of modelling a classical bi-material layered 3D tension problem and a more realistic heterogeneous sample having spherical inclusions of different radii. The final discussion evaluates both model performance and ease of implementation. [ABSTRACT FROM AUTHOR]

A major difficulty in finite element analysis is the preparation of an effective mesh leading to a good response solution. In engineering analyses of complex components, oftentimes, much more time is spent to arrive at an adequate mesh than to obtain the solution of the established finite element model. The difficulty in meshing is due to the fact that finite elements need to abut each other and cannot overlap. This can lead to highly distorted elements, e.g. sliver elements, reducing the accuracy of solution. In practice, to establish an effective mesh, frequently, significant expertise in building meshes is needed, because great care must be taken in cleaning up the CAD geometry and preparing an effective mesh. The objective in this paper is to present a new finite element solution scheme including meshing in which the elements can overlap. The property that finite elements can overlap removes many of the meshing difficulties, leads to an effective meshing procedure and an overall easy-to-use solution scheme for an analyst or a designer. We first present the meshing scheme that we propose, which combines the use of traditional finite elements and overlapping finite elements. A particular feature is that the meshing procedure can be directly embedded in CAD driven solutions. We then present the theory used for the formulation of the overlapping finite elements and the coupling with traditional finite elements. We consider spherical and brick-shaped overlapping finite elements for which the theory is largely based on the formulation of the method of finite spheres. Finally, we illustrate the complete solution scheme in the analysis of some two-dimensional problems using the CAD geometry as the starting point. While the paper presents a new paradigm for analysis in CAD environments, with much potential, we realize that much further study and research is needed on some of the important ingredients of the method to render the complete procedure effective for general practical engineering analyses of static and dynamic problems. [ABSTRACT FROM AUTHOR]

The aim of this study is an alternative approach to structure response or modal analysis. The structure consists of one-dimensional bars with continuously distributed mass and stiffness. The analysis is considered on an abstract basis as a problem of a differential system on an oriented graph. This graph is a geometric representation of the investigated mechanical system, where elements of the graph are individual bars of the system, recti- or curvilinear. The system as a whole is fixed through boundary conditions or interconnected with other sub-systems. Hence the paper can be taken as a follow up to earlier works presented at the CC2013 and CST2014 Conferences, where full mathematical background dealing with a general problem has been discussed. This paper is focused on the problem of dynamics of a system with straight prismatic bars with uniformly distributed mass. Dissipation of energy is omitted in order to keep the formulation in the real domain. The detailed assembly procedure of the Dynamic Stiffness Matrix (DSM) and transformation from local to global coordinates is outlined and demonstrated. Conventional way of eigenvalue searching by means of discrete alternative of the Newton-Raphson method is sketched out and later two possibilities based on polynomial and hyperbolic approximations of the DSM elements are pointed out. Lambda matrices as a tool are introduced together with a couple of application possibilities. The Wittrick-Williams algorithm is discussed and applied to localize and facilitate the eigenvalues searching on the whole frequency interval investigated. Finally, an illustrative example of the eigenvalue analysis of a structure is included. Strengths and shortcomings of the approach are discussed. Some open problems and orientation of further investigation are briefly outlined. The rapid development of analytical and semi-analytical methods being based at DSM idea is commented and justified in comparison with FEM analysis. Even if a certain overlap exists, the DSM approach represents an irreplaceable alternative of analysis especially in broadband dynamic problems stemming from area of the stochastic dynamics, like wind, earthquake or traffic engineering. Problems of dynamic stability are easier to be managed as well, when using DSM oriented approaches together with related algorithms of the analytical and semi-analytical method family. Moreover, many problems of physics at the nano-scale emerged recently where these procedures are preferable to be used. [ABSTRACT FROM AUTHOR]

In the paper by Ham et al. (2014), the method of finite spheres enriched for transient wave propagation problems was used and a monotonic convergence of the calculated solutions with decreasing time step size was seen. This is an important property for practical analyses and different from what is seen in traditional finite element solutions. In this paper we explicitly show and study this characteristic through a dispersion analysis of the solutions calculated by the method of finite spheres using an implicit time integration method, the Bathe method. Another important property identified is that in uniform spatial discretizations for the problems solved, the calculated solution accuracy is almost independent of the solution direction considered. Numerical solutions of some wave propagation problems are given to demonstrate these attributes and show that with the schemes discussed very accurate response predictions can be obtained. [ABSTRACT FROM AUTHOR]

• A novel stochastic homotopy method is proposed to solve static problems involving arbitrarily distributed random variables. • The optimal homotopy series expansion is obtained by minimizing the stochastic residual error. • The proposed method shows better convergency and efficiency than the aPC method in the case of non-Gaussian random field. • This method is successfully applied to the stochastic static analysis of the large-scale engineering structure. The modelling of realistic engineering structures with uncertainties often involves various probabilistic distribution types, which bring forward higher requirements for the generality of stochastic analysis methods. This paper proposes a novel stochastic homotopy method to evaluate the static response of the structure with random variables of arbitrary distributions. In this method, a homotopy series expansion is used to approximate the stochastic static response, and the optimal form of the expansion can be determined by minimizing the residual error about the stochastic static equilibrium equation regardless of distribution types of random variables. The numerical results of a logarithmic function and a thin plate show that the new method exhibits excellent accuracy and stability compared to the homotopy stochastic finite element method depending on the sample selection. Compared with the arbitrary polynomial chaos method (aPC), the proposed method is more efficient under equivalent accuracy. On the other hand, as the expansion order increases, this new method shows better convergence than the aPC method and the perturbation stochastic finite element method in the case of non-Gaussian distributed random variables of large fluctuation. In addition, a cable-stayed bridge example illustrates the application of the proposed method on the large-scale structure. [ABSTRACT FROM AUTHOR]

• The wave input method has a deficiency in expressing complex incident site waves. • A random non-uniform wave input method is proposed for seismic response analyses. • The new method is expressible for complex incident waves and motion differences. To improve the deficiency of the wave input method (WIM) in expressing complex incident site waves, in this paper, a random non-uniform wave input method (RNUWIM) is derived from the use of the random incident waves and the visco-elastic boundary condition. The random incident waves are generated and adjusted by the comparison of the horizontal ground motion's response spectrum and the target spectrum, so that they are representative of the actual wavefield in project sites. Furthermore, the RNUWIM is implemented and verified by the finite element method. The result shows that the RNUWIM inherits the advantages of the WIM in simulating the wave propagation effect and the radiation damping effect. In the new method, the spatial variation of ground motions such as time delays, peak variation, and time-history shape differences can be simulated synthetically. The errors of RNUWIM in ground acceleration, velocity, and displacement at the reference point are respectively 3.23%, 0.66%, and 0.07% in one case, which proves the efficiency of the RNUWIM in seismic response analyses. [ABSTRACT FROM AUTHOR]

• The paper proposed a physics-informed DNN framework for forecasting the hysteretic curves of S-shaped dampers. • The proposed physics-informed DNNs were calibrated numerically and experimentally. • The physics-informed DNNs can be used for the seismic design and mechanical characteristics assessment of S-shaped dampers Numerical simulation that combines finite element methods and experimental data has been recognized as effective in modeling hysteretic behaviors and capturing the principle mechanical trend of passive energy dissipation devices. However, the seismic design and mechanical characteristics assessment for passive energy dissipation devices require a laborious effort and massive computational resources to tune mechanical-oriented parameters. Particularly, the potential risk of departure from the actual system is rising for the desirable seismic design of dampers due to the numerical model simplification and assumption. To eliminate the potential weakness of numerical models, this paper explores a surrogate model by implementing physics-informed deep neural networks (DNNs) to approximate hysteretic behaviors of S-shaped steel dampers. The proposed physics-informed DNNs mainly consists of recurrent neural networks (RNNs) and long short-term networks (LSTMs), which can encode the Bouc-Wen model into the direct graph and incorporate the effect of design-oriented geometry parameters. To validate the generality of the network, the optimization model was calibrated numerically and experimentally, respectively, which exhibits good performance in predicting nonlinear behaviors of different dampers with reasonable accuracy. The proposed physics-informed DNNs can be an alternative to relieve the laboriousness of the seismic design and mechanical characteristics assessment of passive energy dissipation devices. [ABSTRACT FROM AUTHOR]

This paper is concerned with the finite element simulation of debonding failures in FRP-strengthened concrete beams. A key challenge for such simulations is that common solution techniques such as the Newton–Raphson method and the arc-length method often fail to converge. This paper examines the effectiveness of using a dynamic analysis approach in such FE simulations, in which debonding failure is treated as a dynamic problem and solved using an appropriate time integration method. Numerical results are presented to show that an appropriate dynamic approach effectively overcomes the convergence problem and provides accurate predictions of test results. [ABSTRACT FROM AUTHOR]

In this paper, a solid finite element formulation applied to the analysis of functionally graded materials (FGMs) under finite elastoplastic deformation with mixed hardening is presented. The main novelty of this paper is the use of gradually variable material coefficients in finite elastoplastic strain regime. The numerical simulations are performed with the same constitutive models but with different variations of the material coefficients. Full integration and high order tetrahedral elements are used. Results show that high order elements and mesh refinement avoids general locking problems. Finally, the differences (regarding final results) between the homogeneous and the FG cases are depicted. [ABSTRACT FROM AUTHOR]

• CS-FEM is applied to the numerical simulation of viscoelastic fluid flow for the first time. • The gradient smoothing for viscoelastic fluid flow is arbitrary within each smoothing cell. • Two smoothed CBS schemes are presented to solve the viscoelastic equations. • The cell-based smoothing procedure is used to evaluate the viscoelastic fluid drag. This paper generalizes the cell-based smoothed finite element method (CS-FEM) to viscoelastic fluid flows under the fractional-step umbrella. The characteristic-based split (CBS) scheme is employed to decouple the Navier–Stokes equations and the Oldroyd-B constitutive model. Moreover, the discrete elastic-viscous-split-stress (DEVSS) procedure and stabilized pressure gradient projection (SPGP) are introduced into a second-order CBS framework. The projections of velocity and pressure gradients therein combine to form the DEVSS-G/CBS(B)-SPGP algorithm that further stabilizes the discretized model. Both proposed schemes are suitable for use in CS-FEM because they accept equal low-order interpolations for primitive variables. The spatial discretization is based upon bilinear four-node quadrilateral elements where all gradient-related items are smoothed without any methodological difficulties. In particular, the cell-based smoothing concept enables accurate evaluation of viscoelastic fluid force. The developed methods are validated against previously published data for several benchmark problems. A satisfactory agreement exists between the present and earlier results. [ABSTRACT FROM AUTHOR]

• Two new functionals are obtained for geometrical nonlinear analysis of beams. • Mixed finite element method is used with boundary terms effectively. • The structure of incremental formulation is given to solve nonlinear equations. • Comparison of mixed formulations for the planar curved beams is presented. In the present paper, two mixed type functionals are presented to analyze the geometrical nonlinear bending behavior of slender planar curved beam by Finite Element Method (FEM). Bernoulli-Euler beam theory is taken into consideration within the context of the small strains but large displacements. Non-linear bending of curved beams in-plane has been investigated in both local and Cartesian coordinate systems. The Euler-Lagrange equations of the Functional I and Functional II which give the equilibrium and constitutive equations are in relation with the Green-Lagrange strain tensor and the first-order approximation (consistent linearization) of the Green-Lagrange strain tensor, respectively. The details of this process are given in Appendix. In the formulation of the problem in Cartesian coordinate, in addition to the boundary terms, that expression has been used, also to obtain the strain quantities which should be in the constitutive expressions in-prior to the equilibrium equations for the first time in this work. For the solution of the non-linear equations of the problems "incremental formulation" has been used. The validity, reliability and robustness of the proposed functionals have been presented by solving sample problems found in the literature. [ABSTRACT FROM AUTHOR]

• A new sectional analysis tool for concrete member torque-twist response is presented. • Elastic shear strain distribution assumed for post-cracked torsional behaviour. • Framework is flexible, permitting any constitutive model for concrete or steel. • The approach is validated against a large set of experiments in the literature. • Detailed results obtained with lower computational costs then nonlinear finite element analysis. This paper presents a sectional analysis tool which can compute the complete torque-twist behaviour of reinforced concrete beams and shells. The cross section is modelled as a 2-D grid of triaxial elements representing the concrete and uniaxial elements representing the longitudinal reinforcement. Fixed strain patterns based on kinematic assumptions are used instead of using the finite element method to calculate strain distributions corresponding to sectional strains. Constitutive models for the concrete are based on the Modified Compression-Field Theory, though tension stiffening is neglected in the proposed model's current implementation. A key assumption used in this model is that the shear strain distribution caused by twists obtained from a linear elastic analysis can be used to describe the nonlinear behaviour of a member following cracking. The validity of this assumption when applied to rectangular sections is confirmed based on a study of 115 tests from the literature. Excellent agreement of peak strength and overall torque-twist behaviour are observed when comparing the model predictions and experimental data from these tests. Areas of future work are to improve the capabilities of the model are identified. [ABSTRACT FROM AUTHOR]

• The behavior of under-designed existing structures is investigated. • Traditional and innovative retrofitting are compared by a finite element analysis. • The innovative strategy provides a better nonlinear performance of the structure. • The presented strategy reduces the dynamic vulnerability of the existing structure. In the present paper the structural behavior of under-designed existing structures is investigated. The analyzed existing structure is assumed as originally designed in the past with respect to old codes and with respect to gravitational loads only. The under-designed structure is considered to be located in Italy in an area now classified by the new codes as high seismicity zone. In the investigation traditional and innovative steel braces are modeled and compared for the retrofitting of the structure by using a finite element analysis with a displacement based approach. With respect to a traditional retrofitting strategy the adopted retrofitting strategy with innovative steel braces allows a localization of the input seismic energy into the innovative steel braces which are endowed with dissipative properties and thus a better performance of the original existing structure is noted in the elastoplastic range. With respect to a traditional retrofitting system the innovative steel braces applied to under-designed existing structures show to provide a more robust nonlinear behavior of the under-designed structure and a significant reduction of the plasticization in the original existing structural frames when they are subject to dynamic events. These features are associated to a vulnerability reduction of the under-designed existing structure also at the ultimate limit state. [ABSTRACT FROM AUTHOR]

• First time a NEFEM formulation is applied in solid mechanics. • HDG-NEFEM enables super-convergence in domains with curved boundaries for linear elements. • HDG-NEFEM passes low and high-order patch tests in domains with curved boundaries. • Reliable error indicators for degree adaptive process driven by both displacement and stress. This paper proposes a new methodology for the solution of two dimensional linear elastic problems in domains with curved boundaries. The proposed method exploits the advantages of the hybridisable discontinuous Galerkin method to obtain an accurate approximation of both the displacement and the stress fields by solving a global problem that only involves the displacement field on the element edges as unknown. In addition, the methodology incorporates the exact boundary representation of the domain by means of the so-called NURBS-enhanced finite element method. Numerical examples are used to illustrate the three main advantages of the proposed method, namely the reproducibility of polynomials in domains with curved boundaries, the super-convergence of the solution even for linear approximation and the effectiveness and reliability of degree adaptive processes driven by displacement or stresses. [ABSTRACT FROM AUTHOR]

• Meshless Radial Basis Functions (RBF) are used to reconstruct the full displacement field. • Analytic differentiation of the RBF field to extract an accurate strain field. • Upscaling of strain and stress at the hot spot without refine and re-run the Finite Element model. • The RBF approach can be extended to experimental evaluated displacement field. • Proven on a wing box aeronautical structure. This paper presents a method of interpolation and regularisation of finite element (FE) results based on radial basis functions (RBF) capable to upscale at a finer resolution the displacement fields and, accordingly, the strains and the stresses obtained with a coarse mesh. RBF interpolation of nodal displacements supplies an analytical approximation of the field over the whole discretized domain, suitable for mathematical differentiation. The stress status can be retrieved starting from the given analytical expression of the strain. The introduced technique is tested on two-dimensional FE problems presenting stress raisers: a plate in traction with a central hole, a plate in traction with nine aligned holes and a compressed wing rib with three lightening holes. For all the investigated geometries we observe that numerical FE results obtained adopting coarse grids and treated with the proposed approach compares well to FE results obtained adopting finer grids. [ABSTRACT FROM AUTHOR]

• The paper describes a novel approach for the coupling of the material point method (MPM) with the finite element method (FEM). • The most widely used approach in the literature consists of treating the MPM body and the FEM body as two separate entities, handling the contact detection and description with appropriate algorithms. • We propose an alternative approach, where the two bodies are treated as a single continuum in a monolithic fashion inspired by fluid-structure interaction problems. • In this way, the interface between the bodies is automatically tracked and the contact can be accurately described with a simple algorithm. A monolithic coupling between the material point method (MPM) and the finite element method (FEM) is presented. The MPM formulation described is implicit, and the exchange of information between particles and background grid is minimized. The reduced information transfer from the particles to the grid improves the stability of the method. Once the residual is assembled, the system matrix is obtained by means of automatic differentiation. In such a way, no explicit computation is required and the implementation is considerably simplified. When MPM is coupled with FEM, the MPM background grid is attached to the FEM body and the coupling is monolithic. With this strategy, no MPM particle can penetrate a FEM element, and the need for computationally expensive contact search algorithms used by existing coupling procedures is eliminated. The coupled system can be assembled with a single assembly procedure carried out element by element in a FEM fashion. Numerical results are reported to display the performances and advantages of the methods here discussed. [ABSTRACT FROM AUTHOR]

Highlights • DQ-based mapping relationship between element nodes and quadrature points is established for each MSGT-based beam model. • The two elements are independent of shape functions and introducing three kinds of strain gradient effects. • The proposed Timoshenko element is able to overcome shear-locking phenomenon. • A lot of valuable analytical solutions are systematically provided for the first time. Abstract In this paper, the superiorities of finite element method (FEM) and differential quadrature method (DQM) are blended to construct two types of beam elements corresponding to modified strain gradient -Bernoulli and Timoshenko beam models respectively. The two elements, being independent of shape functions and introducing three kinds of strain gradient effects, possess 3-DOFs (degrees of freedom) and 4-DOFs separately at each node. The Lagrange interpolation formula is employed to establish the trial functions of deflection and or rotation at Gauss-Lobatto quadrature points. To realize the inner-element compatibility condition, displacement parameters of quadrature points are converted into those of element nodes through a DQ-based mapping strategy. Total potential energy functional for each beam model is discretized in terms of nodal displacement parameters. The associated differential quadrature finite element formulations are derived by the minimum total potential energy principle. Specific expressions of element stiffness and mass matrices and nodal vector are provided. Numerical examples concerning with static bending, free vibration and buckling of macro/micro-beams are presented to demonstrate the availability of the proposed elements. [ABSTRACT FROM AUTHOR]

• The spatially uncertain parameters are described by the interval field model with upper and lower bounds. • An interval iterative method is proposed for static response analysis of structures with spatial uncertainties which can effectively avoid the overestimation existing in traditional interval analysis. • The upper and lower bounds of the responses obtained by the proposed method could envelop the actual response bounds. Parameters with spatial uncertainties are very common in practical engineering, which could have a significant effect on structural performances. Different from the framework of probability theory, the interval field model describes the spatial uncertainty with upper and lower bounds rather than the probability distribution function or other statistical characteristics. By introducing the interval field model into finite element analysis and solving the resulting interval finite element equilibrium equation, the upper and lower bounds of structural responses can then be obtained. This paper proposes an interval iterative method for static displacement response analysis of structures with spatial uncertainties, effectively improving the overestimation existing in traditional interval analysis due to dependency problems and obtaining a compact interval envelope of the exact response bounds. Firstly, the spatially uncertain parameters described by the interval field are represented by the interval Karhunen–Loève (K-L) expansion, based on which the interval finite element equilibrium equation is formulated. Secondly, the interval finite element problem is decomposed into several subproblems, and an interval iterative method is developed for an envelope solution of the structural response bounds. Finally, the accuracy and efficiency of the proposed method are verified by several numerical examples. [ABSTRACT FROM AUTHOR]

• A finite element formulation for time analysis of free layer damping plates is proposed. • The damping layer is modelled by a five-parameter fractional derivative model. • The proposed formulation is based on the homogenisation of plates. • Grünwald-Letnikov and Newmark methods are used for the numerical integration. • The validation is completed by means of a model with hexahedral finite elements. This paper presents a finite element formulation specifically developed for the transient dynamic analysis of free layer damping plates, in which the viscoelastic behaviour of damping layers is modelled by means of fractional derivatives. This formulation is based on the homogenisation of thin multi-layered plates, adapted to consider fractional derivatives. Then, a fractional matrix equation system is reached, and the time response is solved by means of the implicit Newmark method in conjunction with the Grünwald-Letnikov discretisation of the fractional derivatives. The validation of the method is carried out by comparing the results of three different numerical applications with the ones provided by a finite element model built with hexahedral elements. As conclusion, it can be pointed out that the proposed finite element formulation based on the homogenisation of multi-layer plates reproduces accurately the time response of plates with surface damping treatments. In addition, this method allows conventional plate mass and stiffness matrices to be used to model each individual layer, without requiring specific formulations. [ABSTRACT FROM AUTHOR]

• Developing seismic demand prediction models via SR methods in a probabilistic manner. • Proposing the MGGP and MBBP based models for seismic demand prediction. • Considering aleatoric and epistemic uncertainties for SR model development. • Using evolutionary correlation coefficient for the feature selection process. • Using closed-form mathematical expressions for generating bridge fragility curves. The resilience assessment of existing structures and designing new resilient ones need to have a precise estimation of seismic responses of the structures traditionally carried out by time-consuming calculations of finite element method. This paper presents interpretable, reliable, and fast prediction models developed by two novel branches of the symbolic regression approach (multigene genetic programming and multi biogeography-based programming) in a probability manner for curved bridges. These newly developed techniques seek to find interpretable equations with arbitrary forms that conform to a specific dataset. The reliable prediction models were developed considering the different uncertainties involved, including mechanical, geometrical, structural, and seismic uncertainties. Due to a large number of input variables for the bridge, the evolutionary correlation coefficient was employed to identify the most influential parameters for seismic demands of bridge components. Parameters with the highest correlation were set as inputs for symbolic regression algorithms, generating closed-form mathematical expressions to predict seismic demands. The resulting explicit models present a fast and accurate model that do not require further simulation or the seismic demand model assumptions to develop bridge fragility curves. In comparison to cloud approach and multiple strips analysis, the new method generates fragility curves with less dispersion. [ABSTRACT FROM AUTHOR]

• Data-driven analysis of geometric and material nonlinear behavior for frame structures is proposed. • A data-driven approach can be used to substitute the classic finite element nonlinear analysis. • The dataset is collected from the nonlinear analysis and a yield surface equation. • Six ensemble learning methods are used to train and validate the dataset of nonlinear frame analysis. • Yield surface values at all plastic hinge locations can be predicted using proposed method. Although many different nonlinear analysis techniques for steel frame structures, they causes the increase in total computational cost. A data-driven approach can then be used to substitute the classic finite element nonlinear analysis including second order methods and inelastic modelling. In this paper, a novel data-driven method using ensemble learning models for geometric and material nonlinear analyses of frame structures is proposed. For this approach, the data acquisition process and machine learning-based models are needed to train a large amount of dataset, which is collected from the nonlinear analysis and a yield surface equation. Ensemble learning algorithms are then used to build the data-driven models. After training, the sequence of plastic hinge formation in steel frame structures can be predicted without any nonlinear analysis process. In that situation, the data-driven model is more effective, especially in practical cases with the lack of experimental information due to limited sensors. The validity of the proposed method has been demonstrated by several numerical examples. [ABSTRACT FROM AUTHOR]

• Optimization framework for the modified shear strength reduction method with the pore pressure. • Evaluation of three different Davis? modifications of the shear strength reduction method. • Mesh adaptive solution for stability analysis coupled with the unconfined seepage problem. • Development of in-house Matlab codes and their validation against Plaxis and Comsol Multiphysics. • Solution of embankment stability problems following from geotechnical practice. In this paper, an optimization variant of the shear strength reduction method is introduced and used for the solution of embankment stability problems with unconfined seepage. The optimization framework is based on approximations of non-associated Mohr–Coulomb plastic models with associated ones, especially by using various Davis' approaches. Next, the finite element method is considered and mesh adaptive solution concepts are developed for both the unconfined seepage and stability problems. In-house codes in Matlab are used for their implementation. Finally, two numerical examples inspired by geotechnical practice are investigated in order to demonstrate the accuracy of the optimization framework and to evaluate three different Davis' approaches. The results are compared with commercial codes in Plaxis and Comsol Multiphysics. [ABSTRACT FROM AUTHOR]

• A three-stage approach to approximate non-stationary and discontinuous functions. • Clustering of joint input–output to identify discontinuities or different behaviors of the system. • Dirichlet process mixture models to automatically infer the number of clusters. • Support vector machines to partition the input space according to identified clusters. • Gaussian process regression to make predictions in identified sub-domains. Surrogate models have shown to be an extremely efficient aid in solving engineering problems that require repeated evaluations of an expensive computational model. They are built by sparsely evaluating the costly original model and have provided a way to solve otherwise intractable problems. A crucial aspect in surrogate modelling is the assumption of smoothness and regularity of the model to approximate. This assumption is however not always met in reality. For instance in civil or mechanical engineering, some models may present discontinuities or non-smoothness e.g. , in case of instability patterns such as buckling or snap-through. Building a single surrogate model capable of accounting for these fundamentally different behaviours or discontinuities is not an easy task. In this paper, we propose a three-stage approach for the approximation of non-smooth functions which combines clustering, classification and regression. The idea is to split the space following the localized behaviors or regimes of the system and build local surrogates that are eventually assembled. A sequence of well-known machine learning techniques are used: Dirichlet process mixtures models (DPMM), support vector machines and Gaussian process modelling. The approach is tested and validated on two analytical functions and a finite element model of a tensile membrane structure. [ABSTRACT FROM AUTHOR]

• A mass source term is derived from the modified continuity equation. • Geometric conservation law is satisfied in edge-based smoothed finite element method. • An efficient and stable semi-implicit coupling scheme is presented accordingly. • Viscoelastic fluid–structure interaction is modeled by variable-node smoothing cells. In this paper, a new partitioned semi-implicit coupling algorithm is developed for unsteady viscoelastic fluid–structure interaction in association with an edge-based smoothed finite element method (ESFEM). Stabilization techniques tailored for the viscoelastic fluid subproblem are naturally integrated into the hybrid explicit/implicit coupling framework to achieve the enhanced stability. In particular, a mass source term derived from the modified continuity equation in the finite-element context enables the ESFEM to fulfill the geometric conservation law on variable-node smoothing cells during the subiterations. Moreover, both multi-field equations and interface conditions are easily discretized with the ESFEM which enjoys high flexibility of smoothed Galerkin weak-form integral. Desirable numerical results are given for two benchmark problems. [ABSTRACT FROM AUTHOR]

• The combined fiber-interface model is established based on the theory of micromechanics. • The multiple cracking behavior of SHCC is achieved via finite element simulation. • The method of critical matrix cracking strength is proposed to estimate the potential of bending strain hardening. Strain-hardening cementitious composites (SHCC) are a class of fiber-reinforced cementitious composites characterized by strain hardening behavior accompanied by multiple cracking, which is designed by tailoring matrix, fiber, and fiber/matrix interface based on the micromechanics theory. At present, there are several numerical simulation methods to study the multiple cracking behavior, such as discrete crack model, unity finite element method, 3D smeared crack model, multiscale constitutive models, and lattice model. However, it is difficult to provide a good balance between accuracy and computational efficiency due to the vast number of fiber and the complex relationship between fibers, matrix, and interface. This paper developed a combined fiber-interface constitutive model based on the single fiber pull-out theory, and established a combined unit. A three-dimensional two-phase finite element model was established, which had been proved to be effective on simulating the cracking behavior of SHCC. Furthermore, the effects of fiber volume fraction and matrix cracking strength on the four-point bending properties of SHCC were investigated. Further study showed that there was a critical matrix cracking strength for a specified fiber and matrix. This value can optimize the design of SHCC to achieve high strength and high toughness performance. [ABSTRACT FROM AUTHOR]

• Beam finite element-based buckling analysis of perforated thin-walled members. • Displacement mode-based beam model allowing for the cross-section discontinuity. • Enhanced approximation of element displacement field by using level-set functions. • Decomposition of any coupled buckled shape to a set of pure buckling modes. • Illustrative examples showing the accuracy and efficiency of the proposed approach. This paper presents a numerically efficient tool for linear buckling predictions of perforated thin-walled bars within the framework of generalized beam theory (GBT). GBT-based beam finite element methods (GBT-FEM) have been well developed for eigenvalue buckling analyses of non-perforated thin-walled bars. The novelty of this paper consists in an extension of the standard GBT to the scope of thin-walled members with arbitrarily shaped and placed holes. This is achieved by combining the standard GBT and the extended finite element method (X-FEM). More specifically, insert a set of locally supported enrichment functions, accounting for the discontinuities on displacement fields arising from the cross-section cut-outs/holes, into the GBT-based finite element approximations of the member configuration spaces, using the partition of unity method (PUM), where a set of level-set functions are used to describe the geometric profiles of hole edges, i.e., with the hole edges being the zero level sets, and also used to construct the enrichment functions. The proposed approach makes it possible to calculate the deformation mode participations for any perforated thin-walled members as the classic GBT for non-perforated ones. Finally, the proposed approach is calibrated against the shell/solid finite element analysis with four illustrative examples. It can be found that the presented approach is of higher computation efficiency than the shell model. [ABSTRACT FROM AUTHOR]