Jaroslav Kruis

Efficient computer implementation of coupled hydro-thermo-mechanical analysis

Hydro-thermo-mechanical analysis of reactor vessels based on the finite element method is a very demanding task due to its complexity as well as the large number of unknowns. This contribution deals with efficient computer implementation of the coupled analysis and attention is also devoted to domain decomposition methods which enable utilisation of parallel computers. The parallel processing leads to very good speedup and it also enables to solve significantly large problems in acceptable time. The proposed strategy is demonstrated on a coupled analysis of an existing reactor vessel.

Adaptive hp-FEM with dynamical meshes for transient heat and moisture transfer problems

We are concerned with the time-dependent multiphysics problem of heat and moisture transfer in the context of civil engineering applications. The problem is challenging due to its multiscale nature (temperature usually propagates orders of magnitude faster than moisture), different characters of the two fields (moisture exhibits boundary layers which are not present in the temperature field), extremely long integration times (30 years or more), and lack of viable error control mechanisms. In order to solve the problem efficiently, we employ a novel multimesh adaptive higher-order finite element method (hp-FEM) based on dynamical meshes and adaptive time step control. We investigate the possibility to approximate the temperature and humidity fields on individual dynamical meshes equipped with mutually independent adaptivity mechanisms. Numerical examples related to a realistic nuclear reactor vessel simulation are presented.

Computational homogenization of non-stationary transport processes in masonry structures

A fully coupled transient heat and moisture transport in a masonry structure is examined in this paper. Supported by several successful applications in civil engineering the nonlinear diffusion model proposed by Kunzel (1997) [16] is adopted in the present study. A strong material heterogeneity together with a significant dependence of the model parameters on initial conditions as well as the gradients of heat and moisture fields vindicates the use of a hierarchical modeling strategy to solve the problem of this kind. Attention is limited to the classical first order homogenization in a spatial domain developed here in the framework of a two step (meso-macro) multi-scale computational scheme (FE^2 problem). Several illustrative examples are presented to investigate the influence of transient flow at the level of constituents (meso-scale) on the macroscopic response including the effect of macro-scale boundary conditions. A two-dimensional section of Charles Bridge subjected to actual climatic conditions is analyzed next to confirm the suitability of algorithmic format of FE^2 scheme for the parallel computing.

Parallel modeling of hygrothermal performance of external wall made of highly perforated bricks

Abstract Coupled heat and moisture transport in highly perforated bricks is indispensable part of design of energy efficient buildings. Geometry of the perforated bricks is very complicated which results in large number of nodes and elements in numerical analysis. Moreover, material model of the coupled heat and moisture transport leads after discretization to nonsymmetric systems of algebraic equations which need large computer memory. In order to reduce the computational time or to solve problems with many degrees of freedom, parallel computers are employed. Parallelization is based on the Schur complement method which is able to deal with nonsymmetric systems. Example of the coupled heat and moisture transport in a perforated brick of HELUZ company is showed. Real climatic boundary conditions for two different locations are used.

Solving laminated plates by domain decomposition

The refined Mindlin-Reissner theory is used to estimate the overall response of composite plates. The difficulties with the solution of a system of algebraic equations, which emerged in analysis of composite materials, are studied and a special version of decomposition is proposed. Similarity between the system of equations derived from the layered theory and from the finite element tearing and interconnecting method suggests a strategy for implementation in the parallel environment. Several applications are investigated and a number of numerical results are presented.

MuPIF - A distributed multi-physics integration tool

This paper presents the design of a multi-physics integration tool with an object-oriented architecture that facilitates the implementation of multi-physics and multi-level simulations assembled from independently developed applications (components). The tool provides high-level support for mutual data exchange between codes, including support for different discretization techniques and specific field transfer operators, being aware of the underlying physical phenomena. Parallel and distributed applications and aspects of the applications are also addressed. Each application is required to implement application and data interfaces, which allow abstract access to solution domains and fields, and provide services for steering individual applications. The Python scripting language is extended by modules representing interfaces to existing codes. The high-level language serves as a glue to tie the modules or components together and to create a specialized application. The capabilities of the tool are demonstrated on two examples that illustrate staggered thermo-mechanical analysis and distributed field mapping.

Efficient methods to visualize finite element meshes

Efficient visualization of large finite element meshes is described.The mesh editor is based on improved winged-edge data structure.Comparison with GiD, ParaView and VisIt is provided. This article describes implementation details of a graphical editor which is intended for efficient 3D visualization of finite element meshes. The purpose of this program is to prepare the input data for the finite element method - a widely known method for numerical solutions of scientific and engineering problems described by partial differential equations. Nowadays, increasing demand on accuracy of the calculation through the boom of parallel computing involves the use of finer and more detailed finite element meshes. However, the meshes with a very large number of elements and nodes cause problems with visualization in commonly used programs that are very slow, unresponsive and for some operations often unusable. Therefore, efficient data structures and algorithms were designed and implemented to enable fast work with very large finite element meshes.

Subdifferential‐based implicit return‐mapping operators in computational plasticity

The paper is devoted to numerical realization of elastoplastic problems. The main goal is to improve implementation of the related constitutive problems. This can be done if plastic flow rules are defined by subdifferentials of plastic potentials. Then just one plastic multiplier is used even if the plastic potentials are nondifferentiable for unknown stress tensors. Further, the implicit Euler time discretization scheme is considered and the standard elastic predictor plastic corrector method is used to find the discretized constitutive solution. Due to the presence of the one multiplier, it is possible to construct a unique system of nonlinear equations within the plastic correction regardless the unknown stress tensor lies on the smooth portion of the yield surface or not. Plastic criteria given by the Haigh-Westergaard coordinates are investigated in this paper (PART I). The suggested method is in detail studied on the problem containing the DruckerPrager criterion, a nonassociative plastic flow rule and a nonlinear isotropic hardening. It is shown that for this problem, one can a priori decide whether the unknown stress tensor will lie on the smooth portion or at the apex of the yield surface. The corresponding tangential (consistent) stiffness matrix is constructed and the elastoplastic problem is solved by the semismooth Newton method. The new method is implemented within the in house software SIFEL and several numerical experiments are introduced.

Coupled shrinkage and damage analysis of autoclaved aerated concrete

This paper is devoted to analysis of shrinkage and damage of autoclaved aerated concrete. Coupled hydro-thermo-mechanical analysis is used for detailed description of drying which causes the shrinkage and damage consequently. The heat and moisture transfer are fully coupled while the staggered approach is used between transports and mechanics. Material parameters were obtained from laboratory experiments and the results of numerical simulations correspond with measured data.

Model of Imperfect Interfaces in Composite Materials and Its Numerical Solution by FETI Method

Analysis of material interfaces in composite materials is in the center of attention of many material engineers. The material interface influences significantly the overall behaviour of composite materials. While the perfect bond on material interface is modelled without larger difficulties, the imperfect bond between different components of composite materials still causes some obstacles. This contribution concentrates on application of the FETI method to description of the imperfect bond.