TY - THES A1 - Luong, Thi Mai Hoa T1 - Identification of the state of stress in iron and steel truss structures by vibration-based experimental investigations N2 - Safety evaluation of truss structures depends upon the determination of the axial forces and corresponding stresses in axially loaded members. Due to presence of damages, change in intended use, increase in service loads or accidental actions, structural assessment of existing truss structures is necessary. This applies particularly to iron and steel trusses that are still in use, including historic and heritage monuments. Precise identification of the stresses plays a crucial role for the preservation of historic trusses. The assessment measures require non–destructiveness, minimum intervention and practical applicability. The axial forces in truss structures can be estimated by static calculations using the method of joints, method of sections or finite element method, if accurate information about parameters such as external loads, geometrical characteristics, mechanical properties, boundary conditions and joint connections are known. However, precise information about these parameters is difficult to be obtained in practice. Especially in the cases of historic constructions, reasonable assumptions about the uncertain parameters may not be acquired. Motivated by the preservation of existing truss−type constructions composed of axially loaded slender members, the present work aims to develop a non–destructive methodology to identify the axial forces or corresponding stress states in iron and steel truss structures. The approach is based on vibration measurements and the finite element method combined with optimization techniques. After a state of the art review, numerical and experimental studies were carried out on three partial systems of truss–type structures. The investigated systems included single bars, a two–bar truss−like system and a five–bar truss. They were developed step–by–step as built–up truss−type constructions that are constituted of individual members connecting at joints. The examined aspects included the effects of structural loading on the dynamic performance of truss structures, modelling of joint connections, mode pairing criteria, selection of updating parameters and definition of an objective function, as well as the use of different optimization techniques. Concerning the axial force effects on the structural dynamic responses, the effects of the stress stiffening become more complicated for multiple–member truss systems with increasing complexity. The coexistence of both compressive and tensile forces in trusses has counteracting effects on the modal parameters. These effects cause variation of natural frequencies and interchange of modes when the loads or corresponding member forces are changed. To examine the axial force effects on the structures at different stress states, in the numerical study and laboratory experiments, loads were applied progressively to the investigated truss−like systems. Regarding the modelling of joints for truss–type structures, the joint flexibility affects the structural dynamic responses. Therefore, the numerical models of truss−type structures include joint models with variable rotational springs to represent semi–rigid connections. Considering the mode pairing criterion, the mode pairing is performed by adapting an enhanced modal assurance criterion with the calculation of the modal strain energy. The criterion allows the selection of desired clusters of degrees of freedom related to specific modes. With respect to the model updating strategies, the selection of updating parameters and the choice of an appropriate objective function are identified to be significantly important. In addition, three different optimization techniques were applied to compare their suitability for the inverse axial force identification and estimation of joint flexibility of truss structures. The results of the numerical study and laboratory tests show that nature–inspired optimization methods are considered as promising techniques. A methodology consisted of a two–stage model updating procedure using optimization techniques was proposed for the determination of multiple member axial forces and estimation of the joint flexibility of truss–type structures. In the first stage optimization, the validation criterion is based on the experimentally identified global natural frequencies and mode shapes of the truss. Additionally, the axial forces in selected individual members of the truss are used. They are estimated from the natural frequencies and five amplitudes of the corresponding local mode shapes of the members using an analytically−based algorithm. Based on the results of the identified axial forces in the first stage, a second optimization procedure for the joint stiffnesses is performed. In this stage, the modal parameters of the global natural frequencies and mode shapes are used as validation criterion. From the results of the laboratory experiments, the identified axial forces by the proposed methodology agree well with the experimentally measured axial forces of the investigated systems at different stress states. Moreover, based on the numerical verification, the identified joint stiffnesses indicate reasonably the joint flexibility in relation to the pinned or rigid conditions. To assess the relevance of the proposed methodology on existing structures in real−life conditions, an in–situ experiment was carried out on a historic Wiegmann–Polonceau truss in the city of Potsdam. The in–situ experiment shows that uncertainties relating the mechanical and geometrical properties of historic trusses as well as the experimental sensor setup can influence the accuracy of the axial force identification. In the present work, recommendations are given for the development of a guideline of measuring concepts and assessment strategies applied to existing truss structures. The intention is to integrate the proposed methodology as part of the Structural Health Monitoring for historic truss–type constructions. T3 - BAM Dissertationsreihe - 159 KW - State of stress KW - Beanspruchungszustand KW - fachwerkartige Stahltragwerken KW - Schwingungsmessungen KW - Finite-Elemente-Modellkalibrierung KW - Optimierungsmethoden KW - Truss structures KW - Vibration measurements KW - Finite element model updating KW - Optimization techniques PY - 2018 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:b43-449615 SN - 1613-4249 VL - 159 SP - 1 EP - 195 PB - BAM Eigenverlag CY - Berlin AN - OPUS4-44961 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - THES A1 - Hauswaldt, Sebastian T1 - Kontinuumsmechanische Werkstoffmodelle zur numerischen Simulation von Stahlbauteilen im Brandfall N2 - Das nichtlineare und geschwindigkeitsabhängige1 Materialverhalten von Stahl wird besonders bei hohen Temperaturen sichtbar. Für Finite-Elemente-Simulationen von Stahlkonstruktionen im Brandfall sollte aus diesem Grund plastisches und geschwindigkeitsabhängiges Materialverhalten beschrieben werden. Die vorliegende Arbeit betrachtet unter diesem Aspekt bisherige Materialuntersuchungen und macht Vorschläge für dreidimensionale Materialmodelle mit entsprechenden Eigenschaften. Es werden zunächst die phänomenologischen Eigenschaften von Baustahl anhand einer Literaturrecherche analysiert, wobei verstärkt auf Untersuchungen des Materialverhaltens bei Belastungen und Aufheizprozessen, wie sie im Brandfall zu erwarten sind, geachtet wird. Die für die Bemessung von Stahlkonstruktionen gebräuchliche Spannungsbeschreibung des EC 3-1-2 wird untersucht. Es werden ihre Stärken, aber auch die zur Entwicklung einer kontinuumsmechanischen Materialformulierung fehlenden Eigenschaften, aufgezeigt. Ein nichtlinear-elastisches kontinuumsmechanisches Materialmodell der Deformationstheorie der Plastizität wird so angepasst, dass es die Spannungs-Dehnungslinien gemäß EC 3-1-2 im einachsigen Spannungszustand beschreibt. Es wird des Weiteren ein thermoelastisch-viskoplastisches Modell vorgestellt, das in der Lage ist, Kriechen und Relaxation bei Aufheiz- und Abkühlprozessen zu beschreiben. Die Struktur dieses Materialmodells wird so gewählt, dass die Parameter an hierfür geeigneten Messergebnissen leicht identifiziert werden können. Der deviatorische Anteil des Modells besteht aus einem geschwindigkeitsunabhängigen, plastischen Anteil und einem geschwindigkeitsabhängigen, viskoelastischen Anteil. Der geschwindigkeitsunabhängige, plastische Anteil wurde als Differentialgleichung auf Grundlage der so genannten endochronen Plastizitätstheorie formuliert. Die Parameter der Materialmodelle werden auf Grundlage der Messergebnisse stationärer Warmzugversuche an Baustahlproben identifiziert. Sowohl das nichtlinear-elastische EC 3-1-2-Materialmodell als auch das thermoelastisch-viskoplastische Materialmodell mit den an Baustahl angepassten Materialparametern wird numerisch für die Verwendung mit Finite-Elemente-Programmen aufbereitet und als UMAT-Subroutine für ABAQUS in der Programmiersprache FORTRAN implementiert. Hierbei wird insbesondere auf die Bereitstellung der konsistenten Tangentenoperatoren Wert gelegt, um eine effiziente numerische Berechnung bei Verwendung der Materialmodelle zu ermöglichen. Abschließend werden erste Simulationsrechnungen vorgestellt, um beispielhaft die Möglichkeiten der Anwendung der entwickelten und implementierten Materialmodelle für Simulationen von Stahlkonstruktionen im Brandfall aufzuzeigen. Stichworte: Brandschutz, Stahl, Brandverhalten, Finite-Elemente-Methode, Materialmodell, Eurocode 3-1-2, UMAT 1Ist das Materialverhalten abhängig von der Prozessgeschwindigkeit, wird es als geschwindigkeitsabhängig bezeichnet. Prozesse können sowohl dehnungs- als auch spannungs- oder temperaturgesteuert sein. Die Begriffe zeitabhängig und zeitunabhängig werden hier vermieden, da diese in der Materialwissenschaft mit Alterungsprozessen (’aging’) in Verbindung gebracht werden. N2 - The non-linear and rate-dependent2 material behaviour of steel becomes particularly visible at high temperatures. For finite-element-simulations of steel structures in the event of fire, plastic and rate- dependent material behaviour should therefore be described. With this in mind, this thesis looks at previous material investigations and makes suggestions for three-dimensional material models with the corresponding properties. The phenomenological properties of structural steel are analysed on the basis of a literature study, whereby increased attention is paid to the material behaviour under loads and heating processes, as can be expected in the event of fire. The stress-strain-relation of the EC 3-1-2 used for the design of steel structures is examined. The advantages of this description, but also the missing properties for the development of a continuum-mechanical material formulation, are shown. A non-linear-elastic continuum mechanical material model of the deformation theory of plasticity is adapted in such a way that it describes the stress-strain curves according to EC 3-1-2 in the state of uniaxial stress. Furthermore, a thermoelastic-viscoplastic model is presented which is able to describe creep and relaxation during heating and cooling processes. The structure of this material model is chosen in such a way that the parameters can be easily identified using certain measurement results. The deviatoric part of the model consists of a rate-independent, plastic part and a rate-dependent, vis- coelastic part. The rate-independent, plastic part was formulated as a differential equation based on the endochronic theory of plasticity. The parameters of this material model are identified on the basis of the measurement results of stationary hot tensile tests on structural steel specimens. Both the non-linear-elastic EC 3-1-2 material model and the thermoelastic-viscoplastic material model with the material parameters adapted to mild steel are prepared numerically for use with finite- element-programs and implemented as UMAT-subroutines for ABAQUS in the FORTRAN code. Particular emphasis is placed on the provision of consistent tangent operators to enable efficient numerical calculation when using the material models. Finally, first finite-element-simulations are presented in order to show the possibilities of the deve- loped and implemented material models for simulations of steel constructions in case of fire. Keywords: Fire prevention, steel, fire behaviour, finite element method, material model, Eurocode 3- 1-2, UMAT 2 If the material behaviour is dependent on the process rate, it is referred to as rate-dependent. Processes can be controlled by strain, stress or temperature. The terms time-dependent and time-independent are avoided here, as these are associated with aging in materials science. T3 - BAM Dissertationsreihe - 167 KW - Brandschutz KW - Stahl KW - Brandverhalten KW - Finite-Elemente-Methode KW - Materialmodell KW - Eurocode 3-1-2 KW - UMAT KW - Fire prevention KW - Steel KW - Fire behaviour KW - Finite element method KW - Material model KW - Eurocode 3-1-2 KW - UMAT PY - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:b43-507662 SN - 1613-4249 VL - 167 SP - i EP - 206 PB - Bundesanstalt für Materailforschung und -prüfung (BAM) CY - Berlin AN - OPUS4-50766 LA - deu AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -