TY  - THES
AB  - The design of bridge structures for earthquake loads is a challenge in civil engineering. On theone hand, this is caused by the poor detectability of the stochastic process and, on the otherhand, by the occurrence of sometimes severe damage in the event of an earthquake, which – if it occurs in a controlled manner – is in some cases entirely desirable, since the earthquakestress is dissipated in the rest of the structure by ductile load-bearing behavior. In addition, the seismic design of bridges is of great importance. In areas of high seismicity, a design that ison the unsafe side can lead to devastating uncontrolled damage or even bridge collapse. Besidesthe fact that in such cases there is usually a high number of fatalities, many bridges represent a central component of the transport infrastructure that is particularly necessary in the eventof an earthquake. In areas of low seismicity, on the other hand, the calculation should also not be overconservative due to strong simplifications to the extent that the load case earthquake becomes relevant for the design and can result in an extremely uneconomical design. A costly seismic analysis can therefore eventually be justified regardless of the earthquake level. For the seismic analysis of structures, numerous concepts with different complexity are available. In general, the structural responses due to earthquake can be determined both linearly or nonlinearly and, in the second case, statically or dynamically. Due to various advantages, nonlinear-static pushover calculations increasingly come to the fore and are becoming more and more normatively anchored. In addition to initial monomodal approaches by assuming almost unchanged mode shapes, numerous extended concepts including higher modal contributions and the adaption of mode proportional load patterns have been developed in the last decades. However, in most cases even these methods are based on considerable simplifications, such as constant damping, modal scaling based on elastic spectral accelerations or the widespread subsequent combination of individual modal contributions, so that damage accumulation effects of these remain unconsidered in the course of the nonlinear calculation. Therefore, on the basis of a comprehensive literature review, an optimal concept according to the current state of science, the Adaptive Multimodal Interaction Analysis (AMI) by Norda (2012), is optimized and extended to remedy the still existing deficits. Key aspects of these adaptions include damage accumulation of different modal components while considering a realistic correlation as well as a reasonable combination of multidirectional earthquake excitations. While the introduction of correlation factors leads to a substantial improvement of the overall results, a comprehensive parameter study shows that the previous multidirectional  combination of the AMI at load level is unreliable. This approach is therefore replaced by a posterior superposition concept, separately for orthogonal and collinear response components, developed in the present work. Furthermore, the procedure is adjusted in such a way that individual adaption steps are always related to the occurrence of new relevant plastic effects, whereby the required number of steps and thus the numerical effort can be considerably reduced. Another essential part of this doctoral thesis is the computational algorithm developed for the automated application of this method. More specifically, it is a wrapper programmed in MATLAB, which uses the computational kernel of the software program SAP2000, widely used in earthquake engineering, and enables a fully automated adaptive pushover calculation. Difficulties of the program development lie in the identification of mode interchanges, i.e. the swapping of numbers for certain modes due to changes in the natural frequencies as a resultof nonlinear load-bearing behavior, whereby the natural mode shapes themselves change qualitatively. Furthermore, ensuring constant directions of mode proportional load distributions as well as determining reasonable numbers of pushover steps in the respective adaption steps is an algorithmic challenge. Comparisons of the optimized AMI method with other linear and, in particular, nonlinear methods widely used in practice and research show that, in the course of monodirectional and, above all, multidirectional investigations, significantly better approximations to the results of nonlinear time history calculations can be achieved, especially for complex load-bearing behavior. Finally, this work is concluded by the successful validation of the method on a horizontally curved bridge with irregular stiffness distribution as well as on a cable-stayed bridge by renewed comparison with nonlinear time history calculations under bi- or three-directional seismic excitation, respectively. However, it should also be mentioned that in case of systems with a huge number of degrees of freedom, such as the investigated cable-stayed bridge, the algorithm in its current form reaches its limits. Overall, the developed modified AMI (mAMI), including the comprehensive and comparative investigations, thus represents an added scientific value. The method is able to consider several modal contributions under realistic mapping of their correlation. Furthermore, with the occurrence of relevant plastic effects, the modal load distributions and maximum modal spectral accelerations are selectively updated, with a reasonable number of such updates in most cases. It is important to note that all bridge piers that may become plastic have to be activated by an appropriate reference mode, since companion modes lead to much lower stresses due to an only partial correlation. Thus, if the base mode exhibits almost no deformation components at a particular bridge pier, a further calculation should be performed in which this pier is more strongly stressed by an alternative reference mode. Under these conditions, deformations, internal forces and normal stresses can be calculated very accurately in most cases and otherwise at least with larger deviations on the safe side. Only the normal stresses in the piers of the cable-stayed bridge investigated could not be correctly estimated with the current version of the mAMI. This can mainly be attributed to the stress level in the area of the material strengths, which means that the subsequent SRSS combination of the stress components due to orthogonal seismic excitations is not valid. In addition to the aspect that self-induced vertical modes due to horizontal excitations, which cannot yet be mapped, could have additionally worsened the results, this is the only application limit of the mAMI that could be identified in this PhD dissertation. In general, it is worth mentioning that the quality of the results strongly depends on the modeling of the inelastic load-bearing behavior. It was found that only fiber hinges in SAP2000 are capable of representing the realistic development of natural modes due to nonlinear behavior. In many cases, these exhibit vertical vibration components in previously exclusively horizontal mode shapes due to the decrease in tangent axial stiffness, which cannot be captured by plastic hinge models. With regard to the calculation effort for the applications of the modified AMI carried out within the scope of this PhD dissertation, it should also be emphasized in conclusion that the consideration of two or three modal load combinations (with varying base modes or modal signs) was sufficient in each case to determine the respective structural responses accurately enough and on the safe side. However, if more load combinations become necessary in certain cases, this method may become less attractive due to the significant increase in computational effort.
AU  - Kämper, Dominik Matthias
CY  - Wuppertal
DO  - 10.25926/BUW/0-885
DP  - Bergische Universität Wuppertal
KW  - Pushover
KW  - multidirectional
KW  - multimodal
KW  - earthquake
KW  - intermodal correlation
KW  - bridges
KW  - SAP2000
KW  - mAMI
LA  - eng
N1  - Bergische Universität Wuppertal, Dissertation, 2024
PB  - Veröffentlichungen der Universität
PY  - 2025
SP  - 1 Online-Ressource (xxvi, 212 Seiten) : Illustrationen, Diagramme
T2  - Fakultät für Architektur und Bauingenieurwesen
TI  - Adaptive multimodal pushover analysis of bridges considering multidirectional earthquake excitations
UR  - https://nbn-resolving.org/urn:nbn:de:hbz:468-2-6069
Y2  - 2025-12-05T04:07:02
ER  -