Betonarme Çerçevelerin Doğrusal Olmayan Analizinde Değişik Modelleme Stratejilerinin Araştırılması

Abstract

Deprem etkisindeki yapıların doğrusal olmayan davranışı konusu çok sayıda deneysel araştırmada incelenmektedir. Gelişen bilgisayar teknolojisi desteğiyle, amaca yönelik gerçekleştirilen deneylerden elde edilen veriler kullanılarak, birbirinden farklı analitik modelleme stratejileri ve hesap parametreleri geliştirilmektedir. Bu tez kapsamında, daha önce pseudo-dinamik deneyleri yapılmış olan 4 katlı betonarme bir çerçeve yapı, doğrusal olmayan davranışı dikkate alınarak modellenmiştir. 1950’ li yıllarda Güney Avrupa ülkelerinin birçoğunda yapılmış olan tasarımı ve uygulamayı temsil eden bu betonarme çerçeve sisteme ait deneysel veriler bulunmaktadır. Bu deneysel veriler ölçüt alınarak, incelenen betonarme çerçeve sistemin doğrusal olmayan davranışını belirleyen hesap parametreleri üzerinde çalışılarak modelleme stratejileri belirlenmiştir. Etkin rijitlik, beton mekanik özellikleri(döküm fazları arasındaki dayanım farkı), viskoz sönüm, plastik mafsal boyu gibi parametreler üzerinde ayrı ayrı çalışılarak numerik model kalibre edilmeye çalışılmıştır. Bu açıklamalar doğrultusunda hazırlanan yüksek lisans tezi 6 bölümden oluşmaktadır. Birinci bölümde konunun tanıtılması, çalışmanın amacı ve kapsamı ile konuyla ilgili çalışmaların gözden geçirilmesi yer almaktadır. İkinci bölümde, yapıların doğrusal olmayan davranışı ve bu davranışı etkileyen faktörler açıklanmıştır. Üçüncü bölümde, performansa dayalı tasarım ve doğrusal olmayan analiz yöntemleri açıklanmıştır. Dördüncü bölümde, pseudo dinamik deney ve deneyin uygulandığı betonarme çerçeve sistemin özellikleri hakkında bilgiler verilmiştir. Beşinci bölümde, betonarme çerçevenin modellenmesi hakkında bilgi verilmiş ve birbirinden farklı modelleme stratejilerine göre elde edilen doğrusal olmayan analiz sonuçları ve bu sonuçların deneyden elde edilmiş olan sonuçlarla karşılaştırılması yer almaktadır. Altıncı bölümde, çalışmadan elde edilen sonuçlar yorumlanarak konuyla ilgili öneriler verilmiştir.

The recent earthquakes have dramatically demonstrated that research in earthquake engineering must be directed to the assessment and strengthening of existing constructions lacking of appropriate seismic resisting characteristics. The very recent 'European earthquakes' (e.g. Italy-1997, Turkey - August 1999, Greece - September 1999) confirm and highlight that also Europe may suffer from the vulnerability of the existing building stock. Data on the real characteristics of buildings that have been subjected to earthquakes are in general difficult to obtain. Hence, the experimental pseudo-dynamic test results on the fullscale RC frame generated an immense amount of records, that were used to corroborate the numerical models. Therefore, the calibrated analytical models can be extensively used in reproducing the real behaviour of existing RC buildings. The experimental tests on full-scale structure models assisted the calibration of numerical models and sustain in the assessment of proportioning and detailing rules for the different structural sub-assemblages. This complementary numerical and experimental approach emphasise the important role of the research for the mitigation of the seismic risk. In this study, 4-storey reinforced concrete frame structure is modelled with considering the nonlinear behaviour which is done pseudo-dynamic test. This full-scale RC frame representative of the building's design and construction practice until the late 1970's in most of south European countries, and currently needing seismic retrofit, were constructed and tested pseudo-dynamically, at the ELSA laboratory (European Laboratory for Structural Assessment). The test frame had been designed without specifically considering seismic action (non-seismic resistant constructions). This experimental study aimed at assessing the original capacity of existing structures, with and without infill masonry, and to compare performances of different retrofitting solutions. The tests have shown that the vulnerability of existing reinforced concrete frames designed without specific seismic resisting characteristics, which are an important part of the existing buildings in Europe, constitute a source of high risk for human life. Furthermore, it was demonstrated that advanced retrofitting methods, solutions and techniques substantially reduce that risk to levels currently considered in modern design. The experimental tests on full-scale structure models assisted the calibration of those numerical models and sustain in the assessment of proportioning and detailing rules for the different structural sub-assemblages. This complementary numerical approach emphasise the important role of the research for the mitigation of the seismic risk. The main objective of the theoretical, experimental and analytical work subject of this thesis is to achieve a numerical methodology, which is experimentally calibrated and able to reproduce rigorously the structural behaviour of existing reinforced concrete frame buildings. The evaluation of the available refined models in analysing the seismic performance of existing buildings, proposing improvements to reach a reliable numerical methodology to predict their seismic response. Initial calculations were performed based on the available refined models commonly used to model the new structures. Due to the unsatisfactory results, parametric analyses were performed, and confirmed with the experimental full-scale test results to identify the discrepancies. The detailed parametric analyses reveal inadequacies of the current models when applied to the existing structures. This effort led to fine-tuning of the model parameters' (as plastic hinge length, effective stiffness, concrete mechanical properties, viscous damping etc.), as well as to the inclusion of the inadequate longitudinal and transversal reinforcement detailing in beams, columns and joints, widespread use of smooth reinforcing steel, lack of concrete confinement, and, inadequate lap-splice length, transverse reinforcement in columns limited to perimeter hoops with 90o hooks, which are insufficient. Various analysis methods, either linear elastic or non-linear, static or dynamic, are available for the performance analysis of existing reinforced concrete buildings. Elastic analysis methods include code static lateral force procedures, code dynamic lateral force procedures and elastic procedures using demand capacity ratios. At the present time, linear elastic analysis remains the instrument of the design profession, for the calculation of forces and stresses, as well as for the proportioning of structural members. Nevertheless, linear elastic analysis inability to reflect the real behaviour of structures under abnormal or ultimate loading conditions has been pointed out. This follows because almost all structures behave in some non-linear manner prior to reaching their limit of resistance. A more realistic evaluation of the strength of structures against the failure conditions, or the factor of safety, can only be achieved by analyses that take into account various non-linear effects. The non-linear time history analysis method, with recorded or simulated ground motion records, provides the most accurate means for predicting seismic demands. This inelastic dynamic method is widely used to model specimens tested in laboratory and real structures with a reduced number of elements. Despite its advantages, it must be admitted that non-linear time history analysis can frequently become overly complex and impractical for general use as a first assessment. An alternative is to use simplified nonlinear static analysis methods. The improved models were found capable to analyse existing reinforced concrete structures, reproducing accurately their non-linear response. This numerical models are calibrated against experimental PsD test results on RC full-scale structures. The tests on full-scale models of existing structures constitute an exceptional opportunity for improvement of knowledge on behaviour and capacity of RC structures designed and constructed in Europe until the late 1970's. The numerical analyses performed in this thesis are based in a non-linear plastic hinge model, and take into account the material non-linearity according to the specific materials properties. The influence of physical phenomena and model parameters in the structural response was thoroughly understood due to the integrated numerical and experimental research approach. In the numerical study, it was proposed a simplified model to account for the slippage of the reinforcement in existing RC structures. These statements prepared in accordance with the master's thesis consists of 6 chapters. In the first chapter; introducing the topic, revision of the studies on the subject with the purpose and scope of the study is situated. The second part describes the nonlinear behavior of structures and factors influencing this behavior. The third section describes the performance-based design and nonlinear analysis. In the fourth section contains information about the properties of reinforced concrete frame system that has been applied to pseudo-dynamic testing. In the fifth chapter, which provides information about the reinforced concrete frame modeling and nonlinear analysis results obtained by different modeling strategies from each other and to the comparison with the results obtained from experiments with these results. In the sixth chapter, the results obtained from the study are given advice on the subject interpreted.