The effect of probabilistic seismic hazard maps on seismic assessment of rc structures in Turkey

Seismic design codes in the world were initally aimed to assure life safety for occupents. The level of seismic loads as well as the structural response limits to be verified under such loads had a certain level of variety depending on countries, level of state-of-the-art and the know-how. The design approach drastically changed however after the 1994 Northridge Earthquake in the USA, where a new concept was introduced, that allows the use of varying levels of seismic hazard and thus seismic loads. Furthermore, structural design limits were also re-defined depending on different seismic design levels. The approach basically defined different performances for varying levels of seismic hazard, thus it was called performance-based design. One of the main aspects of this approach was the use of varying seismic hazard levels. These hazard levels were defined as percentage of probability of exceedance in, typically, 50 years. The seismic hazard levels have taken various names depending on the code, but the most used names are DBE (design basis earthquake) and MCE (maximum considered earthquake), which correspond to 10 and 2% probability of exceedance in 50 years, respectively. In this thesis, the effect of seismic hazard maps, which are subject to changes in time, on the seismic assessment of RC structures in Turkey have been investigated. Structures designed to resist earthquake intensities corresponding to 2% probability of exceedance in 50 years are important structures. 10% probability of exceedance in 50 years corresponds to regular structures. In this thesis, the effects of seismic hazard maps both for 2% and 10% probabilities of exceedance have been investigated. In order to achieve this, the Turkish Seismic Zonation Map of 1996 and the Turkish Seismic Hazard Map published in 2018 (in effect as of 1s of January 2019) have been compared. This study consists of 7 chapters in total. In the first chapter, the aim and scope of the thesis are given. Then in the second chapter, the theoretical, experimental and numerical studies and the recommendations related to the previous probabilistic seismic hazard analysis, shear walls with coupled beams and the base isolator systems in the literature are summarized. Selected 9 bulding models are introduced in Chapter 3. These model stocks consist of 2D sample frames and 3D buildings. Five of these structures are 3 to 8-storey buildings which are mostly built in our country, one of them is a base isolator model, 2 of them are high-rise buildings (20 and 24 storeys) and one is a prefabricated building which is mostly constructed in industrial areas. Modelling details of the structures, shear walls with coupled beams and base isolator design parameters are also given in this section. The sensitivity analysis are conducted in Chapter 4. In order to have more accurate results, the number of the element subdivision and the integration points in one member have to be inreased accordingly. Unfortunately icreased numbers of the elements and the integration points cause longer analysis times. It is really important to have reasonable results with the minumum numbers of the subdivision and the integration points to shorten the analysis times. The aim of this chapter is to carry out non linear push over analysis on a RC frame model to find out the optimum conditions of the number of the integration points, subdivision number and use forced based or displacement based formulation to achive the reasonable results with th shortest analysis time. Nine different analyses have been run on the case study frame to examine the effect of element formulation, as well as the number of subdivisions and integration points. A simple load-controlled conventional pushover analysis has been run with an inversed triangular loading pattern. According to the results, the optimum modelling is choosen and applied for all models during the calculations of this thesis. In Chapter 5, the basics and the fundemantals of the 2019 seismic hazard map are summarized, the selected 20 regions are introduced and the procedures to select earthquake records according to the acceleretion spectra are explained. A set of criteria and a comprehensive algorithm have been prepared so that an in-house developed Matlab script could be devised for selecting a suit of accelerograms for the demand level of relevance. The record selection has been done by using the PEER NGA database where 7025 recorded motions were available. An in-house developed software was used to list and download the record automatically and plot the spectra for acceleration at 5% damping, velocity and displacement. In terms of the selection algorithm, first the acceleration spectrum of the original record is compared to that of the target, in the period window of 0.2T to 1.5Tsec. The scale factor needed is calculated by equating the area below and above the target spectrum. The scale factor is applied to the ordinates, so only to the acceleration of the original record. Than The near field vs far field comparison is made where the distances above 15km are assumed as far field. Finally a comparison is made in terms of the soil type where the records taken on soil with Vs30 higher than 300m/sec are assumed to be recorded on firm soil while records taken on soils with Vs30 lower than 700m/sec are assumed to be recorded on soft soil. In Chapter 6, the results of the time history analysis of the systems introduced in Chapter 3, that are made in the OpenSees program, are given in this section. 20 earthquake regions were selected in Turkey. Ten of them are in the first degree earthquake zone and the other ten are in the second degree earthquake zone. The names of the selected regions are given in Chapter 5. 20 earthquake records, that are used on 9 selected sample structures, were chosen as earthquake level 1 (2% chance of exceeding in 50 years and return period of 2475 years) and earthquake level 2 (10% chance of exceeding in 50 years and return period of 475 years). Thus, a total of 7920 time history analysis were performed and the results were compared in this section. A program, that integrates with the Opensees structural analysis program, was developed in Matlab. So that the analyzes can be done in an accurate, systematic and controlled manner, the results can be editable and the data can be prepared for the Opensees program. For each building, 20 earthquake records were selected using the acceleration spectrum in the 2007 Seismic Code, based on the 1996 Map, with the building importance coefficient 1.0 and 1.5. Then, acceleration spectra were created with the coefficients in the draft earthquake code and 20 earthquake records per each spectra were selected. Time history analysis of the models were conducted using the selected earthquake records. Comparative graphs of maximum chord rotations and floor drift ratios in columns, which are calculated as the result of applying these earthquake records to building systems, are plotted. In these graphs, the average and standard deviations of the analysis results are shown, and the value ranges were determined by adding and subtracting a standard deviation values to the average. As a conclusion in Chapter 7, when the base shear values are taken into consideration, it was observed that the values calculated according to the 2019 seismic code (seismic hazard map of 2018) for the 1st degree earthquake region in Turkey, and for the earthquake level 1 (2% in 50 years), base shear can be 65% smaller than the one calculated based on the 1996 map. Similarly, for the 1st degree earthquake region and earthquake level 2 (10% in 50 years), it was observed that the base shear values calculated according to the 2019 seismic code (based on the seismic hazard map of 2018) could be almost 70% smaller than the existing 1996 seismic map. The seismic resistance of the buildings in in province centers of Bartın, Burdur, Kırşehir, Van Erciş, Antalya, Diyarbakir, Kars and Van as well as in Levent in Istanbul are under-estimated when the existing 2007 seismic code and the 1996 seismic hazard map are used. The base shear demand on buildings in Sakarya and Erzurum, two provinces used in this study, is found to be considerably larger in the 2018 seismic hazard map.