Liquefaction potential of the Riva region in Istanbul and the analysis of the soil by fuzzy logic method

"In the Northern Marmara Motorway Project Odayeri-Paaköy (including 3rd Bosphorus Bridge) in the city of Istanbul, Beykoz Region, field observations, drilling works and laboratory experiments were performed on the V26 viaduct legs that are found between Poyraz-Paşaköy. The results show that pile foundations are situated in sandy clay, alluvium, and sandstone mid-level siltstone layers. In the light of realized field and laboratorty experiment results,, the effects of softening and liquefaction, which occur as a result of earthquake load , were analyzed in detailed based on the varios methods, and then were subjected to the linguistic algorithmic equation using developed fuzzy logic method. In the present study, definition of liquefaction the liquefaction process and the concept of liquescent soil were explained initially. As is known, clean sands could be liquidized, however an agreement has not yet been reached on the liquefaction of silty soil, clayey soil and even gravelly soil. Earthquakes in the history; 1994 Northridge, 1999 ChiChi and 1999 Kocaeli, occurred on fine-grained soil, which is structurally different from sand soil, liquefaction studies have been increasing.In the liquefaction literature, several studies claim that the areas regarded as non-liquescent according to the classical methods turn out to be liquescent and therefore the Chinese Criteria is not really reliable. The important consequences of liquefaction; sinking or overturn of heavy structures, floating of lightly buried structures, displacements on retaining structure or collapse of retaining structure, collapse of slope, collapse of settlement and flow/yield, lateral spread etc. are demonstrated by photographs. Furthermore the pre-evaluation of the dynamic behavior of fine-grained and coarse-grained soil was emphasized to be necessary as deformation and loss of strength occur on these soils after earthquake leading to the loss of stability. Besides, it was demonstrated that the bearing capacity loss and collapse types differ by the soil type. On the line of the project route, a total amount of 11 boring of different depth has been performed and the data of 3 boring in which pile foundations are placed on have been analyzed. According to the groundwater level data, it is at 1.5-2 m, therefore it can be said it entails a risk for liquefaction. Due to the fact that liquefaction take an important place on geotechnical engineering and in particular liquefiable characteristics of fine-grained soil have been intensively studied, this area was chosen for the study. The area examined in this thesis is found in the second seismic belt. The boring and laboratory data show that soil may critical under cyclic loading cyclic load, further analysis have been conducted. In this study, a literature review was conducted on the liquefaction potential and the overview about the liquefaction process was provided. Then the liquefaction analysis was performed accordingly. In this thesis ,using the approaches suggested by Marcuson vd. (1990), Seed, Idriss (1982) and Andrews and Martin (2000), samples were collected from the alluvium unit. The results show that soil is found in risky zone, liquefiable zone and zone where required a detailed inspection. The STP data, index characteristics (water content , Attebergs limit) parameters of fine-grained and coarse-grained soil were determined and following that the shear wave velocity ,and the maximum shear modulus were obtained using empirical methods. The simplified analysis proposed by Seed Idris (1971) from STP-based methods, was used to determine the liquefaction potential whereas different approaches were used to obtain the dynamic resistance ratio for fine-grained soil. In addition, due to the inadequate data, clays are assumed to be normally consolidated. It was concluded that, in order to determine undrained shear strength ,liquefaction potential of fine-grained soil, dynamic laboratory experiments (cyclic triacial test ,oedometer test ) should be performed. Considering the disadvantages of the limited information (experiment results that couldn't be obtained), the scope of liquefaction evaluation was chosen. Furthermore, following the performed analysis, fuzzy logic method was developed in order to express liquefaction potential using linguistic terms. Fuzzy logic, is a mathematical method that is used to obtain approximate results, differently from that of Aristo logic, to solve ambiguous problems. From this point, it is regarded as a simple and practical method that can be used widely in Geotechnical Engineering which naturally involves many fuzziness. Besides, f it can be applied to many fields such as engineering, medicine, sociology, psychology, business, expert systems, artificial intelligence, signal processing, transportation, intersection signaling in general. Fuzzy logic was developed by Lotfi Zadeh in the mid-1960s. An element belongs to a subset or it does not in Aristo logic; wheras in fuzzy logic an element is interval function. In fuzzy logic the degree of membership is expressed by a real number in the interval [0, 1]. In this study system of fuzzy were explained in detail. Firstly real input values processed fuzzifier operation by applying fuzzy rules. And then outcomes are applied inference engine operation. Inference engine methods are mentioned in thesis, also. They are Mamdani method, Larsen method, Sugeno method and Tsukamato method. Finally, the fuzzy values obtained inference engine operation transformed into reel date by defuzzifier process. The mamdani method engine operation is the most commonly used method, so it was preferred in study. In 1975, Professor Ebrahim Mamdani of London University applied a set of fuzzy rules steam engine and boiler combination. It was called the Mamdani method. The fuzzy logic membership functions were selected considering important parameters of liquefaction and were obtained through experimental results conducted in this region. Therefore; the parameters affecting liquefaction were selected initially. SPT-N value shear wave velocity, groundwater level and soil depth were used for coarse-grained soils as input variables whereas for fine-grained soils, in addition to these parameters plasticity index was used as well. Later on, fuzzy subset intervals were determined for them based on the previous studies and ABYYHH. Using fuzzy logic toolbox in MATLAB (MATrix LABoratory), the selected parameters were inserted as inputs of the system. Then, by adopting Mamdani method, linguistic fuzzy identification output was derived based on the fuzzy rules. Membership functions for SPT, shear wave velocity Vs , plasticity index are rated as low, medium,and high. The probability of liquefaction for this data is set at "HIGH" for "LOW" range, "MEDIUM" for "MEDIUM" range and "LOW" for "HIGH" range. Membership functions for underground water is rated as shallow and deep. The probability of liquefaction for this data is set at " SHALLOW " for " HIGH "range, "DEEP " for " LOW " range. And for soil depth is rated as little, middle, and great. The probability of liquefaction for this data is set at "LITTLE " for " HIGH " range, " MIDDLE " for " MEDIUM " range, "GREAT"for "LOW" range. As a result the analysis, it has been found that the linguistic results obtained through Mamdani method using a total number of 54 rules for coarse-grained and 162 rules for fine-grained soils match up with the potential of Iwasaki liquefaction results by 85%. The critical point here is that the fuzzy logic input and output functions are selected appropriately. Overall, it can be said that there is a need for more detailed laboratory experiments; so that relevant results could be used to make this method much more reliable. In future studies, it can be analyzed defuzzifier. The methods of clarification max-membership, center-of-gravity/area, weighted average, mean max membership are mentioned thesis in detail. The percentage of liquefaction can be obtained numerically with using the clarification methods.