Improvement of Stochastic Linearization Technique to Analyze Structures with Nonlinear Fluid Viscous Dampers Ghasem Asadpour g.asadour@cv.iut.ac.ir Date of Submission: July 2020 Department of Civil Engineering Isfahan University of Technology, Isfahan 84156-83111, Iran Degree: M.Sc. Language: Farsi Supervisors: Dr. Payam Asadi (asadi@cc.iut.ac.ir) Viscous dampers are used to control the vibration of structures and energy absorption of the structure. Laboratory studies have shown that their force function in terms of power is the speed of their two ends in general. Many studies on linearization technique of behavior of nonlinear dampers are concentrated. By linearization, it is possible to analyze the structures equipped with them at a higher speed, as well as the modal analysis of the structures equipped with them in accordance with regulations. In this study, possible optimal functions of Non-Gaussian for Linearization of nonlinear dampers Using Linear method randomization are optimized. These functions are for structures with a degree of freedom in two modes: structures based on firm soil and based on soft soil and also in three modes of completely linear behavior, two-linear behavior and linear behavior in terms of soil interaction and structure are obtained. Optimization was with the help of genetic algorithm. For the first case of completely linear structures, the error of calculating the standard displacement deviation with the optimal functions obtained under earthquake with PGAs of 0.1 g to 0.6 g has been reduced by about 4% compared to the case of using the possible Gaussian function and is limited to about 5%. Also the standard speed deviation error is limited to about 5.4%. The relative error of displacement of the roof of a 6-floor structure under natural accelerometers with the help of statistical linearization with the proposed non-Gaussian functions was about 24% less than other compared statistical methods .For the second case of structures with elastoplastic behavior, error calculation of the standard deviation of the ductility of claim with the optimal functions obtained under earthquake with PGAs of 0.2g to 0.6g compared to the mode of use of the probable Gaussian function is reduced by about 7.5 % and limited to about 8 percent. Also derived standard deviation error of the ductility of the solvent is limited to about 9.1%. The amount of roof displacement of the studied structure under natural accelerometers with the help of statistical linearization with the proposed non-Gaussian functions in both cases of structures based on hard soil and soft soil had the lowest error. The relative error of displacement of the roof of a 5-floor structure under natural accelerometers with the help of statistical linearization with the proposed non-Gaussian functions was about 41% less than the compared gaussian statistical method.For the third case of structures with linear behavior in terms of soil-structure interaction error of calculating the standard displacement deviation with the optimal functions obtained under earthquake with PGAs of 0.1 g to 0.6 g compared to the mode of using the function Gaussian probability is reduced by about 1.6% and limited to about 8.5%. Also standard speed deviation error is limited to 6.4%. The relative error of displacement of the roof of a 5-floor structure under natural accelerometers with the help of statistical linearization with the proposed non-Gaussian functions was about 28% less than the compared Gaussian statistical method.Based on the results obtained for each of the soft soil and hard soil conditions, an optimal non-Gaussian function for all different ?s and PGAs has been suggested. Keywords : Nonlinear viscous damper; Stochastic linearization technique; Non-Gaussian probability density function; Monte Carlo simulation method; Nonlinear power-law damping; Optimum probability density function