STUDY OF RELIABILITY OF SIMPLE STRUCTURAL SYSTEM FOR CORRELATED FAILURE

• Main Author: SASANGKA ADJI (NIM: 25016008), TRI
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/31373
• Institution: Institut Teknologi Bandung

<p align="justify"> In probability theory, the capacities and risk level are expressed by probability of failure (Pf). The probability of failure (Pf) calculation in a structural system is not simple because it involves several component failure events. This calculation process is an integration of performance functions of several components with an integration area that depends on the correlation between the failure events of these components. The exact existing equations can only be used to determine the possible failure of a structural system whose components are perfectly correlated or not correlated at all. Where as to determine the probability of failure of a structural system with components that are not perfectly correlated, required equation with approach method. <br /> <br /> The purpose of this study in general is to determine the reliability of the structure system with components that are not perfectly correlated failure and collapse. <br /> <br /> The method used in this study is the Interpolation of Slice Method with First Order Second Moment (FOSM) Method to determine upper and lower limit of Unimodal and Bimodal and Monte Carlo Simulation as a comparison method of the calculation. <br /> <br /> The result of study of the structural system reliability for probability of failure (Pf) f or some cases with the Interpolation of Slice Method shows that the result is still in the range of Unimodal and Bimodal and also close with Simulation Monte Carlo result. It could be said that this method can be applied for calculating Pf with unperfectly correlated components. <br /> <br /> Interpolation of Slice Method with correction factor r(rho) can be used as an alternative method to calculate the reliability of the structural system with correlated components with some conditions specified those are the difference of PFi and PFj is not too far, i.e. less than 10 times and the data used to create the graph r (rho) vs rho must be complied to create rho values with the similar range of target rho values. <p align="justify"> <br />