임의의 분포를 가진 시스템 변수의 표본을 이용한 시스템의 성능 불확실성 예측

Abstract

Characteristics of mechanical system parameters which are length, mass, stiffness, damping coefficient and etc should always have uncertainties in real engineering problems due to manufacturing tolerance and material irregularity. Since the parameter uncertainties have an effect on a system performance uncertainty, the performance uncertainty has to be analyzed and considered in design of the system for reliability. In order to analyze the performance uncertainty, population distribution of each system parameter should be known. However, in real engineering problems, the distributions are unknown. Therefore, the performance uncertainty should be estimated based on finite number of system parameter samples from each unknown distribution of the system parameters. In this paper, in case each system parameter has unknown and arbitrary distribution, a system performance uncertainty estimation method based on samples was proposed.