Structures and infrastructure systems suffering from deterioration and damage due to corrosive actions and severe loads may eventually become unsafe. In order to prevent premature failure, monitoring can be implemented. Since structural monitoring essentially has to be nondestructive, direct information about the current strength or resistance of the structure is generally not available. Hence, resistance data must be inferred from observable or measurable data such as vibration characteristics or stiffness values, typically on a statistical basis. This introduces uncertainty into the process which can be modeled in terms of probability density functions. The paper presents a computational approach to combine Monte-Carlo simulation with analytical probabilistic models to assess the influence of observed stiffness values on the strength distribution model and on the prediction of expected failure rates for the future life time of the structure.
展开▼
