Reliability Assessment and Prediction using Monitoring Information

The reliability and performance assessment of engineering structures by monitoring systems in its nature is a complex challenge, since not only a couple of physical quantities, such as material strength properties, environmental conditions, and local stresses but the comprehensive time dependent behavior of structures are of interest. This behavior has to be deducible by sensor readings and associated descriptive mechanical functions. In addition, the layout/design of monitoring systems and the monitored properties are accompanied by uncertainties. Nevertheless, down to the present, probabilistic-statistic principles are hardly involved in the design of monitoring systems and in the assessment of monitoring readings. However, such principles can significantly contribute in the optimization of location and number of sensors, and in the reduction and efficient assessment of recorded data. Therefore, the purpose of this paper is to present (a) stochastically analytical assessment methods usable for the adjustment of monitoring systems, (b) possibilities to merge visual inspection methods with monitoring methods, and (c) to present Bayesian strategies for the incorporation of past information from visual inspection results or from analytical prediction models in monitoring.