How simulation of failure risk can improve structural reliability - application to pressurized components and pipes

Probabilistic methods for failure risk assessment are introduced, with reference to load carrying structures, such as pressure vessels (PV) and components of pipes systems. The definition of the failure risk associated with structural integrity is made in the context of the general approach to structural reliability. Sources of risk are summarily outlined with emphasis on variability and uncertainties (V&U) which might be encountered in the analysis. To highlight the problem, in its practical and analysis perspective, a short account is given on the nature of failures encountered in pressure vessels and pipe systems, with essential statistical data of service failure. Two engineering analysis tools are invoked: probabilistic fracture mechanics (PFM) and quantitative non-destructive inspection (QNDI). Probabilistic models for risk estimation, in terms of failure probability, are introduced in the classical view of full distribution approach (convolution integral computation) and direct Monte Carlo random simulation of the governing input variables implied in the failure model. The considered end-failure criterion operates with a performance function derived from the elastic-plastic Dugdale crack model, formalized in the failure assessment diagram (FAD) methodology. In the analysis are also integrated probability of detection (POD) models which quantify the chance of flaws/cracks detection by non-destructive inspection, applied intermittently or continuously in a structural health monitoring (SHM) system. he technique of construction of nonparametric confidence intervals on POD vs. crack size rule, by simulated bootstrap re-sampling is presented, in view of setting, presumably safe initial crack size used in fracture mechanics computations. By merging PFM and QNDI via POD, the benefit of applying non-destructive inspection for the purpose of increasing structural reliability is estimated in terms of the decrease of failure probability. The influence of the quality of non-destructive inspection is reflected by POD rule (model). The quantitative approach, structured as a computer code pvRISK is provided. A case study of failure risk assessment in PVs is presented and discussed in the context of failure risk management by intermittent inspection or by a SHM system. By sensitivity analysis, as concerns the variability of size of cracks that might subsist in the structure the benefit of applying NDI expressed in terms of reducing the probability of failure by fracture is evaluated.