Probabilistic and Reliability

The influences for the normal use of structures are regulated in the relevant standards. However, for exceptional influences through terrorist attacks or industrial accidents there are no existing distinct regulations defining these influences. Therefore the CWE concentrates on the examination of the methodical development of risk analysis with the identification of hazard scenarios. Here the probabilistic security analysis for power plants are consulted, which have proved themselves in the practical experience. The probabilistic security analysis is a systematic procedure for the integral evaluation of the system security based on the failure and event tree analysis. The research priorities in this area at the CWE can be summed up as followed:

Probabilistik_1

  • Risk analysis
  • Probabilistic safety analysis
  • Fragility curves and consequence analysis
  • Response-surface method
  • Polymorphic uncertainty modelling
  • Numerical design methods
  • Reliability analysis
  • Security measures

In the context of the failure and event tree analysis fragility curves are indispensable for the description of the relation between the influence parameters and the failure probability with cumulative allocative functions.

The objective of the CWE is the development of reliable methods, which allow the systematic determination of fragility curves based on the experimental analyses or simulations. For the classical solution approach Monte-Carlo-Simulations are used, in which on the load side the load intensities are considered in a stochastic manner. The variances on the resistance side are considered through appropriate distribution functions, which can be described continuously or with random fields.

Probabilistic

To reduce the computing time the CWE uses the concept of the “Design of Experiments“ (DOE). For every considered intensity a „Response Surface Model“ (RSM) with respect to the input parameters is set up, which represents a functional relationship between the input parameters and the response variables. From the described methods multidimensional fragility curves may be derived, which can describe the probability of failure of various components at the same time. The implementation of the methods within probabilistic safety analyses is performed at the CWE by coupling common program systems (ANSYS, Abaqus, LS-Dyna…) with specifically developed evaluation software and software packages.