ABSTRACT

The PDS method for calculation of reliability of safety instrumented systems has been developed over more than two decades and harmonized with the IEC 61508 standard as far as possible. The PDSTool is a computer tool where it is possible to draw rather arbitrary reliability block diagrams combining elements by means of k∘oN structures, where serial and parallel structures are special cases. To calculate the Probability of Failure on Demand (PFD) minimal cut sets are identified and processed. A challenge with many computer programs processing the minimal cut sets is to treat stochastic dependencies between the basic events, but also to treat that elements are tested simultaneously. The paper reviews briefly the main principles of the PDS method with main focus on the so-called C kooN correction factor. Then a generalization of the common cause modelling applicable for so-called double voting is presented. Since a generalized cut set approach is used, logical kooN structures are lost causing inaccurate calculations based on the minimal cut sets. Several approaches are described to overcome this challenge. Finally we discuss why a numerical integration is required to cope with simultaneous testing.