CUDA accelerated fault tree analysis with C-XSC

Abstract

Fault tree analysis is a widespread mathematical method for determining the failure probability of observed real-life systems. In addition to failure probability defined by wear, the system model has to take into account intrinsic and extrinsic system influences. To make allowance for such factors, we draw on an implementation by Rebner et al. to compute the lower and upper bounds of the failure probability of the top event based on interval analysis implemented in MATLAB using INTLAB. We present a new verified implementation in C++ to reduce the trade-off between accuracy and computation time, describe the new implementation by giving an illustrative example based on work by Luther et al. and show the advantages of our new implementation. © 2012 Springer-Verlag.