Performability of signal validation systems using stochastic activity networks
B.C. Shen, John Gordon Williams, W.H. Sanders · Transactions of the American Nuclear Society · 1993
Traditional performance and dependability modeling techniques require the separate evaluation of system performance and dependability, disregarding any dependencies that exist between them. The concept of [open quotes]performability[close quotes] overcomes this limitation. Informally, performability quantifies a system's ability to perform in the presence of faults. The complexity of signal validation systems (SVSs) requires a sophisticated representation to-account for performance and dependability in a unified way. A variant of stochastic Petri nets, stochastic activity networks (SANS) have been used successfully to evaluate a wide variety of computer systems and networks. In this paper, we explore the application of SANs as a modeling tool to evaluate an SVS in a nuclear power plant. As an illustrative example, the SAN model of a pressurizer sensor system as well as its SVS is presented. The performability of the system is evaluated and analyzed.