Quantitative fault trees and Probabilistic Risk Assessments (PRAs) are traditionally solved using probabilistic solutions, exponential random variables and rare event approximations. Unfortunately, these methods can provide unrealistic probability of failure predictions, especially when modeling long duration missions with limited data. When comparing the predictions over many years to future performance, the traditional methodology is inadequate to represent space systems, or any system, that typically runs until failure. The models should focus more on life instead of probabilistic solutions as there are many things that can and will fail. Previous work [Paulos and Smith (2016) and Paulos et al. (2022)] demonstrated that using component life simulations instead of probabilistic solutions yielded vastly different results in both the overall risk number and risk drivers. In this paper, the simulation concept described previously is further developed into a simulation that ingests the minimal cut sets and the basic event failure rate information from a SAPHIRE (minimal cut set) model, then processes these into a simulation, runs the simulation, and provides results.

The previous work focused primarily on a single-phase model; this updated version allows for multiphase models and components with multiple failure modes. This current tool is faster and provides more useful information for insight, such as mission times to failures and their quantiles.

The paper discusses several reasons for the differences between the two types of solutions. In particular, the effects of using rare event solutions on non-rare event models is explored and discussed. Potential issues and future work in this area will also be discussed.