Reassessing Nuclear Risk Assessment Using Marked Point Processes and Discrete-Event Simulation: Challenges and Opportunities

subsequent industry practice, has provided a foundational framework for nuclear risk quantification.
However, its reliance on steady-state availability metrics, infinite time horizons, memoryless initiating
event processes, and static fault-tree logic is increasingly difficult to reconcile with modern operational
and regulatory realities.
This paper examines the reformulation of nuclear risk assessment as a time-domain stochastic process
problem using Marked Point Process (MPP) and Discrete Event Simulation (DES). In this framework,
system behavior evolves as a stochastic state process driven by failures, external hazards, maintenance,
testing, and operational decisions. Accident occurrence is represented as a counting process derived from
these state trajectories, enabling direct evaluation of Expected Time To Accident (ETTA) distributions,
exposure durations, and configuration-specific risk under time-varying and dependent conditions—
quantities not readily available from classical cut-set methods.
The framework relaxes key PRA assumptions by introducing finite-horizon evaluation, state-dependent
initiating events, and explicit modeling of repairable systems with imperfect maintenance. Maintenance
efficacy is represented stochastically, capturing degraded and “better-than-new” outcomes and their effect
on future failure exposure. The framework also reframes common-cause failure as a dynamic phenomenon
subject to detection and correction over time, rather than a persistent latent characteristic.
These capabilities support direct evaluation of tradeoffs between prevention and mitigation. Historical
experience from the Emergency Core Cooling System (ECCS) hearings [1] underscored uncertainty in
protection-system performance and the importance of containment as a final barrier. Within the MPP/DES
formulation, both accident occurrence and post-accident consequence exposure can be quantified over
finite horizons, supporting a regulatory strategy that relies on established standards for core-damage
prevention while evaluating containment under the assumption that core damage has occurred.
The transition to MPP/DES carries practical challenges: increased computational demand for rare-event
simulation, expanded data requirements, and questions of model interpretability. At the same time, the
approach aligns naturally with risk-informed, performance-based regulation—including frameworks
envisioned under 10 CFR Part 53—and with the transparency requirements reinforced by recent judicial
developments such as Loper Bright Enterprises v. Raimondo. Future work should address efficient
simulation methods, parameter estimation from operational data, and regulatory metrics suited to
time-domain risk representations.