Improving the Reliability of Decision-Support Systems for Nuclear Emergency Management by Leveraging Software Design Diversity

Abstract

This paper introduces a novel method of continuous verification of simulation software used in decision-support systems for nuclear emergency management (DSNE). The proposed approach builds on methods from the field of software reliability engineering, such as N-Version Programming, Recovery Blocks, and Consensus Recovery Blocks. We introduce a new acceptance test for dispersion simulation results and a new voting scheme based on taxonomies of simulation results rather than individual simulation results. The acceptance test and the voter are used in a new scheme, which extends the Consensus Recovery Block method by a database of result taxonomies to support machine-learning. This enables the system to learn how to distinguish correct from incorrect results, with respect to the implemented numerical schemes. Considering that decision-support systems for nuclear emergency management are used in a safety-critical application context, the methods introduced in this paper help improve the reliability of the system and the trustworthiness of the simulation results used by emergency managers in the decision making process. The effectiveness of the approach has been assessed using the atmospheric dispersion forecasts of two test versions of the widely used RODOS DSNE system. ACM CCS (2012) Classification: Information systems → Information systems applications → Decision support systems → Expert systems

Communication and Decision Support Systems

“…To prevent a system, such as a DSS, from supplying wrong data, it is necessary to foresee both the presence of output verification algorithms and the possibility of reaching the same results with different models and comparing the data provided by them. Essentially there must be a voting algorithm capable of evaluating the reliability of the output by comparing it with forecasts, probabilistic distributions and historical data [22].…”

Simulation, Epistemic Opacity, and ‘Envirotechnical Ignorance’ in Nuclear Crisis

“…Upheld by the collectively shared belief that major nuclear accidents on German soil are highly unlikely and moderate ones controllable, this knowledge claim built upon the visual metaphor of radiation maps. Prompted by nuclear accidents, these maps 4 Residual risks are hazards that are unknown or have a very low likelihood of becoming a threat and therefore are not accounted for in the design of reactor safety systems (Ionescu 2013). The term is routinely used by nuclear experts to bundle all potentially hazardous factors that cannot be represented using numerical risk assessment methods.…”

“…2 Visualization of a simulation result produced by the RIMPUFF code (Thykier-Nielsen et al 1999) using a linear scale (left) and a logarithmic scale (right). Source: Ionescu (2013) systems being compared) concluded that different models using identical input and calibration parameters may lead to different recommendations of countermeasures in a real emergency. Therefore, they called for the harmonization of dose projection models on an international level, especially in cooperation with Germany's neighboring countries (BfS 2016).…”