Final position paper on advanced C&I substantiation: Future assurance of programmable C&I systems in UK nuclear plants

Research output: Book/ReportCommissioned report

Abstract

Report for BEIS Nuclear Innovation Program, Project 3.9.11 in the “Reactor Design: Safety & Security Research & Development Programme” theme led by Fraser Nash

Widespread use of computation (software, or ‘programmable systems’ in nuclear parlance) in control & instrumentation will be a defining advance in the evolution of nuclear power generation systems. Software facilitates construction of complex system functionality, and makes it possible to build flexible and efficient systems at reasonable cost. Software is used in nuclear systems which are individually critical to safety, such as application software in reactor primary protection, and also in numerous diverse components as small as smart sensors. Some of the latter are critical in themselves, whilst others are important to the plant safety case in an important collective sense i.e. due to the cumulative effect of multiple components’ failure probabilities.

Quantitative assessment of reliability is inherent to a scientific understanding of the risks associated with systems. Correspondingly, use of software within nuclear systems has an important side-effect, namely, it is necessary to confront the well known challenges in achieving robust scientific assurance of software reliability. Assurance here refers to pre-operational demonstration of reliability, an essential part of the overall plant safety case. Traditional reliability assurance methods are not well suited to analysis of software, and this is already an issue in current safety cases for contemporary UK nuclear plant designs. Looking forward to future plant designs, the amount of software threaded throughout a plant is likely to increase, and new approaches to assurance of software reliability will be needed. This task should be pre-empted to: 1) facilitate the licensing of new nuclear technologies; 2) provide an important commercial differentiator for future UK nuclear technology, namely, demonstrable enhanced safety, a field in which the UK is already regarded as a world leader.

Methods for assuring the reliability of computation do exist, including some which provide the quantification required in a scientific approach to risk. These can be moderately costly to apply but there is scope to improve both their effectiveness and efficiency. This report discusses the implications of a partnering review of ideas in this space which identifies the main existing methods, new proposed approaches, and further research which would markedly improve reliability demonstrations for critical nuclear software. The ideas are at various levels of maturity and some require further development. In addition, a key remaining challenge is to understand how, and under what circumstances, to combine methods to make their use routine, cost-effective, convenient, and proportionate to risk.
Original languageEnglish
Number of pages18
Publication statusAccepted/In press - 2023

Fingerprint

Dive into the research topics of 'Final position paper on advanced C&I substantiation: Future assurance of programmable C&I systems in UK nuclear plants'. Together they form a unique fingerprint.

Cite this