- Site: Barrow
- IR number: 20-111
- Date: January 2021
- LC numbers: 14, 23
Executive summary
Purpose of Intervention
This inspection focussed on the BAE Systems Marine Ltd. (BAESML) arrangements for compliance with Licence Condition (LC) 14 (Safety Documentation) and LC 23 (Operating Rules) and on their application. This inspection formed part of ONR’s Integrated Intervention Strategy for 2020/21.
The purpose of the inspection was to establish whether BAESML has adequate processes in place to support development of future safety cases for Dreadnought build activities and, in preparation for this, that there is evidence these are currently delivering adequate safety cases for Astute build activities.
Interventions Carried Out by ONR
During the inspection, we sought to confirm that the arrangements implemented by BAESML are adequate, when judged against the regulatory expectations for LC 14 and LC 23. We also sought to establish how they had been applied to form an adequate safety case. Specifically we:
- Inspected the arrangements for the identification and definition of conditions and limits necessary in the interest of safety.
- Inspected the arrangements for the categorisation of safety functions and the classification of safety measures.
- Inspected the arrangements for the monitoring, review, improvement and implementation of safety case processes.
- Explored the responsibilities for ownership of safety cases and involvement of users in safety case production and the adequacy of the process for the independent assessment of safety cases.
- Explored, using faults from a safety case, how conditions and limits are used to define a safe operating envelope and how they facilitate safe operations.
Explanation of Judgement if Safety System Not Judged to be Adequate
NOT APPLICABLE THIS WAS NOT A SYSTEM BASED INSPECTION
Key Findings, Inspector's Opinions and Reasons for Judgements Made
We found that BAESML arrangements for LC 14 set out clear requirements for safety case production. We found evidence of clear ownership of safety case production by the safety case team, who maintain a good system for competence management of those involved in the production. There was good evidence that improvements to the arrangements for LC 14 and LC 23 have been made, particularly with respect to the process for integrating Human Factors (HF). Compared to previous inspections, we found improved evidence that BAESML can articulate the golden thread from the safety functions required by the safety case to the operational claims identified to support them. We also found evidence that appropriate scopes are being identified for the Independent Peer Review (IPR) of safety cases, and identified that the reviews are being conducted in line with defined arrangements, with good evidence of active ownership by the Independent Nuclear Assurance (INA) function. It was not clear why Human Factors outputs are not considered within the IPR scope.
During the inspection, BAESML identified that reviews are currently ongoing for several of the key processes and guidance associated with LC 14 and LC 23 and that these reviews are identifying improvements necessary to meet good practice. This assertion is consistent with our findings, in that we identified several deficiencies in the arrangements when compared to regulatory expectations outlined in the ONR technical inspection/assessment guides.
For categorisation and classification, we identified issues with respect to the definition of safety functions and the transparency within the safety case such that is possible to establish the measures, and their importance, necessary to achieve specific functions. We also found gaps in the implementation arrangements in that they do not adequately consider implementation beyond including the requirements into operator instructions e.g. there is insufficient consideration of training. Further, we found that the arrangements to verify that safety case requirements are implemented do not require confirmation that the measure is implemented in a way that ensure it is achievable and meaningful to end users. This later deficiency has been identified by the INA function, with improvements planned, which we support. We also identified gaps in the arrangements for Operating Limits (OL). Currently BAESML arrangements do not require OLs to be derived into actions that are achievable or substantiated. OLs are not treated in the same way as safety actions (SA), although they can have equal importance to nuclear safety.
With respect to ownership of the implemented safety case, we found issues in that ownership is retained by the safety case production team, rather than sitting with those responsible for compliance against it. Again, this is an area where the INA function has identified improvement and has oversight of it.
When examining a fault from a recent safety case, we identified several areas for improvement. Firstly, we found insufficient justification for the categorisation of nuclear safety significant functions. We also identified that the classification of some operating rules was not consistent with the importance of the function it supported, and not consistent with BAESML guidance. Additionally, we found one instance where the required safety function was not fully met by the wording of the identified SAs (e.g. the SAs could be complied with, without the required safety function being achieved). Further, we found that the evidence presented to underpin the SAs in the HF substantiation did not adequately confirm that the function required by the safety case was met, or that the operating rule will be reliably delivered. BAESML accepted these findings. We discussed them with the safety case and operations team. During this, we gained some confidence that the function required by the safety case was met in the way the operations are controlled. However, whilst we did not identify an immediate risk to nuclear safety, we consider that together these shortfalls are significant and undermine the adequacy of the case presented.
Conclusion of Intervention
BAESML has several activities ongoing to improve their arrangements for LC 14 and LC 23 including updating processes and guidance, changing the ownership of safety cases, and improving positive compliance against operating rules. As BAESML have acknowledged the gaps and have identified improvements, we rated LC 14 GREEN.
However, the pace of the improvements is relatively slow and BAESML was unable to provide a detailed plan as to how the individual aspects are pulled together into a coherent plan. We considered that the gaps are impacting BAESML’s ability to produce an adequate safety case, and whilst we identified no immediate specific risk relating to the nuclear safety of operations, we consider that the gaps identified increase the likelihood that they could exist. As such, we have rated LC 23 AMBER.
We have raised one new level 4 regulatory issue against LC 14 to ensure that improvements to safety case processes in general are identified and delivered in a timely manner to support development of future Dreadnought safety cases. We will also raise a level 3 regulatory issue against LC 23 to ensure the development and application of methods for the deriving and implementing safety case requirements.