the requirements regarding training increase, with the result that there may be more players
involved who need to share identical information or that there will be technical requirements which require the systems to collate different information in a single management system. This requires both provision for strategic control from many levels, as well as the development of standards to achieve the type of interaction e.g. from an eco-system perspective (Manikas et al., 2013; Johnsen et al., 2017) or the prioritisation of open innovation (Blankesteijn, et al (2019). There is a need for MTO measures, i.e. both integrated system design and expertise relating to interaction between distributed players / organisations. It is proposed that the sector identify needs regarding integration, possible benefits and challenges (technical and organisational obstacles) in order to identify
opportunities when integrating a number of safety-critical systems. At the same time, experience gained concerning remote control should be systematised and structured. This is an area which over the past 30 years has seen the increasingly centralised control of complex processing facilities from larger centres, which often cut across national borders (with improved safety and efficiency and cost reductions as a result).
Findings and measures [V4]In-depth study from RNNP, analyse events involving autonomous systems. There is no systematic collation of incidents and near misses relating to
automated/autonomous systems. A preliminary study should be carried out in collaboration with PSA to determine whether it would be possible/appropriate to analyse incidents and near misses from autonomous systems. This could help to clarify the framework conditions needed to carry out such a study, and possibly the need to develop defined hazards and accidents involving
automated/autonomous systems.
Findings and measures [V5]Lessons learned from successful transformation projects (digitalisation/automation)
Many aspects of the projects we have examined have had a positive impact on safety levels,
efficiency and quality in connection with the introduction and development of technologies. This is for example linked to early user participation, the follow-up of training, delimitations linked to having a single responsible supplier, and the management prioritising and supporting the project.
The industry should give greater priority than it does at present to identifying successful factors and processes that other players and projects can learn from, e.g. with a perspective aimed at learning from proved to be successful (Buckingham et al., 2019).
This appears to be difficult to share across suppliers and operators, e.g. as a result of commercial considerations such as competition. From a safety perspective, there should be good mechanisms in place for sharing best practice between the stakeholders involved. We propose placing a spotlight on the development process rather than specific products that have been developed. What can we learn from each other's processes in order to develop systems that address human and organisational factors?
The stakeholders and PSA could develop a form of best practice meeting series or initiatives in this area. Similar work has been carried out in the field of well control incidents through "Sharing to be better". The aim here could be to initiate a period of maturation during which PSA makes the industry aware of the issue, which in turn could lead to specific measures being implemented by various stakeholders over time.
Findings and measures [V6]Supervisory series linked to the use of HF in connection with automation/digitalisation. Given the lack of attention being paid to HF in some projects, PSA could carry out a series of supervisions in cooperation with specialist HF groups in order to review
compliance with regulations, look at relevant knowledge concerning the use of HF (e.g. in drilling and wells, process control) and evaluate the effects of projects that have been carried out both with and without input from the specialist community which has a knowledge of HF. As mentioned previously, the definition and prioritisation of such a series of supervisions should be formulated in cooperation with stakeholders with experience of using HF and the challenges that are faced within the industry in complying with the current requirements. It is therefore proposed that knowledge concerning HF be drawn in not only from the stakeholders and the groups that work with human factors and safety amongst the companies, but also consultants with a high level of expertise relating to HF (e.g. from design groups which have a broad multidisciplinary industry knowledge).
The aim should be to document the status in relation to the regulations, influence projects that are in their early phase, and disseminate knowledge concerning the effects of using HF.
Findings and measures [V7]Detailed analysis of the development of automated drilling operations.
In this report, we have discussed developments and described example projects, which illustrate the work that is under way relating to remote control and automation of aspects of the drilling process.
A detailed analysis of whether fully automated drilling operations are possible from a technical, safety and economic perspective was outside the scope of this report. There are no specific technical obstacles to achieving this, but increasing complexity and vulnerability will present challenges that require good analysis, good interaction design, good processes for user involvement and good testing and ongoing learning. We therefore propose that a somewhat more detailed investigation be carried out which exclusively focusses on analyses of more fully automated drilling operations, with a review of relevant automation levels (including analyses of adaptive automation), analyses of safety-critical operations, the design of interaction design, the establishment of HMI for high performance and other relevant issues. An evaluation of AI should form part of such a project, via a
"state of the art assessment". AI can be utilised very effectively in many areas (Bello et al., 2015;
Kirschbaum, et al., 2020). At https://www.ai-safety.org, there are discussions concerning the framework for the use of AI. It is not only models that can be discussed, but also the surrounding framework; see Figure V7.1.
Figure V7.1 Theme for AI – see https://www.ai-safety.org
On the basis of the specific use of methods within HF, the contributions that HF has made concerning design could be reviewed and documented with regard to highlighting the interaction between automation and humans. One could discuss the principles for interaction design with "High
Performance HMI", practice for alarms, the interaction between stakeholders and the
responsibilities involved in drilling and wells. Here, one should carefully review key aspects of the distribution of responsibilities between operator, contractor and supplier, i.e. who does what with what equipment in order to detect and manage a well control incident? How does automation alter work processes and the distribution of responsibilities based on different strategies for automation (e.g. adaptive automation and AI solutions)? What specific methods are used to identify potential risk as part of projects, and what are considered to be the key risk elements?