Asset Reliability and Risk Interoperability to Optimise Maintenance Execution and Improve Risk Management of Energy Producers (21.RP3.0106) – Completed

The Objectives

The Asset Reliability and Risk Interoperability (ARRI) project is improving traditional risk management by integrating previously siloed reliability and risk models into a unified, dynamic framework. By harnessing real-time operational parameters and asset health estimates, the project aims to develop adaptive risk models that evolve in response to changing conditions. The primary goal is to improve decision-making and risk mitigation through integrating domain models with live asset data and advanced modelling techniques. This integrated approach enables a more predictive and responsive risk management capability, ultimately supporting safer and more efficient operations across complex industrial systems. Through this approach, the ARRI project seeks to establish a new benchmark for dynamic risk modelling, bridging the gap between static analysis and the variability of real-world operations.

Operating offshore processing equipment safely and reliably presents significant challenges. These arise from the complexity and volume of components, rapidly changing environmental conditions, and strict constraints on maintenance activities due to limited access, concurrent operations, and regulatory requirements. In such demanding environments, it is essential to prioritise maintenance tasks effectively to ensure that the right work is performed at the right time.

Before any changes to maintenance plans are approved, a comprehensive risk assessment must be conducted. This involves identifying potential hazards, evaluating the operating context, and understanding the implications of proposed changes. Risk mitigation strategies must also be clearly defined and implemented. In general, current approaches to risk identification and maintenance prioritisation are manually driven with some assistance from data analytics on occasions where suitable data are available. As a result, there exist a number of challenges to the improvement of risk management, including:

1. explainability of risk drivers and root causes when using simple analytics methods or limited data, or both;
2. complexity in the characterisation of how risks change over time due to the complexity of the underlying systems, relationships, and interactions (such as delayed impacts as a result of mitigations designed into systems);
3. understanding cumulative risks at different levels of a system and across interconnected systems;
4. managing changes in risk when tasks are delayed or reprioritised due to the uncertainties inherent in risk analysis (i.e., it is impossible to guarantee that taking one course of action over another will not result in an undesirable outcome at some time in the future).

Evaluating trade-offs for infrequent but critical tasks, such as once-in-a-decade maintenance, is particularly difficult. Moreover, current approaches typically assess different risk dimensions independently to handle the complexity of the interactions. Such simplifications rely on conservative assumptions that affect business viability, struggle with evolving definitions of maintenance activities and their associated risks, and offer limited traceability of risk in relation to constraints like scheduling, capacity and planning criteria.

To address these challenges, there is a clear need for models capable of identifying and assessing cumulative risk and its impacts, including where such impact is delayed, through the complex relationships and interactions that exist in real-world processing plants and facilities. The resulting models will support informed maintenance decision-making and strategic optimisation, allowing simulation of proposed changes and evaluation of their risk implications before implementation.

Results and Conclusions

To manage the complexity of risk modelling in large-scale systems, a model-driven approach was adopted which integrates conceptual and data models incorporating both static and dynamic system models with reliability, risk, and maintenance data. This led to the development of the ARRI Framework: a dynamic, extensible platform designed to support integrated and interoperable risk analysis.

The ARRI Framework comprises a set of conceptual models which are used to represent and integrate the information required for dynamic risk assessment and maintenance schedule optimisation, including models representing system structures, their interactions, and their reliability, maintenance schedules, risk (including impact across multiple risk dimensions), criticality and maintenance optimisation.

By combining asset modelling with time-dependent reliability, impact, criticality, and risk analysis, the framework delivers a comprehensive view of risk across complex systems. It integrates static asset hierarchies with real-time health estimates to reflect current asset conditions, leading to dynamic risk assessments of inter-connected systems based on reliability projections of assets at different levels. The framework also evaluates changes in risk across maintenance plans, supporting the analysis of changes in impact, criticality, and risk as asset conditions evolve. This enables simulation of alternative maintenance strategies to compare total risk and cost before implementation. Moreover, the models are used to perform risk-based optimisation of scheduled work to help lower overall risk.

Through this unified and adaptable structure, the ARRI Framework supports stakeholders to make informed, data-driven decisions in dynamic operational environments.

A prototype implementation of the ARRI framework was applied to a case study of a large facility for which the framework generated dynamic risk and criticality scores that were subsequently used in the generation of various optimisation scenarios to evaluate the behaviour of the optimisation process. Thus, demonstrating the potential for improved maintenance prioritisation using the ARRI framework.

Next Steps

Improving risk management for complex industrial plants requires the adoption of dynamic and adaptive risk analyses that evolve in response changing conditions. Doing so requires integrating models of systems, their relationships, their interactions, multi-dimensional risks, reliability, and maintenance activities in a coherent fashion. Adopting a framework such as the ARRI framework would then allow dynamic assessment of changing impact and risk through and across systems as well as their associated maintenance activities, thus enabling risk-based optimisation of maintenance plans.

Evaluation of the ARRI Framework prototype has identified two major challenges: scaling the framework to assess and optimise dynamic risk across entire facilities, and the limited availability of reliability data required for accurate risk evaluation. Scalability is essential, as dynamic risk assessments and optimisation must accommodate the complex nature of industrial systems. At the same time, large-scale industrial environments suffer from varying degrees of reliability data completeness and consistency across different asset classes, for example common equipment with well-studied and (comparatively) predictable failure modes versus long-lived assets with minimal failure history, which makes it difficult to predict future risks and maintenance needs.

To address these limitations, the next phase of the ARRI project will analyse reliability data requirements using the baseline implementation of dynamic risk assessment and maintenance optimisation as a testbed. This analysis will determine what types of data are needed, how much is required, the necessary level of precision, and where data has the greatest impact on risk and optimisation outcomes. These insights will guide future enhancements to the framework.

The project will also explore the broader operational context to identify practical solutions for resolving reliability data gaps. These may include integrating third-party reliability data catalogues or improving the way risk assessments are conducted when maintenance notifications are raised by personnel.

As development continues, the ARRI Framework will be refined iteratively. Supported by evaluations with FEnEx CRC industrial participants, this process will incorporate solutions to the reliability data challenge and expand the scope of dynamic risk assessment and maintenance optimisation while ensuring that the framework can scale effectively to meet the demands of complex, facility-wide applications.