Our News

From across our global network

  • News

Harnessing Data Science For Advanced Pipeline Integrity Management: A Novel Approach To Reliability Centered Maintenance (RCM)

01 Jul

Predictive Modelling at the Core

Pipeline leaks can lead to significant consequences, including threats to human life, damage to the environment, property, and reputation, as well as financial losses. With a significant amount of pipeline assets globally being over 40 years old, the risk of failure is increasing as managing these aged assets becomes more challenging. The continuing demand for oil and gas will mean these ageing systems will need to function safely and efficiently for the foreseeable future.

For some operators, recurrent failures due to pre-existing and ongoing integrity issues are already a reality. To address these challenges, a proactive approach is needed, such as using data science and reliability modelling to support proactive maintenance and inspection activities.

Central to this innovative strategy is the adoption of predictive modelling techniques, which utilise historical data to forecast the occurrence of corrosion-related failures. This predictive foundation allows for the precise estimation of when and under what conditions pipeline segments are likely to experience corrosion, by examining a comprehensive set of influencing factors including operational conditions, environmental exposures, material properties, and historical maintenance records.

By incorporating Monte Carlo simulation, the strategy further embraces the complexity and variability inherent in pipeline systems, enabling a robust analysis of how uncertainties in key parameters might affect overall system reliability. This dual approach not only offers a granular view of potential future states of the pipeline system but also enables the crafting of maintenance strategies that are both proactive and highly targeted.

Pillars of the Predictive Data-Driven RCM Strategy

  • Predictive Corrosion Modelling
  • Embracing Uncertainty with Simulation
  • Data-Driven Insights for System-Wide Analysis
  • Addressing Data Limitations
  • Predictive Maintenance Strategy Optimisation
  • Risk Assessment and Proactive Decision Making
  • Continuous Improvement through Learning

Implementing Predictive Data-Driven RCM

The adoption of this predictive data-driven RCM approach facilitates a shift from reactive to proactive maintenance, allowing organisations to prioritise and optimise maintenance actions before failures occur. This not only minimises downtime but also extends the lifespan of pipeline assets. Continuous improvement mechanisms ensure that the model and strategies remain relevant and effective, adapting to new data and insights.

The Evolution of Predictive Maintenance

This methodology leverages the latest in predictive data science to offer a sophisticated toolkit for infrastructure management. By predicting future maintenance needs and assessing potential risks with a high degree of accuracy, this approach empowers decision-makers to act pre-emptively, ensuring the safety, reliability, and efficiency of pipeline systems.

The predictive, data-driven approach to RCM marks a significant advancement in how infrastructure maintenance is conceptualised and implemented. It underscores the critical role of predictive data science in transitioning from merely managing maintenance to strategically preventing failures, paving the way for more resilient and sustainable infrastructure systems.

Case study

AIE has recently supported a prominent oil operator in conducting a study on their pipeline system to prioritise inspection planning as well as repair and replacement scope prioritisation. The onshore pipeline system was experiencing frequent and recurrent failures in the forms of both leaks and ruptures due to severe internal corrosion.

To address these challenges, a robust data-driven prediction model was developed which was capable of discerning correlations between various pipeline parameters and survival rates. The intent of this approach was to shift from reactive to proactive maintenance, allowing the client to prioritise and optimise maintenance actions before failures occur.

A prioritised inspection and replacement plan was generated for the client which was driven by historical data and predictive analysis. As part of the simulation modelling, AIE also identified key “Find and Fix” pipelines which showed improved availability if they were inspected and subsequently repaired to restore pipeline integrity

The traditional method involves a collaboration between the operation and subsurface teams to identify flowline replacements that add the most value to the system. The integrity management team is also consulted on the proposed list.

On the other hand, AIE’s data-driven approach combines operations, subsurface, and integrity elements with data science survival modelling to enhance the process and achieve best results. The generated plan included a list of pipelines for replacement, inspection, engineering assessment, and strategic planning. The plan included:

  • Inspection candidates, which are high-risk pipelines that can be repaired rather than replaced.
  • Engineering assessment pipelines, which are predicted to fail imminently but have not experienced any failures to date.
  • Strategic planning pipelines, which are those that could be better placed in replacement plans in the next few years.

In summary, our data science approach has:

  • Enhanced value over engineering-led planning both in terms of managed interventions and budget required.
  • Allocated replacement resources and costs strategically to prioritise the most critical pipelines.
  • Allocated 82% of the original budget to replacement plans, with the remaining budget split between targeted inspection and strategic future planning.
  • Enabled failure and reliability analysis to address issues from limited data, surpassing the capabilities of prognostic methods.
  • Improved compliance through proactive risk assessment and decision-making.

Conclusion

The integration of predictive data science into Reliability Centered Maintenance represents a forward-thinking approach to pipeline management, embodying the principles of efficiency, reliability, and proactivity. This methodological advancement promises not just to respond to the challenges of today but to pre-empt those of tomorrow, offering a strategic blueprint for the future of infrastructure maintenance and management. As predictive modelling and data science continue to evolve, their potential to revolutionise the maintenance of critical infrastructure becomes increasingly clear, setting a new standard for excellence in the field.

Stay tuned for more articles on our Data Science services.

To learn more about our pipeline Integrity and data science offerings, contact us today.



Recent articles: