A Data-Driven Decision Support Tool for Offshore Oil and Gas Decommissioning

A growing number of oil and gas offshore infrastructures across the globe are approaching end their operational life. It is a major challenge for industry to plan make decision on decommissioning as processes resource exhaustive. Whether facility completely removed, partially removed or left in-situ, each option will affect individual parties differently. Stakeholders’ concerns needs collected analyzed obtain most compromised decision. Engaging with hundreds stakeholders extremely complicated, hence time-consuming costly. This issue can be addressed using predictive model provide suggested options based data previously approved projects. However, lack readily available relevant datasets main hindrance such an approach. In this paper, we introduce new dataset extensively covering all types in UK landscape over 21-year period. An experimental framework several learning algorithms predicting presented. Various resampling methods were applied tackle imbalanced class distribution improved classification. Promising results achieved despite exclusion some stakeholder-related features used traditional shows signs potential solution significantly reduce time cost spent project, encourages more efforts put into researching timely topic.