Decision-support for decommissioning offshore platforms.

An estimated 2,500 offshore decommissioning projects are expected to be completed between 2018 and 2040 with significant accompanying challenges. In this research, a decision model for decommissioning offshore platforms is developed. The decommissioning decision model (DDM) aids logical determination of the optimal option for decommissioning a platform through a multicriteria decision analysis of the considered options with respect to safety, cost, environmental impact, technical feasibility, and public perception. It synthesizes information about a platform's features with expert opinion to identify the best option for decommissioning the platform from a list of available options. It also facilitates the progressive integration of historical data to replace subjective human opinion and improve the quality of decision-making as this becomes available. A case-study approach was used to demonstrate the DDM's applicability with information from an industry survey of decommissioning practitioners. Five decommissioning options were considered for the case study platform, and these were evaluated with a hybrid of Likert scale and Analytic Hierarchy Process (AHP). Using this technique, the optimal option for decommissioning the case study was determined with a 60% efficiency savings in time taken to complete the analysis as compared to the traditional AHP process. Results showed that partial removal is the preferred option for the case study, and the platform features with high relevance to options selection are substructure weight, water depth and age. Moreso, respondents from the North Sea were observed to be more averse to leaving platform materials in place as compared to people from Offshore USA, Africa, and Asian Seas. These findings were seen to agree with literature and industry practice through a comprehensive validation process. Thus, evidencing the DDM's flexibility and robustness and making a case for its industry adoption. After its validation, the DDM's capability to support integration of historical data was investigated with the aid of a prediction model for estimating the costs of using different options for decommissioning offshore platforms. This costing model was developed by applying machine learning regression to historical decommissioning cost data. The model predicts decommissioning options costs for five different scenarios with reasonable accuracy as indicated by an r-squared value of 0.935, implying that it is reliable for predicting decommissioning costs. It was used to predict decommissioning options costs for the case study. These costs were then integrated into the DDM to replace the input data for cost criterion as obtained from the survey. The models developed in this research improve upon the existing works in decommissioning optimisation. Industry adoption of the decision model will result to significant reduction of time, resources and efforts spent in decision-making during decommissioning. By acting as an unbiased basis for justifying the choice of a decommissioning option for an offshore asset, the DDM mitigates the traditional conflict between stakeholders of decommissioning projects. The costing model aids early estimation of decommissioning costs for budgeting, asset trading and other preliminary cost evaluation purposes prior to detailed engineering cost estimation. Therefore, both models represent a significant contribution towards the advancement of the current offshore decommissioning practice.

Name dc:type.qualificationname

PhD

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

Robert Gordon University

Year dc:date.issued

2023

Author dc:creator

• Eke, Emmanuel Chibudike

Advisor dc:contributor.advisor

• I. Iyalla, R. Prabhu and J. Andrawus

• Decommissioning
• Oil and gas industry
• Offshore platforms
• Decision-making

Language dc:language

en

Identifier

oai:rgu-repository.worktribe.com:2270718
https://doi.org/10.48526/rgu-wt-2270718

Author Identifier

0000-0003-0434-5658

OAI identifier oai:identifier

oai:rgu-repository.worktribe.com:2270718