Risk modelling of safety critical systems for life extension in offshore oil and gas
Publication Date
December 13, 2022
Creator
Abstract
On the United Kingdom Continental Shelf over half of the platforms are operating beyond their design life in a stage that is known as Life Extension (LE). Their extended life is due to the fact that fields that the platforms are extracting from are productive enough to continue to produce oil and gas at a rate that remains profitable to the operator. The rate of production is typically reduced in this stage and therefore the financial pressure on the platform is increased. Through the ageing of the installation the assets on the platform have undergone deterioration, to manage this operators have performed inspections and maintenance actions to control the condition of the assets, this is known as Asset Management (AM). Through the LE stage the deterioration rate of the assets has the potential to increase, meanwhile the financial pressure is increasing and as such there arises a problem of performing AM tasks such that it is optimised financially. Furthermore, due to the volatile fluids handled on-board it is imperative to ensure that the operation of the platform remains safe, therefore, a decision on performing AM tasks must also consider the probability and consequences of a loss of containment.
This thesis aims to construct a model for the production, generation and firewater deluge systems of an offshore oil and gas platform that simulates the AM of the assets involved in these systems. This will measure the performance of the platform through metrics representing different key indicators, these include the whole-life production availability, the AM expenditures and the risk of fires and explosions. A case-study platform known as Beryl B, is selected, this platform is an appropriate target as it is currently operating in a LE stage. The design of the systems are taken from this platform such that they appropriately reflect a real-world design. The key outcome of the research is the construction of a model for determining the performance of AM strategies and furthermore, an optimisation algorithm for the finding optimal strategies.
The primary technique of this research is a Petri Net (PN), this was selected as it can easily model system structures and interdependencies between arbitrary nodes. A novel extension to the traditional PN was created for this work which combined discrete and continuous elements to produce a hybrid technique. The net is used to model the degradation, AM and production of the platform throughout its life. The discrete net simulates instantaneous, constantly timed and stochastically timed events, which include periodic inspections, maintenance, and stochastic degradation. Additionally, the system structures are represented in the discrete net such that faults can be connected to represent redundancies and system failures. Whereas, the continuous net simulates the constant flow of oil and gas through the production vessels throughout the platform’s life to determine production quantities. The hybrid net allows for these parts to be connected, this is utilised such that the flows are dependent on the asset’s conditions and failures. This forms a powerful technique for simulating the assets of an oil and gas processing platform, considering degradation, inspection, maintenance and production.
The oil and gas within the production vessels are volatile fluids that can ignite and cause damage or loss of life, it is therefore imperative that the frequency of leaks and the possible escalation are well understood. To this end, a consequence analysis technique is applied to the problem with the purpose of evaluating the risk of possible escalations following a loss of containment event. Specifically, three escalations are considered which are caused by ignitions in different situations, these are jet fires, pool fires and explosions. The consequence analysis is integrated into the PN through its simulation of asset degradation, this is such that when a leak failure mode occurs for a production vessel the consequence analysis evaluates risk profiles based on the initiating conditions. The initial conditions of the analysis are dependent on the vessel’s dimensions and fluid contents and thereby give unique results for each vessel. Combined, this can evaluate the performance of a given AM strategy on the platform’s through metrics relating to production quantities, AM expenditure and consequence frequencies and magnitudes.
This model is extended by the addition of an optimisation algorithm to determine optimal AM strategies for the platform from a whole-life perspective. Specifically the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) is applied to the problem, this is a multi-objective evolutionary algorithm. A multi-objective algorithm was selected as the platform’s objectives are not directly comparable, these were selected to quantify the production, cost and risk associated with a strategy. The algorithm produced a set of solutions that approach the Pareto surface, this does not require any preference to be made between the objectives, instead a user can select after the solutions have been generated. By examining the solutions through the generations the evolutionary progress can be seen to produce a solution set which is simultaneously diverse and near-optimal.
Item Type
ethesis
Thesis Type
PhD
Supervisors
Subjects (LC)
Associated Schools / Departments
Department of Civil Engineering (UK)
eprints ID
71750
UoN Repository URI
Except where otherwise noted, this item's license is described as
File(s)![Thumbnail Image]()
Name
thesis.pdf
Type
Full-text
Description
Examined
Size
7.8 MB
Format
Adobe PDF
Checksum (MD5)
4d361856ce9e4303a4e408f9fb627c68