Keywords: offshore, fires, explosions, risk, reliability, Monte Carlo simulation, sensitivity analysis.

Incidents involving fires and explosions present a major hazard to the workforce on offshore oil and gas platforms. Following the Piper Alpha Disaster in 1988, platform operators for the UK sector are required to submit safety cases for approval by the Health and Safety Executive. A key requirement of these safety cases is that hazards associated with an accidental release of hydrocarbons have been demonstrated to be as low as reasonably practicable.

This paper describes an analysis using the existing SAROS (Safety and Reliability of Offshore Structures) software, developed to estimate the expected frequency of fatalities on offshore platforms with open-sided modules. The analysis involves identification and variation of a number of key input parameters within the model to determine the effect on the estimated frequency and magnitude of jet fires, pool fires and explosions and consequent fatalities. The aim is to identify propositions that can be implemented on offshore platforms in order to minimise the risk of fatalities.

There are two fundamental types of module, open and enclosed, categorised according to the module construction. Open modules are open-sided and ventilated by the wind, these module types will be considered within this paper.

The configuration of equipment and inventory within each process module on a platform depends on the function of that module, two types of module are considered in this paper, Separation and Compression. Oil, gas and water mixture is passed to the separation module, where the water is drained and the oil separated from the gas. The gas then passes to the compression module where it is repeatedly pressurised and condensate removed, before leaving the platform.

Each module contains a number of isolatable process sections containing hydrocarbon fluids. These sections have the potential to depressurise, or blowdown, routing gas to the flare. On detection of a hydrocarbon release the isolation valves located between each section will close, reducing the inventory available for release into the module. Both systems function together to reduce the magnitude of a discharge.

A release into a module will be detected either by an automatic gas or fire detection system or be detected manually dependent on the nature of the release. On detection the deluge system will be activated. This system releases a spray of water across the affected part of the module with the intention of suppressing the severity of an ignition.

Safety walls are constructed between the modules on the platform, designed to minimise the impact of an explosion or fire through containment within the originating module.

SAROS models fires and explosions using the Monte Carlo Simulation method by determining the attributes of the initial release, calculating the fire or explosion characteristics and predicting the frequency of fatalities per year [1].

A sensitivity analysis has been developed, that varies a number of key parameters in order to determine the sensitivity of the model to each system and determine whether improvements to any system or systems would benefit the platform by reducing the predicted frequencies of fatalities per year. The results for the following input parameters are presented within this paper:

• Unavailability of isolation and blowdown valves
• Unavailability of gas detection systems
• Unavailability of deluge systems
• Resistance of blast and fire walls

Reducing the unavailability of the detection system is predicted to have the greatest effect on the reduction of the frequency of fatalities per year. The largest variation is predicted when the unavailability of the beam detector is varied.

The effect of reducing the unavailability of the deluge system appeared to vary dependent on the type of module. A module containing a significant proportion of oil predicted an increase in the frequencies of fatalities per year when the unavailability was reduced, however a reduction in the fatalities was predicted when the module contained little or no oil.

Variation of the unavailability of the isolation and blowdown systems within the compression module resulted in minimal variation in the fractions of fatalities. A larger variation was predicted within the separation module.

A reduction in the unavailability of the fire and blast walls has the smallest effect in the frequencies of fatalities.

The results from this paper suggest that individual improvements to safety systems would not have a significant effect on reducing the frequencies of fatalities however a combination of in particular the gas detection and deluge systems, could lead to a more significant decrease in the frequencies of fatalities within open sided modules per year.

1

J.L. Lewthwaite, J.D. Andrews, C.A.J. Gregory, R. Smith, "Risk Modelling of Fires and Explosions in Open-Sided Offshore Platform Modules", Proceedings of the 16th Advances in Reliability Technology Symposium, 243-259, 2005.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £140 +P&P)