Quantitative risk assessment of the performance of barriers controlling fire and explosion risk

Abstract

The growing population of the world places an enormous demand for energy in some form or the other. Essentially, this means that the struggle to hunt and retrieve fossil fuels has to continue for the smooth functioning of the world. Since the early 1900’s when drilling for oil and gas started moving away from shores to some remote locations that we are at today, the size and complexity of machinery and the interconnected systems have grown exponentially. Despite investing enormous time, effort and money into implementing technically advanced and innovative barrier systems to prevent Hydrocarbon (HC) leaks on offshore installations in the Norwegian Continental Shelf (NCS), the statistics still stands far from the target zero. Hence, the efforts towards minimizing the number of leaks can never be ceased due to the catastrophic nature of its potential consequences to humans, environment, asset or the reputation of a firm.

This master’s thesis features one such effort to visualize the effectiveness of some barrier functions proposed in Modelling Instantaneous Risk for Major Accident Prevention (MIRMAP) report. Some of the most critical barrier systems are incorporated into the chosen generic module, which is then subjected to numerous gas leak simulations using the latest version of the software, Kameleon FireEx - Risk and Barrier Management (KFX-RBM), a Computational Fluid Dynamics (CFD) based simulation tool. A base case scenario is a set-up using the chosen module with a 100% functional gas detection system, Emergency Shutdown (ESD) system, with no temporary weather cladding attached and closed fire-proof doors. This set-up is subjected to simulation with six different leak rates, four different wind speeds and two different wind directions. The results of this being a base for comparison, simulations are carried out with partially isolated gas detection system, presuming on-going hot works in the near vicinity with a temporary weather cladding and with fireproof doors left in open position. The variations in the total ignition probability of the module compared to the base case would be the main objective of this thesis. As additional objectives, since the barrier systems could be directly or indirectly influenced by humans, an attempt is made by simulating the base case with a practically acceptable delay in manual activation of the shutdown. Furthermore, to reduce the simulation time, numerous simulations are run to arrive at the optimal grid resolution and courant number with the quality of results remaining undiminished.

The case with delayed shutdown due to manual activation of ESD depicted the highest influence to the complementary cumulative frequencies, followed by the case simulated with a temporary weather cladding erected on one of the open ends of the module, with second highest influence. The comparison of the two base case simulations with different leak scenarios (direction and location), revealed that the release point and release direction can significantly influence the probabilities of ignition, either positively or negatively. The simulations that were carried out to optimize the grid resolution resulted in a significant reduction of simulation time with a grid resolution of 125000, while the quality of the results was undiminished. Further analysis by extracting the highest and the least release rates separately resulted in an insight that the grid resolution is a function of release rate.