(386f) When Does a Vessel Become a Pipe: A Simplified Approach to Pressurised Outflow Simulation

Pressurised pipelines are by far the most widely used method for transporting oil and gas across the globe. In the US alone, almost 40% of its energy demand is supplied through such means. In some other countries this figure approaches 100%.

Given that such pipelines can be several hundreds of kilometres long conveying thousands of tonnes of inventory at pressures as high as 150 bara, their accidental rupture can lead to catastrophic consequences. Indeed there have been numerous incidents involving pipeline ruptures which have resulted in significant number of casualties, property damage and environmental contamination. It is now a statuary requirement to assess all the major hazards associated with pressurised pipelines prior to their commissioning. Pivotal to this exercise is the prediction of mass outflow and its variation with time in the event of pipeline rupture. Such data dictates all the consequences associated with pipeline failure including fire, explosion and toxic release. It lays the foundation for pipeline operators and safety authorities to determine minimum safe distances to populated areas and formulate emergency response planning.

Given the highly unsteady state flows involved during pipeline rupture, the modelling of outflow has been confined to numerical techniques. Despite their success in accurately predicting outflow, a major drawback has been the long computational run times; typically a few days for the simulation of the complete depressurisation of a long (e.g 150 km, 0.7 m i.d) pressurised (e.g 100 bara) pipeline following its puncture.

This study describes the development of a simplified analytical model based on approximating the punctured pipeline as a vessel discharging through an orifice designed to address the above problem. The efficacy of the vessel model in terms of computational run time and accuracy is determined by comparison against an established but computationally demanding rigorous numerical technique based on the solution of the conservation equations using the Method of Characteristics [1?3].

The discharge of different pressurised inventories such as permanent gases (e.g methane, ethane and their mixtures) and two-phase mixtures (e.g. mixture of methane and pentane) are simulated to identify the range of applicability of the vessel model as a function various design and operating parameters These include the ratio of the puncture diameter to the pipeline diameter (0.1 ? 0.5), initial line pressure (21.6 bara and 100 bara) and pipeline length (100 m ? 5 km). In the case of the permanent gases, the results indicate that the accuracy of the vessel model in terms of the degree of agreement with the numerical model increases with decreasing diameter ratio, line pressure, and pipeline length. The finite differences (±11%) are shown to be primarily due to the isothermal gas expansion assumption incorporated in the vessel model development.

Much the same as the permanent gas inventories, for two-phase mixtures, the vessel model accuracy increases with increasing diameter ratio. However no obvious trends in the degree of agreement are obtained with changes in the line pressure or the pipeline length. This is believed to be due to the complicated relationship between the speed of sound with pressure for the two-phase mixtures in turn controlling the expansion wave velocity within the pipeline and hence the pipeline depressurisation rate. Such phenomena are not considered in the simplified analytical vessel model. Despite this, surprisingly the maximum error in the prediction of the vessel model in the case of the two-phase mixtures is only ±5% which is significantly better (c.f ±11%) than that for the permanent gas predictions.

These acceptable errors coupled with the negligible computational run time indicate the clear advantages of using the vessel model presented in this work for simulating the puncture of long pipelines containing pressurised hydrocarbons for the ranges tested.

References

1.Mahgerefteh, H., & Denton, G., & Rykov, Y., A hybrid multiphase flow model (2008), AIChE Journal, 54(9), 2261-2268.

2.Mahgerefteh, H., & Abbasi, M., Modeling blowdown of pipelines under fire attack (2007), AIChE Journal, 53(9), 2443-2450.

3.Mahgerefteh, H., & Oke, A., 'An efficient numerical simulation for highly transient flows' (2006), Chem.Eng.Sci., 61(15), 5049-5056.