Page 386 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

-
Appendix W361

Appendix B
Leak Rate

Determination

Fluid flow through pipelines is a complex and not completely Release duration is arbitrarily chosen at 10 minutes for a gas
understood problem. It is the subject of continuing research by and 60 minutes for a liquid.
engineers, physicists, and, more recently, those studying non- Complete line rupture (guillotine-type failure) is used.'
linear dynamic systems, popularly called the science of chaos. Operation at MAOP is taken as the initial condition.+
In a relative risk assessment, we are less concerned with exact 0 Initial conditions are assumed to continue for the entire
numerical solutions, and more interested in comparative release duration (except for flashing fluids).
values. 0 Depressurization, flow reductions, etc., which occur during
In general, fluid flow in pipes is assigned to one of two flow the release scenario, are generally ignored.
regimes, turbulent or laminar. Some make distinctions between An arbitrary transition point from liquid to gas is chosen for
rough turbulent and smooth turbulent, and a region termed the flashing fluids.
transition zone is also recognized. However, in simplest terms, 0 Pooling of liquids and vapor generation from those pools is
the flow pattern will be characterized by uniform, parallel ignored.
velocities of fluid particles-laminar flow-a by turbulent 0 Temperature effects are ignored in the equations but should
eddies and circular patterns of fluid particle velocities-turbu- be considered in choosing the liquid calculation versus the
lent flow--or by some pattern that is a combination ofthe two. gas calculation. The evaluator should assume the worst case.
The flow pattern is dependent on the fluid average velocity, the for example, a butane release on a cold day versus a hot day.
fluid kinematic viscosity, the pipe diameter, and the roughness 0 Pressure due to elevation effects is considered to be a part of
ofthe inside wall of the pipe. MAOP.
Several formulas that relate these parameters to fluid density
and pressure drop offer approximate solutions for each flow These are often conservative and appropriate assumptions for
regime. These formulas make a distinction between compressi- risk modeling. However, using any simplifying parameters
ble and non-compressible fluids. Liquids such as crude oil, must not mask a worst case scenario. The parameters are
gasoline, and water are considered to be non-compressible, selected to usually reflect conservative, worst case scenarios.
whereas gases such as methane, nitrogen, and oxygen are con- The evaluator must affirm that one or more ofthe above param-
sidered to be compressible. Highly volatile products such as eters does not actually reflect a less severe scenario.
ethylene, propane, and propylene are generally transported as Again, almost any consistent modeling of a leak quantity will
dense gases-they are compressed in the pipeline until their serve the purpose of a relative risk assessment. Consistency is
properties resemble those of a liquid, but will immediately absolutely critical, however. One approach that is currently in use
return to a gaseous state on release of the pressure. involves the above parameters and model releases as follows:
For purposes of a relative risk assessment, any consistent
method of flow calculation can be used. Because the primary 0 Gas-the quantity of gas released from a full-bore line rup-
intent here is not to perform flow calculations but rather to tured at MOP (or normal operating pressure) for 10 minutes.
quickly determine relative leak quantities, some simplifying
parameters are in order. Original (first two editions of this 'Reasoning behind selection of this parameter is provided In Chapter 7.
book) suggestions for calculations of a Leak Impact Factor tAs an alternative, the evaluator can use a pressure profile to determine
(Chapter 7) used the following modeling simplifications: maximum expected pressure.