Seismic risk assessment for oil and gas pipelines - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 693

springs, user-defi ned stress-strain curves for the pipe material, and large
changes in pipeline geometry. The analysis software should include a pipe
element in its element library with the capability to model internal pressure
and provide output at various circumferential locations.
The length of the pipeline model should be suffi cient to adequately
capture the anchoring effects of the soil outside the zone of ground move-
ment. The pipe element length in regions where the pipe strain is expected
to exceed the yield strain (typically at abrupt transitions in ground displace-
ment or locations with abrupt changes in soil restraint such as elbows)
generally should not exceed one pipe diameter. The one-diameter rule is
related to the fact that a gauge length of approximately one pipe diameter
was the basis for reporting strains in tests used to establish strain acceptance
criteria.
Available methods for modeling soil restraint with soil springs assumes
that the spring forces always act in the axial, horizontal, and vertical direc-
tions relative to the pipeline. In most analyses, the soil springs are defi ned
in a global coordinate system. As a result of this, in the direction of the soil
spring forces do not maintain an axial, horizontal, and vertical orientation
relative to the pipeline if the pipeline undergoes large rotations. However,
the error introduced by this misalignment is acceptable considering other
assumptions and uncertainties inherent in the analysis.
In a case where loading and unloading of the soil springs occur, unloading
characteristics of the soil springs need to be modeled. Relative pipe-soil
displacements are permanent in a sense that the soil does not 'spring back'.
Therefore, soil spring forces should quickly drop to zero or unload along a
path parallel to the initial soil spring stiffness.
There is considerable uncertainty regarding the relationships used to
compute soil spring properties. Most of this uncertainty is related to esti-
mates of the soil strength parameters. The uncertainty in estimating soil
strength parameters for pipeline analyses in fi ne-grained soils is further
complicated by the fact that pipelines are typically located above the water
table and within the desiccation zone of the soil. The strength of partially
saturated desiccated soils is not well defi ned in soil mechanics.
Current analysis techniques assume the equivalent soil springs act inde-
pendently, a common assumption for analytical representation of pile foun-
dations and similar buried structures. This assumption can introduce errors
related to the potential for different soil restraints for oblique soil displace-
ment relative to the pipe (e.g., a combination of horizontal and vertical
displacement) and the dependency on soil spring force at a particular point
along the pipeline on adjacent relative pipe-soil displacement. The error
associated with the assumption of independent soil springs is generally
small relative to the overall uncertainty typically associated with defi ning
the soil strength properties in the equivalent spring formulations.