Determine pipeline behaviour
Through this design function the overall pipeline behaviour is determined as a result of the specified loads acting on the specified pipeline model and the options to the method of calculation chosen.
In the previous functions the overall pipeline model has been specified together with the load components that may be applied on the pipeline. Through the input data table LOCASE (or LOADSET in case of calculations according to the Belgian Law or ASME B31.8) a loading combination is composed from the load components by means of load factors. The input data from the tables SOILCTL and GEOMCTL are for iteration process control only. Both tables have default values which can be used in general. The maximum of the allowed number of iterations from both tables is decisive.
The additional output tables contain iteration behaviour results and the applied loads (including load factors).
Various options for the type of pipeline calculations can be selected in design function 1 (Project Setup). The relevant options for this function are:
|•||Project phase, providing options for initialisation of the pipeline behaviour|
|•||Analysis type, providing a choice of rules against which the data is compared|
|•||Model options (options may be combined, but not in all combinations):|
|o||Bend angle, switching between infinite and limited bend angle|
|o||Geometry model, switching between geometrically linear and non-linear analysis|
|o||Section model, switching between 1st and 2nd order ovalisation effect|
|o||Material model, switching between material elastic and elasto-plastic behaviour|
|o||Soil ring-stiffening, taking into account or ignore the cross-sectional lateral soil support|
|o||Soil model, switching between "Standard","Upheaval Buckling" or "Pipe on Seabed"|
|o||Ovalisation redistribution, causing or ignoring redistribution of peak (soil) loads along the pipeline|
|o||Loading redistribution, causing or ignoring additional bending moments and soil reactions due to additional top loads|
General method of calculation:
The method of calculation is based on the finite element methodology.
The function starts composing the overall structural stiffness matrix from:
|•||structural stiffness matrix of the pipeline structure based on element distribution specified in design function 2 (Pipeline Configuration).|
|•||structural stiffness matrix from specified branch connections.|
|•||structural stiffness matrix from external supports at nodes.|
|•||structural stiffness matrix from external soil support at nodes.|
After this has been done, the overall load vector is composed from:
|•||pressure load component|
|•||equivalent temperature difference load component|
|•||deadweight load component|
|•||equivalent soil settlements load component|
|•||nodal point load systems|
|•||equivalent elastic bend load component (in geometr.non-linear option only)|
|•||wave/current generated point loads (in case wave/current has been input)|
From the overall stiffness matrix and the overall loadvector the displacements of the nodes are calculated using a matrix decomposition method.
During the linear calculation process there is a check on the value of the main diagonal element, which shall remain positive. If this is not the case calculations cannot be performed due to singularity of the matrix (structural instability). This shall be improved by structural alteration of the pipeline model. After the calculations have been performed the calculated displacements are multiplied by the overall structural matrix and the resulting force vector is compared with the original force vector. If the deviation is more than 0.001 per mille a warning is given (EQUILIBRIUM NOT / NEARLY CORRECT).
Messages may result during the calculation process as well after each soil iteration from the linearity check on the end points. After the iteration process a warning may be given. If this is the case it is advised to locate the end points further away from the pipeline part under consideration.
Additional output data tables:
The additional output data tables are filled with detailed deformation concerning the applied pipeline loadings including load factors and the iteration behaviour. Through these additional output data tables, detailed information on the iteration behaviour of pipeline and soil (tables ITMON and ITCHECS) can be obtained. Moreover all the (beam) loadings applied on the pipeline and generated on nodes and elements have been stored in 'overview list'-output tables (E-LOAD, S-LOAD, N-LOAD and W-LOAD).
The unlock-facility (see also the overview panel help screen) enables to restart the calculation process from the last iteration performed. This may among other things be required if after processing of the function the iteration results are unsatisfactory because:
|•||accuracy is not within specified accuracy bounds due to too low specified maximum number of iterations, or|
|•||accuracy obtained, although within specified limits, appears to be too low|
Because after unlocking, the input tables of design function 5 are open to modification, the iteration control tables can be modified, influencing the rest of the calculation process. But also the LOCASE table can be modified. This enables for instance to make a series of calculations with a changing loadfactor for one or more load components.
The unlock-facility can be used even if design function 6 has already been processed. In that case design function 6 will be set back!
Stated differently, a set back of design function 5 results in a loss of all calculation results obtained sofar and the iteration sequence number will be 1 again. In case design function 5 is unlocked, the iteration sequence numbering continues. The maximum number of iterations as specified in the control tables (SOILCTL and GEOMCTL) starts counting down anew each time.
A criterion for the accuracy of the bend behaviour modelling is the bend parameter λ. This parameter is calculated and reported under the name BENDPAR in output table PIPEDIM in Design Function 3.1.
t = wall thickness of bend
R = bend radius
r = pipe radius (half of diameter)
ν = Poisson’s ratio of bend material
If the parameter λ is less than 0.15 a warning is given in Design Function 5. For λ < 0.1 the function process is stopped with an error unless the OVAL (Section model: Ovalising) or MAT-NLN (Material model: Non-Linear) option has been selected (required module Q or Q and M). The first order bend behaviour modelling according to NEN 3650 and the second order approach according to Reissner (GENERAL analysis) both become too inaccurate for λ < 0.1. If the advanced OVAL or MAT-NLN option is switched on, the influence of higher harmonics is taken into account up to a limit value λ = 0.01. In the advanced options it is reckoned with the decrease of the diameter and the bend radius due to ovalisation and bending moment.
H5003, last changed: 25/04/2017