Properties Tab Best Practice
1.As per the Support component best practice, for efficiency/simplicity try to use the Simple contact option without any Swivel to begin with. A trial analysis will tell you whether there is a need to introduce more complex contact modelling behaviour.
2.The Select Stages option is useful for ensuring that you only run what you need. It also works for post-processing only runs. Always check the Installation Stages Summary List to see what is included/excluded in the run.
3.The following best practice applies to the Dynamic section:
a.Use selected time steps for database writing. The start time for writing should generally correspond to the end of the ramp and a recording interval of 0.2s to 0.5s is usually fine. All of this will help keep the analysis databases at a reasonable size and also help improve run time as there is less frequent hard drive access.
b.Be cautious in terms of damping. Mass damping for scenarios where there is significant vessel/stinger support contact should be avoided. Stiffness damping can be specified, but how much is realistic is generally unknown and varies from model to model. It is recommended to start with zero damping to begin with and only if there is a lot of high frequency noise in predictions then consider incrementally increasing stiffness damping to see if there is any improvement (e.g. start/increment at 0.001). Remember too much damping can be unrealistic as it can attenuate the response of important aspects of the model and so adversely affect predictions.
c.A lot of pipeline installation scenarios are dependent on support contact and contact analysis typical requires a small tolerance measure to realistically capture surface penetrations and reactions. With this in mind the recommended tolerance inputs for dynamic analysis are shown below. Note that the large number of iterations reflects the small tolerance measure (typically only a few time steps during the ramp time require a large number of iterations). Also note that the large small torque value is to instruct PipeLay to ignore torsional behaviour when it comes to convergence checking. If torsional behaviour is a target/desired output for your analysis then you may wish to consider reducing the small torque input (not often the case hence the recommended large value). If in doubt you can always request custom plots of torque for your analysis, to check if the torque levels are acceptable.
d.For scenarios where there is no support contact then the default tolerance values are fine and no inputs on the tolerance dialog are needed.
e.If convergence problems occur during a dynamic analysis then review the convergence ratios in the Detailed Output file (look for the table after the Solution Parameters section) to see if refining the tolerance inputs further will resolve the convergence issues. For example, in the sample extract below it is possible to see that if the maximum number of iterations was increased from 20 to a large value then the convergence ratio in DOF3 could eventually fall below the tolerance measure of 0.001.
4.The Time Stepping option is useful in dynamic stages where there is a lot of high frequency noise generated, making convergence very difficult to achieve. Typical sources of high numerical noise include intermittent pipeline-support contact and variations in friction forces. Note, that this setting determines the level of solution robustness in PipeLay without compromising on accuracy and should not be confused with mass or stiffness damping. The default value of 0.4 is suitable for most cases. A higher value suppresses more high frequency noise, while a lower value allows high frequency noise in the solution.