| Applies To | |
| Product(s): | AutoPIPE |
| Version(s): | ALL; |
| Area: | Analysis |
| Original Author: | Bentley Technical Support Group |
| Date Logged & Current Version | Sept. 2016 10.01.00.09 |
Problem:
When analyzing a model the following occurs:
---------------------------
Confirm
---------------------------
The load case did not converge. Do you wish to continue iterating?
---------------------------
Yes No Cancel
---------------------------
What can be done to avoid this message from appearing?
Solution:
Please see the following AutoPIPE help location.
Help > Contents> Contents Tab> Reference Information> Analysis Considerations> Static Analysis> Nonlinear Analysis - Gaps, Friction and Soil Yielding Included > Failure to Converge.
A nonlinear analysis will usually converge quite rapidly. However, some analyses may fail to converge, even if extra iterations are allowed. Failure to converge after »15 iterations will usually indicate that the system is unusually sensitive, suggesting that the supports ought to be modified.
If the analysis does not converge after a reasonable number of iterations, the QUIT option, which cancels the analysis, should usually be chosen. If the accept option is chosen, the results should be examined carefully to ensure that they are acceptable. During a nonlinear analysis, an ASCII text file SYSNAME.LOG is created, where SYSNAME is the piping system name (specified on the System dialog). This file can be used to identify the support(s) which failed to converge.
Open the SYSNAME.LOG file that was created, and scroll down to "ERROR SUMMARY FOR NONLINEAR SUPPORTS". Just above this line, see detailed information: Load case causing the problem, values calculated (compare to tolerance settings), node point where the non-convergence occurred, etc.
There are 5 solutions offered. In addition, consider reviewing how the model is moving under the offending load case. Sometimes support gaps need to be decreased / increased. In other cases, additional supports are needed in key locations around the model to prevent the movement to cause a model to not converge. Or a combinations of all 5 solutions is needed to resolve the issue.
Typical Solutions:
#1. Press Yes to continue iterating
a: Model does not converge after 5 iteration, should it continue to iterate or not?
b. How to change number of iterations for static analysis?
#2. Judging by the location and type of error, modify/remove the support.
b. Make sure there are no unrealistic huge loads in the load cases that are not converging
b. Piping model with Wave does not converge
c. For submerged piping, may need to add concentrated or distributed weight to prevent it from uplifting (in practice concrete mats are added in some cases.
d. For submerged piping on seabed, model using V-Stops instead of soil at a reasonable spacing (that will make it pass the sustained stress check). Also, may have to use large friction factors (0.5 to 1.0) for these supports.
e. For submerged piping on seabed, review soil properties with very small K2 value, and with a huge concrete mattress load acting on the pipe in the downward direction. Making these values slightly more realistic seems to have resolved the issue.
#3. Model with soil; modify soil properties and/or modify the "Max. Soil Spacing" setting.
a. Confirm soil spacing is reasonable, Use AutoPIPE's help calculations as a guideline to this value.
b. " Some more converging solutions for buried piping are to first try using nonzero K2 values in order to avoid unlimited displacements. Using larger K2 values temporarily (K2=K1/1000 or so) can help you see if P1 values are sufficient by checking displacements. By setting K1, K2, P1 to 0.01 N/m and 0.001 M/mm/mm, the model should converge with no problem.
c. Look at the displacement plot for the non-converging load case. If there are spot huge displacement values, that may point towards an unrealistic stiffness of support at that location, or soil yielding at a certain location
d. Crank up the soil stiffness values, including the K2 values. Huge displacements at locations which is solely supported by soil may indicate soil yielding under that load case. With a very low K2 value, the point would continue to displace even with a very small amount of force.
e. Piping model with beams and soil does not converge
f. Piping model with underwater piping resting on the seabed does not converge
#4. Reset the static analysis to uncheck "Gaps/Friction/Soils", run the analysis, and look at the support reactions for abnormal large values that are fighting each other.
#5. Modify non-linear analysis tolerance settings (i.e. Displacement Tolerance, Force Tolerance, Friction Tolerance (ratio), or Friction Scale Factor).
Suggested tolerances for Victaulic models:
Models with multiple Victaulic fittings seems quite sensitive and exerts a huge amount of forces. One setting that may help in converging an analysis would be to uncheck the Nonlinear Analysis option, "Pressure Before Temperature". Also suggest adjusting the displacement and force tolerances (0.1 in displacement, 100 lbf force tolerance).
Learn how to review AutoPIPE's log file:
See Also
The load case did not converge...