| Applies To | |
| Product(s): | AutoPIPE |
| Version(s): | ALL; |
| Environment: | N/A |
| Area: | Analysis |
| Original Author: | Bentley Technical Support Group |
| Date Logged & Current Version | Dec. 2016 10.01.00.09 |
Problem:
How does friction affect Non-code User combinations in AutoPIPE?
example: For a specific model with the following conditions, pipe on a continuous support (V-stops at each node with 0.5 friction) subject to different loads.
The Load sequence (“non default”) is set as follow:
- P1 after U41
- T1 after U41
- U41 after GR
The Gravity load is resolved by Autopipe as frictionless (“ignore friction in Gravity case” option is enabled). However, friction seems fully developed at load U41 which develops a Combined Friction force = Friction Factor * Vertical Support Force (213N = 427N*0.5)
The supports friction nearly disappears for subsequent loading (Pressure 1 and Temperature 1) despite the pipe moves over the V-stops (Displ. 2 and Displ. 3 >0 particularly in T1) and the weight is substantially unchanged (see Figure 2 and Table 1) To be noted that this seems to be the reason why the model presents different results from an equivalent Abaqus model with the same load sequence made of the following Steps:
1) Gravity (frictionless)
2) Displacements at anchor (U41)
3) Pressure & Temperature
Appears that the support loads for the combination are consistent as resolved by AutoPIPE within its method. However, combining (summing) the response of the single loads Gr+U41+P1+T1 (resolved in isolation with AutoPIPE load sequence option) is not equivalent to have a full load sequence where the final status of one step (internal forces, reactions, friction and geometry) is the initial status for the next step. In other words, although AutoPIPE seems able to introduce some element of non-linearity it is not appear to capture them all.
Solution:
Confirm that the friction factors will apply to the thermal and pressure load cases in the model provided. To prove they have an effect, set the friction scale factor in the nonlinear analysis dialog to 0.0, the orthogonal friction on the support drops to 0.0, as we expect. Further, if we assume that the normal (GR) force on the support and the friction factor is the same between U41 and T1/P1, we would expect that the orthogonal friction on the supports would be proportional between the load cases, which they seem to be.
See Also
"Non-Code Comb." tab questions