I am trying to compare the differences in large diameter pipe modeling using FE pipe and Caesar.

What are the typical support conditions for large piping, such as 72" duct piping?

By using FE pipe, what are the techniques used for large diameter pipes?

I am confused! Can anyone please help me with this information?

CAESAR or by and large present day commercial Pipe stress programmes are built upon the so called "stick elements ". For these elements each node is having 6 degrees of freedom ( 3 translational and 3 rotational).This element is not having any volume.It is infinitely thin. Properties like moment of inertia etc. are used to compute the element stiffness matrix.Then the eqn. {p}={k}{d} is solved for the displacement vector{d}.

Now coming to FE Pipe and other Finite emement software, you have choice of elements like trangular, quadrilateral,Plate, shell , ring etc. for 2d axisymmetric/plane-stress-plane strain problems as well as 3D elements like 3D beam,hexahedral( 8 node brick, 20 node brick etc.) as well as tetahedral elements.Choice of elements are sometimes programme generated or sometimes at the user disposal . Choice of elements require some basic knowledge of finite element analysis.Results will vary depending on choice of elements.

In displacement based Finite element analysis again {P}={k}{d} is solved.

For general purpose pipe stress analysis, commercial pipe stress programmes using the stiffness method is sufficient. This holds good for higher dia pipes also. However for higher dia pipes you may get local stress problems at support locations, trunnion attachments ( this holds good for smaller dia. pipes too)or ,pressure deformation in large dia. thin walled pipes etc.Under those circumstances it is better to resort to a rigorous Finite element analysis ( rather than finite element analysis with stick elements where such problems won't be addressed )using FE-Pipe and other software like ANSYS,ABAQUS etc.


Hope this answers your question.

Best regards

Anindya Bhattacharya

Stress Analyst

Bechtel Corporation

Although we are still conducting numerical tests, there is much that we've learned about large diameter piping. D/T > 100.

I've tried to summarize the most germain of the issues below:

a) Area replacement rules may not provide intended safety factors for pressure designs. Elastic FEA tends to be more realistic and conservative. WRC 464 elastic-plastic limit analyses tends to produce less conservative, but more realistic results.
b) SIF's from B31 Appendix D may not be accurate, (Notes to Appendix D warn that this is the case), and results from FE/Pipe or NozzlePRO can be used in place and should be checked against the default Appendix D values from CAESAR in the cases where the system is subject to cyclic loadings, (i.e. # of cycles greater than about 5000) (Sometimes the CAESAR default values from Appendix D are adequate. When it isn't then the values from an FEA should govern.)
c) Pipe removed from supports and intersections is modeled reasonably well in CAESAR providing the flanged bend specification is used to describe the deformation restrictions that may exist close to bends because of valves, supports, equipment or OTHER BENDS. (This is one of the places where FE/Pipe can be used.) When FE/Pipe is used to help adjust the bend specification we find that errors in forces and moments on the order of 25% can be avoided. I can provide more on the approach to do this if there is interest.
d) Supports should be full encirclement when possible so that loads are transmitted to the full cross section of the pipe as is assumed in the CAESAR model.
e) Flexibilities of intersections should be used from an FEA analysis. (NozzlePRO generates these automatically from the geometry.) WRC and NB approaches do not perform well for large d/D, large D/T openings.

Good support design, and including the branch flexibilities from an FEA analysis is typically all that's needed because often these systems don't cycle much. When that is the case the safety factor against fatigue failure is actually much higher than it is at say -- 5000 cycles.