Hello everybody. I would like to comprehend, how cII calculates the displacements for a bend (90° & 45°). I read some things about the Karman factor and tried to calculate a simple example by hand,but got different results comparing with cII. If someone can help me solving my problem, I will give you further information about the chosen parameters of my example. Regards, Peter

Dear Peter,

I believe that first of all the Karman factor is used to affect the stiffness matrix, considering the extra flexibility of a pipe bend. After that, based on the imposed loads the displacements will be calculated from CII. For further discussions feel free to send me your model / parameter.

All the best,
Veit
vb@esn-gmbh.com

CAESAR II Version 4.50 has a new entry item - the bend flexibility factor. In writing a newsletter article about it, I wanted to see where all the bend values came from. Part of that work resulted in constructing a bend flexibility matrix and then inverting it to a stiffness martix that we use in CAESAR II. A pdf of my work is found here .

Hallo Dave,
Great! This is exactly what I needed, many thanks. Now I can understand more about the mechanical backround. However, I have another question to the bends. Maybe you have a more clear explanation for the topic "Bend Length Attachment Percent" in the cii configuration. It would be very nice to have a small sketch or anything else of this "inserted elements".
Best Regards
Peter

The "bend length attachment percent" is a "fudge factor" to assist in insuring that the user's in/out tangents are long enough to contain the bend. If "floating point" numbers were exact, and precision wasn't an issue, we wouldn't need this parameter.

We use it to prevent the inclusion of very, very short and therefor very, very rigid pipe elements after elbows where the element leaving the bend is nearly equal to the length of the bend tangent. Such numbers don't work well in the stiffness matrix operations. This paramenter creates extra pipe length in the stiffness matrix where things get too tight.

Users typically want to ignore this setting and not change it.

Thanks a lot for answering, however I don´t really understand, what´s meant f.e. with "length of the bend tangent", and how the stiffness values get increased by the additional pipe lengths. I tried it by modeling the given example of your attachment and got the results with a "bend length attachment percent" factor of 0.01.
With the default of 1.0 I got different values, although I have only one elbow, no bend leaving element or further in it.
Another question of me is, where does the alpha factor come from ? (4/3* (Ro³-...) and where can I find the MEC-21 ?
Regards,Peter

Here is the Help text:

Bend Length Attachment Percent
Whenever the element leaving the tangent intersection of a bend is within (n)% of the bend radius on either side of the weldline, CAESAR II inserts an element from the bend weldline to the "TO" node of the element leaving the bend. The inserted element has a length equal to exactly (n)% of the bend radius. The user may adjust this percentage to reduce the error due to the inserted element, however, the length tolerance for elements leaving the bend will also be reduced. To obtain more accurate results the user must include less "slop" in the system dimensions around bends. The default attachment is 1.0 percent.

The "inserted element" is placed at the end of the bend. Although you do not "see" it there is a straight run added to the end of your bend. Your second job had a long run added.

I have had no reason to adjust this program setting.

MEC-21 is a program, one of the first for pipe stress analysis, developed by the US Navy for piping. (Quoting the Preface of the MEC-21/7094 User Guide) "MEC-21 was written in 1959 in the PRINT language for the IBM-705 and is an expanded and modified version of MEC-5 written in 1958. It was revised and improved in 1960 to reflect application experiences and solve more complex problems encountered in nuclear submarine piping. ... MEC-21 ... was conceived and written by John A. Olson and Robert V. Cramer of the Mare Island Naval Shipyard Engineering Computer Application Branch, Code 246, under the supervision of David. B. Mitchell and Stanley R. Williamson."