Good morning,

can you pls explain us the meaning of Rf=final centerline radius and R0=original cl radius of UCS-79 of a ellipsoidal head 2:1.

Many thanks in advance.
josil

Rf= Final centrerline radius of the formed product. For 2:1 ellipsoidal head it is 0.17D and 0.9D respectively for knuckle and spherical radius (refer UG 32, (d).

R0= Original CL radius of the material from which you are forming your component. For flat plate radius is infinity.

Regards

Sanjeev

Sanjeev,

thanks for reply.
We have calculated an ellipsoidal head with our compress software.
The value given by the program does not comply with the dimensions in UG-32 (d).
Are there any other rules to adhere e.g. corrosion allowance, forming tolerance ?

Best regards
josil

Josil,

Generally these commercially available softwares take care of all the code requirements but we have orserved some bugs in the various softwares in the past.

I would go with the code requiremeents. If you are confident that your input values are correct and compress result are not in line with the code requirements, you need to raise a query to their technical support team.

For corrosion allowance and forming tolerance there is no firm guidelines. Corrosion allowance is based on your process media, MOC and the corrosion rate. Code however provide you a minimum required CA. Forming tolerance are based on your forming process.

Hope this clarifies

Regards

Gentlemen,

Perhaps this may assist:

COMPRESS applies the average knuckle radius as "Rf" in the formula for strain from UCS-79. For 2:1 heads the knuckle radius is determined by using 0.17 * mean head skirt diameter (for SE heads of other proportion, Table 1-4.4 is used).


Consider a 2:1 head with 60.0" ID, 0.125" corrosion allowance, minimum thickness = 0.375", and straight flange of 0.4375" thickness.

Then,

knuckle IR = use 0.17 * D as per UG-32(d) = 0.17 * 60 = 10.2000

Rf = "final centerline radius" per UCS-79
= knuckle IR + head thk / 2 (COMPRESS uses straight flange thk if one exists...even if thinner than head min thk)
= 10.2 + 0.4375 / 2 = 10.4188

% Extreme fiber elongation = (75 * t / Rf) * (1 - Rf / Ro) (initial radius Ro = infinity for flat plate)
= (75 * t / Rf) * (1) (COMPRESS uses straight flange thk if one exists)
= (75 * 0.4375 / 10.4188) * (1) = 3.1494


If there were no straight flange for this example, then the fiber elongation would be calculated:

knuckle IR = use 0.17 * D = 0.17 * 60 = 10.2000

Rf = knuckle IR + head thk / 2 = 10.2 + 0.375 / 2 = 10.3875

% Extreme fiber elongation = (75 * t / Rf) * (1 - Rf / Ro)
= (75 * t / Rf) * (1)
= (75 * 0.375 / 10.3875) * (1)
= 2.7076


Tom Barsh
Codeware Technical Support

I think this is a right approach since in the starting since you do not have the minimum thickness after forming.

To start with, straight flange thickness ie nominal plate thickness can be considered and after forming, can be verified with actual minimum achieved thickness.

Since there is no thinning takes place in SF, thickness of SF will be same as nominal plate thickness.

Tom,

Can you advise further on the case where you expect a SF thinner than head min thk

I do not think that there will be much difference in the final result.

Regards

Sanjeev

Sanjeev,

COMPRESS "assumes" that the nominal thickness of the head straight flange is the same as the nominal thickness of the flat plate that the head is manufactured from. Thus when a straight flange exists, COMPRESS will use its thickness as the basis for the maximum strain calculation.

In most cases, it is expected that the knuckle region of the head will become thinner than the original plate. But this is not true for all cases. In some cases, depending upon the proportions of the head and the manufacturing process, the thickness in knuckle may actually increase in thickness! This would be the case when the material is "worked" and the excess winds up in the knuckle area.

Tom Barsh