Mill tolerance and Piping Stress Analysis

“Where does mill tolerance come from?” Does the mill tolerance affects weight, Fy and sustained stress for B31.3?”, “What should be taken as a tolerance on plate? “, “What is the significance of “-12.5% Mill Tolerance”. “What if I check “all case corroded” in Caesar?”. These are samples of questions about mill tolerance found on forums.


First answers are contained in the above picture, from www.tmk-group.com, because the best way to understand mill tolerance is to understand where it comes from. Steel pipes are produced by two main manifacturing methods. Basically welded pipes are produced by joining together two ends of steel strips. Seamless pipes are produced by piercing a billet (step 8 of above picture).

Mill tolerance comes from the manufacturing process of seamless pipe. Welded pipes have a tolerance as well, but usually is not a mill tolerance and it is much lower because it’s more easy to get high accurancy when producing steel plates. On the other hand welded pipes need to consider the weld joint efficiency factor, to take in account the weakness in the welding.

Corrosion allowance is calculated from the corrosion rate per year (depending on pipe material, temperature, atmosphere, type of fluid, concentration,…) and the expected life of the pipe.

Values of mill tolerance are defined in the material standards. For example paragraph 16-3 of ASTM A 106 states “the mimimum wall thickness at any point shall be not more than 12.5% under the nominal wall thickness specified”.
Someone wrote around “According to my understanding, Mill Tolerance is the tolerance for variation in the thickness of pipe from nominal pipe thickness which is 12.5% according to B31.3 “. This is not correct. EN 13480, ASME B31.1 and B31.3 often address to pipe tolerance, but they normally don’t define its value. There some exceptions; for example B31.3 gives some tolerance “for machined surfaces or grooves, where the tolerance is not specified“. What they do is to specify when nominal thickness has to be considered in calculations and when mill tolerance has to be considered. For example tolerance has to be considered for minimal pipe thickness calculation.
ASME B31.3: “The minimum thickness T […], considering manifacturing minus tolerance […] shall be not less than […] sum of mechanical allowances plus corrosion and erosion allowances […] plus […] pressure design thickness”. To be more conservative it is better to subtract the mill tolerance from nominal pipes first, and after to subtract the corrosion.
Many codes use the worst scenario approach.
For example in the Basic assumptions for the analysis of sustained load B31.3 requires to use nominal thickness for weight calculation (“The loads due to weight should be based on the nominal thickness“), and nominal thickness minus tolerance and erosion to calculate section modulus to use in sustained and occasional stress (“Section moduli used to compute the stresses in this paragraph shall be based on nominal pipe dimensions less allowances“). Removing corrosion will minimize the section modulus Z and then maximize the M/Z value. For the expansion load case B31.3 doesn’t mention corrosion and tolerance, so nominal thickness could be used.

ASME B31.1 is similiar to B31.3, with corrosion and tolerance a little more hidden. Tolerance is included in minimum thickness: “After the minimum pipe wall thickness tm is determined […] this minimum thickness shall be increased by an amount sufficient to provide the manufacturing tolerance allowed in the applicable pipe specification or required by the process.” Corrosion is included in additional thickness A “to provide for corrosion and/or erosion”.
On the other hand ASME B31.1 mentions only nominal thickness in section modulus calculations and flexibilty analysis.
This is very important for corrosion. Caesar II doesn’t use the specified corrosion values when using B31.1, because B31.1 doesn’t mention it. Using nominal thickness for sustained load design may appear a little unconservative. Someone says the reason why B31.1 doesn’t mention corrosion in sustained case is because material with no significant corrosion should be selected. That’s why many  users prefer to force Caesar II to use corrosion by checking “All case corrored” in Caesar II configuration. It is possible to set this parameter when using Asme B31.3 as well: the consequence will be that Caesar will use corroded pipe in section moduli calculations not only for occasional and sustained cases, but for thermal cases too.

Regarding tolerance and corrosion EN 13480 is very similiar to Asme B31.1. After stating “the minimum thickness shall be determined with regard to the manufacturing process for pipes and fittings”, EN 13480 shows how to take in account corrosion allowance c0 and pipe thickness tolerance c1 in mimimum pipe thickness calculation. Tolerances and corrosions are not mentioned in section moduli calculation and stress analysis.
The nominal dimension shall be used in the calculations and the tolerance regarding thickness shall be fulfilled“.
Again, when facing with EN 13480 and using Caesar II many users prefer to use “all case corroded” option turned on.

There are some exceptions. For example the British Gas Code IGE/TD/12 addresses both the +mill and -mill tolerance condition. B31.8 Chapter VIII requires to consider mill tolerance when compunting the “combined stress”.

Mill tolerance in mentioned in paragraph UG 16 of Asme Code, Section VIII, Div.1 too: “ the manufacturing undertolerance on wall thickness shall be taken into account except for nozzle wall reinforcement area requirements in accordance with UG-37 and UG-40.The  manufacturing undertolerances are given in the several pipe and tube specifications listed in the applicable Tables in Subsection C. After the minimum wall thickness is determined, it shall be increased by an amount sufficient to provide the manufacturing undertolerance allowed in the pipe or tube specification.

Another question is spring hanger design: for this calculation Caesar II uses nominal thickness.
Many approaches are used to solve these issues.
First rule is : never reduce wal thickness at the input level. This would reduce the system stiffness – makeing it too flexible – and reduce the weight of the system.
Another consideration came from the Caesar forum, where a user wrote he had to comply with Asme B31.1, but the pipe vendor stated his pipe could have a +25% mill tolerance. This situation is not very common, since many people suspect vendor try to use all the -12,5% mill tolerance to save steel. By the way the answer gave by Intergraph was quite interesting:
Assuming your pipe fluid is a gas (i.e. density = 0.0), and you’re only concerned about an increase in system weight, you can evaluate the +mill_tolerance condition by simply setting up an additional load case, and applying a mulitplier to the “W” component. For example, say you compute that the +mill_tolerance will increase the (pipe) weight of the system by 5%, then you could setup another load case like this:

case n) 1.05W + P1 + T1 (OPE)

Be careful with this multiplier on “W”, it will increase all system weight – rigids, insulation, pipe, and fluid”.

 

Lascia un commento