We have a vessel with MOC as SA 516 Gr 70 with a requirement of Normalizing the heads after forming.

Material was ordered to plate mill with the requirements of simulated Heat treatment cycle for normalizing. As pet MTC, properties of material meet all the code/ specification requirements.

Even though not required by the code, as per the standard practice of equipment fabricator test coupons were taken from all the plates and kept inside the furnace along with the heads for normalizing. Mechanical testing was carried out on these test coupons which underwent the same normalizing cycle as of heads but unfortunately most of the coupons failed in tensile testing (UTS is lower than the required). Normalizing cycle followed for the heads is same as of simulation cycle followed in the mill and therefore we are totally unclear on cause of failure.

All the properties of these failed coupons including the UTS are however meeting requirement for SA 516 Gr 60 and fortunately thicknesses on the heads are also suitable for the lower UTS/ allowable stress values corresponding to SA 516 Gr 60.

Is there any provision in the code to re-certify the material to a different grade (MTC is for SA 516 Gr 70) or use the actual tensile values achieved during the mechanical testing.


Regards

Hi Sanjeev,

Cause of failure may be Softning of material due to repetative heating cycle.

You can very wel refer UG10 and get it Recertified for Gr 60 value.

I don't see any reason why u can't - Yes you need to change all basic documents for the job.

_________________
Niraj

It looks like your equipment fabricator was right to use test specimens because normalizing is a heat treatment above the lower transformation temperature, see UCS-85 HEAT TREATMENT OF TEST SPECIMENS.

UCS-85(b) Heat treatment as used in this section shall include all thermal treatments of the material during fabrication exceeding 900°F (480°C), except as exempted below.

UCS-85(c) The material used in the vessel shall be represented by test specimens which have been subjected to the same heat treatments above the lower transformation temperature and postweld heat treatment except as provided in (e), (f), (g), (h), and (i) below.

UCS-85(f) Materials conforming to one of the specifications listed in P-No. 1 Group Nos. 1 and 2 of QW-422 and all carbon and low alloy steels used in the annealed condition as permitted by the material specification are exempt from the requirements of (c) above when the heat treatment during fabrication is limited to postweld heat treatment at temperatures below the lower transformation temperature of the steel.

_________________
AI-OR-US

The reason you can't is in the title of UG-10
UG-10 MATERIAL IDENTIFIED WITH OR
PRODUCED TO A SPECIFICATION
NOT PERMITTED BY THIS
DIVISION, AND MATERIAL NOT
FULLY IDENTIFIED

"Escamotage" You can if you consider the material not identified But in this case you have to repeat all tests including chem.analisys.
Probably an improper normalization has caused the softening. Normalization shall be performed with temperature and holding time same as that stated in the MTR of the Steelmaker. To recover UTS an accelerated cooling helps.
Mauro

Low UTS is very unusual for SA-516-70. Check the Mn content, you may have SA-36 instead of SA-516-70.
In support to Mauro comment, yes a reason of the low UTS may be an improper normalization process. The critical parameter is a cooling rate of the test specimens. Yes, your heads UTS may be still OK, but if they placed test specimens in such a way that their cooling rate was substantially low you may end up with a low UTS (on the specimens!).
To fix the problem, I would just repeat the normalization process making sure you have a proper holding temperature (around 910 C), a proper holding time, and a cooling rate reasonably fast both for the heads and for the specimens. Air fans on the cooling stage may help.

Good luck!

_________________
AI-OR-US

Mill certificate we received for the material comply with the requirements of SA 516 Gr 70 including the chemistry. Moreover this material was ordered with guaranteed mechanical properties with simulating normalizing cycle also and MTC reflect a decent mrgin (on UTS & Yield) over minimum required values (~ 25 MPa).

Normalizing carried out in the fab shop followed the same cycle as followed in simulation however tolerances are applicable for gas fired system and that is the reason we are are unclear on the reason of failure.

UCS 85 (c) require a test specimen which was already complied by doing a simulation test in the mill and code requirements were met even if another sample was not taken up for actual normalizing.

Something has definitely gone wrong some where. Since test results are meeting requirements of SA 516 Gr 65, we want to use the same material as Gr 65. From earlier posts I understand the following;
- Re-certification is not possible however certification of the material as Gr 65 is possible by fabricator by hiding the earlier identification (Gr 70) of material (AI known about this failure so do not know whether they will allow this but we are in touch with AI for resolution)

AI proposed another solution that datasheet/ UDS revised for Gr 65 so for AI, required material is Gr 65 and fabricator certifiy this as Gr 65. I however prefer to handle this as a non conformity since original ordered material was Gr 70 which is not complied. The reason for this preference is to keep all the records for this change otherwise if I change datasheet/ UDS as per AI proposal, there is no record for the change since datasheet/ UDS will indicate ordered material as Gr 65 and supplied MOC as Gr 65 and therefore every thing looks perfect.


Request your opiniopn on this.

Opinion:

UG-10 provisions were written not for your case. As it was noted by Gonzaga, your material is fully identified and was produced to the Code permitted material specification. A good engineering practice is to rework the material by conducting a proper normalization. What you are trying to accomplish is to find a loophole in the Code. This is not a good engineering practice.

_________________
AI-OR-US

.

Its not an attempt to find out loopholes and get away with the responsibilities for the safe design. Its only an attempt to reduce the schedule impact as first priority and obviously cost impact (but definetely not priority no 1) also keeping the design within technically acceptable.

From loophole point of view, If we would not have done normalizing of test coupon again in fab shop, we had already met code requirements by doing simulation test in the mill. Failure test results could have also been kept under the carpet (as already meeting code requirements by doing simulation).

Unfortunately do not have enough coupons also to carry out test again and also not sure about the material behaviour after another normalizing cycle (as already mentioned before we followed same simulation cycle for actual normalizing also and therefore unclear on reason for failure).


Regards

Sanjeev

It looks like you do not have much choice

One more possibility may be (if it is not too late): you may perform a chemical analysis of broken specimens that produced low tensile strength values.
If the chemical composition does not confirm SA-516-70, your test cupons were mixed up by the material supplier and you may discard the failed tests. You even may demand a substitution of the test cupons if you need.

As for additional normalization process, bleleve me, for this steel one more normalizing when properly done would only improve the mechanical properties.

_________________
AI-OR-US

AI-OR-US,

Thanks for your advise. In the mean time we got some feed back from plate mill. It appears that the only parameters which is not in line with the simulation test is Cooling rate. Mill informed that cooling rate was 12 Deg C/ min (720 Deg C/ min, looks like accelerated cooling, not specified in MTC) where as in fab shop it was cooled in still air and suspected slow cooling rate probably resulted in failure.

Other parameters (heating rate/ holding time & temperature) are however same as used for simulation test.

These are 3300 mm dia x 55 mm Thk, incoloy cladded hemi heads for 4 equipment and therefore all the efforts to save it.

Regrading the material to Gr 65 may be one of the technically acceptable option, possibility of taking help of code provision for manufacturer's certification of material.