1 25 Cr Reformer Reactor Flange Brittle Fracture R2

Reformer Reactor Inlet Fange Brittle Fracture

Isaac Pabon & Clay WhitePhillips 66Houston, Texas

Incident Summary On April 14, 2014 at 1230 hours during startup from a turnaround,

an Operator was investigating the cause of a sudden nitrogen pressure drop at a reactor in the Reformer unit. Upon investigation, the Operator discovered a large crack in top inlet flange.

The 36 inch diameter flange was an integrally reinforced “nut stop” design, fabricated from a 1¼ Cr- ½ Mo material (ASTM A 182 Grade F 11). The pressure and temperature that the reactor was under at the time of the failure was approximately 80 psi and 147 F respectively. ⁰

Initial Findings The reactor is the 4th reactor in

the unit, and was added approximately 8 years after original start-up in 1979.

During the TAR complete internal visual inspection was completed for this reactor, including;

– UT reading taken on all pressure containing elements;

– inspection of the RTJ gasket surfaces on both nozzles and;

– WFMPT was on the top nozzle to head ID weld

No indications were found from any inspection.

Initial Observations

The crack observed by the Operator was located in the nut relief cut.

Upon disassembly, it was discovered that the crack ran approximately 270o around the flange.

For most of the failure, the crack was located at the apex of the nut relief cut to flange juncture.

However, prior “damage” was observed at multiple locations around the flange and within the bolt holes

Exterior Crack

Hole Thru (interior)

Hole Wall Crack (Interior view)

Bolt Hole

Removed the elbow and found additional cracks in bolt holes and thru wall fracture

Prior damaged area along fracture path (arc gouge)

Microstructural Examination

Several sections of the cracked flange were separated for analysis, including multiple areas where arc gouge damage had occurred

Microstructural examination and hardness testing revealed a hard heat affected zone around each gouge examined

In addition there was evidence of a pre-existing flaw (crack) that had arrested previously around several of the gouged areas

Initiation Site For Through Wall Crack In Gouge(metallography provided by Stress Engineering Services)

High Hardness Of The Gouged Surface

13

Local Sub-Critical Crack/Oxidation Of Secondary Crack

14

Physical Failure Analysis

Gouges served as initiation site for the through wall crack due to high hardness of the gouged surface.

Metallographic examination confirmed that a local sub-critical crack had existed for some time prior to the through wall crack.

Oxidation of a secondary crack also observed during metallographic examination indicate the possibility of some thermal fatigue cracking at the gouged location prior to the through wall crack as well.

Mechanical Properties

Mechanical testing, tensile test and charpy impact test, were also performed.

Flange had a very low toughness, approximately 3 ft-lbs when tested at ambient temperatures (70 F).

The 15 ft-lb transition temperature would be close to 180 F (by extrapolation).

Operating conditions for the Reformer reactor are ideal for strain age embrittlement, generally there is a more pronounced affect for 2 ¼ Cr materials, but 1 ¼ Cr are also susceptible

Impact Energy / Low Toughness

Absorbed Energy must be > 15ft·lb

Minimum Pressurization Temperatures

P66 Guidelines based on age for 2 ¼ Cr– Pre 1973 - 350F– 1973 to 1980 – 250F– 1980 to early 2000’s – 200F and sometimes a little

lower– New (last few years)– as low as, but not below

125F

Stress Analysis

Performed Stress Analysis (FEA) to determine the location, distribution and magnitude of stress applied to the flange from bolt and piping loads.

Piping loads were determined to be minimal

Bolting loads, however, were substantial at the location of the fracture (40 ksi+)

FEA Results in 40ksi at Crack Initiation(analysis provided by E2G)

Conclusions (industry issues)

Impact strength of 1 ¼ Cr materials can be significantly affected from operation in the temper embrittlement range

Toughness appears to be below the current “conventional” wisdom for managing minimum pressurization temperatures (MPT’s)

Applied bolting stress for this flange produced stress high enough to easily drive a brittle fracture (40 ksi vs 8 ksi threshold)

Recommendations

Modify startup procedure to include new MPT(Minimum Pressurization Temperature) of 180 ̊F at 80 psi

Modify maintenance torqueing procedure MI-65300 for Class D flange assembly– Specify final ring pass temperature– Modify Start-up procedure to reflect MI-65300

requirements

Investigate best option for removing bolts from flange and implement into procedure

24