Sunday, July 14, 2013

Can PRESSURE BALANCING helps in slowing down existing cracks growth in Ammonia Converter faulty weld joint?

"The Opinion/idea expressed in this Topic has no experimentation/documentations proof, its entirely a personal idea with no approval in any forum. Keeping in view of critical condition of the ammonia converter weld joint, such ideas may emerge time and again and will be placed here just for sake of discussion and brain storming. Reader may agree/disagree."

Recently, Plant-II experienced a shutdown in July-2013 due to crack in R4501 shell weld joint. After insulation removal, one thorough transverse crack was observed along with three more transverse subsurface cracks that were revealed in Ultrasonic Testing of faulty weld seam of  the Ammonia converter. The thorough crack was boxed up using plate (110mm x 50mm x20mm) , however, it was decided nothing to do with the subsurface cracks and plant was put into operation.
My opinion was also the same to keep those subsurface crack untouched. In my point of view, if the patch-up plate was welded on these subsurface cracks then it would cause more harm to the weld joint. In my opinion, due to high pressure and temperature, the patch-plate may restrict the growth of   transverse crack but the possibility of reorientation of these subsurface crack toward the circumference of the weld joint increases which may result total failure of complete weld joint. So let the subsurface crack loose their residual stresses by appearing on the surface as transverse crack. The point will come in trailing discussion about the advantage of installation of patch-up plate on open crack and disadvantage of installation of patchup plate on subsurface crack.

"Weld Cracks are not friendly but in the given condition(in case of faulty weld joint where we all are handicapped for the time being) the transverse cracks are far better than the circumferential cracks."

I am sorry if reader of this blog not agree what i am going to elaborate below. The idea just comes in mind and we have to brainstorm on it in a logical way keeping all merits and demerits in mind. I know that the cracks formation and their growth is a very complex phenomena (Nitriding, Hydrogen induced cracking, poor heat treatment, Fatigue, poor workmanship  etc). May be some among us directly reject what am i writing about now but any silly idea may be thoroughly reviewed with open mind so i am putting this idea with a very simple assumptions.
 Let us explore following three cases. In these cases suppose a sample segment from faulty weld joint of R-4501 shell. Please click on the following picture so that all three cases can be viewed.

Again i am fully aware that the crack growth and initiation is a complex phenomena and their are many many variables involves. But,for the sake of discussion,  lets proceed further:
Considering above cases, lets compare case-II with Case-III, In my opinion, case-III may at least restrict or minimize the growth of crack and  initiation of new cracks as compare to case-II. So what comes in mind..PRESSURE BALANCING?... If you are not getting bored then keep reading as i am going to disclose some real example to prove this fact.
How the pressure balancing can be done on the R-4501 shell weld joint? Here is a clamp design with factor of safety =9. Detailed engineering can be carried out on the clamp but my FEA results shows some major dimension of the clamp that can be further reduced as FOS =9 is a bit on higher side which is a positive sign.

The idea is to enclose the circumferential weld joint with a segmental  PRESSURE BEARING clamp which is pressurized externally with a inert gas to the same pressure as within the R-4501 shell.(All the segments of the clamp are welded together)



Many of us had seen the Nozzle welded on the 1st crack on the same weld seam in Turnaround-2012. Look the following picture:

Can you see a bigger crack inside the nozzle;) yes that one which is 1 inch width of slot, milled during TA-2012. In my Opinion, that 1inch slot is also a type of crack wide open inside the nozzle.
Following questions comes in mind:
1-  Is the initial crack really eliminated by slot drilling/milling?
2- Are we sure there are no fissures/microcrack left in the slot periphery?
3- Why the plate was not welded on the above slot? Why the slot kept open?
4-What may be the disadvantage of welding of plate on the slot?
5- Are they getting the advantage of PRESSURE BALANCING with the open slot?

Some More Questions:
1- Can we install The PRESSURE BALANCING clamp instead of nozzle?
2- What is the difference between the Nozzle and the Properly designed box-up clamp?
3- Nozzle was welded on the shell membrane, can't we weld the boxup clamp similarly?
4- Nozzle has top blind with ring joint gasket, the boxup clamp does not require gasket as well as blind flange.

Now, we knew that it will be difficult to install nozzle arrangements if the cracks appears time and again due to limitation of space.
At the time of first crack appeared on the weld of R-4501, we have welded 100mm x 50mm x20mm low-alloy plate. Actually that plate was also doing PRESSURE BALANCING as the crack was wide open and the plate was pressurised to reactor pressure after plant startup. The plate remained upto few months and crack didn't enlarged and escaped through the plate boundary. However, its better to cover more healthy area when we are talking about pressure balancing.
Just after TA-2004 plant-I, a crack appeared on the downstream piping (ASTM A335 P1) of waste heat boiler (E208) at pipe-elbow weld joint . A clamp was welded on-line  on that piping-elbow interface and was then removed in the next TA-2006!!!.. It gave two year service and didn't propagated.

We can see thin partition plates inside high pressure vessel/exchangers that hold there shape even at higher pressure due to the fact of PRESSURE BALANCING.

1- Retard the growth of existing cracks.
2- Retard the initiation of new cracks.
3- Pressure bearing, In case crack appear on surface within clamp periphery.
4- Can borrow us time for coming up to long term solution.

The discussion is kept open. Comments are appreciated.

Saturday, June 15, 2013

Interim Boxup of Crack weld Joint of Ammonia Synthesis Converter (R-4501)

This study is based on self-motivation and exclusively for learning purpose by comparing the real-life vs the software simulation results. It is based on the design evaluation (using modern software) of the low alloy plate used for interim repair box-up carried out on cracked shell weld joint of Ammonia Converter R-4501 back in 2013.

Plate welding was carried out using temper bead welding technique. The interpass temperature were maintained above 250C. The root pass was carried out using GTAW technique. After DPT, SMAW technique was adopted. After final DPT, the plate was heated above 400C and put under insulation for cooling.

The purpose of this evaluation was to check the results from the modern FEA software and compare them to the well-evaluated and proven short-term boxup. The results from this study will help us using modern software as an another great tool for validation and selection of box-up material thickness and its dimensions and help us making bold decisions for such type of critical boxup activity.

The CAD model of the plate was prepared as per actual plate dimensions (110mm x50mmx20mm) used for box up at that time. The fillet weld was integrated in the plate CAD model.  

 A very fine solid Jacobian mesh was generated with Total elements= 107737Ea to get accurate results.

 The fillet weld was constrained and a pressure of 150Kgf/cm2 was applied on the bottom of the plate in the software. 

Considering the Yield Strength of the material , the maximum Von mises stresses observed toward the tip of the longest fillet weld found to be 1457 kgf/cm2. This stress value is at the root pass interface between the plate and the shell of the converter.

 The minimum Factor of safety (FOS) that is (Yield strength/Max von misses stress) was found 4.34 (Normally FOS above 1 is consider to be a safe design).

The maximum displacement of 0.009mm was observed at the centre of the plate.

The software result validates the selection of plate material and its dimensions used in R-4501 box-up.
FEA results also shows that the fillet weld is the weaker part of this boxup. The fillet weld of this dimension plate can withstand pressure upto 500 kgf/cm2 before failure.

Friday, June 7, 2013

Falling of Zblock in Primary Reformer Convection Section and FEA interpretations

Primary Reformer (F-4201) is a critical equipment that reform process gas at elevated  temperature.  Total 364 burners installed on the side wall of the furnace heating 168 catalyst tubes in the radiation chamber. The hot flue gases pass through many tube bundles inside the convection section and transfer heat to process fluid before emitting into the atmosphere.
The smoke hood roof of the furnace is covered with the Z-block installed on 6mm thick carbon steel plates. More than 1500 zblocks of dimension 305mm x 305mm x 305mm installed on the smoke hood roof. Recently we had an experience of falling of z-blocks resulted in huge distortion in the carbon steel plates due to intense heat (~1150C) at that portion. Due to fall of one zblock, CS plate crossed its yield point at that particular zblock location and a small portion of it fell down.
The following FEA study is based on the non-linear analysis carried out to see the effect of high temperature on CS plate. The study is based on 12sec exposure of 1100C at the center (305mm square area) of the CS plate of dimension 2000mm x 1000mm x 6mm. Temperature of 1100C was applied considering falling of 305mm zblock  at the center of the CS plate. The solution took 6minutes to complete on 64bit system core i3 computer.
Gravity, the overall skin temperature of the plate (~170C) and the weight of Z-block at the remaining portion of the plate was considered in the study. Minor vacuum pressure was also applied to meet the real world requirement.
Following is the picture of the carbon steel plate after fall of Zblock. The bulge portion can be seen clearly. The plate crossed the elastic limit and the deformation you see is permanent deformation (Plastic region).
 At the center portion of the following mesh 1100C temperature was applied for 12sec. Also the plate was constrained all around the edges.

The following picture shows the Von mises stress plot at 12sec of the study.The stresses clearly shows the weaker portions of the plate at the fallen zblock location.

The displacement plot shows 51mm deformation at the fallen Zblock location which matches with the actual plate installed.One can also see in the following picture  how much deformation took place at each 1sec interval.

Below picture shows the weakest portion (Red one) that fell down. The shape of the created holes almost resembles comparing FEA and the actual.

Conclusion at your end.