Thermal fatique failure and mechanical loading at the same time

[skarap7092]

Dear all,

We face an issue with a regenarator vessel, where in two years of operation in an emergency shutdown for other reason we found a crack that is propagating around the knuckle between a grid floor and the join. With the knuckle.

The metal is 304h and the width of the weld is double v 57 width...diameter of grid is 6690
The weld was under refractory.

I would like to share and hear your approaches regarding causes and ways of repairing the grid. I am available for any other information you may need regarding design conditions.

Kind regards

 

[metengr]

What is the thickness of the knuckle where the crack is located? How long is the crack? What is the crack depth? The cause of the crack needs to be explored by either removal of a material sample containing a portion of the crack or field replication (metallography) of the crack in the vessel wall prior to removal.

The second part of this is repair and this requires surface NDT to verify removal of the crack before welding, weld repair itself and post repair NDT.

 

[bxtsafe]

If the crack is propagating I imagine there is a cyclic loading in the structure. I think it would be very useful to simulate the structure/joint using Linear Elastic Fracture Mechanics to understand which loads are causing the crack growth and to propose a new joint geometry which avoid crack nucleation/propagation.

I understand by your post that the crack nucleate at the weld (probably at the weld toe), which is very common. Was the weld toe grinded after welding? This procedure avoids quite a lot crack nucleation in welds.

 

[Tmoose]

Can you post a picture?

 

[skarap7092]

Dear all,

Thickness of the knuckle is 57 and is covering around 180 deg all over periphery (diameter 6700).crack has went through weld depth, and regarding propagation of crack i hace the opinion that is extending since there is 35 tones on the bed ncluding lining and the vessel cooled-shrickaged.
For the pic i dont have it with me now....In short future..

 

[metengr]

So, skarap7092, what you are describing would appear to be thermal fatigue crack propagation. Do you know how to repair this location in the vessel?

 

[skarap7092]

Actually is under repair now,
We were looking for 2 options.

Either Arc gouging only the cracked area and welding, either or lifting the grid by hydraulic snubbers and making bevel and welding afterwards..

 

[weldstan]

I assume that the crack is in the load bearing, non-pressure internal attachment to the pressure part (shell/head) of the vessel. Is the nonpresure part and pressure part 304H or are they dissimilar? Because the shell is refractory lined, it will be at a considerably different temp than the non-pressure attachment. What is the temperature differential? While the crack can be repaired, resolution of temperature diffeential must also be done; otherwise, cracking will recur under the same operating conditions.

 

[skarap7092]

You are right that the crack is in non-pressure internal attachment but not to the shell! Shell is 516Gr70 and grid where crack is located is 304H.The thing is that we have a knuckle design, and the crack is between knuckle-grid-joint where both materials are 304H.Both are non pressure and the different temperature comes from lining portion of grid. Besides this this kind of supports gives freedom for exapnsion in skirt during operation.

 

[weldstan]

Are you truly getting freedom of expansion? What is the difference in thickness at the weld joint? You will still need to address the temperature delta during start-up/shut down. Can T/C's be mounted after repair to better understand the problem?

A picture/drawing would truly help.