Large Bore 316H Stainless Welded to Refractory Lines Carbon Steel

[slhutzley]

A client is requesting a Stress analysis for a proposed design that would involve welding a short spool of 316H piping between two sections of refractory lined Carbon Steel piping.

Fluid operating temperatures are between 1300°F and 1400°F. I do not have details on the refractory lined pipe's external temperature yet.

Thermal expansion of 1400°F SS vs. much cooler CS is a primary concern.

What other potential issues should I be concerned with? What is the best way to analyze the stresses due to differential temperature & dissimilar materials at the weld? I appreciate any input.

 

[GD2]

The weld joints may be overstrained due to different material, temperature and thermal expansion. You need to do an engineering assessment for both longitudinal and radial strains.
Equation for radial strain can be found in Roark’s book. You might have to put a transition piece of different material to keep the radial strains within allowable level.

You can find transition piece in many process plant between process piping and furnace tube.

GDD
Canada

 

[Snickster]

I have not come across that design but just thinking about it I have the following comments:

If SS is not internally lined then by heat conduction the temperature in the SS and CS will be about the same at the weld and then gradually decrease to the temperature of the refratory lined CS at some distance from the weld. Therefore the differntial radial expansion calculated at same 1400 deg temperature would cause a differential radial expansion at the weld and a discontinuity stress. Not sure how you would model the stress for this case or if there is a cook book calc for this- maybe need finite element analysis.

Not a material expert but there could be degredation of corrosion resistant properties of SS due to contact with carbon steel at high temperature to change chemical composition of the SS.

 

[Stick.]

In addition to the steady state stresses from the different thermal expansions, you may need to look at transient thermal stresses during warmup and cooldown cycles, depending on how fast the temperature changes happen (potentially including unplanned cooldowns if those apply). Assuming I'm understanding correctly, the stainless piece will be bare inside and the carbon steel will be insulated inside with refractory, which may lead to different thermal response times, leading to even higher stresses during thermal transients. If you have to go down this route, you're probably in finite element analysis territory (at least that's what I'd do since I have access to it).

 

[weldstan]

Thermal cycling will definitely lead to failure. How many cycles? Not many. From some of my distant past work, less than a hundred should be considered.

 

[LittleInch]

I've seen FEA done on welds for dissimilar metals and sometimes we needed to put in a band of material which was between the two extremes for thermal expansion to avoid getting too high a stress in the weld itself.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[slhutzley]

All great input. I appreciate it.

It's apparent I need to involve a materials specialist, and an FE model is expected as well.

One thought I have is to suggest extending refractory beyond the dissimilar weld so the temperatures at the weld are lower, and the temperature gradient curve is less steep. I made a crude sketch that I attached. I have limited experience with refractory-lined piping. Does this seem feasible? Worthwhile?

 

[weldstan]

Definitely a sound idea.

 

[EdStainless]

The lining has to extend past the weld some in order to protect the weld from the fluid.
That sketch looks promising.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed