High Temperature in Cyclones of FCCU Regenerator 1

[Lebpower]

Hello,

I'm Assessing a failure in the cyclones of the FCCU Regenerator. It is most likely to be related to high tempeture in the cyclones according to the data gotten from the tempeture intruments and according to the findings (sigma phase was found in the rupture zone).

I found the design data and analysis model used by the manufacter.

For the evaluation they used 3 models

1.-) The first model consists of one pair of cyclones including their part of the top dome, the hanger rods and the bracings. Constraint equations are used to model the rotational symmetry.

2.-) The second model consist of three pairs of cyclones to model the situation when one cyclone is filled with
catalyst and the situation one pair of cyclones has a 100 ºC higher temperature as the others. Constraint equations are used to model the rotational symmetry.

3.-) The third model consist of all cyclones and the top dome to calculate the seismic load.

and they examined 4 load cases.

Load Case 1: pressure and weight.
Internal pressure :2.9 bar
Differential pressure primary cyclones :0.08 bar
Differential pressure between pr. cycl. and sec. cycl. :0.1275 bar
Differential pressure secondary cyclones+plenum :0.175 bar
For this load case the first model is used.

Load Case 2: pressure and weight and one primary cyclone full of catalyst during 1000 hour
at 815 ºC.
Pressure and weight are the same as in load case 1. Additional one primary cyclone is filled with catalyst.
For this load case the second model is used.

Load case 3: Pressure and weight and one cyclone set 100ºC higher in temperature.
Pressure and weight are the same as in load case 1.
In the second model one pair of cyclones have a 100 ºC higher temperature as the other two.
The duct is also modelled with a 100 ºC higher temperature but not the plenum because in the plenum the high
temperature gas is mixed with the gasses with normal temperature.
For this load case the second model is used.

Load Case 4: pressure and weight and a seismic load.
Pressure and weight are the same as in load case 1.
For seismic conditions the normal gravity acceleration is increased to 11.157 m/s2 and a horizontal acceleration
of 1.51 m/s2 is added.
For this load case the third model is used.


DESIGN DATA.
Design code
2001 ASME BOILER AND PRESSURE VESSEL CODE VIII Division 2 + addenda.
Design conditions
Design pressure : 2.9 bar(g)
External pressure on the:
plenum inside the vessel: 0.175 bar
Secondary cyclone : 0.175 bar
Duct between prim and sec cycl.: 0.1275 bar
Primary cyclone : 0.08 bar
Design temperature : 755 ºC
: 815 ºC for 1000 hour (load case 2).
: 855 ºC for 1000 hour (load case 3).
Corrosion allowance : 3 mm for carbon steel
: 0 mm for stainless steel



They calculated the number of cycles and they got 525 cycles. They say a number of 500 cycles is the minimum required number of cycles, so the construction satisfies.

Concerns:

1.-) I'm not quite sure about when a "cycle" can be considered finished and if this can be associated to the failure in the cyclone legs. So, I'd really appreciate if anyone can give me more information about the Cycles in orden for me to understan and know more about them.
2.-) How can I estimate the remaining life of the cyclones considering the high tempetures that we have been witnessing inside the regenerator?

Greetings,

R.A.

 

[TGS4]

You're asking a lot for someone to do for free, anonymously. I recommend hiring an expert to assist you.

 

[Lebpower]

Well I don´t think i'm asking a lot.

I think I gave a lot of information in order to make it easier to answer my doubts.

I only asked 2 questions.

The harderest one could be the second one and I'm only asking for guideness.


R.A.

 

I agree with TGS4.
What you are asking for requires a lot of analysis, well beyond the scope of a forum such as this.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[mk3223]

Should you contact the original Engineer(s), who designed the FCCU Cyclones, for the design criteria and the RCA of the equipment failure?
IMO, the questions may be related to the equipment manufacturer proprietary information and trade secrets.

 

[georgeverghese]

Would hazard a guess to this materials failure:
Regenerator temps of up to 750degC seem typical. The high temps seen at these cyclones is most likely due to afterburn due to insufficient air coupled with high carbon laydown on the feed catalyst (which in turn would be due to processing of heavier stocks). Rupture at the cyclone diplegs may be due to gas carryunder, if these cyclones are not refractory lined. Thermal fatigue may be a small contributing factor, but the major reason may be high temp( >750degC) carbon monoxide (and possibly H2 also) attack of the cyclone materials.
Thermal cycles that can induce fatugue can be due to complete shutdown and restart, but in this case may also be due to temperature cycling in excess of some permissible limits as defined in ASME for this material of construction. For example, ASME may say either of these is also a cycle:
(a)any single excursion of design temp + 150degC
(b)any excursion of design temp +100degC, provided DT/dt is say >100degC/hr.

 

[Lebpower]

Thanks georgeverghese,

In our case the high temps were due to low levels of catalyst and also low levels of feed.

I wasn't sure about what a cycle is and when I can say a Cycle is finished. But your post really helped me. Can you tell me where I can find those permissible limits in ASME?

I found in ASME VIII 2, the definition of "operational cycle" is that the same?

 

[Lebpower]

Another question:

Does a cycle finish after a calculated number of hours?

I mean, In load Cases 2 and 3 they made the calculations based on 1000hours. Are these 1000hours related to a cycle?
Can I say after 1000hours a cycle is finished?

 

[Geoff13]

I know nothing about cyclone design, but I have designed cyclic pressure vessels. I will see if I can shed some light on what fatigue cycles are. Div 2 clause 5.5.2 provides more detailed guidance.

Fatigue is caused by the application, and subsequent removal, of stress in a vessel (or part). Coke drums would be the most obvious since the pressure varies from zero during cutting and filling all the way through design pressure during the actual coking. The amount of time at the maximum stress is irrelevant to fatigue. Fatigue only requires a count of the number of minimum-to-maximum stress range cycles. A coke drum might be one cycle per day, but other vessels can experience many cycles per hour.

Uniform temperature does not generally cause fatigue (though it could cause creep if it's hot enough). Differential temperature between two "close" areas will cause different thermal growths which, depending on the details of the parts and temps, could cause stresses between the areas. The subsequent return to uniform temp would then be a cycle. Where differing materails weld together (i.e. carbon and stainless) it is possible the different rates of thermal expansion could create stresses, and then changes in the uniform temp could cause a cycle, but this would be a very, very unusual situation.

I can't guess what a fatigue cycle is for your cyclones.

 

[Lebpower]

According to ASME VIII 2.

Operational Cycle – An operational cycle is defined as the initiation and establishment of new conditions
followed by a return to the conditions that prevailed at the beginning of the cycle. Three types of
operational cycles are considered: the startup-shutdown cycle, defined as any cycle which has
atmospheric temperature and/or pressure as one of its extremes and normal operating conditions as its
other extreme; the initiation of, and recovery from, any emergency or upset condition or pressure test
condition that shall be considered in the design; and the normal operating cycle, defined as any cycle
between startup and shutdown which is required for the vessel to perform its intended purpose.

They mention 3 cycles. I'm pretty much clear about the 2 first cycles.. But I'm not quite sure about the "normal cycle".

What can "a normal cycle" be on the cyclones?

georgeverghese mentioned a very good point about 150ºC above the design temp. I'd like to know where can I find that in ASME.

 

[georgeverghese]

Have sometimes been asked to estimate the no of these thermal cycles for vessels/ heat exchangers on past projects, and the criteria for this estimate given by mechanical engineers is similar to that in my previous response - but I dont know where in ASME mechanical engineers find / develop these criteria - suggest checking with a mechanical design engineer who has done thermal fatigue analysis.
Yes, transient low catalyst levels in the regenerator could lead to gas carryunder at the cyclone diplegs.

 

[Lebpower]

Ohhh Ok. Well Thank you so much George.

I hope a mechanical design engineer shows up soon and helps me.

 

[Dhurjati Sen]

Dear Leb,

Did your cyclones have linings (often RESCO & Hexmesh)?
The failure took place in the primary or the secondary cyclones?
Have you ever been inside a regenerator?

Regards.

DHURJATI SEN

 

[Lebpower]

Hi Dhurjati,

8 out of 10 primary were found fallen
and 6 out of 10 secundary were found fallen.

They have an erosion resistant linning.

And yes I've been inside the regenerator.

 

[jtseng123]

Lebpower,

I have purchased many reactor and regenerator cyclones for FCCU from experienced vendors in US. Due to the nature of the cruel operation and unpredictable factors, if it can serve more than 5-7 years without shut down for repair, you are lucky. I also have reviewed many of their FEA analysis similar to yours. The reason they gave you "fatigue cycles" is because it will be overstressed due to extreme temperature difference between CS head/plenum and SS cycles, that forcing vendor to run the fatigue analysis and let you know: hey ! our design is robust and it "may be good" for so many start up and shut down cycles. In years they all come up with a mutual agreed cycles: 300 or 500 or something else, depending on normal operation or upset condition, to "satisfy" customers.

So my opinion is, put that calculation aside. It's not going to predict the remaining usable life. You will be on your own to keep running the cyclones. Too many factors to cause cyclone failure. If your cyclones have been more than 10 years, contract an experienced supplier who has decades of failure records to pin down the issue and revamp for you, and don't be too serious about the fatigue analysis. They will all produce the same even you are buying new cyclones.

 

[Dhurjati Sen]

Great post JTSENG123

However, just a few things I want to add.

1. The operating procedure needs to be revisited as such failures would lead to unexpected and lengthy shutdowns affecting the profitability severely.

2. Shutdown frequency needs to be revisited and for the first few years, the cyclones are to be inspected at every available opportunity, even before a catalyst replacement.

3. The quality control of the lining needs to be full proof at the time of replacement of the cyclones as a good lining would enhance the life of the cyclones.

Regards,

DHURJATI SEN

 

[jtseng123]

By the way, the fatigue analysis is for vessel head to plenum, not the cyclone itself. Cyclone is non-pressure part and at very high and almost uniform temperature, no fatigue analysis required for it. Vendor is only to guarantee the major parts: head and plenum, will not fail for the life of the vessel, not the life of the cyclone. There shall have a color stress profile that you can see where is the maximum stress.

 

[Lebpower]

Hey jtseng123 and Dhurjati,

Thanks for your time invested on my post. I really appreciate it.

In fact, there is a color stress profile.

As you see, the maximum stress ocurrs at the lower bracings.

-------

I'm now part of the investigation team and what we want is to prevent/avoid this failure by proving to the operation dept. the effects of operating cyclones at tempetures above design temps.
I wanted to relate these effects to the end of cycles or something like that. But maybe what jtseng123 said is right, the cyclones could easily have failed due other factors and not due to the cycles.

 

[jtseng123]

The overstress of the bracing is not going to impact the performance of the cyclones. By the way, it looks like your supplier has wrong design, or even wrong analysis such that overstress happened at the bracing. Bracing is to connect diplegs from moving away by turbulence inside the vessel and to damp vibration, at the same time shall allow it free thermal growth by slot design or using sleeve. Such a simple thing to design bracing without overstress, wondering who is the supplier.