ASME Section VIII Div 2 Fatigue Assessment 1

STOP!

You are using technology that is >4 years old.

There is a new Edition of ASME Section VIII, Division 2. It was completely updated for the 2007 Edition. Many of your questions may be answered there.

With respect to the adjacency requirement: there is a reference in the 2010 Edition of Division 2, Article 5.5.2.3.d.1 that should clarify the matter.

After you have updated yourself to the new technology, read the new Code, and read the relevant articles, please feel free to return with any clarification questions.

bricklayer (Mechanical)

Read Scope 2010 SECTION VIII, DIVISION 2

Do you have User’s Design Specification Section VIII Div 2?

The fatigue screening criterion shall be performed in accordance with paragraph 5.5.2.


L S THILL

TGS4
I have looked at the 2007 Edition where they changed it to Method A & B. Reviewing 2007 did answer my first question about what is considered integral vs. non-integral. However the customer we are doing this for is very big and is stuck in 2004 when it comes to this code so we have to complete the assessment per 2004 per their requirements. After reading through sections 5.5.2.3 and 5.5.2.4, Method A & B, I still have the same questions. I do not have access to the 2010 Division 2 artical you mentioned though.

L S THILL
unfortunately at this time i do not have access to any 2010 codes.

bricklayer (Mechanical)

Do you have a copy: "ASME Section VIII – Division 2, PART 5 DESIGN BY ANALYSIS "Example Problems Manual", ASME PTB-3-2010

Also: VCESage Vessel Analysis software.



L S THILL

OK -I understand that you are stuck with the 2004 Edition of the Code. No problem. We can help...

1) Yes, integral means full-pen welds.
2) The adjacency requirement is detailed in Note 3, under AD-160.2 Condition A. It says:

Note 3 said:

Adjacent points are defined as follows:
(a) For surface temperature differences:
(1) shells and dished heads in meridional direction, L=2.5sqrt(Rt)
(2) flat plates, L=3.5a
where:
L = minimum distance between adjacent points
R = radius measured normal to the surface from the midwall to the axis of revolution
a = radius of hot spot or heated area within a plate
t = thickness of the part at the point under consideration

If the product Rt varies, the average value of the points is used.
(b) For through-thickness temperature differences: Adjacent points are defined as any two points on a line normal to any surface.

Click to expand...

So, in your case, let's take two adjacent points - ones that are 2.5sqrt(Rt) apart. What is their temperature difference during the beginning of start-up? If they will truly be all at ambient, then the temperature difference is zero. What is the temperature difference during normal operation? If the temperature of the exchanger shell is uniform, then the difference is still zero. If the temperature of the exchanger shell is not uniform, then calculate the temperature difference based on a distance from the adjacency calculation (2.5sqrt(Rt)).


3. For the range of temperature differences during normal operation, you need to examine the thermal profile of your shell during normal operation. If it is part of your normal operation to have a sudden influx of cold water on the shell side, where the shell-side temperature gradient is high, then you would put that number here. However, if changes to the shell-side temperature profiles are slow and essentially uniform, then the value here is zero.

Here's an example of both 2) and 3):
Suppose your shell temperature ranged from 70°F to 200°F over it's entire length of 20 feet. Assume that the radius is 100in, and the thickness is 1in. 2.5sqrt(Rt) is 25in. Since your temperature gradient is 130°F in 240in, the gradient is 0.5417°F/in. So, the temperature difference between adjacent points is that gradient multiplied by 25in, or 13.54°F. That number would go into (c), because at start-up, the difference between those points is zero, because it is uniformly at ambient.

If a rush of water axially caused the cold-end to cool quickly (to 40°F, say), then the worst-case scenario for (d) would be when the wold water reaches the hot end. Then, the value for (d) would be 200°F-40°F, or 160°F. Most exchangers don't operate like this. Most likely, your value for (d) will be zero.

Does that make sense, or have I confused things more here?

L S THILL
I do not have Do you have a copy: "ASME Section VIII – Division 2, PART 5 DESIGN BY ANALYSIS "Example Problems Manual", ASME PTB-3-2010


TGS4
Thanks for the information it is starting to make a little more sense. Let me provide you with a little more detailed information. This is a spiral plate HX used for a batch process. For every start up we are to assume that the HX is all at ambient 40F. Then the cold water 50F will start to flow through and then then slurry, initially 212F. The 35000 gallons of slurry will keep cycling through until it reaches 77F, The water on the other hand will always be comming in at 50F. This unit is small, 36" in length and 60" in diameter with flat heads.

It looks to me like you'll need to do some sort of transient thermal analysis (and I don't mean that you have to do an FEA). Depending on your thickness, 2.5sqrt(Rt) could be a significant portion of the vessel, or it could be a small amount.

Based on your description of this as a batch process, I doubt that you will gain an exemption from a fatigue analysis. More than likely, you will have to do a fatigue analysis. Do you have the capability to do such a fatigue analysis? (I recognize that your customer is stuck in the 2004 Edition of the Code. However, there have been some SUBSTANTIAL improvements to the area of fatigue analysis in the new Code - particularly around fatigue of weldments.)

Ok, I have been able to complete sections (a) thru (e) for section AD 160.2 Condition B, however I am stuck on section (f).

(f) The specified full range of mechanical loads, excluding pressure but including piping reactions, does not result in load stress intensities whose range exceeds the Sa value obtained from the applicable design fatigue curve for the total specified number of significant load fluctuations. If the total specified number of significant load fluctuations exceeds the maximum number of cycles defined on the applicable design fatigue curve, the Sa value corresponding to the maximum number of cycles shall be considered to be significant if the total excursion of load stress intensity exceeds the value of S, where S is defined as in (b) above.

I am unsure exactly what to do to determine if my vessel meets this requirement or not. I am assuming that the "piping reactions" are the specified nozzle loads, which consist of Dead, Thermal and Seismic; where I can neglect the Seismic portion. The only other cyclic information that may fall into this category is a change in specific gravity (1.0 to 1.6) for a specified number of cycles. I am unsure how the specific gravity really plays into mechanical loads though. So I guess what I am looking for is a more detailed explanation of exactly what (f) is after and how to find it without running a complete FEA analysis on the unit.

Thanks