Thermal Cycles for Fatigue Exemption

[Paulettaa]

TGS4, Please help me.
I want to know how to take thermal cycles into account when doing a fatigue exemption screening. My question is how can I understand the temperature distribution throughout the vessel. Is it my responsibility to somehow calculate the temperature distribution all over the vessel or should the owner give me information about that.
Well,for the equipment that I am looking at, there is a long period during which the gas flow has the temperature of 47 C and suddenly the inlet gas changes temperature to 280 C for regeneration cycle. The UDS specifically states that the transient thermal gradient across the vessel is negligible, however I think this applies to the adjacent points that are on the surface of the shell not for the through thickness adjacent points.
How can I measure the temperature difference between the inner and outer surfaces of the equipment? Shall I run a transient thermal finite element analysis? and if so shall I model the whole vessel or only a portion of the cylindrical shell is enough? Furthermore what value for the coefficient of the convection shall I consider? Since this convection coefficient is not mentioned in the UDS.

 

[TGS4]

In general, this work would be performed by the owner (or the owner's designated agent). However, when the owner abrogates that responsibility, it may shift to manufacturer.

You will certainly need to perform a transient thermal analysis. The longitudinal and through-thickness temperature differences will need to be calculated. The typical region of interest is the inlet nozzle (nozzle, head, and top portion of the cylinder).

Regarding the connective heat transfer coefficients, you should ask for that from the owner. But if they are unable to provide that, you may need to calculate it yourself. There are procedures in textbooks and standards such as ASTM C-680.

 

[Paulettaa]

TGS4, I checked ASTM C-680. Unfortunately the methods introduced there are for steady state heat transfer in a fully developed thermal and hydrodynamic flow. For the fatigue exemption I need transient thermal properties and the flow is not fully developed in the corresponding length of the shell (with the Reynolds number the flow will be turbulent). I see in ASTM the textbook "Fundamentals of Heat and Mass Transfer" by Incropera and Dewitt is extensively referenced. In the same book the range of heat conductance for gases is mentioned to be between 25 to 250 W/m2K. Is it reasonable to take the h to be the maximum value in this range since the higher the value of the heat conductance is, the higher transient thermal temperature difference through the thickness of the shell will be. If this leads to the exemption of the vessel from detailed fatigue calculations then in the real situation it is exempted also since this is a conservative way of dealing with the value of heat conductance.

Warm Regards

 

[TGS4]

I wouldn't use a value of the convective heat transfer coefficient for a gas as high as 250 W/(m^2*K) as that seems excessive to me.

Nevertheless, if you were to use such a high value, and you could still exempt the components from a fatigue analysis, then you would probably be on the conservative side.

 

[georgeverghese]

If this vessel is in mole sieves regen service, then the mole sieves manufacturer / supplier will usually have time vs bed temperature profiles over the entire length / width of the vessel of the mole sieves bed. Would imagine the vessel wall will approximate the bed temp at the outer edges of the bed, since these vessels are usually insulated. Wouldnt be suprised though, if you find the mole sieves supplier to be reluctant to pass on these profiles to you though, as this is licensor technology - no harm asking.

 

[Paulettaa]

TGS4, for one special case I have a vessel which is operating at 48 C and suddenly a gas of 280 C starts to enter the vessel for regeneration process. I assume the vessel wall and all metal being in 48 C initially.

The above image is the contour of temperature distribution in a transient thermal analysis. The model is actually the inlet nozzle and spherical head for a pressure vessel subject to cyclic loading. When the temperature through the thickness of the nozzle neck and the self reinforcement is considered versus time, the temperature difference starts to increase at the beginning and then starts to decrease after some seconds. so I consider the maximum temperature difference to be used in the exemption calculations. The through thickness temperature difference will be maximum 8 C for different locations of the model. However my problem is when I want to consider the temperature difference between the inner and outer edges of flange face. First of all it takes a much longer time for the temperature difference to reach its peak value (more than 500 seconds of transient thermal analysis). Second is that I do not know if these two points are considered as adjacent points as defined by VIII-2.

(a) For surface temperature differences, points are considered to be adjacent if they are within the distance L computed as follows: for shells and dished heads in the meridional or circumferential directions,
L=2.5(Rt)^0.5
and for flat plates,
L=3.5a
(b) For through-the-thickness temperature differences, adjacent points are defined as any two points on a line normal
to any surface on the component.

What are these two points? are they through thickness points or are they surface meridional or circumferential points ?
Another question is what is meant by ΔT_Ein method A of screening for fatigue calculations. Suppose that I have two cycles for the first one the inner surface is 8 degrees warmer than the outer surface and for the second one the inner surface is 15 degrees colder than the outer surface. Do I have to consider the value to be the maximum of the two cycles (i.e. max(15,8)=15) or the sum of them (i.e. 8+15=23).

 

[TGS4]

This is getting to the point of consulting. I recommend that you contact me directly.