API 620 conical roof thickness calculation questions

[TXTank]

Should the roof live load and snow load be included in the roof weight when calculating the unit force T1 (equation 8 of 5.10.2.5)?

For example, suppose I have a tank with a 100 foot diameter tank with conical roof with alpha = 80, plate density = 0.2833, and internal pressure of 0.1 psi. The roof plate weight is 81337 lbs. T1 using this weight is 48.516 and T2 is 345.5. The required thickness per 5.10.3.2 is 0.0466. If I add a roof live load of 20 lb/ft^2, the weight jumps to 238416.5 and T1 is now negative at -191.433. After iterating several times for 5.10.3.3, I came up with a required thickness of 0.273. Quite a big difference.

A second question: What if the internal pressure is zero? T2 is then zero. Which section then applies for calculating thickness?

 

[JStephen]

For the internal pressure case, the pressure is positive, the roof plate weight is negative, and you'd omit the live load and snow load.
If the roof is a self-supporting cone (not expected with 100' diameter!), then you could take the design vacuum as a negative number, and roof plate weight, snow load, live load, all as negative numbers, and calculate a negative T1 from it. I don't think API-620 address the allowable stress for T2 in compression in a cone, though.

If internal pressure is zero, you just have a self-supporting or supported cone roof.

 

[TXTank]

Speaking of supported cone roofs, I assume the roof weight should include any rafters, girders, and columns.

For the zero internal pressure case, it doesn't look like any of the 5.10.3 sections apply. Would it be best to use 5.10.3.3 in this case?

 

[TXTank]

Another question, while I'm on the subject. If I use 5.10.3.3 to determine a required thickness, should the corrosion allowance be added to it?

 

[IFRs]

Think about what happens when the tank sees internal pressure. The roof plate tries to move upwards and then it tries to lift up the tank shell because it is welded to the tank shell. If the pressure is high enough to lift the roof and tank shell, the corner weld is at risk. This is what we want to avoid. When evaluating internal pressure, we assume the tank is not experiencing snow or live loads to be conservative. Obvious if you now think about it in this light, only structure supported by the roof plates or shell is included since only these dead weights are always available to resist the internal pressure in the vertical direction. Structure supported by the columns will not help resist uplift, it will just stay there and the roof plate will lift off of it.