I work at a company which mainly sells, designs and produces heat exchangers and air cooled coolers. Sometimes we also sell heaters with parts of inlet- and outlet duct. I have some questions about the wall thickness determination of duct.
Our current way of work is as follows:
- We make a rough sketch of the duct, let's say a transition piece from 1000x1000mm rectangular to 600mm circular, with a length of 700mm (see attachment).
- We determine the largest flat surface, in this case the flat surface is a triangle (due to the transition) of 1000mm width and a length of 700mm.
- We calculate the thickness using the formula from ASME VIII div. 1 UG34: tr = d*√((Z*C*Pd)/(Sb*E))
+ d=700mm (short span)
+ D=1000mm (long span)
+ Z=1.72 (factor depending on d/D)
+ C=0.33 (support factor)
+ Pd= 0.02N/mm²(design pressure)
+ Sb= 115.1N/mm² (allowable stress, 316L, 100°C)
+ E= 1 (joint efficiency factor)
Result: tr=6.95 -> use 8mm
My feeling is that with this way of work we design way to conservative: we consider the plate as a rectangular of 1000mm by 700mm, resulting in a total surface of 2x the triangle surface. Also this calculation does not include stiffening.
Now my questions are:
1. What formula's are used in the HVAC industry for determining thicknesses and on which standards are they based?
2. Can you help me with some rules about how to take the stiffening from bended plates into consideration? Or are the plates only bended to prevent vibration of the plates?
(Example picture)
3. What software do you prefer, aspecially when also nozzle loads are applicable. Is finite element method the best solution?
Our current way of work is as follows:
- We make a rough sketch of the duct, let's say a transition piece from 1000x1000mm rectangular to 600mm circular, with a length of 700mm (see attachment).
- We determine the largest flat surface, in this case the flat surface is a triangle (due to the transition) of 1000mm width and a length of 700mm.
- We calculate the thickness using the formula from ASME VIII div. 1 UG34: tr = d*√((Z*C*Pd)/(Sb*E))
+ d=700mm (short span)
+ D=1000mm (long span)
+ Z=1.72 (factor depending on d/D)
+ C=0.33 (support factor)
+ Pd= 0.02N/mm²(design pressure)
+ Sb= 115.1N/mm² (allowable stress, 316L, 100°C)
+ E= 1 (joint efficiency factor)
Result: tr=6.95 -> use 8mm
My feeling is that with this way of work we design way to conservative: we consider the plate as a rectangular of 1000mm by 700mm, resulting in a total surface of 2x the triangle surface. Also this calculation does not include stiffening.
Now my questions are:
1. What formula's are used in the HVAC industry for determining thicknesses and on which standards are they based?
2. Can you help me with some rules about how to take the stiffening from bended plates into consideration? Or are the plates only bended to prevent vibration of the plates?
3. What software do you prefer, aspecially when also nozzle loads are applicable. Is finite element method the best solution?