Differential Expansion in large Ducts

[SteveTricarico]

I'm trying to determine if an expansion joint is required to separate two sections of flue gas ductwork in a utility sized SCR system. I need to justify the decision with a quantitative analysis. The flue gas temperature is at 650F and is introduced into a vertical duct about 70 feet wide and 6 feet deep. After rising about 100 feet, the duct turns and runs horizontal for about ten feet and then turns down and opens into a large "reactor" box with dimensions of 70 feet by 30 feet. The reactor is about 60 feet long (top to bottom) and then eventually tapers back to a 70 foot by 6 feet outlet. The entire structure is supported from below and can grow "up" to accommodate thermal expansion. The reactor box is full of structural members and hundreds of catalyst modules. I'm concerned that the significant difference in thermal mass between the inlet section and the reactor section will result in different thermal growth and will over-stress the carbon steel duct. It is an unsteady state phenomenon with the flue gas initially cooling as the metal is warmed. The flow rate is relatively high, about 2,000,000 lbs/hr, and the fuel source is coal. Any suggestions on how to "model" this system without doing a full CFD analysis?

Thanks Steve