In all of the various SWMM models the side outlet orifice and sump orifice are modeled in the same way: by equating the full flow orifice equation to the full flow manning's equation and making an equivalent pipe. Once the equivalent pipe is created the flows in the two orifices should be the same if the area, the type of orifice, and the upstream invert levels of the orifice are the same.

SWMM 4.4 and Mike SWMM have additional orifice features for real time control such as gates. Visual SWMM has the the option of actually using the orifice equation instead of the equivalent pipe if the user so designates in the interface.

The inlet capacity in XP-SWMM and Visual SWMM is designed to limit the inflow from the runoff layer into an inlet in Extran. It simulates the flow restriction of a grate inlet, curb inlet or combination inlet. The bypassed flow if the inlet cannot handle all of the flow goes down an overland road section to the next inlet. You have to model the road section explicitly.

Thanks for the reply, it helped. However, I am still a bit confused on the fact the invert for sump orifices is actually dropped by 0.96D. How can modeled flow be the same for both orifice types if this is the case? Wouldn't the equivalent pipe actually be at a lower invert than that for a side orifice thereby making the flow in the two pipes different, albeit not by much?

This assumes that the bottom orifice invert elevation after the drop was the same as the side outlet orifice elevation. As you state if both started at the same elevation then the bottom outlet orifice would have a new invert elevation. My statement that the flow would be the same assumes that the head difference across the orifices are the same. Given that the bottom outlet orifice has a "head start" this is unlikely.

Regardless, the flow calculations and all the other computations that go on in the calculation of the flow are the same in both types of orifices.

I agree that in the real world different orifice types would not produce "100% same" flows. Their modeling technique (equation, algorithm, etc.) should be revisited if they do.