If the culvert is under outlet control and this a steady flow model, you might consider doing the entire model as regular cross sections with lids. You can also do this with unsteady flow, but unsteady flow cross sections with lids are sometimes less stable (make sure to check the priessman slot option). If the inlet is under inlet control, the cross sections with lid approach won’t give you inlet losses based on inlet shape (rounded, square, etc.).
Yes, you can model it as two or more culverts. You will need to have two ‘regular’ cross sections between the culvert. These cross sections should have lids and the cross section bottom and top should generally match the culvert shape. Make the top of the deck high enough to stay above the energy grade line.
If you want to model the loss as a minor loss with your own minor loss coefficient, there is more than one way to do it but one method:
For the upstream culvert set the exit loss to ‘0’. Set the contraction and expansion losses for the two cross sections to zero and set the inlet loss for the downstream culvert to the minor loss coefficient. For culverts, the exit loss is based on a difference in velocity between the culvert and the downstream cross section. The inlet loss is just the coefficient and the inlet velocity (it does not take a difference in velocity). So the inlet loss will give you the minor coefficient multiplied by the inlet velocity.
If you wanted to model the losses based only on the contraction/expansion and let the program decide that, you could make the two cross sections match the downstream culvert and enter an exit loss for the upstream culvert (the inlet coefficient and cross section should be zero). Or you could model the first cross section to match the upstream culvert and the second cross section to match the downstream culvert and compute the losses between the cross sections (the other losses should be zero).
If the change in culvert shape is gradual and not too extreme, letting the program compute the contraction/expansion might be reasonable. For a sudden change in shape and/or a bend and/or change in slope, the minor loss would probably be better.