Don't let the structural scare you, I know what I'm talking about (15 years of hydraulic engineering practice on top of 10 years as structural):
In the standard step method, the calculations proceed from downstream to upstream for subcritical flow and vice versa for supercritical flow.
If the calculation assumes that the flow is subcritical and the flow is actually supercritical, the iterations will not balance. The reason is that the projection of the energy line, used to estimate the energy line at the next section, will be divergent from the actual one.
In those cases, backwater calculation models (including HEC-2 and HEC-RAS) will force an answer by assuming that the flow at the section is critical. This allows the model to continue the calculations in cases where a short reach of supercritical flow occurs within a generally subcritical reach.
If your results show several sections of critical flow, you should run the model as supercritical. You would have to use different model.