One solution to this problem is to increase the Manning's n value in the area where the program is first going to critical depth and in the steeper portions of the reach.

This sharp increase in the water surface slope will often cause the program to overestimate the depth at the next cross section upstream, and possibly underestimate the depth at the next cross section downstream (or even the one that went to critical depth the previous time step). If you are running the software in the default mode (mixed flow option not turned on), and if the program goes down to critical depth at a cross section, the changes in area, depth, and velocity are very high. However, the HEC-RAS software does have an option to run the 1D solution scheme in a mixed flow regime mode, which allows it to solve through these types of flow transitions. Venant equations and their derivatives tend to cause model instabilities (generally in rapid flow areas the derivatives are over estimated).

As Froude number approaches 1.0 (critical depth), the inertial terms of the St. Additionally, the assumption of a hydrostatic flow distribution may not be valid. Areas of rapidly varied flow, such as flow profiles transitioning from subcritical to supercritical flow, and hydraulic jumps, tend to cause the 1D solution scheme to have difficulties in remaining stable. The default solution methodology for the 1D unsteady flow routing option within HEC-RAS is generally for gradually varied flow. Steep streams tend to have very high velocities and rapid changes in depth, area, and velocity, which make it more challenging to obtain a stable model solution through these areas. Modeling a dam break flood wave through a steep stream system is even more difficult. Steep streams are very difficult to model with an unsteady flow model in general.

Download steep to