Mixed flow over broad crested weir

[Anonymous]

I have modelled an earth/gravel type broad-crested weir at a dam with the following properties: 5m wide in section (perpendicular to flow), manning’s n = 0.028, no slope. The weir has a gradual upstream slope coming from the dam and a vertical drop at the downstream end.

For the steady state flow analyses I have used a mixed flow analyses with the boundary conditions as follows: upstream a known water surface level and downstream a normal depth slope.

We are expecting the water surface level to be sub-critical upstream of the weir and then become critical over the weir before it becomes supercritical over the drop. However, our results are showing a sub-critical water surface all the way to the vertical drop, i.e. there are no transition to critical or supercritical flows over the weir.

Is HEC-RAS able to model these supercritical flows over the weir and is HEC-RAS able to accurately model flows over a vertical drop?

[Anonymous]

If you set aside the situation where you have a high tailwater influence:

The weir equation is based on critical depth at the weir. RAS will compute the energy grade required upstream of the weir in order to pass the given flow over the weir at critical depth. The cross section upstream of the weir will be subcritical.

Again, setting aside the situation where the tailwater is high enough to influence the flow:

The weir equation assumes a waterfall into a pool. The pool is controlled by the downstream situation. RAS does not report any answers on the downstream side of the weir where supercritical flow is occurring.

Yes, RAS can accurately model flow over a vertical drop given the standard limitations for a 1D steady state back water. No, the weir equation is not the way to model supercritical flow.

If you have a broad-crested weir/steep region that you want to model supercritical flow, just model the entire region with regular cross sections and don’t use a weir at all. Although I don’t think any 1D model is going to accurately model the supercritical part of vertical drop. That sounds like a 3D problem.

[Author]

Posts