Surge wave analysis

[Caroline]

I’m trying to analyse the surge wave or backwater effect in a channel after an abrupt closing of an inline gate.

The result after gate closing first shows a heaving of the upstream water surface, next a decrease of the water surface and so on. I.e. the “unsteady flow time series”-plot shows an fluctuation with increase/decrease of water depth over all sections at exactly the same time.

I was expecting a travelling wave where the wave would appear at the cross section closest to the gate and then, over time, the sections further away would show a wave pattern.

I’ve tried to use both stage and flow hydrograph as boundary condition for the upstream side but it doesn’t seem to affect the result.

Does this behaviour seem reasonable to you? I assume HEC-RAS can deal with travelling waves?

/ Caroline

[Caroline]

I just realised, that when changing the upstream boundary condition to constant flow instead of constant water surface level, there is a shift, the water rises at some cross sections simultaneous at is it sinking in other.
(When using constant water surface, the result is as described above.)

But, the cross sections 200 m upstream still show a wave directly as the gate closes and not after a while, for instance after the time it would take a wave to travel that distance for the specific channel depth I’ve used.

[Chris G.]

I’ve noticed that before as well. Remember that hydraulic transients are dynamic and the flood wave travels very fast through your canal (gh)^0.5. A depth of 3 meters gives a wave celerity of about 5.5 m/s. If your time step and detailed output intervals are very coarse and/or your canal length is relatively short, you may not notice a traveling wave at all when you animate in HEC-RAS. I’ve played around with this in the past, and by making the canal very long, using a fine time step and especially a small detailed output interval, and closing the gate very fast, you see a very well-defined wave traveling in the upstream direction. I can send you an example HEC-RAS project if you leave me your email.

Chris G.
@RASModel

[Caroline]

Hi,
Chris, thanks for your help and hec-ras model.

I’ve moved on now to analyzing the downstream side of the channel, i.e. the wave downstream the gate. The effect of the gate-closing is larger than on the upstream side. The water surface closest to the gate sinks approximately 5 times the rise on the upstream side, (~0.5 m compared to ~0.1 m).

As downstream boundary condition I use a flow hydrograph. A few sections upstream, I put an inline structure to prevent the channel from being dried out.
I do adjust the flow response in the flow hydrograph after looking at the xs flow in the unsteady flow spatial plot. I keep the flow constant until the decrease in flow translates to the channel-end. Then I start to reduce the outgoing flow.

Does it seem reasonable to use my boundary conditions?
Thanks!

[Chris G.]

Why a flow hydrograph for the downstream boundary? I’ve had very little success in the past using flow hydrographs for the downstream boundary, because if you don’t exactly match what RAS is calculating for flow, then you get errors at the d/s boundary that can lead to instabilities. And, as you noticed, sometimes the How about normal depth for your downstream boundary? This should work very nicely for a canal, as long as your downstream boundary is not in the middle of a backwater curve from a downstream control (if so, you could still use normal depth, just move the downstream boundary out of the backwater influence).

[Caroline]

I have had some problems using the flow hydrograph, but the solution is actually more stable than the stage alternative.

Since my channel connects two water levels (upstream controlled lake, downstream sea) with larger deep upstream than downstream my slope gets a negative number which HEC-RAS refuses to use. Is there some way to overcome that problem?

By the way, what did you mean to write in the sentence ” And, as you noticed, sometimes the”? It ends pretty abrupt.

Have a nice weekend

[Anonymous]

hi, I have a culvert that empties into the sea, as I analyze the backwater. Thank you.

[Anonymous]

Agree with Chris that the output should be more frequent. Typically, the computational interval and the output interval should allow for at least 20 to 30 points within the time scale of the event. So, if the gate is moved in 10 min. the computational and output interval needs to be on the order of 20 sec. That’s just for resolution of the computed results. Stability may require a smaller computational time step.

Use of stage as a downstream boundary is common. Many applications of HEC-RAS use stage downstream and flow upstream. I have used RAS several tidal situations where the downstream boundary is a tidal stage hydrograph and the upstream BC is a flow hydrograph. In these applications, RAS computes negative flows and velocities correctly.

[cristian16]

Hi all,
I have a similar situation, where I want close the gates very fast (10 seconds). What boundary condition suggest to me? It’s possible simulate a close so quickly?
Best Regards-
Antonio

[Author]

Posts