HEC-RAS Model Calibration w/Measrured Flow Data 2

[Ryb01]

Hello All,

I'm in the process of modelling several reaches in HEC-Ras. For each site(reach) we've obtained measured flow data (from a third party). We are to calibrate the HEC-Ras model to produce the same water surface elevation produced in the field.

I've attached a sample data set of the measured flow in excel format for reference;

The total calculated flow for this section is 739.74ft3/s, approximate profile grade where the section was measured is approximately 0.54%. The This reach gently meanders for several miles without obstructions. Accordingly, for tailwater conditions I've assumed that Normal Depth would be appropriate.

I've tried a couple of different methods in an attempt to calibrate the model (with a coulple of assumptions) but have been rather un-sucessful. I would assume that in a steady state model, the only variable to play with would be the mannings coefficient as we have a measured flow, channel geometry, normal depth and standard expansion and contraction (0.1 0.3). I've tried the following;

1) Attempted to calibrate the HEC-Ras model with the X-Section data above. I've completed iterations by adjusting the channel manning's coefficients to re-produce a depth at 327.67ft. I've also tried varying the channel n by station. In short, the manning's coefficient that I'm getting is in excess of 0.09. This reach and several others are in rockey type areas, however, based on past experience of calibrated models I would anticipate the manning value to max out at approx 0.05-0.06.

2) Attempted to reproduce results using flow master and the cross section tool in the FHA Hydraulic Tool Box. These methods have also produced a high mannings coefficient within the channel.

I was wondering if anyone has had experience with calibrating HEC-Ras models based on measured flow and could provide any feedback on the above? I'm wondering one of two things 1) Is there is another way to calibrate the data in which I have not considered? 2) There is an error with the data we've received?

Is there any reference material for calibration you could provide?

Thank you in advance for your feedback, it is greatly appreciated.


Ryb01

 

[beej67]

I must be missing something in the attached spreadsheet. Are each of the points along the same reach? If so, shouldn't the flow be the same, or continuously increasing downstream?

Also, how were they measured?


Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[Drew08]

Looking just at the plotted X and Z coordinates, it looks like the section is at the meander bend and likely in a pool (i.e., left bank has a shallow slope, point bar; right bank has a steep slope).

If your model sections are cut pool to pool, the average depth of flow will be larger (and also the calibrated 'n') than if you cut sections riffle to riffle. Riffle sections typically occur in the straight sections or points of inflection of a meandering stream and are usually more symmetric than what you have shown. I believe most published values for Manning's ‘n’ are based on riffle to riffle depth estimates.

 

[Ryb01]

Beej67..the points in the spreadsheet have been measured at a cross section in the reach, measured from the left bank to the right bank perpendicular to the flow.

The flow has been measured(or so I have been told)using a series of flow slices along the cross section. The velocity has been measured in the river and flow derived from the Q=VA relationship at each flow slice. The total flow would appear to be a summation of all the individual flow slices.

This particular section I've calculated a manning's coefficient of 0.146 for the channel. This does not seem right to me even if the x-section was measured pool to pool or riffle to riffle as Drew08 had pointed out. I should mention that this measured section is "in-water only".

Thank you for you comments so far. Please let me know if you can think of anything else.

 

[beej67]

I see.

Are those surface velocities then? If so, it could be your error. Average velocity is going to be less than surface velocity, and you'll need to correct for that before going to continuity (Q=VA) to determine flow.

Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[Ryb01]

beej67,

Interesting question...no sure, but will look into it further.

 

[francesca]

You need to block out ineffective flow areas from the model. Typically where you have <3 ft depth and <0.5 fps velocity your flow is ineffective. Have you tried building a composite Manning's from the equation presented here:

http://www.fhwa.dot.gov/BRIDGE/wsp2339.pdf

(page 9)?

The USGS has photos and lat/long coordinates of places with calibrated Manning's values. You can look at them on Google Maps. For instance, this channel with n=0.065

http://wwwrcamnl.wr.usgs.gov/sws/fieldmethods/Indirects/nvalues/0065.htm

is here:

http://maps.google.com/maps?ll=37.737853,-119.561741&spn=0.006033,0.009645&t=h&z=17