From: Matthew Barnes [mbarnes@uq.edu.au]
Sent: Thursday, 23 August 2007 11:29
To: Gray Duncan
Cc: Nielsen Ole
Subject: RE: sensitivity results [SEC=UNCLASSIFIED]

Hi Duncan,

 

I'm well thanks, but busy once again now the teaching semester is well underway. Research is certainly more productive when the undergrads are on holiday! How are you and the family?

 

Please find attached the data requested. Details of the initial conditions are included in the excel file.

 

Please note:

  • Measurement sampling frequency = 50Hz (0.02s); ANUGA output frequency = 10Hz (0.1s)
  • Ux measurements start at t = 6s (this is the first sensible measurement after filtering for instrument communication errors)

 

I have other data sets of equal quality at various measurement locations and for bed slopes = 1:50; 1:20; and 1:10. Let me know if you would like anything further.

 

Cheers,

Matt.



From: Duncan.Gray@ga.gov.au [mailto:Duncan.Gray@ga.gov.au]
Sent: Wed 22/08/2007 3:49 PM
To: Matthew Barnes
Cc: Ole.Nielsen@ga.gov.au; Tom Baldock
Subject: sensitivity results [SEC=UNCLASSIFIED]

Hi Matt,

How are you going?

 

About a year ago you sent these results comparing an experiment against an ANUGA simulation.

 

Could you send up the data and the scripts used to create the simulations?  Ole is considering a new flux limiter and this is the best dataset to validate the new limiters against, but we can’t do the testing without the data.

 

I’ll be away from the 27/8 – 18/9, so it would be great if you could cc Ole into your reply.

 

Cheers

Duncan

 

+++++++++++++++++++++++++++++

 

Attached is the sensitivity plot as bitmap. The .emf extension stands for

enhanced metafile. I usually save Matlab plots that I plan to use for

presentations in .emf format as the compression is very good.

 

 

 

Enjoy NZ and speak to you soon.

 

 

 

Matt.

 

 

 

 

 

Matthew Barnes

 

PhD Research Scholar, Dept. Civil Engineering (Coastal Engineering)

 

The University of Queensland, Brisbane QLD 4072, Australia

 

Ph: +61 7 3365 4170

 

Email:  <mailto:mbarnes@uq.edu.au> mbarnes@uq.edu.au

 

http://www.uq.edu.au/coastal/main.htm

 

-----Original Message-----

From: Duncan.Gray@ga.gov.au [mailto:Duncan.Gray@ga.gov.au]

Sent: Wednesday, 13 September 2006 10:21 AM

To: mbarnes@uq.edu.au

Subject: RE:

 

 

 

Hi Matt,

 

Thanks for sending the files.  I suspect the unrealistic values would be due

to the drop to the catchment not being vertical.  2.0 m is at the bottom of

the catchment.  I think 1.99 would be on the slope down. This could be why

the water has high velocity in the x. You're right as well, when you note

that in ANUGA the water does not project over the edge.   

 

 

 

Actually, with regards to the files, the velocity and depth files had no

extension.  I added .csv and could open them in excel.  I can't open the

sensitivity file.  I don't know what application uses the extension .emf.

Maybe this info could be sent as an .xls file?

 

 

 

It's great that the simulated shear stress compares well to the measured

shear stress. 

 

 

 

Hope it goes well in Japan and the Uni of Aberdeen.  I'll be in New Zealand

next week.

 

 

 

Cheers

 

Duncan

 

 

 

--------------------------------------------

Duncan Gray             Software Engineer  

Risk Research Group     P: +61 2 6249 9077

Geoscience Australia    F: +61 2 6249 9986

E-mail: Duncan.Gray@ga.gov.au                   

--------------------------------------------

 

-----Original Message-----

From: Matthew Barnes [mailto:mbarnes@uq.edu.au]

Sent: Tuesday, 12 September 2006 5:45

To: Gray Duncan

Subject:

 

Hi Duncan,

 

 

 

I hope you're well. I've attached a few files that may be of interest.

 

 

 

The velocity x, velocity y, and depth overtop files refer to the UQ dam with

overtopping at the edge into a catchment. Gauge points are at the flume edge

(2.0), 0.01 m back from the edge (1.99) and 0.1 m back from the edge (1.9).

The modelled velocities and depths close the edge (2.0 & 1.99) are

unrealistic. I assume this is due to ANUGA modelling the overtopping flow as

"falling" down into the catchment rather than "projecting" over the edge as

the observed overtopping behaves. Model data comparisons with a Dirichlet

boundary at the edge are much better.

 

 

 

The ANUGA sensitivity file is for the same initial conditions (horizontal

PVC bed, depth 0.2) but varying Manning's n. The depth and velocity

model/data comparisons for the smaller n values are excellent. Predicting a

shear stress using the modelled velocity and allowing the Darcy friction

factor f to vary (an approach I hope will work for the swash) also compares

well with the shear stress measured by the shear plate. I'm very happy with

this result!

 

 

 

I'll be away for the next couple of months, but will continue working with

ANUGA. I'll keep you posted with any interesting comparisons.

 

 

 

Cheers,

 

 

 

Matt.

 

 

 

Matthew Barnes

 

PhD Research Scholar, Dept. Civil Engineering (Coastal Engineering)

 

The University of Queensland, Brisbane QLD 4072, Australia

 

Ph: +61 7 3365 4170

 

Email:  <mailto:mbarnes@uq.edu.au> mbarnes@uq.edu.au

 

http://www.uq.edu.au/coastal/main.htm

 

 

 

 

 

 

--------------------------------------------
Duncan Gray             Software Engineer   
Risk Research Group     P: +61 2 6249 9077
Geoscience
Australia    F: +61 2 6249 9986
E-mail: Duncan.Gray@ga.gov.au                    
--------------------------------------------