| 1 | ANUGA CONVERGENCE STUDY USING TRUE-SCALE VERSION OF THE |
|---|
| 2 | OKUSHIRI ISLAND TSUNAMI WAVETANK EXPERIMENT |
|---|
| 3 | |
|---|
| 4 | This directory currently contains code to scale-up to true scale, the 1:400 |
|---|
| 5 | wave tank simulation of the 1993 Okushiri island tsunami as described at |
|---|
| 6 | the Third International Conference on Long Wave Runup: |
|---|
| 7 | http://www.cee.cornell.edu/longwave/index.cfm?page=benchmark&problem=2. |
|---|
| 8 | |
|---|
| 9 | The "up-scaled" files are then used for conducting a convergence study |
|---|
| 10 | in ANUGA. |
|---|
| 11 | |
|---|
| 12 | Data files available in this directory are |
|---|
| 13 | |
|---|
| 14 | okushiri_truescale_bathymetry.txt: |
|---|
| 15 | The true-scale digital elevation model |
|---|
| 16 | |
|---|
| 17 | okushiri_truescale_input.txt: |
|---|
| 18 | The true-scale timeseries applied at the western boundary |
|---|
| 19 | |
|---|
| 20 | okushiri_output_truescale_ch5-7-9.txt: |
|---|
| 21 | Experimental data measured at three gauge locations in the original wavetank experiment |
|---|
| 22 | which has been up-scaled to true-scale |
|---|
| 23 | |
|---|
| 24 | |
|---|
| 25 | The ANUGA scripts to run are |
|---|
| 26 | |
|---|
| 27 | project_truescale.py: |
|---|
| 28 | This script contains project filenames and is called in the create and run scripts below. |
|---|
| 29 | |
|---|
| 30 | create_okushiri_truescale.py: |
|---|
| 31 | This script will convert the text files to native |
|---|
| 32 | ANUGA netcdf formats and also create a suitable triangular mesh. |
|---|
| 33 | |
|---|
| 34 | run_okushiri_truescale.py: |
|---|
| 35 | This script will run a numerical simulation based on the the bathymetry |
|---|
| 36 | and the given boundary condition and store the model output in an ANUGA |
|---|
| 37 | sww file which can be viewed using animate, or further interrogated by ANUGA. |
|---|
| 38 | |
|---|
| 39 | compare_timeseries.py |
|---|
| 40 | This script will extract timeseries from the sww file and plot them |
|---|
| 41 | together with the experimental data provided. Numerical similarity |
|---|
| 42 | measures will also be computed. |
|---|
| 43 | |
|---|
| 44 | get_timeseries.py |
|---|
| 45 | This script extracts timeseries from the sww file and writes them to individual |
|---|
| 46 | csv files for each gauge location. |
|---|
| 47 | |
|---|
| 48 | export_results.py |
|---|
| 49 | This script extracts a number of variables from the sww file (for eg stage, |
|---|
| 50 | depth, momentum) in three regions at three different but fixed regular grid |
|---|
| 51 | resolutions (at either discrete time indices or maximised over the whole |
|---|
| 52 | simulation). |
|---|
| 53 | |
|---|
| 54 | Methodology for true-scale transformation: |
|---|
| 55 | |
|---|
| 56 | Positions (ie bathymetry, polygon definitions, gauge locations) were derived by |
|---|
| 57 | carrying out a scalar multiplication of all x, y, z values in the original files by 400. |
|---|
| 58 | |
|---|
| 59 | Input waveform: Truescale input wave with period T' is assumed to have a wavelength |
|---|
| 60 | and amplitude 400 times the original 1:400 waveform. Given the relationship: |
|---|
| 61 | |
|---|
| 62 | T = wavelength / sqrt (g * h) |
|---|
| 63 | |
|---|
| 64 | Then T' = 400 * wavelength / sqrt (g * 400 * h) |
|---|
| 65 | = 20 * wavelength / sqrt (g * h) |
|---|
| 66 | = 20 * T |
|---|
| 67 | |
|---|
| 68 | Therefore, time (s) is multiplied by 20 and water surface (m) (ie amplitude) is multiplied by 400. |
|---|
| 69 | |
|---|
