Changeset 8941

Timestamp:

Jun 28, 2013, 11:04:40 PM (12 years ago)

Author:

mungkasi

Message:

Adding some descriptions.

File:

• trunk/anuga_core/source/anuga_validation_tests/report.tex (modified) (1 diff)

trunk/anuga_core/source/anuga_validation_tests/report.tex

@@ -45 +45 @@
 
 
-
-% ****Add ***
-How to run, using produce results,
-
-change defaults using local parameters.py format
-
-alg = '2_0'
-
-would use flow_algorithm '2_0' See manual for description of flow algorithms.
-
-validate all will rerun a subset of the tests (which have sensible "correct"
-results to test against.
-
-One by one testing 
-
-
-
-
-
-
 The results in this report were produced by \anuga{} version \majorR{} from svn
 repository revision \minorR{} at time \timeR.
-The flow algorithm was \alg{} and CFL condition \cfl, unless otherwise stated explicitly.
+The flow algorithm was \alg{} and CFL condition \cfl, unless otherwise stated explicitly. Based   on this version, 26 tests are available in the subversion. To get an automated report, we can run either an individual run of the available tests or the complete (whole) test.
+
+To do an individual test, we can run the python module \\
+\verb produce_results.py \\
+available in the corresponding test directory. The module will do the numerical simulation of the given problem, plot the simulation results in png files, and type-set the corresponding individual automated report. The individual automated report is in pdf file and saved in the same directory.
+
+To do the complete test, we can just run the python module \\
+\verb produce_results.py \\
+available in the directory \\
+\verb anuga_validation_tests \\
+Similar to the module for an individual test, this python module will do the numerical simulations of all the given problems, plot results in png files and save them in its corresponding directory, and finally type-set the complete report. The complete automated report is saved in this directory.
+
+The simulation results can be analysed qualitatively and quantitatively. Qualitative analysis can be done by investigating the plots of the results whether they are physical or not, and whether the behaviour is the same as we expected. Quantitative analysis can be conducted by checking the numerical error. We have provided python module with the name begun by the word ``validate'' in each of individual tests. We can also run validate$\_$all.py to measure the numerical errors from the directory \verb anuga_validation_tests . This validate$\_$all.py will run a subset of the available tests having sensible ``correct'' results to test against.
+
+The main parameters in the validations are the Courant--Friedrichs-Lewy (CFL) number and the flow algorithm. They are spelled ``cfl'' and ``alg'' respectively in the python module \\
+\verb parameters.py \\
+which is available in the \\
+\verb anuga_validation_tests \\
+directory. In the default setting, we set the CFL to be $1.0$ and the flow algorithm to be 1$\_$5 (second order in space and first order in time). The complete available flow algorithms are as follow:
+'1$\_$0', '1$\_$5', '1$\_$75', '2$\_$0', '2$\_$0$\_$limited', '2$\_$5', 'tsunami', 'yusuke'.
+%\begin{enumerate}
+%\item '1$\_$0', 
+%\item '1$\_$5', 
+%\item '1$\_$75', 
+%\item '2$\_$0', 
+%\item '2$\_$0$\_$limited', 
+%\item '2$\_$5', 
+%\item 'tsunami', 
+%\item 'yusuke'.
+%\end{enumerate}
+They can be found in \\
+\verb \anuga_core\source\anuga\shallow_water\shallow_water_domain.py .
+
+
 
 The report is organised as follows. We collect a number of tests against