diagnostics

This might be subtitled "when things go wrong!". Within the TEXSTAN source code is approximately 2000 FORTRAN lines devoted to checking the input data file for possible mistakes in the setup of the input data.

The code for checking is continually added to, on a yearly basis, and with each passing year more is added. Such is the way it goes for a computer code that has been designed to be robust and general.

In spite of all the input error checking protocols, there will be problems, and some problems are just not very easily resolved, especially if they are runtime errors in which the code is in execution (the input error-checking did not detect errors), and suddenly the code STOPS!

So, what to do if there is a problem? Through use of the flag k8 the user can cause various diagnostics files to be printed. For the user, the most important options are k8=9 and k8=36.

The k8=9 flag allows the user to stop TEXSTAN immediately after the input data file has been read and printed to file in.txt. By examination of the printed input file, the user can be sure the input data has been interpreted the way the user intended it to be.

To help the user ascertain whether certain boundary condition data arrays are smooth, the k8=36 option has been developed. With this option TEXSTAN analyzes the boundary condition files in a form convenient for plotting and then stops before integration begins. The file tex_edge.txt contains the free stream velocity array analysis, and the file tex_wall.txt contains the transverse radius, wall temperature, and radius of curvature analysis. This analysis is explained in detail in the overview: bc smoothness section.

Some of the more common "early user" mistakes include

• inconsistent units in the input variables
• when a variable is not needed in the input data, leaving it "blank" instead of using "0" or "0.0"

We strongly encourage study of the benchmark datasets described in the external flows and internal flows sections of this website.