Numerical Computations Database for Supersonic and Hypersonic Fluid Flow Simulations

Richard Sanders, Ph.D., Associate Professor, UH
Chen P. Li, Ph.D., JSC
Eric Morano, Ph.D., Post-Doctoral Fellow, UH

ONE OF THE MOST CHALLENGING PROBLEMS of modern Aerospace Engineering is the accurate prediction of very high speed fluid flows typically encountered by spacecraft during hypersonic atmospheric entry. Understanding such flows necessitates the efficient comparison and validation of results reported from a wide range of scientific disciplines. These disciplines include, fluid mechanics (complex flow analysis), basic physics and chemistry (models), experimentation (validation and analysis), numerical analysis (algorithms), and computer science (high-performance computing and visualization).

Right. Dr. Eric Morano, post-doctoral fellow, (l.) and Dr. Richard Sanders (r.) solve problems of fluid flow in hypersonic atmospheric entry.

Historical Overview
During the past few decades, high-powered supercomputers have been exploited to a greater degree in order to simulate complex hypersonic flows via numerical methods. However, results from numerical simulation do not always completely agree with actual physical data. The disparity may be due to discretization error of the underlying numerical techniques and/or, more profoundly, from imperfect or not well understood physical modeling. Therefore, an international cooperative effort was initiated in the 1980s to study various aspects of high speed fluid flows. Short courses on hypersonics were held in Paris (France), Colorado Springs (USA), and Aachen (Germany). In the early 1990s, three sessions, Parts I, II and III, of a scientific workshop series organized by INRIA (Institut National de Recherche et en Informatique) took place in Antibes, France. Following Part II of the Antibes series, researchers deemed vital to the preservation of the essential scientific knowledge base acquired during the HERMES reentry vehicle R&D program into an electronic ftp archive. The European Hypersonic Data Base (EHDB) was initiated in 1993 to store numerical and associated experimental contributions to standardized test problems collected during the Antibes series.

In 1995, the first United States sequel High Speed Flow Database Workshop was organized by the University of Houston. This workshop on the UH campus provided an incentive to employ more powerful world-wide web technology to further develop the European archive concept. With support from NASA-JSC and ISSO at the University of Houston, the Houston High Speed Flow Database (HHSFD) was announced at a dedicated special session of the 1997 American Institute of Aeronautics and Astronautics Snowmass meeting and officially opened July of the same year.

Overview of the HHSFD
Scientists today are often solicited to contribute to multipartner programs having technological goals that involve integrating results from varied areas of basic science. HHSFD will doubtless offer assistance to scientists from specialized research areas engaged in the larger scope problems involved in the study of very high speed fluid flow. Technically, the HHSFD is a collection of complementary technical components that allow specialist from one field to take advantage of developments by other specialists at a single site and within a common framework. The Houston High Speed Flow Database is a world-wide web accessible assemblage of information on the numerical and experimental simulation of high speed fluid flows. The url is: http://hhsfd.math.uh.edu. The contents of the HHSFD include:

Standardized test problem specifications.

To assess a given simulation technique, scientists must compare apples to apples. A wide variety of fully specified test problems are supplied specifying flow geometry, free stream conditions, and appropriate physical scales.

Computational solutions and grids.

Numerical solutions created by internationally recognized specialists provide the backbone of the HHSFD. Most test problems have several solutions that may be studied and/or downloaded over the net.

Accompanying experimental data.

Several test problems are supported by high quality experimental data.

Methodological articles.

Contributors to the HHSFD have supplied either methodological abstracts or full multi-page articles that explain in detail the techniques employed to solve a given test problem.

Standardized data formats.

Multi-block structured and unstructured form solutions within the database are stored in a common well-defined file format. A special data compression routine is available for downloading.

Network flow visualization and tools.

Real time with interactive flow visualization is an integral component of the HHSFD. Two and three dimensional contributions may be visualized over the net. Soon, various visualization tools will be made available for downloading.

Operational Details
The HHSFD is organized in a hierarchical file structure, as depicted in Fig. 1. Directories are laid out in layers that correspond to the available workshops, followed by test cases, then eventually subproblems and contributions. The database control program herl, presented below, has the capability of browsing through these directories where it finds various informational files used to create on-the-fly web pages sent to the database client's web browser. Information files contain various details concerning specific workshops, test case definitions, and contributors' biographical information.

Figure 1. HHSFD Directory Structure.

Client access to the HHSFD and the primary database interface is controlled by a common gateway interface (CGI) program called herl. Herl is a C-language program specifically designed by UH investigators to provide a command language to dynamically build web pages based on the client's specific query with respect to the current state of the database. The created hypertext markup language (html) page is then presented to the web client's browser. Currently, there are more than 20 herl commands, some of which are:

cd--Change the current working directory.
dir--Assign to a named variable the names of files (or directories) present in a specified directory.
fset--Read a specially formatted file and assign to variables named within the file their corresponding values.
goto--Conditionally branch to a specified label of the form label_name: in the current input file.
grep--Perform pattern matching on a specified string using a specified regular expression and assign the result to a named variable.
input--Include a specified file for runtime processing.
set--Assign a specified value to a named variable.
subst--Substitute a specified regular expression with a specified string in a given string. The result is assigned to a named variable.
exec--Start a subprocess.

Right. HHSFD graphical interface.

An excerpt from a herl program which produces a database web page is shown in Fig. 2. This portion of code creates the list of test cases available for a given workshop. Herl commands are delimited by backslashes "\" while herl variables are delimited by dollar signs "$." All other text in the example is standard html.

<B> $wstitle$ test case links: </B>

<TABLE>
  \dir tcdir* = . "T.*" "11.1" \
  \arith i = 0\
  \NextTestcase:\
  \  arith i = $i$+1
  \  goto DoneTestcase ifnot $tcdir[i]$\
  \  finfo check* = $tcdir[i]$/tcdef.info\
  \  grep check = "01.." $check1$\
  \  goto NestTestcase ifnot $check$\
  \  fset $tcdir[i]$/tcdef.info testcase testname\
  <TR> <TD>
  <A HREF=$HERL_HOME$:Workshops/desc.htmlp$--
  --$&wsdir=$wsdir$&tcdir=$tcdir[i]$&uid=$uid$>
  <IMG SRC="$NEXT_BUTTON$"></A>
  </TD> <TD>
  $testcase$ &nbsp $testname$
  </TD> </TR>
  \  goto NextTestcase\
  \DoneTestcase:\
</TABLE>

Figure 2. Excerpt of an herl program.

Further Developments
Further workshop data will be added to the HHSFD starting with the November, 1997, workshop in Naples, Italy, titled "The First Europe-US High Speed Flow Field Database Workshop, Part II." Several new test cases will be introduced from labs and/or universities in Europe, Russia, Japan, and the United States. A mirror site of the HHSFD will be installed in Europe at INRIA Sophia-Antipolis early in 1998. In fact, the database infrastructure developed at the University of Houston will very possibly serve as a general purpose tool by INRIA Sophia-Antipolis to support the numerous and varied professional meetings and workshops sponsored in this southern France locale. Further detailed discussions concerning these matters will continue between HHSFD personnel and their INRIA counterparts this November in France.

NASA Langley researchers sponsored a visit by HHSFD personnel to their lab September, 1997 to demonstrate and initiate HHSFD software. A small version of the HHSFD will be set-up at the agency for internal use. This set-up is planned for the NASA Ames Research Center.

Contents
ISSO -- Institute for Space Systems Operations
1996-1997 Annual Report