README file for the fstrsta.alt alter package

1. INTRODUCTION

The purpose of this alter is to take a large modes database (large db)
that has been operated on by the run0a alter package, and perform a fast
restart for modal frequency or modal transient response analysis.  The
large db is attached in a read-only status.  Most of the modeling input
data is extracted from the large database, allowing branching over of
expensive operations such as matrix generation and eigensolution.  This
alter applies to SOLs 111 and 112 only.

2. INPUT

Conventional EXEC and CASE CONTROL input files are used.  Use ASSIGN and
DBLOC FMS statements of the form shown in the example problem.  They use
the large db in a read-only status, and bring over only the data blocks
needed for this type of run.  Place " include 'fstrsta.alt' " before
the CEND entry.  Two SET entries are required, one selected by a DISP=
command, and the other by a STRESS= command.  Grid points and elements
listed on these sets are extracted from the large db.  Loads may be
applied to them, and output requests may be made from them.  When used
in conjunction with Interactive NASTRAN, these SETs are generated in the
PATRAN session automatically.

The Bulk Data file includes conventional dynamic loading requests.  A
new parameter is provided:

PARAM, DEBUG, NO.  For this default value, very little output is sent to
the .f06 file.  In particular, output results such as displacements or
stresses are not printed in the .f06 file.  It is expected that this
data will be obtained from the .xdb results database, which is
automatically requested by the alter package.  For the value "PRINT",
printed results are provided in the .f06 file.  For the value "YES", all
printed results and many intermediate tables and matrices are sent to
the .f06 file.  This value is practical for only small size models
during training exercises. 

3. OUTPUT

Very little printed output is provided unless param, debug is used.  An
.xdb results database is automatically provided.  

4. FUNCTION

The SET command data is used to extract grid point data from the large
database by making a smaller version (shadow structure) of the grid
point file, GEOM1.  A small version of the element file is generated for
a local GEOM2 file. All other matrix generation data such as element
properties, material properties, etc. are extracted from the large
database and used without making smaller local copies.  The intermediate
files that are needed to support dynamic load generation and results
output are generated, then processing jumps to the area where modes are
calculated.  Rather than calculating modes again, the mode shapes and
eigenvalues for the shadow structure are extracted from the
corresponding matrices and tables in the large db, and reduced to a
small local set.  Dynamic load generation through results processing is
conventional.  If a limited number of points and elements are selected
for output, run times are small, even for a large model.  The NASTRAN
run typically completes in less than 10 seconds on a modern NT machine.

5. LIMITATIONS

There are many limitations on what data can be input on this run,
because by intent this type of run is not allowed to change the basic
model description in the large read-only db.  In particular, any input
that would change the modes of the structure is ignored.  Parameters
that are supported or not supported are more difficult to define.  Most
input not supported is simply ignored.  Some types may cause fatal
errors.  This list of supported input types is likely to change as
development continues.  See the fstrstd.doc file for details.

6. EXAMPLE FILE

See fstrst1.dat.  It is designed to run after the run01.dat file makes
the large db.  May be run scr=yes.  The run01 output file contains
results that may be used to check the results from this run.  The
results from this run are computed independently, and not copied from
the large db.

7. RELATED FILES

Both the run0* and fstrst* alters have other files that can be added by
included files.  They are intended for use by the PATRAN version of this
option.  They are not required if you are writing your own version of
these alters, but you may inspect them for items that may be of interest
in your work.  They include *f.dat files, containing FMS statements, and
*b.dat files, containing bulk data entries.