Skip to content

Analysis Flow

FrontISTR analysis proceeds by preparing input files, running the solver fistr1, and checking the output files. This page summarizes the overall flow, the types of input and output files involved, and how to run serial and parallel analyses. The detailed procedure for each step is provided in the corresponding pages of the execution guide.

Overall Flow

Serial and parallel runs differ in whether domain partitioning and result merging are required. The basic flow is as follows.

  1. Obtain FrontISTR and its accompanying programs and make them ready to run. → Installation
  2. Prepare the input files (overall control data hecmw_ctrl.dat, mesh data, and analysis control data) according to the specified format.
  3. For parallel runs, split the single domain mesh into distributed domain meshes with the domain partitioner hecmw_part1. This step is not required for serial runs.
  4. Run fistr1.
  5. Check the output log, results, and visualization files. For parallel runs, merge the results if you want them collected into a single file.

For the detailed serial procedure see Serial Analysis, and for the parallel procedure see Parallel Analysis.

Input and Output Files

FrontISTR Input/Output File

Input/output files of the analysis

Input Files

FrontISTR reads the following three types of input files.

  • Overall control data (hecmw_ctrl.dat): specifies the input file names of the mesh data and analysis control data, and the output file names of the results and visualization data.
  • Mesh data: defines the finite element mesh, material data, section data, and group data. Use a single domain mesh when analyzing with a single CPU, or a distributed domain mesh split by hecmw_part1 when running in parallel on multiple CPUs.
  • Analysis control data: defines the analysis type, boundary conditions, loads, solver control, and visualization control.

All of these are text files that can be created and edited with an editor. For the format of the contents of each file, see Input Rules.

Output Files

When FrontISTR is executed, the following files are output depending on the analysis content and the specifications in the analysis control data.

  • Message and log files: output the execution progress and a summary such as the maximum/minimum values of displacement, strain, and stress, and the eigenvalue analysis results.
  • Results file: output when !WRITE,RESULT is specified in the analysis control data, containing the physical quantities per node and element. Used for post-processing and for reading into another analysis.
  • Visualization file: output when !WRITE,VISUAL is specified, in a format for visualization tools.

The naming conventions and the way to check the output files are given in Serial Analysis and Parallel Analysis, and the details of the physical quantities and formats that can be output are given in Output and Restart.

Execution

Place fistr1 (fistr1.exe on Windows) in a directory on the path, or in the directory containing the input files. Running the following in the directory containing the input files starts a serial run.

fistr1

The mesh data and analysis control data specified in hecmw_ctrl.dat are read, and the specified results and visualization files are output.

Parallel Execution

For parallel runs, prepare an MPI-enabled fistr1 in advance and input the distributed domain mesh split by hecmw_part1. Execution follows the MPI runtime environment. An example with four domains is shown below.

mpirun -np 4 fistr1

The Windows binary packages bundle an MPI runtime, so parallel execution is possible without installing MPI separately. For details see Installation with Binary Packages. The concept of domain partitioning is described in Domain Partitioning and Parallel Computation.

Until Ver.3.x, a single domain mesh had to be input only for parallel contact analysis, but from Ver.5.x a distributed domain mesh is input for parallel contact analysis just like other analyses.

See Also