Before performing the analysis, it is recommended that you perform a check on the model. This is done by clicking the Analysis
Analysis Check Model command. The purpose of checking the model is so that the user can confirm that the FEA model is as intended. There are no holes, no missing boundary conditions, and so on. The checking is done in the Results environment.
- The date and time when the model was last Checked and last analyzed are stored. If a change is made to the model that affects the results, such as changing the material properties or input in the Element Definition, then the Checked date is updated. When the Checked date is newer than the analyzed date, the following message is given whenever the model is displayed in the Results environment:
- The existing results do not match the current model. If the changes have no affect on the results, then it is reasonable to proceed. (Changes include loads, mesh, materials, node numbering, and element numbering.) If the model has changed, the old results should not be used.
- If the model has been changed, or if a different version of the software is being used (which could result in different node or element numbering), it is possible that the existing results will not match the new FEA model. If you proceed to use the existing results, check the validity of the results carefully. Alternatively, re-analyze the model to be certain that the results are for the current model. Changes to the model include any geometry changes which can either be deliberate (for example, deleting lines and redrawing them) or transparent (for example, transferring to PV/Designer).
- This message appears when doing a restart analysis for those analysis types that show the results while the analysis is running. This behavior is expected.
The geometry decoder checks that all the lines in the model are connected in such a way that valid elements can be formed out of those lines. If there are lines present in the model that cannot be used to form legitimate elements or other decoding problems, a warning dialog will appear. This dialog will allow the user to review the warnings in a text file.
Depending on the severity of the geometry warnings the program will proceed to the Results environment or not according to the list below:
1. Clean Model: No warnings, and model appears in the Results environment. After inspecting the model, go to back to the FEA Editor environment.
2. Minor Warnings: A warning file is produced, but the model will appear in the Results environment. Use the warnings in the text file and Results environment to determine the cause of the problem. The sources of this kind of error are:
- Some of the lines in a part do not form elements as defined in that part. Consequently, nodes are generated, but then are not used to form elements. The warning indicates that nodes were dropped, and usually you will see holes in the model.
3. Severe Warnings: A warning file is produced. No model is produced, so the Results environment does not appear. Use the text in the warning file to determine the problem, then return to the Interface to fix them. The sources of this kind of error are as follows:
- No combination of lines in a part can form elements. For example, when a part is defined to contain brick elements, but no brick elements can be formed out of the lines in that part, then the geometry translator reports an error.
- Improper material properties.
Perform Analyses on Windows Operating Systems
Once the model is set up, you can perform the analysis by clicking the Analysis Analysis Run Simulation command. The Analysis Window dialog will appear.
Options Analysis tab. Some options, such as scheduled analyses, require the Automated Analysis to be deactivated (unchecked). You can view the progress of the analysis by pressing the Details >> button. The dialog will expand to show a log file with details of the analysis. The frequency at which the log file is updated is controlled by the Monitor Rate field; use a longer Monitor Rate to make it easier to scroll through the log file before it updates. (After the analysis completes, the log file and summary file can be reviewed from the Report tab.)
Some analysis types will display the model in the Results environment during the analysis (automatic monitoring). This permits the user to monitor the results for expected behavior during the analysis. All the normal features of the Results environment can be used while the analysis is continuing. The analysis types that provide automatic monitoring include:
- Nonlinear Structural Analyses (Mechanical Event Simulation, Static Stress with Nonlinear Material Models, and MES Riks)
- Thermal (Steady-State and Transient Heat Transfer)
Load Case Options Automatic Advance command. For other analysis type, once the analysis is complete, the model will be loaded into the Results environment.
Scheduled Analyses:
Instead of performing the analysis live, you can schedule the analysis to be performed at a later time. See the page Scheduling Analyses for details.
Submitting Analysis to Other Computers and Clusters:
It is also possible to send the analysis from the current computer to a different computer for the actual solution, or use a cluster of computers to solve the analysis. This is typically done to take advantage of a more powerful system at another location.
- If sending the analysis to a single remote computer, the remote computer can be running a Windows 32 or 64 bit operating system, or a Linux 32 or 64 bit operating system. The Options Analysis Configure Remote Execution must be set up first. During the analysis, the current computer is free while the remote computer performs the analysis. After the analysis is complete, the results are copied back to the original computer, at which time the Results environment will show the results. The automatic monitoring of the results described above is not available during remote submission.
- If sending the analysis to a Linux cluster, Linux 32 or 64 bit operating systems can be used. The Options Analysis Configure Remote Execution must be set up first. During the analysis, the current computer is free while the remote cluster of computers performs the analysis. After the analysis is complete, the results are copied back to the original computer, at which time the Results environment will show the results. The automatic monitoring of the results described above is not available during remote submission.
- If using a Windows cluster, the Options Analysis Configure Windows MPI Clusters must be set up first. During the analysis, the current computer manages the distribution of the solving to the other machines in the cluster; thus, the current computer is not free. After the analysis is complete, the results are shown in the Results environment.
You can select the remote computer or cluster as follows:
- Make sure the Automated Analysis check box on the Options Analysis tab is deactivated (unchecked). Otherwise, the analysis will begin immediately on the current computer when the Run Simulation command is given.
- From the current computer, use the Analysis Analysis Run Simulation command. This will open the analysis dialog.
- Select the name of the remote computer or cluster in the Target Computer drop-down box.
- If the target computer is a cluster, indicate the Number of Nodes to use in the analysis.
- Click the Analyze button. If the message An error occurred while copying the Autodesk Simulation model to the remote system. occurs, double check that the shared directory on the remote computer is accessible from Windows.
-
- On the Solution tab of the Analysis Parameters, set the Type of Solver to Sparse.
- In the Type of sparse solver section, choose the MUMPS sparse solver.
If the target computer is a cluster (distributed memory processing), then make sure the following options are set up in the model:
Perform Analyses on Linux Operating Systems
Otherwise, to run the analysis manually on a Linux machine, the following procedure needs to be followed: (The same manual procedure can be used on Windows, without needing to copy the files if the analysis is performed locally.)
- Perform the Analysis Analysis Check Model command on the model in Autodesk Simulation.
- Copy all the model files into a directory on the Linux machine. It includes the directory named \modelname.mod and all the files it contains.
- Run the appropriate processor from the list below using the appropriate runtime command.
- Linear static stress: ssap0 filename -run
- If the model includes gap elements or linear contact, replace ssap0 with ssap0g
- If the model includes composite elements, replace ssap0 with ssap0c
- If the model includes both composite elements and gap elements or linear contact, or if using distributed computing on a Linux system, replace ssap0 with ssap0gc
- Natural frequency (modal): ssap1 filename -run
- If the model includes composite elements, replace ssap1 with ssap1
- Transient stress (modal superposition): ssap2 filename -run
- Response spectrum (modal superposition): ssap3 filename -run
- Transient stress (direct integration): ssap4 filename -run
- If the model includes composite elements, replace ssap4 with ssap4c
- Frequency response (modal superposition): ssap5 filename -run
- Critical buckling load: ssap6s filename -run
- Random vibration (modal superposition): ssap7 filename -run
- Natural frequency (modal) with load stiffening: ssap8s filename -run
- Weight and center of gravity: ssap9 filename -run
- Static stress with nonlinear materials, MES with linear materials, MES with nonlinear materials, and MES Riks Analysis: apak4 filename -run
- Steady-state heat transfer: ssap10 filename -run
- Transient heat transfer ssap11 filename -run
- Electrostatic: esap filename -run
- When running the analysis on a Linux cluster, add the runtime option -nmpi=N where N is the number of nodes to use in the analysis. For example,
- ssap0gc modelname -run -nmpi=N for linear stress analysis
- apak4 modelname -run -nmpi=N for MES/nonlinear stress analysis
- If a task-specific cluster machinefile is needed, it must be named machinefile and be placed in the current directory. The settings in this file will override the machine list managed by the multi-purpose daemon (MPD). This is only required on the master computing node. This file can be used to control which machines are to be used during the distributed processing.
Note: - Linear static stress: ssap0 filename -run
- When the analysis is completed, copy all the files back to the original computer and folder. The results can then be viewed with the Results environment.