Tips to help you select an FEA solution
In selecting a finite element analysis (FEA) software solution, it is crucial that you consider the pre- and postprocessor, which can be critical for analysis speed and accuracy. FEA analysts rely on the pre- and postprocessor to work with an assortment of data files, provide a variety of ways to idealize the model, support one or more solvers and produce the data and reports that are needed to meet both internal and external requirements. Engineering managers rely on the pre- and postprocessor solution to reduce the risks associated with accuracy while meeting time-critical product development deadlines.
Although FEA solvers are tasked with providing accurate results quickly, the role of the pre- and postprocessor should also be considered. The pre- and postprocessor enables you to idealize a product model based on geometric information and then simulate how that model will behave under certain real-life conditions. The degree of control offered by the pre- and postprocessor is vital to increasing model quality while decreasing analysis time. The risks of poor model quality and accuracy increase as models and analyses grow in complexity.
This guide presents the critical role of the pre- and postprocessor in improving the accuracy of FEA results and making the most of engineering resources. Read this guide to find out more about these top considerations:
- User interface
- Accessing CAD data
- Understanding the results
- FE model creation and Idealization
- Automation and customization
- Solver support and solution scalability
- Overall value and support
Role of the pre- and postprocessor and the FEA
The greatest challenge for engineering managers is mitigating the risks inherent in any new product design. FEA technologies help enable significant reductions in risk, which is why they are so widely used. Industry-leading pre- and postprocessing can provide an additional order-of-magnitude gain in analyses. The gains are in accuracy and control – simplifying and cleaning up the geometry and the discretized data that goes into creating the finite element model and ensuring that the calculated results are both understandable and relevant. Figure 1 illustrates the role of the pre- and postprocessor and solver in the product design process.
The role of the preprocessor, as seen in figure 2, is to import the geometry data, correct the geometry and discretize or mesh it in order to idealize a physical design and create an FE model for analysis. Automation and customization capabilities provided by the preprocessor can help to speed up this process. Following analysis by a solver, the postprocessor imports and displays the results in a graphical format and helps the understanding of model behavior. With a good understanding of the performance of the design from the analysis, the analyst may return to the preprocessor to further refine the model if necessary, and rerun the analysis.
The importance of meshing the FE model cannot be understated for both accuracy and speed. Figure 3 illustrates a mesh convergence analysis in the preprocessor which can help determine an ideal mesh size for an analysis thereby minimizing run-time while conserving accuracy. In a report titled “Cost Saving Strategies for Engineering: Using Simulation to Make Better Decisions,” researchers at the Aberdeen Group, Boston, found that engineers at 61 percent of the report’s “Best-in-Class” companies had significant control over meshing elements and his control was a contributing factor to the companies’ success. See a link to this report in the Other resources section.
Top pre- and postprocessor considerations
These top considerations provide more information and answers to the key questions to ask your software vendor when selecting an FEA solution.
The preprocessor should provide the ability to fully control creation and adjustment of the finite element mesh. That is, meshing toolboxes should be available to help create properly sized and shaped elements in the right locations to ensure that the final model will produce accurate results efficiently.
Accessing CAD data
It may be possible to take advantage of the CAD model of the design to be analyzed, and use it to create a finite element model. To this end, CAD data can be imported for meshing and idealization. CAD neutrality is important for a standalone pre- and postprocessor, as there are many different CAD programs and data formats. The preprocessor should be able to import geometric data from all leading CAD software packages including Solid Edge software, SolidWorks, Autodesk, NX software, Pro/ENGINEER, CATIA and I-deas software. The preprocessor should also be able to import CAD geometry represented by industry-standard data formats, including Parasolid, ACIS, STEP, IGES, VDA and DXF.
FE model creation and idealization
FE model creation is a crucial part of the simulation process and impacts both analysis accuracy and efficiency. Problematic CAD geometry, such as the existence of small curves or sliver surfaces, is one of the greatest obstacles encountered in generating an FE model from a CAD model. Left untouched, these and similar geometry errors will degrade mesh quality and ultimately reduce results accuracy and solution efficiency. The preprocessor should be able to efficiently detect all such geometric irregularities and repair or remove them from the model completely, while ensuring that there is no consequential loss of other associated model data, such as boundary condition definitions.
In addition to CAD data access and import, the pre- and postprocessor should also have the ability to create and manipulate geometry and finite element entities in the absence of any geometry.
The time taken to perform an analysis is directly proportional to the model size as shown in figures 4 and 5. Model size is dependent upon the number of node points and their associated degrees of freedom in a model. Certain topologies can be idealized to significantly reduce model size without compromising accuracy. For example, thinwalled structures can be represented by fewer two-dimensional shell elements instead of many solid elements. Similarly, long slender topologies can be modeled using one-dimensional beam elements which produce a much smaller FE model without losing accuracy. The preprocessor should provide the means to idealize geometries through capabilities such as mid-surface extraction that allow thin-walled solids to be meshed with shell elements. Beam modeling tools should also be provided.
Solver support and solution scalability
Solver support and solution scalability To solve for various types of physics, such as mechanical, fluid flow, or crash analysis, a pre- and postprocessor should integrate with and support the main industry solvers including NX Nastran® software, NEi Nastran, MSC Nastran, TMG, Adina, LS-Dyna, Ansys, Abaqus and Sinda. The pre- and postprocessor should also support FE model definition and analysis control parameters in the creation of the solver input file, as well as importing results data after solution.
Each of the solvers mentioned above provide a number of different solutions. Typically, analysis types include:
- Linear statics, i.e. static loads and constraints
- Normal modes, i.e. natural frequencies of vibration
- Linear buckling
- Dynamic response to a transient or frequency based loading
- Heat transfer, both steady state and transient
- Optimization of model parameters to minimize weight
- Nonlinear to solve for effects such as large displacement, nonlinear materials and contact
- Explicit solutions for nonlinear crash analysis
- Rotor dynamics for rotating parts
- Composite materials
- Aeroelasticity to simulate effect of an airstream on a structure
- Fluid dynamics and fluid flow analysis
With some solvers much of this advanced functionality is presented in modules available for purchase in addition to a basic FE package that is usually limited to linear statics, modes and buckling solutions. The more advanced functionalities should be supported by the preprocessor. The more such capabilities are supported, the more scalable is the solution. Scalability allows you to attempt more advanced analyses as your knowledge and expertise increases. It also allows the purchased product configuration to be expanded when the need to perform more advanced analyses arises.
The user interface of the pre- and postprocessor should be easy to learn and use, promoting productivity. Adoption of popular and commonly used interface types such as Windows can help the usability of an interface. Customization is also important, such as the ability to tailor the user interface to your needs and allow commonly used tools and functions to be reached easily while deemphasizing seldom used features.
Understanding the results
Each time an FE model is solved, it can help create a vast amount of results data. The ability to process the data and quickly gain an understanding of the model behavior is important for a fast analysis turnaround. The postprocessor should therefore allow full control of results selection and include a robust and varied set of tools to manage and display results, while at the same time facilitate easy comprehension of the data. Results viewing becomes more complex with highly idealized models, so the postprocessing tools should provide the ability to easily view appropriate results quantities on shell and beam elements.
Automation and customization
More advanced and involved solutions invariably demand the ability to modify or enhance the simulation approach. Also, in setting up a model for analysis, there are often repetitive sets of commands that would be laborious to complete without some method of automation or knowledge capture. Interaction or data transfer with third party software products such as Word and Excel is also important. The customization capabilities of a pre- and postprocessor application programming interface (API) and macro programming capabilities are invaluable in managing these challenges.
Overall value and support
Separate and distinct from questions about pre- and postprocessors is an assessment of the company behind the software. This goes far beyond financials and similar statistics. For example: does the company offer a readyto- go system delivered with software, manuals, guides and security devices? Will additional purchases be required?
Keeping a pre- and postprocessor up to date and operating it effectively raises additional questions beyond the initial buying and installation concerns. What is included in maintenance and support packages? What are the hours of application engineering (AE) support by telephone? What are the terms and costs for an AE onsite? Is there a trial version with free support? Is there a useful web portal? What about the availability of bug fix releases of the software?