Structural Dynamics Modification (SDM) is a design-oriented tool to rapidly simulate effects of structural changes on the dynamics of structures.
FEMtools SDM uses a modal domain method to rapidly estimate the influence of many structural changes on the modal parameters. Because only a model geometry and modal parameters are used, SDM works as well on finite element data, test data and hybrid models. The structural changes can be modeled as simple springs, masses, dampers, or any type of finite element (bar, beam, shell, volume).
The benefit of using SDM is that all analysis is done in modal space. All structural modifications are transformed to equivalent changes of mass, stiffness and damping expressed in modal coordinates. The resulting set of equations can be solved at minimal computational expense and can therefore be done real-time. For example, using a slider control, FEMtools can dynamically change parameter values and the user immediately sees the effect of the change in tables and graphics.
Structural Modifications
Rapidly apply and analyze the effect of structural changes to the dynamic response of structures. These changes can be applied to a modal test or finite element model. For test engineers this offers the possibility to measure and quickly evaluate possible design changes in the field, for example, to avoid or shift some resonance frequencies. For FE engineers, using the SDM solver offers a way to rapidly evaluate many design changes before applying them on the full FE model and use the spatial solver.
Modal Sensitivity Analysis
Sensitivity coefficients are rates of changes for a perturbation of a parameter value. SDM can be used to quickly sample a number of parameter values between a extremes, and obtain information on how modal properties like resonance frequency change. The fast SDM technique could also be used to approximate higher-order sensitivity coefficients. This information can be presented, for example, in the form of tables or design curves that serve as reference information for designers.
Design of Tuned Absorbers
A tuned absorber is an added substructure that, with the right mass, stiffness and damping, is able to absorb a specific vibration. With SDM techniques it is possible to identify the optimal characteristics of the tuned absorber.
Inverse Analysis
In this application, the user specifies the design targets in terms of resonance frequencies and selects the design parameters (i.e. physical properties of modification elements). A least squares inverse analysis solves for the optimal parameter values to realize the targets.
Substructure Coupling
Analyzing the effects of coupling several substructures using modification elements is similar to SDM. The modal parameters of each substructure (obtained from modal test or FEA) together with the stiffness, mass and damping properties of the modification elements that join the substructures, are used to compute the modal parameters of the assembly.
Modal Solver
Techniques like finite element model updating (see FEMtools Model Updating) require iterative changes and re-analysis of a finite element model. Instead of iteratively solving in physical space, an SDM solution can be used to quickly prototype a parameter selection and investigate the model updating results. If accepted, the iterations can be repeated in spatial coordinates.
Finding Starting Values for Rigid Connections Converted to Springs
In model updating, it is sometime required to replace a connection that is initially estimated to behave as rigid by a spring in order to update the spring stiffness for a more physically correct value. Before updating this spring stiffness, it is required to introduce a starting value that does not significantly change the modal characteristics but at the same time is sufficiently low so that the structure become sensitive to changes of the spring stiffness value. SDM is an ideal technique to sample different values to find a suitable starting value.
Sampling Design Space
The gap between observed modal behavior and finite element analysis can be too wide to overcome by gradient-based model updating. In that case the engineer should not only investigate the physical element properties, but the FE model itself. In some cases this may lead to trying and comparing different versions of the FE model based on different assumptions regarding level of geometrical simplification, equivalent mass and stiffness, boundary conditions etc. SDM can be a fast technique to try out many modifications on a base FE model and thus sample design space with coarse structural changes before refining those models that are closest to real behavior. This is a similar approach as what is often used in design optimization.
Training Neural Nets for Damage Detection
There are other classes of problems than design optimization that require solutions to the inverse problem. That is, given changes in a structure's modal properties, what corresponding changes in its mass, stiffness, and damping properties have taken place. An application like structural damage detection requires solution of the inverse problem in a situation where the number of parameter exceeds the number of equations (underdetermined) leading to non-unique solutions. Neural networks offer promise for solving this class of problem because of their pattern recognition and interpolation capabilities. In order to solve an inverse variational problem, however, a neural network must be "trained" using a set of solutions to its corresponding forward variational problem. Training a neural network typically requires hundreds, even thousands of solution sets. Because it is so fast, SDM is a good choice to run these many cases and to train a neural network
Select the items that apply, and then let us know how to contact you.
Send FEMtools 3.0 literature Send company literature Interested in a 30 day Software Trial Version Please contact me
The following issues are of interest:
Comments
Please validate the submittal by typing "info" into the following field:
