x
Our website uses cookies. By using the website you agree ot its use. More information can be found in our privacy policy.

Sensitivity Analysis of non-linear models

The aim of this project which is funded by the Free State of Saxony and the European Union is the research and development of methods for sensitivity analysis of nonlinear models.

It is a joint project with TU Dresden,  where some preliminary work  is already done in this area. The research is being conducted at DYNAmore office in Dresden.
The results of the research project will be presented in the form of a software prototype, which will support an engineer to perform simulation-based design of a vehicle for crash load cases by identifying the most significant design parameters.The development process will be simplified and accelerated by this software tool, even for very complex cases. Methods of Sensitivity analysis are used e.g. in the optimization for the reduction of design parameters to significant parameters. Therefore, the design objectives of the significantly influencing parameters are used. The identification of the significant parameters is not trivial, but can be made possible by means of sensitivity analysis methods. Existing methods and algorithms, however, have significant disadvantages. They are either fundamentally not suitable for complex non-linear problems (such as the design of a vehicle crash) or their use requires tremendous amount of  computational effort.
The aim of this project is to research and develope methods for global sensitivity analysis of nonlinear models in FE-simulation and their realization in an innovative software prototype.

Its use in the product development process would have the following advantages:

  • Identification of the design parameters which have large influence on design goals.
  • Reduction of the model i.e. the problem complexity by reducing the number of parameters/degrees of freedom.
  • Acceleration of the subsequent product optimization processes based on both experience and intuition based optimization and also specially the numerical optimization.
  • Reduction in cost due to reduced computing or development time, resulting in shorter product cycles.
  • Demonstration of correlations between parameters and design goals in a sense of cause and effect, there by, Increasing the understanding of the model for the design engineer.