#FEA #MBD #SensitivityAnalysis #DesignOptimisation

We are often required to perform several sensitivity analyses before using a computational model as part of the validation process to investigate its behaviour thoroughly, and improve and quantify its accuracy.

​

This is not always necessary but it is highly recommended as a design tool to clients in cases where a significant amount of modelling parameters are completely unknown. For this purpose, the modelling inputs of interest are modified by a percentage before rerunning the model and the effect of this modification to the model's outcome is recorded. The modelling inputs might involve material properties, contact properties, boundary conditions, mesh density (mesh convergence) and minor geometry changes.

​

Due to the number of models we are developing and the increasing need for this type of analysis, we have developed an automated tool that can perform any type of sensitivity analysis automatically and back out the findings in an annotated excel spreadsheet, all running in the background until the study is completed. Then, from the data, we can provide information on each modelling parameter, potentially dependent variables and how to improve the model's accuracy while reducing its running time.

​

The latest application of this tool was for investigating the effect of different contact parameters, including friction coefficient, contact separation criteria, mesh at the area of contact and remeshing parameters, to the simulation of power transmission through a chain drive.