Abstract
As a key part of numerical analysis, the modeling process has a tremendous influence on the quality of the results. While there is general awareness concerning uncertainties that arise during modeling, their quantity and sensitivity rarely are known. Hence, modeling quickly can become inaccurate and inefficient. The scope of the present paper is to innovate predictive modeling processes concerning the dynamics of real complex structures by means of linear modal analysis with the finite element method (FEM). The aim is to offer a transparent design catalog relating specific uncertainties to each model component in order to achieve error prevention for engineers dealing with comparable systems. A complex system is simplified and investigated for different levels of detail. Only after the model uncertainties for one level of detail are obtained, the next level of complexity is approached. For each level of detail, the vibrating monolithic structures and assemblies are examined and sensitive influencing parameters are subjected to uncertainty quantification with respect to experimental results. The eigenfrequencies and mode shapes are obtained from an experimental modal analysis (EMA). Using the presented course of action for the validation process, the confidence interval of the numerical results is significantly enhanced compared to previous research.
Original language | English |
---|---|
Journal | SAE Technical Papers |
Volume | 2018-June |
Issue number | June |
DOIs | |
State | Published - 13 Jun 2018 |
Externally published | Yes |
Event | SAE 10th International Styrian Noise, Vibration and Harshness Congress: The European Automotive Noise Conference, SNVH 2018 - Graz, Austria Duration: 20 Jun 2018 → 22 Jun 2018 |
Keywords
- experimental modal analysis
- finite element method (FEM)
- finite element modeling
- interval method
- parameter identification
- uncertainty quantification