Two-Wheeler Fatigue and Random Response
This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada
Two-wheeler vehicles (motor bikes) are complex mechanical systems. The design of such systems requires sophisticated analysis and test techniques to be used in an integrated durability process. Traditionally, analysis techniques for such systems have focused on time-based methods primarily because no viable alternatives existed. However, it has been known since the 1960's that the frequency domain method for structural analysis offers superior qualitative information about structural response. The potential performance advantages of working in the frequency domain are listed below,
• The frequency domain approach requires one frequency response function (FRF) per structural configuration, compared with one for each event in the time domain.
• With careful planning, very large models can be easily handled in the frequency domain.
• The frequency domain approach provides better response information such as Power Spectral Densities (PSD's) and frequency response functions (FRFs or transfer functions) related to qualitative behaviour.
• The frequency domain approach can be easily used to establish diagnostic information such as damage per frequency, damage per mode or damage per event. This can then be used as a very useful diagnostic technique.
• Additional response statistics like displacement and acceleration response can be obtained and used for correlation purposes.
• Relative random response calculations can be used for component gap or collision detection.
• Application of the frequency domain approach is far simpler for users and offers the possibility for significant improvements in terms of system integration and design optimization.
However, computational and technological issues have held back the use of frequency domain methods for fatigue calculations. Recent technological developments have now enabled the practical implementation of the frequency domain approach through advances in the following areas:
• Extended the use from a single input PSD to multiple load inputs.
• Correlation between inputs now properly dealt with (both complex terms).
• Both a modified von Mises and/or Principal stress approach have been incorporated.
• Mixed random and deterministic loads allowed.
• Improved to allow very large solver FRF files; i.e. very large models.
• Both a Nastran modal (SOL111) and direct frequency response (SOL108) analysis allowed.
• New developments allow Stress-Life (S-N), E-N and random response only (including peak and rms elastic-plastic strain) to be calculated in one run.
• Simultaneous random and relative random response statistics possible.
• Spot weld and seam welds now supported.
• Loads cascading and surrogate loads analysis now possible.
This paper demonstrates two new applications from the list above. Firstly, a relative random response calculation is performed between two parts of a motorbike rear cowl in order to assess the possibility of collision (rattle) between parts. Secondly, a measured multi-channel loading on the front cowl of the motorbike is used to derive a simpler (surrogate) load that has a similar affect in terms of structural response (stress and fatigue), and so can be used more efficiently in a testing laboratory.
|Date ||18th June 2019|
|Order Ref||NWC_19_162 Download|
|Non-member Price || £5.00 | $6.34 | €5.59 |
Back to Search Results