The Hybrid Road Approach for Durability Loads Prediction 2005-01-0628
To reduce vehicle development cycles it is necessary to perform numerical durability analyses in an early development phase. Typically there is no physical prototype available at that time hence there are no measured data, either from the proving ground or from test rigs. This paper presents an alternative method to predict the required loads.
Using Multi-Body Simulation (MBS), the loads prediction process is performed for an unconstrained vehicle, which means that vehicle body position and orientation are allowed to change. Of particular interest are the time series of the loads acting at components of the front-and the rear-suspension, as well as on the body structure of the vehicle.
For the loads prediction BMW uses the so called Hybrid-Road-Approach developed by LMS. After an initial pilot project demonstrating that approach's feasibility and potential, the project presented below is the first run of that approach by BMW in their productive environment.
In this approach, loads prediction is based on measurements for a similar production or predecessor vehicle (it's the current 3-series model, referred to as vehicle A), a MBS model of this variant A and a MBS model of the vehicle under development (the next generation 3-series, referred to as vehicle B). Selected measurements were spindle forces and moments, steering angle and accelerations at several locations on the body. The loads prediction approach depends on tackling the following two issues:
Back-calculation of tire-road contact point displacements, which can be seen as an approximation of the underlying road profile. The back-calculation is performed by an unconstrained vehicle simulation of variant A combined with a virtual drive-file-iteration
Transferring back-calculated effective road profile and measured data from variant A to variant B and performing an unconstrained full vehicle simulation based on this data. The unconstrained vehicle simulation of variant B then delivers the required component and body interface loads.
As an extra bonus, this approach avoids the roll over problem of full vehicle simulations, which may occur by excitation with measured spindle forces. The key point to solve this problem is the substitution of vertical spindle force excitation. This is achieved by excitation with back-calculated vertical tire-contact-point displacements using a simple tire model. That quantity, as a function of time, represents the underlying road profile in combination with the actual driving maneuvers (hybrid road profile). This approach, including the method of back-calculation, is similar to the drive file iteration process on full vehicle test rigs and is therefore called ‘virtual drive file iteration’.