Improving collapsibility robustness of an EPS-CD by means of simulation and six sigma

Steering columns for column assist Electric Steering are one of the main elements of the passive safety restraint system in a passenger vehicle. It is designed to collapse applying an optimized ride down force to reduce chest and head injuries of the driver during a frontal crash event, making this requirement one of the most important ones for its development. The system crash response is generated by several elements and interfaces due to the fact of many other different product requirements that a steering column must fulfill. This could lead into a complex validation process throughout the development, as it is needed to integrate the product into the vehicle in order to meet the passive safety requirements. As part of ZF lean Validation approach a new method has been developed to improve the system robustness in ZF Active safety division by means of virtual simulation and six sigma disciplines. This resultant method is also applicable for a model based development process and also contributes to improve frontloading development techniques. An analysis of variance has been done using a correlated virtual model for crash simulation, where the main parameters of the system are known (normally the case when same product generation is reused in a new vehicle). Obtaining as a result of the analysis the regression equation of the system response. This equation defines and quantifies the importance of each factor, and consequently is defining the component contribution and how to develop a column from bottom to top. Moreover, as a second step, a Monte Carlo analysis has been applied using the equation obtained generating statistical results for a number of cases (based in the design tolerances) not manageable from the hardware testing perspective neither from the finite elements methods. In that project Monte Carlo analysis has been applied for 100,000 cases generating response confident intervals instead of the classic FEM deterministic response where only the nominal case is considered.