Method for the consideration of statistically distributed geometrical deviations in the design of cylindrical roller bearing arrangements

Abstract

Rolling bearings are of great importance in all areas of mechanical engineering due to their precise and low-friction guidance of rotating components at low cost. The design of rolling bearing arrangements is usually based on approximately ideal geometries of bearings and adjacent parts on which the bearing seats are located. However, real components inevitably show statistically distributed deviations from their nominal shapes, which are limited by tolerances. Tight tolerances result in high manufacturing costs while on the other hand their expansion leads to an increase of occurring geometrical deviations which cause changes in the bearing-internal contact and load conditions and can thus impair the operating behavior of the bearing, e.g. with regard to fatigue life, frictional torque or vibration behavior. By taking these effects into account in the design stage, potential problems in operation can be identified early during development and countermeasures can be taken, e.g. by adjusting bearing and seat tolerances. This enables an improvement of overall product quality and a reduction of costs. This paper presents a method for the integration of component deviations in the computation of the fatigue life (Lnmr) and dynamic behavior of cylindrical roller bearing arrangements—both of which are crucial design criteria—based on a combination of statistical tolerance analysis and multi-body simulations. The method is applied to an exemplary use case for illustration.