A computation strategy based on neural network for stiffness determination of deep‐groove ball bearings

Abstract

All deep‐groove ball bearings have similar features in geometry, mechanism, and structure. Stiffness of this type of bearings is related to geometry, dimensions, and operating conditions by a very complex, high‐order and coupled‐variable function. This paper has verified that the stiffness function for all deep‐groove ball bearings can be replaced by a back‐propagation neural network (BPNN) which is trained by using some (not all) samples.

Design optimization of MR‐compatible rotating anode x‐ray tubes for stable operation

“…The stiffness of a commercial sturdy ball bearing is roughly at least 1×10 7 N/m, and cost-effective off-the-shelf bearings have standard sizes. 40 The inner diameter of the currently used bearing determines the outer size of the MACOR ceramic inner sleeve due to the fixed size of the wave spring which lies on top of the inner race of the bearing, and the inner sleeve size affects the center shaft diameter. The weakness and brittleness of the ceramic material used for the sleeves also restrict the maximum size of the shaft diameter.…”