In this article, a novel, non-contact method is proposed to detect cage slip of rolling bearing. The proposed method can obtain the speed signal of cage and inner raceway by the introduction of weak magnetic field detection technology simultaneously. Meanwhile, to enhance the identified accuracy of cage slip, an adaptive filtering method is developed to suppress noise in weak magnetic field sensor output signal. The method adopts complementary complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to decompose the weak magnetic field sensor output signal into a set of intrinsic mode functions adaptively. The intrinsic mode functions contain noisy component of detection signal (noisy intrinsic mode functions) as well as corresponding signal component of the roller and inner raceway (signal intrinsic mode functions). The noisy- and signal intrinsic mode functions are identified by the development of reconstruction criterion. The criterion is determined based on upper boundary of rotational frequency of roller under pure rolling conditions. Then, the noisy intrinsic mode functions are forced to zero, and signal intrinsic mode functions are reconstructed to extract the speed information of roller and inner raceway. An experiment system is developed to verify the effectiveness of proposed method. The experiment result shows that the cage slip decreases with the increment in load under the given speed conditions and increases with the increment in speed under the given load conditions, as well as the cage slip is more serious under given no-load conditions. It shows that the weak magnetic field detection technology can be used to detect the cage slip of rolling bearing.

The rolling bearing failure from cage slip under fast speed variation is an important issue that has drawn a wide interest. To ensure the safe operation, cage slip needs to be detected. However, due to the limit of bearing working environment of high temperature, oil mist, and small mounting space, the conventional methods such as optical fiber, ultrasound, and eddy current detection are not always feasible. This paper presents a novel method to detect cage slip based on weak magnetic detection under the speed variation. The rolling bearing, which is in the earth magnetic field, can be magnetized weakly. By the detection of the weak magnetized bearing on the earth magnetic field disturbance, the roller speed information can be obtained. This method is not affected by the bearing work environment. Then, the signal processing technology based on the ridge detection is proposed. By the speed spectrum of inner raceway, the searched path can be re-grouped. And the searched boundary is determined by assuming the roller pure rolling. Combining the searched path and the boundary, the local neighborhood of the roller rotational instantaneous frequency (IF) is determined. By finding the magnitude maximum of the local neighborhood, the roller IF can be extracted. Finally, the time-varying spectrum of the cage slip can be evaluated under the speed variation. By the comparison with the traditional optical method, it illustrates that the cage slip goes through dramatic changes under speed variation and is almost unchanged under the stabilization stage.