Bolted connections have been widely applied in engineering structures, loosening will happen when bolted connections are subjected to continuous cyclic load, and a significant rotation between the nut and the bolt can be observed. Combining deep learning with machine vision, a bolt loosening detection method based on the fifth version of You Only Look Once (YOLOv5) is proposed, and the rotation of the nut is identified to detect the bolt loosening. Two different circular markers are added to the bolt and the nut separately, and then YOLOv5 is used to identify the circular markers, and the rotation angle of the nut against the bolt is calculated according to the center coordinate of each predicted box. A bolted connection structure is adopted to illustrate the effectiveness of the method. First, 200 images containing bolts and circular markers are collected to make the dataset, which is divided into a training set, verification set and test set. Second, YOLOv5 is used to train the model; the precision rate and recall rate are respectively 99.8% and 100%. Finally, the robustness of the proposed method in different shooting environments is verified by changing the shooting distance, shooting angle and light condition. When using this method to detect the bolt loosening angle, the minimum identifiable angle is 1°, and the maximum detection error is 5.91% when the camera is tilted 45°. The experimental results show that the proposed method can detect the loosening angle of the bolted connection with high accuracy; especially, the tiny angle of bolt loosening can be identified. Even under some difficult shooting conditions, the method still works. The early stage of bolt loosening can be detected by measuring the rotation angle of the nut against the bolt.

Signal processing is important in the balancing of the motor armature, where the balancing accuracy depends on the extraction of the signal amplitude and phase from the raw vibration signal. In this study, a motor armature dynamic balancing method based on the long short-term memory network (LSTM) and zero-phase filter (ZPF) is proposed. This method mainly focuses on the extraction accuracy of amplitude and phase from unbalanced signals of the motor armature. The ZPF is used to accurately extract the phase, while the LSTM network is trained to extract the amplitude. The proposed method combines the advantages of both methods, whereby the problems of phase shift and amplitude loss when used alone are solved, and the motor armature unbalance signal is accurately obtained. The unbalanced mass and phase are calculated using the influence coefficient method. The effectiveness of the proposed method is proven using the simulated motor armature vibration signal, and an experimental investigation is undertaken to verify the dynamic balancing method. Two amplitude evaluation metrics and three phase evaluation metrics are proposed to judge the extraction accuracy of the amplitude and phase, whereas amplitude and frequency spectrum analysis are used to judge the dynamic balancing results. The results illustrate that the proposed method has higher dynamic balancing accuracy. Moreover, it has better extraction accuracy for the amplitude and phase of unbalanced signals compared with other methods, and it has good anti-noise performance. The determination coefficient of the amplitude is 0.9999, and the average absolute error of the phase is 2.4°. The proposed method considers both fidelity and denoising, which ensuring the accuracy of armature dynamic balancing.

In order to improve the dynamic balancing accuracy of the micromotor armature, a method of V-shaped milling based on a discrete vector model for unbalance correction is proposed. The discrete vector model is fitted according to the parameters of the milling cutter and rotor, and then all the unit unbalance vectors in the discrete vector model are added to the milling center. The numerical relationship between the milling depth and the removal of the mass unbalance vector is obtained, and the accuracy of the model is verified via comparison with the data of the simulation experiments. The complexity of the integral formula of the numerical milling model makes it difficult to apply in practice. The discrete vector model does not require integration of the numerical formula and only considers the milling area as being composed of countless discrete blocks, which greatly simplifies the process of solving the unbalance vector. In view of the different thicknesses of the tooth surface of the armature, in order to avoid damage to the armature during milling, the unbalanced vector is decomposed at the center of the tooth surface by force decomposition. The experimental results show that this proposed method can effectively improve the dynamic balancing accuracy of the micromotor armature.