In response to the high missed detection rates and low efficiency of traditional methods in detecting tiny defects on the machining surfaces of mechanical parts, this study proposes an efficient defect detection method based on deep learning. Initially, referencing the network architectures of Resnet and Yolo, an image detection network was designed featuring a shared encoder, a classification decoder, and a localization decoder. The shared encoder is used to extract a unified feature representation; the classification decoder accomplishes efficient data classification; and the localization decoder achieves precise defect localization. Furthermore, upon acquiring high-resolution images of the machining surfaces with dimensional features, this study introduces a real-time sliding window method to perform segmented detection and classification of these images, transforming most of the target detection tasks into image classification problems, thereby further enhancing the efficiency and accuracy of defect detection and target localization. Practical results demonstrate that this method outperforms traditional approaches in terms of missed detection rates and detection efficiency, effectively addressing the challenge of detecting complex machining surface defects, and providing a high-precision, high-efficiency defect detection solution for the mechanical part machining field.

Conventional linear isolators struggle to combine high static load-bearing capacity with effective low-frequency vibration isolation. To address this limitation, this study proposes an inclined rhombic crab-inspired quasi-zero-stiffness (I-QZS) isolator. A generalized static model is established based on the segmented linkage of crab walking legs. A physics-constrained NSGA-II algorithm is used to optimize the key geometric parameters while preventing bistable snap-through by imposing a positive-stiffness constraint over the full stroke. A stiffness-compensation strategy bridges the gap between the ideal rigid-body model and the actual 3D-printed compliant structure. The dynamic response is represented by a cubic polynomial restoring-force model, and the resulting equations are solved using the incremental harmonic balance method with SVD (singular value decomposition)-based null-space continuation. Large-amplitude excitation experiments show that the I-QZS shifts the resonance peak to 0.77 Hz, reducing the peak frequency by 64.19% and the peak transmissibility by 67.06% relative to a linear isolator, while substantially broadening the isolation bandwidth. Bifurcation analysis further identifies stability boundaries for engineering design. These results provide an integrated theoretical and experimental framework for ultra-low-frequency passive vibration isolation.

Variable-speed operation of scroll compressors is a prevailing trend for energy saving in refrigeration systems; however, complex electromagnetic excitation induces prominent vibration and noise, yet its dominant timing, spatial distribution, and action mechanism remain unclear. An electromagnetic–structural–acoustic sequential coupling model of a scroll compressor is established and validated at three speeds (3600–6600 rpm), and a dominance identification method integrating harmonic–modal matching, variational mode decomposition, and electromagnetic correlation identification is proposed. Predicted frequencies agree well with experiments; even-order harmonics migrate linearly with speed, with harmonic–modal matching exceeding 80% at low and medium speeds. At 5400 rpm, the 24th-order harmonic (2160 Hz) coincides with mode 2 (2162 Hz), causing resonance and a threefold amplitude increase. At low and medium speeds, vibration dominance indices range between 0.68 and 0.75, while noise dominance indices decrease from 0.55 to 0.48, dropping to 0.35 and 0.28 at 6600 rpm, indicating noise source transition. Vibration at S1 through S4 shows spatial variation, and far-field noise at F1 through F4 is non-uniform. These findings clarify how electromagnetic excitation dominates the vibration and noise of scroll compressors, providing a theoretical basis for speed-segmented and zone-specific noise source identification and control.