This paper addresses the issue of the difficulty in obtaining inter-turn fault (ITF) samples in electric motors, specifically in permanent magnet-synchronous motors (PMSMs), where the number of ITF samples in the stator windings is severely lacking compared to healthy samples. To effectively identify these faults, an improved fault diagnosis method based on the combination of a cycle-generative adversarial network (GAN) and a deep autoencoder (DAE) is proposed. In this method, the Cycle GAN is used to expand the collection of fault samples for PMSMs, while the DAE enhances the capability to extract and analyze these fault samples, thus improving the accuracy of fault diagnosis. The experimental results demonstrate that Cycle GAN exhibits an excellent capability to generate ITF fault samples. The proposed method achieves a diagnostic accuracy rate of up to 98.73% for ITF problems.

Exploring ways to improve the performance of rotating bands is of great importance for enhancing the power of modern artillery. This study prepared graphene oxide-coated Nylon (GO-Nylon) and Nylon samples based on nylon rotating bands in artillery systems to investigate the feasibility of introducing GO-coated nylon rotating band materials to enhance their tribological and thermal properties. The friction behavior and thermal effects of these two surfaces were analyzed under different external loads and surface roughness conditions. The results show that the excellent thermal conductivity of GO effectively reduced temperature accumulation during friction. Under an external load of 8 N, the surface temperature of GO-Nylon decreased by 14% compared to the Nylon surface, and the coefficient of friction (COF) decreased by 21%. At the same time, a simulation model was established, and its calculation results were consistent with the experimental trends, providing a further explanation of the experimental phenomena. This research provides a basis for the application of graphene-based coatings in the defense industry and presents new ideas for the development of high-performance rotating band materials.

Visual detection of unmanned aerial vehicles (UAVs) is an essential perception component of urban low-altitude security systems. However, UAV targets in complex urban scenes appear small and exhibit weak visual features. Existing convolutional structures inadequately extract their contours and fine-grained details, and background clutter such as building textures and trees attenuates target feature responses. To address these problems, this paper proposes a lightweight detection model, SA-YOLO (Small-target Adaptive YOLO). A bilateral detail refinement convolution (BDRConv) strengthens target details while preserving primary features through dual-branch collaboration. An adaptive SimAM (ASimAM) attention mechanism modulates attention response strength via a learnable scaling factor. DySample dynamic upsampling achieves sub-pixel-level reconstruction of cross-scale features through content-adaptive sampling offsets. A weighted joint loss based on the normalized Wasserstein distance improves bounding-box regression for small targets. On the DUT-Anti-UAV and the self-constructed C-UAV datasets, SA-YOLO achieves mAP50-95 values of 52.1% and 43.2%, respectively, with only 2.41 M parameters at 171 FPS. Across five training seeds, the mAP50-95 improvement over the baseline is statistically significant on DUT-Anti-UAV, whereas on C-UAV the gain is directionally consistent but not statistically significant. Compared with the accuracy-leading detectors, SA-YOLO delivers comparable accuracy in a more compact model, achieving a favorable accuracy–model-size tradeoff at a real-time inference speed.