Hunan Xiangdian Test Research Institute (China)

The wide bandgap semiconductor increases the power density of switching power supplies while bringing about serious common-mode (CM) electromagnetic interference (EMI) problems. Compared to conventional passive EMI filters (PEFs), active EMI filters (AEFs) cancel the noise by sensing and injecting, which significantly reduces the volume of the filter. However, there is a challenge in simultaneously achieving high transient voltage immunity and compact design in the sensing part. Moreover, the high-frequency insertion loss is limited largely due to stray parameters and insufficient bandwidth of active components in the injecting part. To overcome these challenges, this paper presents three key contributions: First, a novel Rogowski coil CM current sensor with lead compensation is proposed. The multivariate function of mutual inductance of the coil is derived and used to reduce the size of sensor while maintaining high sensitivity and high transient voltage immunity. Second, a high-order lead compensation network (LCN) is adopted to counteract the phase lag of active components at high frequencies. Third, the degradations of the injecting part caused by stray parameters are analyzed, and a high-frequency insertion loss enhanced injecting circuit is proposed to mitigate the degradation by closed-loop variable gain control (VGC). The effectiveness of the proposed AEF has been verified on a 3kW inverter experimental platform. With a small size of 38mm*23mm, the proposed AEF achieves an insertion loss of 22.8dB-30.3dB at high frequencies (10MHz-30MHz) and a maximum of 44.7dB at 1.4MHz.

This paper sets out to observe the comprehensive clinical effect and advantages of cryopreserved autogenous skull replantation. We collected the data of 141 patients with skull defect. Among them, 111 patients underwent cryopreserved autogenous skull replantation, 68 cases were frozen without cryoprotectant and 43 cases were frozen with cryoprotectant. In addition, 30 patients received artificial skull repair. We collected several variables related to the occurrence of complications and explored their correlation, followed by clinical experiments. After treatment, all incisions of 141 patients healed in the first stage, and there was no infection or need for a second operation. During the follow-up period, bone resorption occurred in some of the 111 patients with autologous bone repair. The hospitalization cost of patients with autologous bone repair is lower than that of patients with artificial material repair. Internal repair within 60 days (the time interval between decompressive craniectomy and cranioplasty, that is the cryopreservation time of the bone flap is within 60 days) and cryopreservation using cryoprotectant can effectively reduce the absorption rating of the skull. Among them, the absorption rate is higher in diabetic patients. As for clinical experiments results, the cell activity decreases and the apoptosis rate rises with freezing time. Revascularization was found in all the cryopreserved autogenous skulls after replantation. Cryopreservation can effectively reduce the absorption of autogenous bone.

To address the challenging requirements of citrus detection in complex orchard environments, this paper proposes ACDNet (Adaptive Citrus Detection Network), a novel deep learning framework specifically designed for automated citrus harvesting. The proposed method introduces three key innovations: (1) Citrus-Adaptive Feature Extraction (CAFE) module that combines fruit-aware partial convolution with illumination-adaptive attention mechanisms to enhance feature representation with improved efficiency; (2) Dynamic Multi-Scale Sampling (DMS) operator that adaptively focuses sampling points on fruit regions while suppressing background interference through content-aware offset generation; and (3) Fruit-Shape Aware IoU (FSA-IoU) loss function that incorporates citrus morphological priors and occlusion patterns to improve localization accuracy. Extensive experiments on our newly constructed CitrusSet dataset, which comprises 2887 images capturing diverse lighting conditions, occlusion levels, and fruit overlapping scenarios, demonstrate that ACDNet achieves superior performance with mAP@0.5 of 97.5%, precision of 92.1%, and recall of 92.8%, while maintaining real-time inference at 55.6 FPS. Compared to the baseline YOLOv8n model, ACDNet achieves improvements of 1.7%, 3.4%, and 3.6% in mAP@0.5, precision, and recall, respectively, while reducing model parameters by 11% (to 2.67 M) and computational cost by 20% (to 6.5 G FLOPs), making it highly suitable for deployment in resource-constrained robotic harvesting systems. However, the current study is primarily validated on citrus fruits, and future work will focus on extending ACDNet to other spherical fruits and exploring its generalization under extreme weather conditions.

Abstract During the winter of 2023-2024, Hunan province experienced four rounds of freezing disaster weather processes (Dec 15-20 th in 2023, Jan 18-25 th , Feb 1-11 th , and Feb 20-29 th in 2024). The maximum wind power de-rated capacity due to icing in the province reached 7850 MW (on Feb 25 in 2024), accounting for 82.3% of the total installed wind power capacity in Hunan province. Using a wind turbine icing and standby withdrawal prediction model, the average prediction accuracy rates for turbine icing de-rating across the four rounds were 83.8%, 78.4%, 83.3%, and 92.2%, respectively. Further error analysis of the prediction results revealed that the average temperature predicted is higher than the actual temperature, which may be the reason for the prediction error of wind power capacity. Therefore, using the Qiujialun wind farm as a sample, which had the largest prediction error, a temperature correction algorithm for individual turbines was established. The meteorological data at the model layer closest to the actual turbine hub altitude are used to reduce the deviation between the model altitude and the actual altitude in this algorithm, which is critical for mountainous wind farms like those in Hunan province. The improved algorithm was applied during the third and fourth rounds. The average absolute deviation (AAD) of the de-rating prediction for Qiujialun decreased from 22~24.1 MW in the first two rounds to 4.7~7.5 MW in the last two rounds, indicating improved prediction accuracy. Based on this correction method, recalculations were performed for the first two rounds. The absolute deviation for the first round decreased from 22 MW to 11.9 MW, and for the second round from 24.1 MW to 13.8 MW, representing an error reduction of approximately 50%.