Fujian University of Technology

Tunnel engineering plays a critical role in modern infrastructure development. During its service life, tunnels are inevitably subjected to dynamic loads, such as earthquakes, Tunnel Boring Machine (TBM) excavation, and rockburst events. These forces can lead to mechanical property degradation of jointed rock mass, thereby affecting the safety of tunnel engineering. Grouting technology, a crucial method for enhancing the mechanical properties of jointed rock mass, has been widely utilized in tunnel engineering. Therefore, investigating the impact mechanisms of filled jointed rocks is of significant importance. This study examines the impact behavior of filled jointed rock specimens using a split Hopkinson pressure bar apparatus. The results indicate that the dynamic mechanical properties of filled jointed rocks exhibit significant strain-rate dependence. With an increase in the joint surface roughness coefficient of composite-filled specimens, higher dynamic compressive strength and lower failure rates are observed compared to specimens filled with portland cement materials. Specimens with higher joint surface roughness coefficients showed lower damage factors and greater stability under dynamic loading. The failure process and mechanical properties of filled jointed rock under dynamic loading were shown and analyzed specifically based on the numerical model built in this paper. At high strain rates, the relationship between the dynamic increase factor and strain rate for filled jointed rocks follows the Grady equation, which can be used to predict and analyze their dynamic compressive strength. This study provides valuable insights into the impact mechanisms of grouted jointed rocks.

Medical professionals rely on MRI scans to evaluate brain tumors, a task prone to fatigue and subjective interpretation. This study enhances the YOLOv10 model by integrating two innovative modules: AvConv module and Bidirectional Pathway, both designed to improve detection accuracy. Experiments were conducted using two Kaggle datasets (Dataset 1: 1116 images; Dataset 2: 153 images), both annotated for tumor presence. The proposed model was compared against eight mainstream object detection methods. Results demonstrate that the proposed AvConv module achieves mAP50 of 88.5% and mAP50-95 of 77.7%, which represent relative improvements of 0.3% and 3.1%, respectively, over the baseline YOLOv10 achieves mAP50 of 88.2% and mAP50-95 of 74.6%. Meanwhile, the Bi-directional detection model reaches 90.3% mAP50 and 83.5% mAP50-95, corresponding to 2.1% and 8.9% gains relative to YOLOv10. These enhancements validate the effectiveness of the introduced modules in advancing detection accuracy while maintaining competitive inference efficiency, supporting the role of customized deep learning architectures in reliable, automated brain tumor diagnosis.