Browsing by Author "Yuan, Hao"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Hyper-Progressive Single Shot Detector (HPSSD) Algorithm for Door Panel Type-B Detection
    Yuan, Hao; Okpor, Samuel Ita; Kong, Xiao; Erukainure, Frank Efe; Wilson, Samuel Britwum (IEEE, 2023-09)
    In an automobile, the door panel type-B constitutes the interior compartment of the door, primarily composed of screws and white installations forming its structural framework. However, automated manufacturing and maintenance procedures often struggle to accurately detect these components due to their pronounced resemblance to other elements on the panel. Computer vision techniques present a viable solution to this challenge. In this paper, we propose the Hyper-Progressive Single Shot Detector (HPSSD), an object detection algorithm designed to address the aforementioned challenge. Our proposed HPSSD builds on the Single Shot Detector (SSD) algorithm and introduces several enhancements to improve its detection capabilities. The first modification involves replacing the VGG-16 backbone with a ResNet-50 module. Furthermore, we incorporated the Residual Convolutional Block Attention Mechanism (RCBAM) to boost the algorithm’s functionality. To enlarge the receptive fields of each pixel–an essential step for enhancing detection accuracy–we executed multi-dilated convolutions. In the final stage of our development process, we embedded a three-stage progressive attention mechanism (PAM). The PAM is instrumental in generating refined feature maps, which serve as the foundation for precise object detection on the door panel dataset comprising 1200 images. After running 50k iterations on the door panel dataset, the HPSSD displayed a promising mean average precision of 98.2% at a speed of 21 frames per second (FPS). Our results suggest that the HPSSD, with its ability to deliver real-time, accurate detection, is an ideal tool for improving the quality inspection of door panels in smart factories.
  • Thumbnail Image
    Optimization of Rib-To-Deck Welds for Steel Orthotropic Bridge Decks
    Yuan, Hao (Virginia Tech, 2011-11-04)
    Orthotropic steel deck has been widely used over the decades especially on long-span bridges due to its light weight and fast construction. However fatigue cracking problems on the welds have been observed in many countries. Rib-to-deck welds need special care since they are directly under wheel loads, which cause large local stress variations and stress reversals. Currently the only requirement by AASHTO bridge code is that the rib-to-deck welds need to be fabricated as one-sided partial penetration welds with minimum penetration of 80% into the rib wall thickness. However considering the thin rib plate thickness, it is very difficult to achieve this penetration without a "melt-through" or "blow-through" defect. Large cost has been caused for the repair. However recent research has found that the fatigue performance of the rib-to-deck weld is not directly related to its penetration. Other factors contribute to the fatigue performance as well. Therefore, alternative requirements which are more cost-effective and rational are desired. The objective of this research is to provide recommendations to the design and fabrication of rib-to-deck welds by investigating their fatigue performance with different weld dimensions, penetrations, and welding processes. Fatigue tests were performed to 95 full-scale single-rib deck segments in 8 specimen series fabricated with different welding processes and root gap openness. Specimens were tested under cyclic loads till failure. Three failure modes were observed on both weld toes and the weld root. Test results showed that the fatigue performance was more affected by other factors such as failure mode, R-ratio and root gap openness, rather than the weld penetration. The failure cycles were recorded for the following S-N curve analysis. Finite element analysis was performed to determine the stress state on the fatigue cracking locations. Special considerations were made for the application of hot-spot stress methodology, which post-processes the FEA results to calculate the stress values at cracking locations with the structural configuration taken into account. The hot-spot stress range values were derived and adjusted accounting for the fabrication and test error. Hot-spot S-N curves were established for each specimen series. Statistical analyses were performed to study in depth the effect of weld dimensions and test scenarios. Multiple linear regression (MLR) was performed to investigate the effects of different weld dimensions; and multi-way analysis of covariance (Multi-way ANCOVA) for the effects of specimen series, failure mode, R-ratio and weld root gap. It was found that the weld toe size was more relevant to the fatigue performance, other than the weld penetration. The failure mode and R-ratio were very influential on the fatigue performance. Recommendations to the weld geometry were proposed based on the MLR model fitting. S-N data were re-categorized based on ANCOVA results and the lower-bound S-N curve was established. AASHTO C curve was recommended for the deck design.