Comparative research on different countries’ structural design codes holds great importance and can gain valuable insights: Awareness of Design Levels, Identifying Code Deficiencies and Optimizing Designs. The crack ...Comparative research on different countries’ structural design codes holds great importance and can gain valuable insights: Awareness of Design Levels, Identifying Code Deficiencies and Optimizing Designs. The crack width of concrete structure is an important design aspect of the civil design. The four highly recognized and widely used crack width theories are systematically summarized. Based on the mentioned theories and project practices, American code ACI system, Eurocode 2 1992-1 and Chinese code GB 50010 have different crack width control requirement and calculation methods. The crack width control method based on ACI system code has evolved from the Z-factor method to the steel bar spacing control method which is simple and easy to be adopted for engineering. Meanwhile, the ACI 224.1R also gives a direct crack width calculation method consistent with the steel bar spacing control method. The Eurocode 2 and GB 50010 based on the bond-slip & no-slip theory consider much more affecting factors than ACI for predicting crack width. Taking the crack width calculation of Tunnel 5 intake as an example, the crack widths of the structure are calculated according to ACI system code, Eurocode 2 and GB 50010 respectively, the results show that the crack width results in various codes are not much different. The EN 1992-1 and GB 50010 results are almost the same which are less than the ACI 224.1Rresults.展开更多
鸭蛋裂纹检测技术对于禽蛋加工工厂实现智能化蛋品检测、分级具有重要意义。该研究针对鸭蛋裂纹检测流程复杂、计算量大、模型尺寸大等问题,提出了一种基于改进YOLOv5l(you only look once version5 large)的轻量裂纹检测算法,通过在黑...鸭蛋裂纹检测技术对于禽蛋加工工厂实现智能化蛋品检测、分级具有重要意义。该研究针对鸭蛋裂纹检测流程复杂、计算量大、模型尺寸大等问题,提出了一种基于改进YOLOv5l(you only look once version5 large)的轻量裂纹检测算法,通过在黑暗条件下使用LED灯照射鸭蛋,根据裂纹蛋壳与完好蛋壳透光性不同产生的图像差异进行检测。通过在YOLOv5中引入Ghost_conv模块,大大减少了模型的浮点计算量和参数量,并在模型的骨干网络中加入ECA(efficient channel attention)注意力机制以及使用多尺度特征融合方法 BIFPN(bi-directional feature pyramid network),增加模型对有效信息的关注度,以提高算法检测精度。同时使用CIoU与α-IoU损失函数融合后替代YOLOv5原始GIoU函数加速回归预测。利用自建的鸭蛋裂纹数据集验证改进后模型的性能,结果表明,本研究提出的改进YOLOv5l网络模型检测精准率为93.8%,与原始YOLOv5l模型相比,检测精度提高了6.3个百分点,参数量和浮点计算量分别减少了30.6%、39.4%。检测帧速率为28.954帧/s,较原始YOLOv5l模型仅下降3.824帧/s。与其他的目标检测常用网络SSD(single shot multibox detector)、YOLOv4、Faster-RCNN(faster region convolutional neural networks)相比,精度分别提高了13.1、12.5、8.2个百分点。本研究提出的方法能够在低硬件资源条件下进行高精度检测,可为实际场景应用提供解决方案和技术支持。展开更多
文摘Comparative research on different countries’ structural design codes holds great importance and can gain valuable insights: Awareness of Design Levels, Identifying Code Deficiencies and Optimizing Designs. The crack width of concrete structure is an important design aspect of the civil design. The four highly recognized and widely used crack width theories are systematically summarized. Based on the mentioned theories and project practices, American code ACI system, Eurocode 2 1992-1 and Chinese code GB 50010 have different crack width control requirement and calculation methods. The crack width control method based on ACI system code has evolved from the Z-factor method to the steel bar spacing control method which is simple and easy to be adopted for engineering. Meanwhile, the ACI 224.1R also gives a direct crack width calculation method consistent with the steel bar spacing control method. The Eurocode 2 and GB 50010 based on the bond-slip & no-slip theory consider much more affecting factors than ACI for predicting crack width. Taking the crack width calculation of Tunnel 5 intake as an example, the crack widths of the structure are calculated according to ACI system code, Eurocode 2 and GB 50010 respectively, the results show that the crack width results in various codes are not much different. The EN 1992-1 and GB 50010 results are almost the same which are less than the ACI 224.1Rresults.
