ISSN:2349-2058
S.No. | Title & Authors | Page No | View | ||
1 |
Title : Evaluation of Emergency Management Capability for Railway Operation Emergencies Based on Variable Fuzzy Sets Authors : Bo Li, Yuan Liang, Delong Zou, Xiangyang Liu
Abstract :
Considering the complexity and fuzziness of railway operation emergencies, this study proposes a variable fuzzy evaluation model to assess emergency management capability. Based on the four-stage emergency management framework—prevention, preparedness, response, and recovery—an index system consisting of 17 evaluation indicators for railway emergency management capability is established, with reference to the "dual prevention mechanism" for value assignment. To ensure scientific rigor and reliability, subjective weights obtained through Principal Component Analysis (PCA) are integrated with objective weights derived from the CRITIC method, thereby forming a combined subjective–objective weighting model. Using the "10·15" Beihe Railway passenger train derailment accident in Heilongjiang Province as a case study, four different parameter combinations of the variable fuzzy evaluation method are applied to calculate the grade characteristic values of emergency management capability. The results indicate that the evaluation level of emergency management capability for railway operation emergencies is Grade III (moderate), with potential for advancement to Grade II (stronger). Recommendations for enhancing the emergency management capability of railway operation emergencies are provided, thereby verifying the scientific validity and rationality of the proposed model. |
1-7 |
![]() |
||
2 |
Title : Application of SSD Optimization Algorithm in the Detection of Diseases of Bridge Subaqueous Structures Authors : Bo Li, Delong Zou, Yuan Liang, Xiangyang Liu
Abstract :
At present, the detection methods for bridge subaqueous structures are relatively single. Nowadays, deep convolutional neural networks are increasingly applied to the field of object detection, improving the traditional detection effect. In this paper, an R-MobileNet-SSD network model with lightweight characteristics is built for the detection of diseases of bridge subaqueous structures. The Retinex algorithm is used for processing to improve the brightness and clarity of images, while retaining the natural color and details of images, and the contrast of images is enhanced through histogram equalization. Subsequently, in order to improve the detection accuracy and training quality, a deep convolutional network DCGAN is built to generate virtual samples, and the SSD model is improved and analyzed. The main VGG model in the original SSD is replaced with a MobileNet model to make it more lightweight. Through the detection and analysis in different underwater environments, this paper verifies that the R-MobileNet-SSD model can detect diseases in various underwater environments, and the detection effect is greatly improved compared with the original SSD model. |
8-13 |
![]() |