Ke Lei

Mr. Ke Lei

Key Laboratory of Urban Underground Engineering of Ministry of Education,Beijing Jiaotong University

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 February 2024 in Symmetry
Verified
Journal Article
Open access
PDF file
KLKe Lei
Ke Lei
Zhongsheng TanZhongsheng Tan
Zhongsheng Tan
Xiuying WangXiuying Wang
Xiuying Wang
ZZZhenliang Zhou
Zhenliang Zhou
Deep neural networks are effectively utilized for the instance segmentation of muck images from tunnel boring machines (TBMs), providing real-time insights into the surrounding rock condition. However, the high cost of obtaining quality labeled data limits the widespread application of this method. Addressing this challenge, this study presents a semi-symmetrical, fully convolutional masked autoencoder designed for self-supervised pre-training on extensive unlabeled muck image datasets. The model features a four-tier sparse encoder for down-sampling and a two-tier sparse decoder for up-sampling, connected via a conventional convolutional neck, forming a semi-symmetrical structure. This design enhances the model’s ability to capture essential low-level features, including geometric shapes and object boundaries. Additionally, to circumvent the trivial solutions in pixel regression that the original masked autoencoder faced, Histogram of Oriented Gradients (HOG) descriptors and Laplacian features have been integrated as novel self-supervision targets. Testing shows that the proposed model can effectively discern essential features of muck images in self-supervised training. When applied to subsequent end-to-end training tasks, it enhances the model’s performance, increasing the prediction accuracy of Intersection over Union (IoU) for muck boundaries and regions by 5.9% and 2.4%, respectively, outperforming the enhancements made by the original masked autoencoder.

1 July 2022 in Symmetry
Verified
Journal Article
Open access
PDF file
XLXiaoxue Li
Xiaoxue Li
Zhongsheng TanZhongsheng Tan
Zhongsheng Tan
Xiuying WangXiuying Wang
Xiuying Wang
KLKe Lei
Ke Lei
Through the research on the effect of underpass construction using the New Tubular Roof Method on the overlying strata and railway tracks, the characteristics and pros/cons of three forms of the New Tubular Roof Method are analyzed. Based on the geological conditions and structural dimensions of the Taiyuan Yingze Street Underpass Project, by analogy to the structural sections of Shenyang Metro Xinle Ruins Station and Seoul Metro Lot 923 Station, three research cases were designed: “Taiyuan method”, “Shenyang method” and “Seoul method”. The numerical models are established, and the construction process of three cases are simulated. The results show that the three forms of the New Tubular Roof Method have different characteristics. The impact of construction on the rail is mainly reflected in the absolute settlement, while the offset of the rail and the relative displacement between the rails are small, which are not enough to pose a threat to the driving safety. “Taiyuan method” has the best control effect on the deformation of the rails in general, but there is a complex superimposed interference effect during tube jacking, and the settlement in the tube jacking stage accounts for a large proportion. “Taiyuan method” and “Shenyang method” adopt the integrated inner-connecting tube roof structure to cover the entire excavation area, which have excellent effect of isolating excavation disturbance. However, “Shenyang method” has the problem of excessive settlement during the stage of steel tube incision. The settlement caused by the construction of “Seoul method” in the tube jacking stage is relatively small, and there is no need to perform the complicated and dangerous tube incision. However, the excavation disturbance of “Seoul method” will partially escape from the side of the structure, and the excavation influence range is significantly larger than other methods.

14 January 2022 in Symmetry
Verified
Journal Article
Open access
PDF file
LLLei Li
Lei Li
KLKe Lei
Ke Lei
When a traffic tunnel passes through special strata such as soft rock with high geo-stress, expansive rock, and fault fracture zones, the traditional supporting structure is often destroyed due to complicated loads, which threatens the construction and operation safety of tunnel engineering. Concrete-filled steel tubular (CFST) structure gives full play to the respective advantages of steel and concrete and has better bearing capacity and economic benefits than traditional support structure, which has achieved good results in some underground engineering applications. In order to promote the application of CFST in the construction of traffic tunnels with complex geological conditions and improve the bearing capacity of the initial supporting structure of tunnels, the influencing factors of the bearing capacity of CFST arch were studied by numerical simulation. The main achievements are as follows: (1) The load-displacement curves of CFST members under different material parameters are basically consistent. CFST members have significant restrictions on displacement in the elastic stage and have high ultimate bearing capacity. Although the bearing capacity decreases obviously after reaching the peak, it shows good extension performance. (2) The height of the steel tube section, the thickness of the steel tube wall and the grade of the core concrete have an approximately linear positive correlation with the bearing capacity of CFST arch, but the influence of these three factors on the bearing capacity of CFST arch decreases in turn, and when the grade of core concrete increases above C50, it has no significant effect on the bearing capacity of members.

17 August 2020 in Symmetry
Verified
Journal Article
Open access
PDF file
LLLei Li
Lei Li
KLKe Lei
Ke Lei
In view of the structural form and common construction methods of traffic tunnels, the bearing performance of the closed-type CFST support designed for traffic tunnels is studied. The closed-type CFST support, which consist of a CFST girder with external shotcrete, is improved from the CFST support used in mine roadways. The reasonable cross-sectional form of closed-type CFST support is analyzed by the FEM. The closed-type CFST support is mainly composed of CFST arches, a shotcrete layer, sleeves, and blind flanges. The post-buckling analysis of the closed-type CFST circular arch members using circular-shaped, rectangular-shaped, triangular-shaped, and trapezoidal-shaped steel tubes is implemented. The result shows that the closed-type CFST support has better performance than the traditional tunnel support. The study also found that for closed-type CFST support, the triangular-shaped steel tube section has the highest bearing capacity, stiffness, and steel utilization rate, which is the preferred cross-sectional form. The bearing capacity of the circular-shaped steel tube section is acceptable. Moreover, the circular-shaped steel tubes are more convenient to obtain and process, so it is also an optional cross-sectional form. The square-shaped and trapezoidal-shaped steel tube sections have neither performance advantages nor economic efficiency, so these two forms are not recommended.