Zhifei Song

Zhifei Song


Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

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

Representative Publications

View All Publications
3 May 2022 in Water
Verified
Journal Article
Open access
PDF file
KMKe Man
Ke Man
RLruilin liu
ruilin liu
Xiaoli LiuXiaoli Liu
Xiaoli Liu
ZSZhifei Song
Zhifei Song
LZLIU Zongxu
LIU Zongxu
立武立文 武
立文 武
In order to solve the problems of long artificial time consumption, the inability to standardize the degree of damage, and the difficulty of maintaining data in traditional tunnel disease detection methods, this paper proposes the use of Residual Network (ResNet) models for tunnel water leakage and crack detection. ResNet proposes a residual learning framework to ease the training of networks that are deeper than those previously used. Furthermore, ResNet explicitly reformulates the layers as learning the residual functions of the reference layer inputs, rather than learning the unreferenced functions. The ResNet model is built on the Tensorflow Deep Learning (DL) framework and transfer-learning is used to optimize the model. The ResNet-V1 can be obtained by pre-training in ImageNet. The fully connected layers of the ResNet-V1 were modified to four classifications of tunnel disease. Then, the SoftMax function is used to recognize the tunnel diseases. Four network structures have been chosen, i.e., ResNet34 and ResNet50, with and without Transfer-learning, respectively. Those models were selected for testing and training on the sample dataset, and these four network structures were compared and analyzed using five types of evaluation indicators, which are the confusion matrix, accuracy, precision, recall ratio and F1. In identifying tunnel cracks and water leakage, the accuracy of ResNet50 and ResNet34 using the transfer-learning were 96.30% and 91.29%, and the accuracy of ResNet50 was 5.01% higher than that of ResNet34; for the network structure without the transfer-learning, the accuracy of ResNet50 was 90.36% and ResNet34’s accuracy was 87.87%. These data show that the accuracy of ResNet50 is higher than that of ResNet34 with or without the transfer-learning, and the deep structure framework is superior in the identification of tunnel diseases; secondly, comparing the network structures with and without the transfer-learning, it can be found that using the Transfer-Learning can improve the ResNet network’s accuracy for tunnel disease identification. The experiments and reliability analysis demonstrate the intelligent tunnel disease identification method proposed in this paper, and its good robustness and generalization performance. This method can be used for the rapid identification of cracks and water leakage in a tunnel survey, construction and maintenance, which has practical engineering implications for tunnel disease detection.

19 December 2017 in Geotechnical and Geological Engineering
Verified
Journal Article
Closed access
PDF file
ZSZhifei Song
Zhifei Song
SYSun Yun-Jiang
Sun Yun-Jiang
LXLin Xuan
Lin Xuan

12 January 2022 in Water
Verified
Journal Article
Open access
PDF file
KMKe Man
Ke Man
Xiaoli LiuXiaoli Liu
Xiaoli Liu
ZSZhifei Song
Zhifei Song
LZLIU Zongxu
LIU Zongxu
RLruilin liu
ruilin liu
ZCZixiang Cao
Zixiang Cao
For Fangshan granite in Beijing, the static compression and dynamic compression tests have been carried out separately under natural air drying and water saturation. It was found that the dynamic compressive strength of water-saturated granite is higher than that of air-dried granite, which is contrary to the result that the strength of water-saturated rock is lower than that of air-dried granite under static load. Furthermore, under the medium strain rate condition, when the strain rate is 85 s−1, the dynamic strength of natural air-dried granite could be increased by nearly 0.5 times compared with its static state. The dynamic strength of water-saturated granite could be increased by nearly 1–2 times compared with its static strength, which shows that water-saturated granite has stronger strain rate sensitivity than natural air-dried granite. Meanwhile, under impact loading, from the perspective of water-bearing granite the Bernoulli effect of fluid, the adhesion effect of free water and the Stefan effect of fluid in water-saturated granite were revealed, and found to be the essential reasons affecting the dynamic strength of water-saturated granite. The dynamic strength in different water-bearing states in the range of medium strain rate could then be analyzed in depth, providing a certain reference value for the strength design of water-bearing rock engineering.

1 January 2021 in Advances in Civil Engineering
Verified
Journal Article
Open access
PDF file
ZSZhifei Song
Zhifei Song
KMKe Man
Ke Man
LXLiu Xiaoli
Liu Xiaoli
Based on wavelet transform, the blasting vibration signals are analyzed here. For millisecond blasting, the blasting effect is mostly affected by the actual delay time. Local characteristics of the analyzed signals could be highlighted by the wavelet transform. The simultaneous initiation of large explosive quantity could be avoided by the use of multistage detonators, while the vibration resistance effect could be better. For the same level of detonator segment, the larger the arranged time interval, the less the possibility of initiation at the same time, which is not conducive to the vibration resistance. Therefore, it is suggested to use high‐level detonators with detonating cord or high‐precision digital electronic detonators to minimize the initiation error. Furthermore, by identifying the delay time using wavelet transform, the interval delay time of different detonator segments could be obtained. Moreover, the nominal delay time, actual delay time, and interval delay time are further compared and analyzed. It is suggested that the millisecond delay series of detonators should be selected in the whole section blasting, and the segment should be jumped as much as possible, so as to increase the secondary breakage time. Detonators with longer interval delay time should be avoided to the full.

22 January 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
WGWenze Geng
Wenze Geng
ZSZhifei Song
Zhifei Song
CHCheng He
Cheng He
HWHongtao Wang
Hongtao Wang
XDXinyi Dong
Xinyi Dong
The type of soil and its compactness significantly influence its permeability coefficient, which in turn affects the drainage difficulty of soil pore water and the distribution of the infiltration line. However, current tailings dam models typically consider only a single soil layer instead of taking into account the differences in soil types and compactness, resulting in a deviation between simulated results and actual conditions. To address this issue, this study proposes three models with a gradually increasing degree of layering refinement based on soil type and compactness. These models aim to simulate the variations in the infiltration line under three different strategies: constant head, rainfall, and drainage. The simulation results indicate that the average increase in the infiltration line of the three schemes after rainfall is 46.2%, 65.88%, 83.52%, respectively; the fitting percentages for each scheme of infiltration line after 720 days of drainage and the constant head stage are 72.38%, 88.27%, and 93.61%, respectively. It can be seen that the higher the refinement level of the layered model, the more sensitive it is to changes in the infiltration line. Furthermore, as the refinement level of the layered model increases, the simulation effect on the changes in the infiltration line improves, and the simulated results become more consistent with the actual situation. This finding provides a strategy and possibility for the study of the tailings dam’s infiltration lines, safety, and stability.

10 November 2021 in Geofluids
Verified
Journal Article
Open access
PDF file
KMKe Man
Ke Man
ZLZongxu Liu
Zongxu Liu
ZSZhifei Song
Zhifei Song
Xiaoli LiuXiaoli Liu
Xiaoli Liu
Based on the sandstone from the slope of Baorixile open-pit mining area in Hulunbuir City, Inner Mongolia, the dynamic uniaxial compression test of sandstone with different freeze-thaw cycles has been carried out by Split Hopkinson Pressure Bar test (SHPB). The test results show that the crushing degree of sandstone becomes serious with the freeze-thaw cycle times and strain rate increases. The dynamic compressive strength increases with the raise of strain rate under the same freeze-thaw cycles, while it reduces with the increases of freeze-thaw cycles at the same strain rate. It is found that the 10 freeze-thaw cycles are an obvious inflection point. When it is less than 10 cycles, the dynamic compressive strength of sandstone specimens decreases rapidly, it is more than 10 cycles, and the strength decreases gradually. This is due to that the evolution progress of pores in sandstone is more uniform after a certain number of freeze-thaw cycles. Meantime, the effect of freezing and thawing is mostly restrained by the pore evolution. On the other hand, the dissipated energy required for sandstone failure grows up with the increase of the number of freeze-thaw cycles. It shows that more energy is needed for the engender of pores and fractures in sandstone caused by freeze-thaw cycle. This led to the deterioration of sandstone structural stability and the decrease of dynamic mechanical properties.