Abstract:The contact state between the shield tunnel rings is related to the internal force. The traditional shield tunnel structure analysis model cannot consider the contact nonlinearity under the combined action of axial force， shear force， and bending moment. Therefore， this paper proposes a shield tunnel longitudinal analysis model considering the nonlinear contact under the combined action of axial force， shear force， and bending moment. This analysis method is used to analyze the mechanical behavior of the shield tunnel crossing a creeping fault in Shantou Bay. It is found that when the longitudinal joint transmits axial force， shear force， and bending moment at the same time， its force transmission effect can be equivalent to the beam considering the reduction of axial and bending stiffness， and its reduction coefficient is related to the displacement eccentricity. The mechanical analysis of the shield tunnel under creeping load shows that the contact nonlinear longitudinal joint has a great influence on the internal force of the structure， and ignoring the influence of nonlinearity may cause more than 1.4 to 3.3 times of internal force difference. The research results in this paper can be used to analyze the longitudinal mechanical behavior of shield tunnels under complex loads such as creeping fault dislocation.

Abstract:Centrifuge shaking table tests and three-dimensional numerical simulations of dynamic interactions among soil， foundation， and structure are conducted for both a traditional pile group foundation of bridge and a novel cushioned pile raft foundation of bridge in sand. By comparing the dynamic responses of the two kinds of foundations subjected to different peaks and types of earthquake waves， the seismic isolation mechanism of the cushioned pile raft foundation is discussed， and the influences of the thickness of cushion and other factors on the seismic response of the cushioned pile raft foundation are studied based on parameter analysis. The results show that when compared with the pile group foundation， the acceleration of the superstructure of the cushioned pile raft foundation decreases， while its peak and residual displacement increases. The constraint of pile head is weakened， and sliding takes place between the raft and the interposed layer in strong earthquakes， thus reducing the inertial force transmitted to the foundation， which leads to a decrease of the maximum pile bending moment and a moving down of the position of the maximum pile bending moment. The pile bending moment increases significantly with the decrease of the thickness of cushion.

Abstract:This paper compares the response spectrum method for seismic analysis of underground structures with the simplified analysis method （such as the response displacement method and the response acceleration method） suggested by the seismic code. First， the response spectrum method， the response displacement method， and the response acceleration method are briefly introduced. Then， taking the two-dimensional cross-sectional seismic analysis of three subway stations as an example， the accuracy of the three analysis methods is compared. The results obtained from the dynamic time-history analysis is used as the reference solution. The results show that the error of the response spectrum method is 0.1% to 14%， that of the response acceleration method is 0.2% to 26%， and that of the response displacement method is 9% to 44%. The accuracy of the response spectrum method is higher than both that of the response acceleration method and the response displacement method.

Abstract:The exceedance probability of a limit state of subway station is deduced based on the novel probabilistic seismic demand model （PSDM） proposed in the present paper using the deep learning method. Principal component analysis （PCA） was used to orthogonalize IMs and reduce the dimension of IMs. The trend model to predict the seismic responses of structure was established based on the back propagation （BP） neural network， which avoids the limitation of the assumption of the traditional PSDM that the demand measure （DM） of structure has a linear relationship with the intensity measure （IM） of ground motion in the log-transformed space. The error model to describe the error between the statistics-based trend model and the physical mechanism-based numerical model was established using the probabilistic neural network， which can expand the limitation of the assumption that the residuals is normally distributed with homogeneous variance. Taking a two-story and three-span subway station in Shanghai as a case study， the fragility curves of the subway station were developed based on the proposed method. The results show that the trend model established based on deep learning well simulates the nonlinear change of the seismic response with the first principal component of IMs. The established error model accurately describes the nonhomogeneous variance of residuals of the seismic responses predicted by using the trend model.

Abstract:This paper investigated the effects of soil layers on the ground motions and seismic performance of tall pier bridges， using finite element models incorporating nonlinear p-y springs. The results show that when proper motions are selected as inputs， the seismic shear force and bending moment of tall piers could be evaluated by linear-elastic six-spring models with acceptable accuracy under earthquake records with a comparatively low intensity. However， when large input intensities are considered， this type of models will overestimate the shear force demands. Furthermore， the displacement demands of pile cap， which serves as an indicator of concrete piles， are always significantly underestimated by the six-spring models， leading to unconservative design for pile foundations.

Abstract:In view of the shortcomings of the current simplified seismic analysis model of immersed tunnel with multi-mass-spring， such as the inability to reasonably simulate the detailed structure and mechanical characteristics of tube joints， a multi-scale analysis model was proposed， which simultaneously characterizes the macroscopic response and the microscopic joint structure of the immersed tunnel. The macroscopic multi-mass-spring-beam model was used to describe the macroscopic seismic response characteristics and the dynamic interaction between structure and stratum. The microscopic refinement model was used to capture the dynamic laws of the relative deformation of joints and the stress of shear keys. Taking an immersed tunnel in Guangzhou as an application example， the corresponding multi-scale model was established. Considering the combination of calculation conditions such as ground motion input direction and environmental temperature change， the seismic response characteristics of the immersed tunnel structure， the joint deformation， the shear key stress， and the influence law of key factors were studied. The results show that with the increase of the angle between the input direction of ground motion and the axial direction of the tunnel， the shear force and bending moment of the structure increase significatly， while the axial force and joint deformation decrease. The peak axial force decreases by more than 85% when the input direction is 90°， while the maximum joint tension is only 17% of that when the input direction is 0°. The change of temperature has a significant influence on the axial force and joint deformation. The maximum opening of joint increases by about 30% due to the temperature drop， and the tensile force of the tunnel is close to 60% of the peak axial force.

Abstract:In order to investigate the influence of sliding-cable modular expansion joints on the seismic performance of a three-span concrete skew continuous beam bridge with a 30° skew angle， sliding-cable modular expansion joints are adopted to restrict the rotation of skew bridge girders， the seismic performance of the skew bridge equipped with sliding-cable modular expansion joints in pulse-like and no pulse-like ground motions is explored using analytical fragility curves， and the post-earthquake repair cost curve is proposed to evaluate the seismic performance of bridge structures. The results show that the fragility of piers and bearings is significantly reduced after cable-sliding modular expansion joints are adopted. Under the influence of velocity pulse effect， the fragility of both skew bridge bearings and piers with cable modulus expansion joints significantly increase but the structure failure mode remains the same， and the use of cable-sliding modular expansion joints effectively reduces the seismic repair cost of the bridge regardless of whether ground motions are pulse-like or not.

Abstract:Ultra-high-performance concrete （UHPC） is known for its ultra-high strength， engineered cementitious composites （ECC） possesses high tensile strain-hardening characteristic， and fiber reinforced polymer bar （FRP bar） has the advantages of high tensile strength， low density， and good corrosion resistance. It is expected that the combination of high-performance matrix and high-performance reinforcement can solve the durability issue long-existed in civil engineering and maintain high strength simultaneously. In this paper， three kinds of matrix materials （normal concrete， ECC， and UHPC） and two kinds of reinforcements （steel and BFRP bar） were selected for fabricating six different composite systems. Tensile tests were conducted to obtain the tension-strain behaviors of the composite system for comparison. The test results indicate that the tensile properties of the composite system are affected by various factors. High-performance matrix can effectively enhance the tensile strength of the system， but the sustained tensile contribution and the strain-softening behavior after peak strength of high-performance matrix may bring premature failure to the system， thus reducing the ductility eventually. It is preliminarily demonstrated that the ductility of the composite system is determined by the components with a weaker deformability. In the design of the composite structural member made of high-performance materials， it is recommended that the individual properties of materials and the interaction between materials be comprehensively considered.

Abstract:Wind tunnel tests were conducted to study the correlation of the internal pressure of opening buildings and the external pressure around the orifice， as well as the effects of turbulence intensity， wind direction and opening ratio on the internal pressures. In addition， the wind tunnel test results were compared with national and international codes， and advantages and disadvantages of these codes were also analyzed. It is found that the internal pressure of the opening building is highly synchronous with the external pressure around the orifice in the time history， but it is quite different in the power spectrum at oblique wind angles. The mean internal pressure responses increase with the increase in turbulence intensity and opening ratio. Both the Chinese load code and the American ASCE7-16 greatly underestimated the internal pressure coefficient， while the internal pressure coefficients proposed by AS/NZS 1170.2：2011 are in good agreement with the wind tunnel test results.

Abstract:Car-following risk status reflects the degree of risk situation in the car-following process. For the sake of real-time prediction of car-following risk status （CFRS）， this paper， by extracting three CFRS indicators： the reciprocal of time to collision， the transverse oscillation coefficient， and the velocity instability coefficient， determines the status division thresholds based on car-following behavior spectrum. Six machine learning prediction models： the precise decision tree， the lifting tree， the linear regression fitting， the support the vector machine， the K-nearest neighbor， and the ensemble tree， are used for CFRS prediction. The accuracy rate， the recall rate of four-level risk status， and the average recall rate are selected to evaluate the prediction effects of the model. The prediction effects in different duration of prediction and different duration of feature index sequence are compared. Based on the driving behavior data of six typical car-following scenarios in different road types and traffic state combinations collected by the driving simulator， the analysis indicates that the precise decision tree is the best way to predict CFRS. The prediction effect of car-following on the branch in congested traffic flow is significantly better than that in other scenarios， and there is no significant difference between the other five scenarios. By increasing the duration of the feature index sequence， the problem that the deterioration of the prediction effects due to the increase of the duration of prediction can be alleviated. The research results provide technical support for early warning and prevention of vehicle active safety.

Abstract:In order to determine the optimal layout elements combination of route guidance variable message signs （VMS） on the expressway in high-density road network areas， a PC-fixed driving simulator and an eye tracker were used to collect the behavior data of 20 participants under 4 VMS schemes and 2 normal traffic signs. An indicator variance analysis was conducted to test the differences among 6 different traffic signs including route guidance results， driving stability， driving load， and visual recognition behavior. The results suggest that there is no significant difference between VMS scheme 4 and normal traffic signs， which means that the driving safety can be ensured by this scheme. The index factor analysis method was applied to propose a parameter S that can comprehensively reflect the characteristics of various driving behavior indicators of drivers， and make the multiple quantitative evaluation of driving behavior. The results indicate that the performance of VMS scheme 4 is the optimal scheme. In summary， the layout elements combination of route guidance VMS for expressway in high-density road network areas is recommended as follows： the shape of the road network should be represented by topological diagrams， expressways should be represented by light strips connected horizontally and vertically， other roads should be represented by short horizontal lines， and the interchanges should be represented by circles. Red， yellow， and green colors should be used to respectively indicate congestion， slow traffic， and smooth traffic conditions. A white triangle should be used to identify the direction of interchange， the hub and general interchange should be respectively marked with the place name and the interchange name， and the driver’s current location should be highlighted.

Abstract:In order to reduce the accident rate and improve the traffic efficiency， the influence of meteorological， traffic， time， and bridge structure data on the traffic accident rate of the estuarine long-span bridge was studied， based on the multi-source real-time data provided by the health monitoring system of Jiashao Bridge. The time matched sampling method and random sampling method were used to extract the accident data within four years of operation， and a conditional logistic regression model was established. The results indicate that the traffic flow and humidity are significant positive factors for the probability of traffic accidents of estuarine long-span bridges while the average traffic speed is a significant negative factor for the probability of traffic accidents. During peak hours， the possibility of accidents increase significantly. According to the existing data， the deflection change of the bridge， structural vibration， and the proportion of trucks in the traffic flow have no significant influence on accident rate.

Abstract:In view of the current situation that the conventional centralized system structure cannot meet the requirement of online extension， online maintenance， and fault tolerance of the system in the operation process of ZPW-2000 track circuit monitoring system， a principle for optimization of the design of the ZPW-2000 rail circuit monitoring system with the autonomous decentralized system（ADS） was proposed. Based on the analysis of the characteristics of the ADS online technique， the detailed optimization plan of the ZPW-2000 rail circuit monitoring system was proposed. In addition， the architecture and elements of the rail circuit monitoring system based on the ADS online technique were designed in detail. The effectiveness of the optimization design method is demonstrated by a case study. The case study proved that the ADS online technique can effectively improve the maintainability， reliability， and flexibility of the ZPW-2000 rail circuit monitoring system， with the obvious advantages of online extension， online maintenance， and fault tolerance.

Abstract:Jinmeng Bay （JMB） is one of the famous bathing beaches in the Bohai Bay. Green tides have occurred in JMB since 2015， which had a serious effect on aquaculture and tourism. Based on the MIKE 21 hydrodynamic and water quality model， the temporal and spatial distribution characteristics of the main pollutants in JMB were simulated and analyzed. A principal component analysis was conducted for assessing the water quality in JMB and its adjacent waters. The results show that the tidal current in JMB is mainly reciprocating flow along the shore， which is very weak. The current direction was southwest in flood tide and northeast in ebb tide. The variation of temporal and spatial distribution of water quality in JMB is significant. In temporal terms， the water pollution is more serious in the wet season than that the normal season and the dry season. In spatial terms， the pollution is more serious in the estuary than that in the offshore area and the water quality in the Tang estuary is obviously inferior to that in other estuaries. Finally， some suggestions were proposed which would provide reference for further prevention and control of pollution in JMB.

Abstract:Due to the abundant rainfall in river network region， the river water quality shows a temporal-spatial dynamic change affected by the point pollution and precipitation-runoff nonpoint pollution discharge. In order to assess the attainment of river water quality in small watershed， a nonpoint pollution model based on digital elevation model （DEM） raster was developed. Coupled with the hydrodynamic and water quality model， the dynamic river water quality response to point pollution and nonpoint pollution discharge was established. The modeling system was calibrated and the main water quality parameters were determined in the small watershed of Wujing River catchment in Changzhou， Jiangsu province. The modeling system developed is able to perform dynamic simulation of water quality in river network on daily scale， and realize an elaborate analysis from annual average attainment to daily attainment. According to the characteristics of regional pollution sources in the site studied， different pollutant load reduction schemes were simulated and evaluated. Under the optimal conditions， throughout the whole year， the frequency of attaining Grade III of surface water quality standard at lake inflow station would increase from 23.0%， 0， 16.4% to 71.8%， 66.3%， 75.9% for COD_Mn， NH₃-N， and total phosphorus （TP） respectively.

Abstract:Sensor noise interference， uneven point cloud density， complex and diverse scenes， and the occlusion between objects pose great challenges to the research of the semantic segmentation of 3D point cloud scenes. In the view of the uneven sampling density of 3D point cloud data and limited depth of graph convolutional networks， this paper proposes a density adaptive method， which uses a multi-layer perceptron to learn a weight function， and uses kernel density estimation to learn a density function. The convolution operation is performed on the evenly sampled point cloud data. At the same time， inspired by deep learning in the image field， residual connection， hole convolution， and other structures were introduced to train deeper point cloud segmentation networks. The algorithm achieves an excellent performance on standard data sets of multiple point cloud segmentation.

Abstract:In order to overcome the inability of traditional methods to effectively evaluate user credibility in the absence of user personal information and published content， a user credibility factor graph model （UCFGM） based on comment feedback and trust relationship is proposed， which formalized the influence of trust relationship and comment feedback on user credibility evaluation into a probability model， and proposed a semi-supervised classification learning method to build the model. The effectiveness of the model is verified on the Extended Epinions dataset， and it is found that the trust relationship is more likely to have a positive impact on user credibility evaluation than comment feedback. Compared with traditional algorithms， UCFGM can improve the accuracy of user credibility evaluation by 12% to 29% in the absence of user description and comment information.