Abstract:Through data mining， quality impact factors were extracted from prefabricated building articles， then filtered into key factors and analyzed by the Apriori algorithm. Results indicate that big data help to reduce the interference on index selection caused by human factors， while a more universal law could be obtained by observing the correlation between impact factors and the quality of prefabricated buildings. Among all 5M1E factors， the “Method” branch gains greater support， indicating a significant impact on prefabricated buildings’ quality. Based on confidence analysis and the proposed strong association rules， it is found that “standardization of component manufacturers” receives not only a higher support degree but also the greatest frequency among the associated factors. Suggestions are put forward and shall be an effective reference for quality impact factors’ evaluation and management formulation.

Abstract:An improved embedded min-max feature selection algorithm was proposed for the nonlinear multi-label classification problem， and in combination with the support vector machine algorithm， a heuristic algorithm was proposed for the complex combinatorial optimization problem. The efficiency and accuracy of the proposed algorithm were verified after a series of experiments conducted on steel faults diagnosis dataset.

Abstract:An improved failure mode and effect analysis （FMEA） method based on fuzzy theory， grey relational analysis（GRA） and technique for order preference by similarity to ideal solution （TOPSIS） was proposed to construct risk analysis model for enterprise digital transformation. Firstly， factors of six dimension that had influence on digital transformation were proposed by literature analysis， expert interview and questionnaire survey. Next， through fuzzy theory， the qualitative evaluation index of expert group was translated into quantitative index. Then， comprehensive weight of risk factors was considered， and failure modes were sorted through GRA and TOPSIS. Finally， the model was applied to a digital transformation enterprise in the testing industry to verify its feasibility. Moreover， management measures were suggested based on the analysis results.

Abstract:A “water-energy-food-land-carbon ”Nexus（WEFLC） analysis framework was established with food production and carbon emissions as expected and undesired outputs， respectively. Typical agricultural production cases of farming， livestock breeding， and aquaculture in Zhoushan islands were selected to conduct the agricultural space assessment and low carbon planning. The order of unit output value and carbon emissions per unit area is livestock breeding， aquaculture， and farming. The total output value of livestock breeding is greatly affected by the trend of carbon trading. While maintaining the same food output， livestock breeding， aquaculture， and farming section can reduce energy， water， and land resource consumption， and therefore， low carbon emissions are realized for agricultural production by employing photovoltaics and improving the efficiency of water use as well as introducing organic waste-disposal technologies .

Abstract:The energy efficiency indicators of 970 WWTPs in the Yangtze River Economic Belt （YREB） were evaluated， including the energy intensity （EI） via normalization and the relative energy efficiency （REE） via data envelopment analysis. Meanwhile， influencing factors of designing， operational conditions， and externalities were analyzed. The results show that the EIs of 970 WWTPs change from 0.10 to 1.84 kWh·m^－3 and from 0.41 to 1.42 kWh·kg^-1 in terms of wastewater treated and COD removed， while the variation range of REE is 0.02~2.35. The impact of scale effect on energy efficiency is significant. The WWTPs configured with activated sludge-based processes tend to own better performance on energy efficiency. Besides， relatively higher pollutant concentrations in influent and less strict discharge standards would facilitate the energy efficiency. Finally， WWTPs in subregions of upstream， midstream， and downstream in YREB have different energy efficiency situations. This phenomenon is resulted from the factors such as designing factors， operational conditions， and externalities of WWTPs. This paper is expected to provide theoretical basis and technical support for energy-saving and emission-reduction of WWTPs.

Abstract:The environmental factors in the water column were investigated along the 2 transects at the Harmful Algae Blooms （HABs） coast of the Yangtze River Estuary and Hangzhou Bay during summer in 2018 and 2019. Based on two datasets of Chl a obtained via in situ sensor-monitoring and simultaneously water sample analysis， it showed that the distribution of Chl a in summer was featured with the maximum concentration not only existing at a surface depth of 2m， but down to a depth range of 10~15 m. The in situ observed Chl a concentrations showed the positively linear regression （r > 0.8，p < 0.05） with those determined from water samples through the water columns， suggesting that the observed Chl a data with high depth-resolution may be reliable to represent the vertical distribution of phytoplankton biomass. Therefore， based on that， the primary productivity of the whole blooming water column was estimated by the vertically generalized production model （VGPM） using the optimized method of the integrated Chl a concentrations with the photic depths. The optimized VGPM pattern may reflect the vertical distribution feature of Chl a and carbon fixation rate of the whole blooming waterbody， to improve our knowledge on the role of HABs coast water under the impacts of severe anthropogenic activities in the carbon neutrality and fixation in the marine carbon cycle.

Abstract:Based on the hierarchical micro-structure of nacre， this study developed a bio-hierarchically stacking method at a macro-scale. A series of beams were fabricated using ultra-high ductile concrete （UHDC）. Uniaxial tensile and compressive tests were conducted to obtain the constitutive model of UHDC. Four-point bending tests were carried out to investigate the flexural property of the stacked beams. The results indicate that the flexural strength of directly stacked beam reaches 16.75 MPa， which is 1.34 times of that of monolithically cast beam. All hierarchically stacked beams exhibit superior deformability， toughness and energy dissipation capacity， even though the deflection/span ration reaches as high as 1/50. Moreover， interlamellar sliding and cracking branching are also observed on the hierarchically stacked beams， which are similar to the failure model of nacre. Finally， a theoretical model was proposed based on the constitutive model of UHDC and verified by comparing the experimental results and theoretical results.

Abstract:The H-B strength criterion was revised from the energy point of view to improve its accuracy and applicability， which was different from the previous methods of improving rock mass characteristic parameters. Based on the fact that elastic strain energy release was the intrinsic mechanism of rock failure，researches were carried out： The general expression of existing H-B criterion was summarized and its energy mechanism in determining material failure was clarified， revealing that the H-B criterion was both an empirical strength criterion and an energy-type strength criterion expressed by elastic strain energy， which led to the key to improving the calculation accuracy of the H-B strength criterion.The H-B criterion was obtained based on the best fit of the strength data the relationship between the criterion and the theoretical expression of elastic strain energy， introducing material deformation parameters to replace Poisson's ratio in elastic strain energy to establish a modified three-dimension H-B criterion including only three parameters.The further theoretical analysis shows that the modified three-dimension H-B criterion is a set of strength criteria with deformation parameters as parameters， when the deformation parameter is equal to 0.5， it degenerates into the existing three-dimensional H-B criterion； Finally， the true triaxial test calculations show that the modified three-dimension H-B criterion can better describe the failure characteristics of the rock， the calculation results are more accurate， and the internal mechanism that produces the above results is analyzed.

Abstract:Eight groups of “Z” joint specimens were designed to conduct direct shear tests on steel keyed joints. Taking shapes，sizes，numbers of steel keys and joint type as experimental parameters， the mechanical properties of joints were studied. The results show that the stiffness of epoxied joints is superior to that of the dry joints. The epoxied joints have an ultimate shear strength of 151.39%， and 69.68% higher than dry joints， for 1 and 2 steel keys， respectively； while the failure of the epoxied joints is found to be more sudden and brittle. Adding 1 or 2 steel keys to a flat joint， shear strength can be increased 43% and 41.58% for epoxied joints， 3.81 times and 6.06 times for dried joints respectively. The failure modes of the dry joint are determined by the cracking strength of concrete and the shear strength of the steel key， while the epoxied joint appears direct shear failure. When the steel key epoxied joint occurs direct shear failure， the shear resistance is provided by the steel key and friction， and its mechanical properties are similar to the steel key dry joints.

Abstract:Based on the MT3DMS program， a numerical model which characterized the migration of groundwater contaminants was constructed. Iterative local updating ensemble smoother （ILUES） algorithm is used as the data assimilation method and Karhunen-Loeve （KL） expansion was introduced to implement the principal component analysis（PCA） dimensionality reduction of hydraulic conductivity field. The joint estimation of contaminant source parameters and hydraulic conductivity field was realized by assimilating measurement data of electrical resistivity tomography （ERT） and conventional observation data （concentration and hydraulic head）. The results show that the ILUES data assimilation method in which ERT data are used as the observation has better estimation efficiency and more accurate simulation results. As for the characterization of the site’s heterogeneous hydraulic conductivity field， the results also demonstrate the obvious advantage of using ERT data as the observation. In summary， the data assimilation framework in which environmental geophysical data such as ERT are used as the observation can effectively improve the effect of parameter estimation.

Abstract:There are significant difference between car trajectories and bus ones at urban long-distance arterials， and so the signal coordination strategy should take it into account to balance the progression needs of both vehicles. Such as to tackle with the problem of signalized intersection grouping at the long arterial with high-demand bus vehicles， this paper analyzed the temporal and spatial trajectory characteristics of cars and buses along the arterial with bus stops and presented a new signal coordination based on differential intersection group partition. The proposed model set the coordinated breaking points of car and bus trajectories at intersections and bus stops， respectively， and built a mixed integer linear programming model to minimize the weighted summation of delay and waiting time for car-users and bus passengers. In simulation tests that used the field arterial of having 13 signalized intersections from Zhongshan， compared with Multiband， the developed model can reduce the average vehicle delay of cars and buses by 34.7% and 13.7%， and the number of average stops by 28.7% and 30.0%， respectively. Further exploration with simulation experiments for sensitivity analysis on saturation， common cycle length and bus dwelling time found that the developed method can provide the synchronous progression for cars and buses on the long arterial.

Abstract:This paper presents an anomaly detection method for middle-low suspension system based on an improved canonical correlation analysis （CCA） by using the multiple types of monitoring data from the suspension system. A better threshold and a higher detection rate can be obtined by the proposed method in comparison with the thresholds by K-medoids -based method and the support vector data description （SVDD）-based method.

Abstract:A fluid-solid coupling numerical model considering the liquid and rolling contact fatigue （RCF） crack in rail based on the Coupled Euler-Lagrangian （CEL） method was established. The pressure distribution by the liquid on the crack internal surface under moving wheel-rail contact load was analyzed. The effect of the liquid with different viscosities on the RCF cracks with different lengths and propagation angles was researched. The conclusions show that the pressure by the high-viscosity liquid （HL） distributed on the entire length of the crack which was 4.8~5.9 times than that of the low-viscosity liquid （LL）. The closer to the crack tip， the greater the pressure generated by HL. And the maximum pressure on the crack tip reached about 690~751MPa in the example. With the wheel load passing the crack mouth， the HL still kept in the crack and the pressure by it on the crack surface lasted for a long time， about 80%~ 90% of a load cycle， with saddle-shaped pressure distribution on the crack surface. While the LL was extruded from the crack by the passing load which made the pressure mainly on the mouth and middle part of the crack for a short time， about 10%~30% of a load cycle， with peak-shaped pressure distribution on the crack surface. The larger the crack angle and the longer the crack length were， the more pressure on crack surface was and the easier the crack would propagate. When the crack angle was 30° and 40°， the hydraulic pressure by HL increased about 38.5%~63.8% and 1.1~1.2 times respectively compared to that of the crack angle with 20°. The maximum pressure by HL with 2.0 mm length was increased about 6.1%~8.3% than that of with 1.5 mm length.

Abstract:Based on a 1.5L hybrid turbocharged direct injection gasoline engine. The effects of different inlet structures on in-cylinder steady and transient tumble intensity and combustion performance were studied by using steady-state inlet test bench and 3D simulation tools. 4 different kinds of inlet were designed in this research. They were compared on a steady-state airway testing system with the simulation results. The in-cylinder transient tumble ratio， transient turbulent kinetic energy （TKE）， velocity distribution， air/fuel ratio and flame front of different inlet crank angle were calculated by simulation. The results showed that higher tumble ratio would improve the in-cylinder fuel air mixing and increase the average TKE at ignition phase， which affected the flame propagation velocity. Then it affected the in-cylinder combustion and performance of the GDI engine.

Abstract:15 characteristic parameters related to driving safety were selected based on the operation data of pure electric vehicles. According to the statistics of the parameters， threshold line with multiple parameters was proposed. Then two clustering algorithms and four dimension-reduction methods were compared. Based on the classification effect， the combination algorithm of t-Distribution stochastic neighbor embedding （t-SNE） and Gaussian Mixture Model （GMM） were used to establish the driving style classification model， which was used to classify the driving style into three classes. On this basis， the vehicle operation characteristics of the three classes were compared. The classification model can effectively reflect the driving behavior and provide reliable evaluation basis for fleet management and road safety.

Abstract:This paper discusses the problem of corporate bankruptcy prediction using unbalanced data by innovatively integrating data preprocessing technology and integration algorithm. Firstly， redundant information processing and different sampling methods are used to preprocess unbalanced data. Secondly， a decision tree with Classifier 5.0 （C5.0） and a single hidden layer feedforward neural network are used as the base classifier to select the optimal sampling method by combining with three kinds of resampling data preprocessing technologies. Thirdly， the self-aggregation method is combined to improve the classification performance， and the integration models of the two base classifiers are compared by the area under the receiver operating characteristic curve with 10-fold cross-validation. Finally， the actual data of more than 10 000 Polish manufacturing companies in the database of University of California Irvine are used for experimental verification. The experimental results show that the integrated model combining under-sampling or synthetic minority over-sampling method with neural network archive the best classification performance， which provides positive support for the enterprises to implement bankruptcy prediction.

Abstract:A flexible repayment method based on the borrower’s asset is considered. By introducing a predetermined asset boundary， the borrower’s asset is divided into high and low asset regions. A restricted free boundary problem model for the remaining loan value is established via different repayment terms in two regions. The explicit finite difference method and shooting method are applied to solve the free boundary and the expected clear off time of the loan. Finally， the influence of the model parameters on the expected clear off time is discussed. The results show that there is a monotonic relationship between the expected clear off time and these parameters.