Rail Internal Defect Detection Method Based on Enhanced Network Structure and Module Design Using Ultrasonic Images

Improving the detection accuracy of rail internal defects and the generalization ability of detection models are not only the main problems in the field of defect detection but also the key to ensuring the safe operation of high-speed trains.For this reason,a rail internal defect detection method based on an enhanced network structure and module design using ultrasonic images is proposed in this paper.First,a data augmentation method was used to extend the existing image dataset to obtain appropriate image samples.Second,an enhanced network structure was designed to make full use of the high-level and low-level feature information in the image,which improved the accuracy of defect detection.Subsequently,to optimize the detection performance of the proposed model,the Mish activation function was used to design the block module of the feature extraction network.Finally,the pro-posed rail defect detection model was trained.The experimental results showed that the precision rate and F1score of the proposed method were as high as 98%,while the model’s recall rate reached 99%.Specifically,good detec-tion results were achieved for different types of defects,which provides a reference for the engineering application of internal defect detection.Experimental results verified the effectiveness of the proposed method.

DNEF:A New Ensemble Framework Based on Deep Network Structure

Deep neural networks have achieved tremendous success in various fields,and the structure of these networks is a key factor in their success.In this paper,we focus on the research of ensemble learning based on deep network structure and propose a new deep network ensemble framework(DNEF).Unlike other ensemble learning models,DNEF is an ensemble learning architecture of network structures,with serial iteration between the hidden layers,while base classifiers are trained in parallel within these hidden layers.Specifically,DNEF uses randomly sampled data as input and implements serial iteration based on the weighting strategy between hidden layers.In the hidden layers,each node represents a base classifier,and multiple nodes generate training data for the next hidden layer according to the transfer strategy.The DNEF operates based on two strategies:(1)The weighting strategy calculates the training instance weights of the nodes according to their weaknesses in the previous layer.(2)The transfer strategy adaptively selects each node’s instances with weights as transfer instances and transfer weights,which are combined with the training data of nodes as input for the next hidden layer.These two strategies improve the accuracy and generalization of DNEF.This research integrates the ensemble of all nodes as the final output of DNEF.The experimental results reveal that the DNEF framework surpasses the traditional ensemble models and functions with high accuracy and innovative deep ensemble methods.

Fabrication and abrasive wear properties of metal matrix composites reinforced with three-dimensional network structure

Reticulated polyurethane was chosen as the preceramic material for preparing the porous preform using the replication process. The immersing and sintering processes were each performed twice for fabricating a high-porosity and super-strong skeleton. The aluminum magnesium matrix composites reinforced with three-dimensional network structure were prepared using the infiltration technique by pressure assisting and vacuum driving. Light interfacial reactions have played a profitable role in most of the ceramic-metal systems. The metal matrix composites interpenetrated with the ceramic phase have a higher wear resistance than the metal matrix phase. The volume fraction of ceramic reinforcement has a significant effect on the abrasive wear, and the wear rate can be decreased with the increase of the volume fraction of reinforcement.

Optimal network structure to induce the maximal small-world effect

In this paper, the general efficiency, which is the average of the global efficiency and the local efficiency, is defined to measure the communication efficiency of a network. The increasing ratio of the general efficiency of a small-world network relative to that of the corresponding regular network is used to measure the small-world effect quantitatively. The more considerable the small-world effect, the higher the general efficiency of a network with a certain cost is. It is shown that the small-world effect increases monotonically with the increase of the vertex number. The optimal rewiring probability to induce the best small-world effect is approximately 0.02 and the optimal average connection probability decreases monotonically with the increase of the vertex number. Therefore, the optimal network structure to induce the maximal small-world effect is the structure with the large vertex number （〉 500）, the small rewiring probability （≈0.02） and the small average connection probability （〈 0.1）. Many previous research results support our results.

Evaluation and optimization analysis of high-speed rail network structure in Northeast China under the background of northeast revitalization

The construction of high-speed rail(HSR)network has promoted the social-economic ties of cities,accelerated the compression of time and space,and changed the pattern of regional development.In this paper,with the adoption of the operation frequency data of HSR from 12306 website,and based on the HSR connection strength model and social network analysis model,as well as according to the HSR connection strength,HSR network density,centrality,agglomeration subgroup,and other indicators,we analyzed the characteristics of HSR network structure in Northeast China.Results show that the number of HSR cities in Northeast China is small,cities in HSR network generally exhibit weak connectivity,and the existence of HSR network marginalizes cities such as Ulanhot,Baicheng,and Songyuan,which significantly reduce the overall network connectivity of Northeast China.The overall centrality of HSR network in Northeast China is characterized by“one axis,four edges”;specifically,the one axis is located in Harbin-Dalian transportation line and the four edges are located on both sides of the main axis of Harbin-Dalian transportation line.Eight agglomeration subgroups(four double city subgroups and four multi city subgroups)have formed in Northeast China.The core status of Shenyang in HSR network is improved significantly,and“one axis and two wings”HSR network in Liaoning Province is improved significantly.With the gradual expansion of Chaoyang-Fuxin,Dandong-Benxi,and Jilin-Yanji branch networks,the“point axis”HSR network mode in Northeast China has gradually developed and matured.In the future,it is recommended to rely on eight agglomerating subgroups to encrypt HSR network structure,create secondary node central cities,and gradually build a new pattern of opening up in Northeast China.

A Tree-Based Data Collecting Network Structure for Wireless Sensor Networks

In a sensor network with a large number of densely populated sensor nodes, a single target of interest may be detected by multiple sensor nodes simultaneously. Data collected from the sensor nodes are usually highly correlated, and hence energy saving using in-network data fusion becomes possible. A traditional data fusion scheme starts with dividing the network into clusters, followed by electing a sensor node as cluster head in each cluster. A cluster head is responsible for collecting data from all its cluster members, performing data fusion on these data and transmitting the fused data to the base station. Assuming that a sensor node is only capable of handling a single node-to-node transmission at a time and each transmission takes T time-slots, a cluster head with n cluster members will take at least nT time-slots to collect data from all its cluster members. In this paper, a tree-based network structure and its formation algorithms are proposed. Simulation results show that the proposed network structure can greatly reduce the delay in data collection.

Two Network Structure Indicators for Conventional Public Transit

The significance of network structure indicators for the planning and management of conventional public transit is widely acknowledged.In order to improve and enrich the conventional public transit assessment system,two network structure indicators are proposed.Firstly,according to the obvious defects lying in the traditional no-linear coefficient,the realistic no-linear coefficient γRNL,a modified no-linear coefficient indicator,is put forward,which takes into account the effects of barriers in a city.Secondly,to cover the gap of an indicator which can reflect the coverage homogeneity of a transit network,the length dimension LDis proposed on the basis of Fractal Theory.Finally,a case study is applied to verify the validity and practicability of the two indicators in problem diagnosis using regression analysis.The results validate that γRNLcan evaluate the detour of bus lines more reasonably than the previous no-linear coefficient because it reflects the layout of bus lines,and LDcan represent the rate of change of the network density,adding a new member to the scheme of network structure indicators for public transit.

Spatial network structure of transportation carbon emission efficiency in China and its influencing factors

Grasping the spatial correlation structure of transportation carbon emission efficiency(TCEE)and its influencing factors is significant for promoting high-quality and coordinated development of the transportation industry and the relevant region.Based on the ideal point cross-efficiency(IPCE)model,the social network analysis method was employed herein to explore the spatial correlation network structure of China’s provincial TCEE and its influencing factors.The results obtained showed the following outcomes.(1)During the study period,China’s provincial TCEE formed a complex and multithreaded network association relationship,but its network association structure was still relatively loose and presented the hierarchical gradient characteristics of dense in the east and sparse in the west.(2)The correlation of China’s TCEE formed a block segmentation based on the regional boundaries,and its factional structure was relatively obvious.The eastern region was closely connected with the central region,and generally connected with the western and northeastern regions.The central region was mainly connected with the eastern and western regions,and relatively less connected with the northeastern region.Besides,the northeastern region was weakly connected with the western region.(3)Shanghai,Beijing,Zhejiang,Guangdong,Jiangsu,Tianjin,and other developed provinces were in the core leading position in the TCEE network,which significantly impacted the spatial correlation of TCEE.However,Heilongjiang,Jilin,Xinjiang,Qinghai,and other remote provinces in the northeast and northwest were at the absolute edge of the network,which weakly impacted the spatial correlation of TCEE.(4)Provincial distance,economic development-level difference,transportation intensity difference,and transportation structure difference had significant negative impacts on the spatial correlation network of China’s provincial TCEE.In contrast,the energy-saving technology level difference had a significant positive impact on it.The regression coefficients of transportation energy structure and environmental regulation differences were positive but insignificant;their response mechanism and effects need to be improved and enhanced.

Study on Cognitive Optical Network Structure and Self-optimization with the Application of Artificial Intelligence Technology

Cognitive optical network is the intermediate to combine artificial intelligence technology with network,and also the important network technology to promote network intelligence level constantly.In the paper,it analyzes the cognitive optical network structure with the application of artificial intelligence technology by starting from the basic conditions of cognitive network and cognitive optional network on the basis of fully understanding the connotation of cognitive network and cognitive optical network,and explores its self-governance functions,so as to better realize the self-optimization and self-configuration of network.

Research on Multi-Layer Distributed HF Radio Network Structure

High frequency（HF）transmission is an important communication techniques.However,conven-tional point-to-point transmission can be easily destroyed,which limits its utilization in practice.HF networking communication has the capability against demolishment.The network structure is one of the key factors for HF networking communication.In this paper,a novel analysis method of the network connectedness based on the eigenvalue is derived,and a multi-layer distributed HF radio network structure is proposed.Both the theore-tical analysis and the computer simulation results verify that the application of the proposed network structure in the HF radio communication can improve the anti-demolishment ability of the HF network efficiently.

Learning Bayesian network structure with immune algorithm

Finding out reasonable structures from bulky data is one of the difficulties in modeling of Bayesian network （BN）, which is also necessary in promoting the application of BN. This pa- per proposes an immune algorithm based method （BN-IA） for the learning of the BN structure with the idea of vaccination. Further- more, the methods on how to extract the effective vaccines from local optimal structure and root nodes are also described in details. Finally, the simulation studies are implemented with the helicopter convertor BN model and the car start BN model. The comparison results show that the proposed vaccines and the BN-IA can learn the BN structure effectively and efficiently.

Thermoreversible Thickening and Self-assembly Behaviors of pH/Temperature Dually Responsive Microgels with Interpenetrating Polymer Network Structure

The pH /temperature dually responsive microgels of interpenetrating polymer network( IPN) structure composed of poly( N-isopropylacrylamide)( PNIPAM) network and poly( acrylic acid)( PAA) network( PNIPAM /PAA IPN microgels) were synthesized by seed emulsion polymerization. The results obtained by dynamic laser light scattering( DLLS) show that the microgels have good pH /temperature dual sensitivities. The temperature sensitive component and the pH sensitive component inside the microgels have little interference with each other. The rheological properties of the concentrated PNIPAM /PAA IPN microgel dispersions as a function of temperature at pH 4. 0 or 7. 0 were investigated by viscometer,and the results displayed that only at pH 7. 0 the dispersions presented thermoreversible thickening behavior. Then the PNIPAM /PAA fibers were prepared by self-assembly of the PNIPAM /PAA IPN microgels in the ice-crystal templates formed by unidirectional liquid nitrogen freezing method. Field emission scanning electron microscopy( FESEM) images indicate that the PNIPAM /PAA fibers are rounded,randomly orientated and interweaved.

Flow Properties of Entrained Flow Gasifier Fine Slag and Network Structure of its Molten Slag

The entrained flow gasification has been identified as the most promising gasification technology.Serious environmental pollution and waste of land resources are caused by the increasing amount of storage and production of coal gasification slag.The aim of this work is to explore the feasibility of high-temperature combustion and melting technology for treating coal gasification fine slag and determine the important parameters of system operation.The flow properties and molten slag structure characteristics of three fine slags from different entrained flow gasifiers were studied.Depending on the melting mechanism of melt-dissolution,the melting time of fine slags is short.Three fine slags all produce glassy slags,which is conducive to slag discharge.The degree of polymerization of silicate melt is proportionate to the amount of SiO_(2)in the slag.A part of Al^(3+)exist in the form of[AlO_(4)]^(5-)because of the effect of CaO and Na_(2)O,as the network former.Finally,the degree of polymerization of the three type molten slag was calculated by considering the role of Si and Al in molten slag and the property of each one.

Influence of Al_(2)O_(3)/SiO_(2) Ratio on the Structure and Properties of Na^(+)/K^(+)Ion Exchange Na_(2)O-MgO-Al_(2)O_(3)-SiO_(2) Glasses

In this work,the structure,viscosity and ion-exchange process of Na_(2)O-MgO-Al_(2)O_(3)-SiO_(2) glasses with different Al_(2)O_(3)/SiO_(2) molar ratios were investigated.The results showed that,with increasing Al_(2)O_(3)/SiO_(2) ratio,the simple structural units Q_(1) and Q_(2) transformed into highly aggregated structural units Q_(3) and Q_(4),indicating the increase of polymerization degree of glass network.Meanwhile,the coefficient of thermal expansion decreased from 9.23×10^(-6)℃^(-1) to 8.88×10^(-6)℃^(-1).The characteristic temperatures such as melting,forming,softening and glass transition temperatures increased with the increase of Al_(2)O_(3)/SiO_(2) ratio,while the glasses working temperature range became narrow.The increasing Al_(2)O_(3)/SiO_(2) ratio and prolonging ion-exchange time enhanced the surface compressive stress(CS)and depth of stress layer(DOL).However,the increase of ion exchange temperature increased the DOL and decreased the CS affected by stress relaxation.There was a good linear relationship between stress relaxation and surface compressive stress.Chemical strengthening significantly improved the hardness of glasses,which reached the maximum value of(622.1±10)MPa for sample with Al_(2)O_(3)/SiO_(2) ratio of 0.27 after heat treated at 410℃for 2 h.

Epileptic brain network mechanisms and neuroimaging techniques for the brain network

Epilepsy can be defined as a dysfunction of the brain network,and each type of epilepsy involves different brain-network changes that are implicated diffe rently in the control and propagation of interictal or ictal discharges.Gaining more detailed information on brain network alterations can help us to further understand the mechanisms of epilepsy and pave the way for brain network-based precise therapeutic approaches in clinical practice.An increasing number of advanced neuroimaging techniques and electrophysiological techniques such as diffusion tensor imaging-based fiber tra ctography,diffusion kurtosis imaging-based fiber tractography,fiber ball imagingbased tra ctography,electroencephalography,functional magnetic resonance imaging,magnetoencephalography,positron emission tomography,molecular imaging,and functional ultrasound imaging have been extensively used to delineate epileptic networks.In this review,we summarize the relevant neuroimaging and neuroelectrophysiological techniques for assessing structural and functional brain networks in patients with epilepsy,and extensively analyze the imaging mechanisms,advantages,limitations,and clinical application ranges of each technique.A greater focus on emerging advanced technologies,new data analysis software,a combination of multiple techniques,and the construction of personalized virtual epilepsy models can provide a theoretical basis to better understand the brain network mechanisms of epilepsy and make surgical decisions.

A new item-based deep network structure using a restricted Boltzmann machine for collaborative filtering

The collaborative filtering(CF) technique has been widely used recently in recommendation systems. It needs historical data to give predictions. However, the data sparsity problem still exists. We propose a new item-based restricted Boltzmann machine(RBM) approach for CF and use the deep multilayer RBM network structure, which alleviates the data sparsity problem and has excellent ability to extract features. Each item is treated as a single RBM, and different items share the same weights and biases. The parameters are learned layer by layer in the deep network. The batch gradient descent algorithm with minibatch is used to increase the convergence speed. The new feature vector discovered by the multilayer RBM network structure is very effective in predicting a rating and achieves a better result. Experimental results on the data set of MovieL ens show that the item-based multilayer RBM approach achieves the best performance, with a mean absolute error of 0.6424 and a root-mean-square error of 0.7843.

Network structure, portfolio diversification and systemic risk

We investigate the effect of portfolio diversification on banking systemic risk,where the network effect is incorporated.We analyze three kinds of interbank networks,namely,random networks,small-world networks and scale-free networks.We show that the effect of portfolio diversification on banking systemic risk depends on interbank network structures and shock types.First,systemic risk increases first and then reduces with the increase of the level of portfolio diversification in the case of the individual shock.Second,in the case of the systemic shock,systemic risk reduces with the increases of the level of portfolio diversification.Third,banking systems with scale-free network structures are the most stable,and those with small-world network structures are the most vulnerable.

Effects of ZnO,FeO and Fe_(2)O_(3)on the spinel formation,microstructure and physicochemical properties of augite-based glass ceramics

Augite-based glass ceramics were synthesised using ZnO,FeO,and Fe_(2)O_(3)as additives,and the spinel formation,matrix structure,crystallisation thermodynamics,and physicochemical properties were investigated.The results showed that oxides resulted in numerous preliminary spinels in the glass matrix.FeO,ZnO,and Fe_(2)O_(3)influenced the formation of spinel,while FeO simplified the glass network.FeO and ZnO promoted bulk crystallisation of the parent glass.After adding oxides,the grains of augite phase were refined,and the relative quantities of augite crystal planes were also influenced.All samples displayed good mechanical properties and chemical stability.The 2wt%ZnO-doping sample displayed the maximum flexural strength(170.3 MPa).Chromium leaching amount values of all the samples were less than the national standard(1.5 mg/L),confirming the safety of the materials.In conclusion,an appropriate amount of zinc-containing raw material is beneficial for the preparation of augite-based glass ceramics.

Single Image Deraining Using Dual Branch Network Based on Attention Mechanism for IoT

Extracting useful details from images is essential for the Internet of Things project.However,in real life,various external environments,such as badweather conditions,will cause the occlusion of key target information and image distortion,resulting in difficulties and obstacles to the extraction of key information,affecting the judgment of the real situation in the process of the Internet of Things,and causing system decision-making errors and accidents.In this paper,we mainly solve the problem of rain on the image occlusion,remove the rain grain in the image,and get a clear image without rain.Therefore,the single image deraining algorithm is studied,and a dual-branch network structure based on the attention module and convolutional neural network(CNN)module is proposed to accomplish the task of rain removal.In order to complete the rain removal of a single image with high quality,we apply the spatial attention module,channel attention module and CNN module to the network structure,and build the network using the coder-decoder structure.In the experiment,with the structural similarity(SSIM)and the peak signal-to-noise ratio(PSNR)as evaluation indexes,the training and testing results on the rain removal dataset show that the proposed structure has a good effect on the single image deraining task.

Facile autoreduction synthesis of core-shell Bi-Bi2O3/CNT with 3-dimensional neural network structure for high-rate performance supercapacitor

Core-shell Bi-Bi2 O3/CNT(carbon nanotube) with 3-dimensional neural network structure where Bi-Bi2O3 nanospheres act as cell bodies supported by a 3-dimensional network of CNTs acting as synapses is designed and prepared by simple solvothermal method and subsequent annealing autoreduction treatment,and this structure facilitates the efficient transport of electrons.It can provide two electron transfer paths due to the double contact of Bi2O3 shell with CNT and metal Bi core which enhances the efficiency of the electrochemical reaction.The Bi-Bi2 O3/CNT electrode shows a high gravimetric capacitance of 850 F g-1(1 A g-1),and the specific capacitance of Bi-Bi2O3/CNT can be still 714 F g-1 at 30 A g-1 indicating excellent rate performance.The asymmetric supercapacitor is assembled with Bi-Bi2 O3/CNT as the negative electrode and Ni(OH)2/CNT as the positive electrode,delivering a high energy density of 36.7 Wh kg-1 and a maximum power density of 8000 W kg-1.Therefore,the core-shell Bi-Bi2O3/CNT with 3-dimensional neural network structure as the negative electrode of supercapacitor shows great potential in the field of energy storage in the future.