A Normalizing Flow-Based Bidirectional Mapping Residual Network for Unsupervised Defect Detection

Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately evaluate sample distributions,mapping normal features to the normal distribution and anomalous features outside it.Consequently,this paper proposes a Normalizing Flow-based Bidirectional Mapping Residual Network(NF-BMR).It utilizes pre-trained Convolutional Neural Networks(CNN)and normalizing flows to construct discriminative source and target domain feature spaces.Additionally,to better learn feature information in both domain spaces,we propose the Bidirectional Mapping Residual Network(BMR),which maps sample features to these two spaces for anomaly detection.The two detection spaces effectively complement each other’s deficiencies and provide a comprehensive feature evaluation from two perspectives,which leads to the improvement of detection performance.Comparative experimental results on the MVTec AD and DAGM datasets against the Bidirectional Pre-trained Feature Mapping Network(B-PFM)and other state-of-the-art methods demonstrate that the proposed approach achieves superior performance.On the MVTec AD dataset,NF-BMR achieves an average AUROC of 98.7%for all 15 categories.Especially,it achieves 100%optimal detection performance in five categories.On the DAGM dataset,the average AUROC across ten categories is 98.7%,which is very close to supervised methods.

The Korteweg-de Vires equation for the bidirectional pedestrian flow model considering the next-nearest-neighbor effect

This paper focuses on a two-dimensional bidirectional pedestrian flow model which involves the next-nearest-neighbor effect. The stability condition and the Korteweg-de Vries （KdV） equation are derived to describe the density wave of pedestrian congestion by linear stability and nonlinear analysis. Through theoretical analysis, the soliton solution is obtained.

Dynamic feature analysis in bidirectional pedestrian flows

Analysis of dynamic features of pedestrian flows is one of the most exciting topics in pedestrian dynamics. This paper focuses on the effect of homogeneity and heterogeneity in three parameters of the social force model, namely desired velocity, reaction time, and body size, on the moving dynamics of bidirectional pedestrian flows in the corridors. The speed and its deviation in free flows are investigated. Simulation results show that the homogeneous higher desired speed which is less than a critical threshold, shorter reaction time or smaller body size results in higher speed of flows. The free dynamics is more sensitive to the heterogeneity in desired speed than that in reaction time or in body size. In particular, an inner lane formation is observed in normal lanes. Furthermore, the breakdown probability and the start time of breakdown are focused on. This study reveals that the sizes of homogeneous desired speed, reaction time or body size play more important roles in affecting the breakdown than the heterogeneities in these three parameters do.

FEATURE PRESERVING IMAGE INTERPOLATION BY ADAPTIVE BIDIRECTIONAL FLOW

Most image interpolation algorithms currently used suffer visually to some extent the effects of blurred edges and jagged artifacts in the image. This letter presents an adaptive feature preserving bidirectional flow process, where an inverse diffusion is performed to enhance edges along the normal directions to the iso-phote lines (edges), while a normal diffusion is done to remove artifacts ('jaggies') along the tangent directions. In order to preserve image features such as edges, angles and textures, the nonlinear diffusion coefficients are locally adjusted according to the first order and the second order directional derivatives of the image. Experimental results on the Lena image demonstrate that our interpolation algorithm substantially improves the subjective quality of the interpolated images over conventional interpolations.

Bottleneck effects on the bidirectional crowd dynamics

The bottleneck effect on bidirectional crowd dynamics is of great theoretical and practical significance, especially for the designing of corridors in public places, such as subway stations or airports. Based on the famous social force model, this paper investigates the bottleneck effects on the free flow dynamics and breakdown phenomenon under different scenarios, in which different corridor shapes and inflow ratios are considered simultaneously. Numerical simulation finds an interesting self-organization phenomenon in the bidirectional flow, a typical characteristic of such a phenomenon is called lane formation, and the existence of which is independent of the corridor＇s shape and inflow rate. However, the pattern of the lane formed by pedestrian flow is related to the corridor＇s shape, and the free flow efficiency has close relationship with the inflow rate. Specifically, breakdown phenomenon occurs when inflows from both sides of the corridor are large enough, which mostly originates from the bottleneck and then gradually spreads to the other regions. Simulation results further indicate that the leaving efficiency becomes low as breakdown occurs, and the degree of congestion is proportional to the magnitude of inflow. The findings presented in this paper match well with some of our daily observations, hence it is possible to use them to provide us with theoretical suggestions in design of infrastructures.

A State Space Modeling of Non-Isolated Bidirectional DC-DC Converter with Active Switch

In this paper, analysis, design and implementation of non-isolated soft-switching bidirectional DC-DC converter with an active switch are described. The proposed topology gives the output voltage as twice as the input voltage and enhances the efficiency up to 94.5% and 92.9% for boost and buck mode operation by proper selection of the duty cycle. Soft switching can be achieved at both steps up and step down operating modes. Small signal analysis based on state space averaging and transfer functions have been presented in detail for the proposed converter. Finally, the feasibility of the desired converter is confirmed to mat lab simulation and investigational results.

利用NetFlow的被动服务发现方法

了解网络中布置了多少服务端和谁在扫描网络对于组织网络的安全管理员是非常重要的,基于NetFlow技术提出一个新的服务发现方法.该方法定义6个启发算法将单向的NetFlow流组装成面向连接的双向流,整理输出3种类型的流,进而提取4种类型的端点,并在大型校园网里对这6个基于流数据集的启发式算法进行测试,测试结果显示本方法可以简单有效地实现组织网络的可视化.

Traffic flow prediction based on BILSTM model and data denoising scheme

Accurate prediction of road traffic flow is a significant part in the intelligent transportation systems.Accurate prediction can alleviate traffic congestion,and reduce environmental pollution.For the management department,it can make effective use of road resources.For individuals,it can help people plan their own travel paths,avoid congestion,and save time.Owing to complex factors on the road,such as damage to the detector and disturbances from environment,the measured traffic volume can contain noise.Reducing the influence of noise on traffic flow prediction is a piece of very important work.Therefore,in this paper we propose a combination algorithm of denoising and BILSTM to effectively improve the performance of traffic flow prediction.At the same time,three denoising algorithms are compared to find the best combination mode.In this paper,the wavelet(WL) denoising scheme,the empirical mode decomposition(EMD) denoising scheme,and the ensemble empirical mode decomposition(EEMD) denoising scheme are all introduced to suppress outliers in traffic flow data.In addition,we combine the denoising schemes with bidirectional long short-term memory(BILSTM)network to predict the traffic flow.The data in this paper are cited from performance measurement system(PeMS).We choose three kinds of road data(mainline,off ramp,on ramp) to predict traffic flow.The results for mainline show that data denoising can improve prediction accuracy.Moreover,prediction accuracy of BILSTM+EEMD scheme is the highest in the three methods(BILSTM+WL,BILSTM+EMD,BILSTM+EEMD).The results for off ramp and on ramp show the same performance as the results for mainline.It is indicated that this model is suitable for different road sections and long-term prediction.

An Enhanced and Extensible Data Structure for Flow Export Engine

Flow-based measurement is a popular method for various network monitoring usages.However, many flow exporting softwares have still low performance to collect all flows.In this paper, we propose a IPFIX-based flow export engine with an enhanced and extensible data structure, called XFix, on the basis of a GPL tool,-nProbe.In the engine, we use an extensible two-dimensional hash table for flow aggregation, which is able to improve the performance of the metering process as well as support bidirectional flow.Experimental results have shown its efficiency in multi-thread processing activity.

Effects of Hub Profiles on Efficiency and Flow of Jet Fan with Single Stage Impeller

The traditional jet fans have equipped with the tandem impellers and the exclusive motors, and such designs are associated with the expensive initial cost. This serial research proposes how to simplify the fan profile for reducing the initial cost without the performance deteriorations. The paper discusses the effects of the hub profiles installed in the traditional/commercial jet fan, on the performances in the several type single-stage work. The hub with the long tail corn gives the best efficiency, and not only the stay vanes but also the cooling fins deteriorate more or less the performances, accompanying the impeller works as follows. The efficiency at the reverse rotation of the original impeller, namely at the flow condition running from the outlet to the inlet, is doubtlessly lower than one at the original rotation. The unique cascade, where the leading and the trailing edges of the blade are alternated in the tangential direction, was prepared in anticipation of improving the performances. These results advise the desirable profile of the jet fan equipped with the single-stage impeller, and the numerical simulation provides the optimum blade profile for the bidirectional flow.

一维管网与二维地表双向耦合的城市暴雨内涝模拟

为快速、准确模拟城市暴雨内涝演化过程,提出一种排水管网与上覆地表动态水力交互方法,构建了一维管网模型(SWMM)与二维水动力(LISFLOOD-FP)双向耦合的模型,解决了一维管网和二维地表的双向流量交换和时间同步难题。以上海外高桥地区为例,采用两次短历时降雨过程对耦合模型进行校准和验证,比较分析了单向、双向耦合的淹没范围与水深变化。结果表明:双向耦合模拟精度较高,在研究区具有良好的适用性;对于占比80%以上的轻度(<0.2m)积水区,单向、双向耦合的模拟积水面积比为1.21;对于中等(0.2~0.3 m)和重度(>0.3 m)积水区,单向耦合模拟结果趋于严重,该比值分别增至1.88和2.1。所构建的双向耦合模型能够揭示城区内涝积水、扩散及消退的全过程,可用于城市暴雨内涝推演,为内涝治理和灾害防御提供科学依据。

汇流角对双向流道灌水器水力性能影响模拟研究

为探明双向流道汇流角对灌水器水力性能的影响,采用数值模拟分析方法,研究6组汇流角度下双向流道灌水器水力性能及抗堵性能变化规律。结果表明:随着双向流道汇流角的增大,灌水器流量系数及流态指数均逐渐增大,当汇流角为45°、60°、90°、105°时,灌水器流态指数约为0.5,当汇流角为135°、165°时,灌水器流态指数大于0.7;灌水器双向流道存在主流区与旋涡区,随着汇流角的增大,流道内低速旋涡区范围明显减小,水流对冲消能作用减弱;流道内固体颗粒运动轨迹随流道结构变化明显。在汇流角小于90°时,颗粒更易进入旋涡区做涡团运动,且主要出现在靠近流道进水口的流道单元,灌水器发生流道堵塞的概率很大。水流汇流角度大于90°时,水流携沙能力增强,颗粒在流道内停留时间减小,流道抗堵能力增强。双向流道灌水器在汇流角为105°时,具有相对优异的水力性能和抗堵性能,灌水器流道设计时可优先考虑。

基于多尺度时空特征与软注意力机制的交通流预测方法

交通流预测在规划交通系统、优化道路资源和缓解交通拥堵等方面具有重要意义。针对交通流预测中时间周期性特征提取不充分的问题,提出一种基于多尺度时空特征和软注意力机制的交通流预测方法MSTFSA。首先,利用图交谈注意力网络(GTHAT)提取空间数据的非欧几里得结构特征,通过分配动态权重表征不同时间相邻道路交通流的影响程度;其次,利用双向增强注意力门控循环单元(Bi-EAGRU)结构提取时间数据的连续性关联特征,增强每个时刻的时间特征与上下时刻的联系;然后,基于软注意力机制融合周周期、日周期和近邻时间3个尺度下的相似交通流趋势,实现对时间周期性特征的充分提取,最后,结合高速公路数据集PeMS04和PeMS08验证MSTFSA的预测精度。实验结果表明,MSTFSA的交通流预测精度表现出良好效果,与基线模型STSGCN和ASTGCN相比,在PeMS04数据集上的预测均方根误差(RMSE)分别降低7.15%和3.8%,平均绝对误差(MAE)分别降低7.79%和3.99%。MSTFSA能较充分地提取并融合交通数据的多时间尺度时空特征,在交通流预测精度提升方面表现出一定的优势。

基于双向流固耦合分析载瘤对颈动脉血管发展及影响

目的研究带有囊形瘤体的颈动脉血管在血液周期性流动过程中的力学特性和瘤体变化,探究囊形瘤体发展的具体机制以及瘤体对血液流动的影响。方法采用双向流固耦合方法对颈动脉血管内囊形瘤体和血液之间的相互作用进行有限元数值仿真研究。分析血管的形变、关键区域的血液速度和力学特性以及瘤体对血管的影响过程。结果在瘤体与血管的交界线上,瘤体呈现较大的形变,瘤体壁面压力低。瘤体对侧血管壁和血管分叉周围的三角区域壁面压力高,易于发生拉伸或破裂。瘤体内部血液速度明显低于正常血管,形成的漩涡使瘤体内部空间未能被充分利用。瘤体在脉动周期内所受的壁面剪切力一直较小,容易导致杂质沉积形成粥样硬化斑块。结论囊形瘤体会干扰血管内血液正常流动;囊形瘤体会促进镜像瘤体产生。研究结果为囊形瘤体的治疗和预防提供理论参考。通过了解囊形瘤体的力学特性和对血管的影响,医生可以更好地制定个性化的治疗方案,提高治疗效果。

校企人才双向流动的效能、困境与实践路径

在当前高职产教融合实践中,高职院校和企业间人才的合理、有效交流有利于缓解高职院校教师队伍结构性失衡问题,有利于提升高职院校教育教学水平和人才培养质量,有利于促进企业创新发展。针对校企人才双向流动过程中存在的政策制度低效、运行机制不健全以及校企价值错位导致双向流动效果不佳等问题,从完善政策制度、深化校企合作、健全运行机制等方面探讨提升校企人才双向流动实效的实践路径,实现职业教育链和产业链的有效衔接。

考虑纵向时空特性的双向交通流组合预测研究

针对现有研究方法未充分考虑现实路网双向交通流的纵向时空特性,道路方向辅助信息特征再组织性不足的问题,提出一种考虑纵向时空特性的双向交通流组合预测模型。采用卷积深度信念网络预训练机制细粒化提取上下游纵向路段间的空间特性,利用双向长短时记忆网络提取双向交通流时序特性,通过二者组合的方式完成纵向时空特性的深层次挖掘,并在此基础上,结合改进的双向注意力机制对模型提取的特征进行方向性自组织规划,从而实现针对不同方向性类别的自适应权重分配过程。实验结果表明:所提模型的平均绝对百分比误差和均方根误差相较于未考虑上下游路段影响的模型的分别降低了3.01%、5.57%,相较于考虑上下游路段影响但未改进注意力机制的模型的分别降低了0.29%、4.87%。

基于电-热流场双向耦合的直流继电器温升特性仿真研究

高压直流继电器体积小、散热能力有限,长期工作下存在局部过热的风险,发热严重时甚至影响到接触稳定性和继电器工作寿命。针对200 A直流继电器,综合考虑导体回路通电下欧姆损耗发热,在大空间的自然热对流、热传导以及热辐射、材料电阻率随温度变化等因素,建立继电器在额定工况下电-热流场双向耦合模型,设置电阻率随温度变化函数,进行耦合计算。最后,搭建对应的样机测温平台,对仿真正确性进行验证。结果表明双向耦合结果比单向耦合结果高约5.1%,与实验测量值误差小于6.2%,验证了电热流场双向耦合的有效性。

基于幂级数系数分向传递的递归型电气热多能流算法

多能流计算是综合能源系统(integrated energy system,IES)分析和优化的基础。在各能源网络运营主体交互信息有限的背景下,为提升多能流计算的收敛性和计算效率,提出一种基于幂级数系数分向传递的递归型电-气-热多能流算法。首先,分别建立电、气、热能源系统的全纯嵌入能流模型;其次,分析全纯嵌入法在多能流计算中的递归求解原理,将幂级数系数的传递划分为横向传递与纵向传递两类,建立可分向传递求解的电-气-热IES全纯嵌入多能流模型;接着,推导各全纯嵌入状态量的幂级数系数递归关系,通过幂级数系数的分向传递,递归求取状态量的幂级数系数,实现电-气-热多能流求解;最后,算例结果表明,所提方法能够以少量的交互信息实现多能流计算,具有更优的收敛性能和计算效率。

高过载软回收系统内弹道双向耦合特性分析

为探究高过载软回收系统中弹前压力对内弹道的影响,在20000 g过载指标要求下,建立了弹前具有高压缓冲气体作用下的双向耦合内弹道仿真模型,采用FLUENT软件计算弹前压力,并利用UDF将弹前压力与经典内弹道进行双向耦合仿真分析。在5.51 kg装药情况下,对质量为90 kg的弹仓进行高过载仿真,结果表明:2 MPa的弹前预充气体压力会造成膛内最大压力增大2.32%;预充气体为5、8、10 MPa的情况下,弹仓的最大速度分别降低7.32%、12.36%和15.61%;弹仓质量为70 kg和80 kg时,弹仓最大速度分别增大9.63%和4.56%;气体介质种类不同对弹前压力造成影响较小,选择空气作为缓冲介质最优。以弹前压力与内弹道之间的双向耦合模型为基础,分析预充气压、弹仓质量、预充气体介质对内弹道特性的影响,为空气阻尼软回收试验的内弹道研究提供了理论参考。

双向无线电能传输技术的研究现状

传统的无线电能传输技术主要面向单向能量传输,随着无线电能传输技术应用领域的拓展,迫切需要双向无线电能传输(bidirectional wireless power transfer,BWPT)技术以实现无线充电设备间的能量交互。首先简述BWPT系统的基本工作原理,主要从BWPT系统的典型双向变换拓扑、谐振网络、同步控制技术、功率控制策略、软开关运行及其应用场景等方面论述其研究成果,分析电容式双向无线电能传输系统的发展现状和该技术亟待解决的关键问题,最后对BWPT系统未来值得关注的研究方向进行展望。