基于神经网络模型的某大坝混凝土弹性模量时变规律反演分析

基于人工神经网络的非线性映射特性,在三维有限元计算的基础上,结合大坝原型观测资料,提出了大坝参数时变规律的反演方法。以某混凝土重力坝的原型观测资料为例,利用该方法反演得出了该坝坝体混凝土弹性模量的变化规律。结果表明该方法是切实可行的,可应用于分析评价大坝材料参数的变化。

基于二次振荡波过程的交流电网相继速动判据

针对现有的交流电网相继速动保护判据的可靠性能受限于运行工况与系统参数,以及其无法耐受较高过渡电阻的难题,该文首先研究了本级线路盲区故障与下级线路故障时由对端断路器开断产生的二次振荡波过程的差异性,并基于行波模量网络分析了特定故障工况下的特殊波过程;然后,提出了基于数学形态学梯度算法的振荡波头极性辨识判据以及特殊故障工况下的辅助判据,以形成全新的相继速动保护判据;最后,基于PSCAD仿真平台验证了所提保护判据的有效性、灵敏性与可靠性,其适用于三相/单相跳闸方式、适应于所有故障类型、不受系统运行工况与系统参数影响,同时耐受过渡电阻能力高达300Ω。

环状柔直配网线路的单端量保护原理

直流线路故障的快速、可靠识别是多端柔性直流配网发展面临的技术难点之一。针对模块化多电平换流器、电压源换流器共存的环状直流配网的中压直流线路,提出利用附加电感电压的故障识别方法。首先,提出基于模量网络的故障后线路附加电感电压初始值计算方法。其次,利用线路附加电感电压初始值在区内、外故障时的差异,实现故障的快速识别;并利用故障极和非故障极上电感电压初始值的差异进行选极。该方案采用单端电气量快速、准确识别故障,无需通信,可靠性高。最后,在PSCAD/ETMDC平台搭建仿真模型,验证所提计算方法的正确性和保护方案的可行性。

广义双回线路反向量行波特性新解

通过理论分析和仿真发现:采用两回线路电流构造得到双回线路反向量电流行波,并非不受双回线路以外的系统接线情况的影响,其故障点反射波和对端母线反射波不仅与故障点过渡电阻有关系,还与两端母线的出线数有关。因此,构建基于行波传播路径的双回线路行波分析体系。在该体系下,首次推导出双回线路等长和不等长情况下反向量故障初始行波、故障点反射波以及对端母线反射波的表达式以及得到反向量故障点反射波、对端母线反射波与故障点过渡电阻以及母线端出线的数学关系,并提出基于广义双回线路反向量行波的概念。大量仿真和实测数据表明:基于行波传播路径分析反向量行波特性的正确性。

直流输电线路差动保护新原理

针对直流输电系统的快速发展亟需研究新原理的直流线路保护问题,本文提出一种基于直流电流正、反向行进波构成的直流输电线路差动保护原理。根据对直流线路两端的电流正、反向行进波间关系的分析结果,推导出由本侧电流正、反向行进波得到对侧电流正、反向行进波的计算方法,利用同一侧保护安装处电流行进波的计算值与实测值来构造差动保护判据,并给出了保护整定值的选取原则。根据对故障点处与保护安装处地模电流间关系的分析,利用地模电流构造了故障选极判据。理论分析和仿真实验均证明了所提的行进波差动判据对区内外故障具有明确的选择性,选极判据对故障极线路的判别正确。

Research on development of urban taxi supply based on influence factors classification

In order to determine the regulations of the development of taxi supply under entry regulations in Chinese cities, an improved neural network model is applied to find the particular years when the government artificially puts new taxis into the market, and then extract the political influence from the taxi supply. The model is also utilized to study the relationships between the adjusted taxi supply and non-policy factors. A case study of Nanjing city is conducted. The results show that 2001 and 2007 are the particular years that the Nanjing government artificially put new taxis into its taxi market, which is in accordance with the five-year plan of China and the local development plans. The results also show that the improved neural network model has a good performance in expositing the evolution of adjusted taxi supply related to non-policy factors.

Early-stage Internet traffic identification based on packet payload size

In order to classify the Intemet traffic of different Internet applications more quickly, two open Internet traffic traces, Auckland I1 and UNIBS traffic traces, are employed as study objects. Eight earliest packets with non-zero flow payload sizes are selected and their payload sizes are used as the early-stage flow features. Such features can be easily and rapidly extracted at the early flow stage, which makes them outstanding. The behavior patterns of different Intemet applications are analyzed by visualizing the early-stage packet size values. Analysis results show that most Internet applications can reflect their own early packet size behavior patterns. Early packet sizes are assumed to carry enough information for effective traffic identification. Three classical machine learning classifiers, classifier, naive Bayesian trees, i. e., the naive Bayesian and the radial basis function neural networks, are used to validate the effectiveness of the proposed assumption. The experimental results show that the early stage packet sizes can be used as features for traffic identification.

Application of BP neural network model with fuzzy optimization in retrieval of biomass parameters

The retrieval of the biomass parameters from active/passive microwave remote sensing data （10.2 GHz） is performed based on an iterative inversion of BP neural network model with fuzzy optimization. The BP neural network is trained by a set of the measurements of active and passive remote sensing and the ground truth data versus Day of Year during growth. Once the network training is complete, the model can be used to retrieve the temporal variations of the biomass parameters from another set of observation data. The model was used in weights and microware observation data of wheat growth in 1989 to retrieve biomass parameters change of wheat growth this year. The retrieved biomass parameters correspond well with the real data of the growth, which shows that the BP model is scientific and sound.

Study on Artificial Neural Network Model for Crop Evapotranspiration

Based on potted plant experiment, BP-artifieial neural network was used to simulate crop evapotranspiration and 3 kinds of artificial neural network models were constructed as ET1 （meteorological factors）, ET2（ meteorological factors and sowing days） and ET3 （meteorological factors, sowing days and water content）. And the predicted result was compared with actual value ET that was obtained by weighing method. The results showed that the ET3 model had higher calculation precision and an optimum BP-artificial neural network model for calculating crop evapotranspiration.

SGuard：A Lightweight SDN Safe-Guard Architecture for DoS Attacks

Software Defined Networking(SDN) is a revolutionary networking paradigm towards the future network,experiencing rapid development nowadays.However,its main characteristic,the separation of control plane and data plane,also brings about new security challenges,i.e.,Denial-of-Service(DoS) attacks specific to Open Flow SDN networks to exhaust the control plane bandwidth and overload the buffer memory of Open Flow switch.To mitigate the DoS attacks in the Open Flow networks,we design and implement SGuard,a security application on top of the NOX controller that mainly contains two modules:Access control module and Classification module.We employ novel six-tuple as feature vector to classify traffic flows,meanwhile optimizing classification by feature ranking and selecting algorithms.All the modules will cooperate with each other to complete a series of tasks such as authorization,classification and so on.At the end of this paper,we experimentally use Mininet to evaluate SGuard in a software environment.The results show that SGuard works efficiently and accurately without adding more overhead to the SDN networks.

Gray correction method of X-ray fusion image

The conventional X-ray gray weighted image fusion method based on variable energy cannot characterize the phys- ical properties of complicated objects correctly, therefore, the gray correction method of X-ray fusion image based on neural network is proposed. The conventional method acquires 12 bit images on variable energy, and then fuses the images in a tra- ditional way. While the new method takes the fusion image as the input of neural network simulation system and takes the acquired 16 bit image as the output of neural network. The X-ray image physical characteristic model based on neural net- work is obtained through training. And then it takes steel ladder block as the test object to verify the feasibility of the mod- el. In the end, the gray curve of output image is compared with the gray curve of 16 bit real image. The experiment results show that this method can fit the nonlinear relationship between the fusion image and the real image, and also can expand the scope of application of low dynamic image acquisition equipment.

Modeling and Characterizing Internet Backbone Traffic

With enormous growth of the number of Internet users and appearance of new applications, characterization of Internet traffic has attracted more and more attention and has become one of the major challenging issues in telecommunication network over the past few years. In this paper, we study the network traffic pattern of the aggregate traffic and of specific application traffic, especially the popular applications such as P2P, VoIP that contribute most network traffic. Our study verified that majority Internet backbone traffic is contributed by a small portion of users and a power function can be used to approximate the contribution of each user to the overall traffic. We show that P2P applications are the dominant traffic contributor in current Internet Backbone of China. In addition, we selectively present the traffic pattern of different applications in detail.

Prediction of resilient modulus for subgrade soils based on ANN approach

The resilient modulus(MR)of subgrade soils is usually used to characterize the stiffness of subgrade and is a crucial parameter in pavement design.In order to determine the resilient modulus of compacted subgrade soils quickly and accurately,an optimized artificial neural network(ANN)approach based on the multi-population genetic algorithm(MPGA)was proposed in this study.The MPGA overcomes the problems of the traditional ANN such as low efficiency,local optimum and over-fitting.The developed optimized ANN method consists of ten input variables,twenty-one hidden neurons,and one output variable.The physical properties(liquid limit,plastic limit,plasticity index,0.075 mm passing percentage,maximum dry density,optimum moisture content),state variables(degree of compaction,moisture content)and stress variables(confining pressure,deviatoric stress)of subgrade soils were selected as input variables.The MR was directly used as the output variable.Then,adopting a large amount of experimental data from existing literature,the developed optimized ANN method was compared with the existing representative estimation methods.The results show that the developed optimized ANN method has the advantages of fast speed,strong generalization ability and good accuracy in MR estimation.

Bottleneck Analysis for Data Acquisition in High-Speed Network Traffic Monitoring

The increasing network throughput challenges the current network traffic monitor systems to have compatible high-performance data processing.The design of packet processing systems is guided by the requirements of high packet processing throughput.In this paper,we depict an in-depth research on the related techniques and an implementation of a high-performance data acquisition mechanism.Through the bottleneck analysis with the aid of queuing network model,several performance optimising methods,such as service rate increasing,queue removing and model simplification,are integrated.The experiment results indicate that this approach is capable of reducing the CPU utilization ratio while improving the efficiency of data acquisition in high-speed networks.

Prediction of uniaxial compressive strength and modulus of elasticity for Travertine samples using regression and artificial neural networks

Uniaxial Compressive Strength (UCS) and modulus of elasticity (E) are the most important rock parameters required and determined for rock mechanical studies in most civil and mining projects. In this study, two mathematical methods, regression analysis and Artificial Neural Networks (ANNs), were used to predict the uniaxial compressive strength and modulus of elasticity. The P-wave velocity, the point load index, the Schmidt hammer rebound number and porosity were used as inputs for both meth-ods. The regression equations show that the relationship between P-wave velocity, point load index, Schmidt hammer rebound number and the porosity input sets with uniaxial compressive strength and modulus of elasticity under conditions of linear rela-tions obtained coefficients of determination of (R2) of 0.64 and 0.56, respectively. ANNs were used to improve the regression re-sults. The generalized regression and feed forward neural networks with two outputs (UCS and E) improved the coefficients of determination to more acceptable levels of 0.86 and 0.92 for UCS and to 0.77 and 0.82 for E. The results show that the proposed ANN methods could be applied as a new acceptable method for the prediction of uniaxial compressive strength and modulus of elasticity of intact rocks.

Soft Sensor for Ammonia Concentration at the Ammonia Converter Outlet Based on an Improved Group Search Optimization and BP Neural Network

The ammonia synthesis reactor is the core unit in the whole ammonia synthesis production. The ammonia concentration at the ammonia converter outlet is a significant process variable, which reflects directly the production efficiency. However, it is hard to be measured reliably online in real applications. In this paper, a soft sensor based on BP neural network (BPNN) is applied to estimate the ammonia concentration. A modified group search optimization with nearest neighborhood (GSO-NH) is proposed to optimize the weights and thresholds of BPNN. GSO-NH is integrated with BPNN to build a soft sensor model. Finally, the soft sensor model based on BPNN and GSO-NH (GSO-NH-NN) is used to infer the outlet ammonia concentration in a real-world application. Three other modeling methods are applied for comparison with GSO-NH-NN. The results show that the soft sensor based on GSO-NH-NN has a good prediction performance with high accuracy. Moreover, the GSO-NH-NN also provides good generalization ability to other modeling problems in ammonia synthesis production.

Dynamic Target Wireless Network Selection Technique Using Fuzzy Linguistic Variables

Even though various wireless Net- work Access Technologies （NATs） with dif- ferent specifications and applications have been developed in the recent years, no single wireless technology alone can satisfy the any- time, anywhere, and any service wire- less-access needs of mobile users. A real seamless wireless mobile environment is only realized by considering vertical and horizontal handoffs together. One of the major design issues in heterogeneous wireless networks is the support of Vertical Handoff （VHO）. VHO occurs when a multi-interface enabled mobile terminal changes its Point of Attachment （PoA） from one type of wireless access technology to another, while maintaining an active session. In this paper we present a novel multi-criteria VHO algorithm, which chooses the target NAT based on several factors such as user preferences, system parameters, and traf- tic-types with varying Quality of Service （QoS） requirements. Two modules i.e., VHO Neces- sity Estimation （VHONE） module and target NAT selection module, are designed. Both modules utilize several ＂weighted＂ users＇ and system＇s parameters. To improve the robust- ness of the proposed algorithm, the weighting system is designed based on the concept of fuzzy linguistic variables.

Support vector regression-based internal model control

This paper proposes a design of internal model control systems for process with delay by using support vector regression(SVR).The proposed system fully uses the excellent nonlinear estimation performance of SVR with the structural risk minimization principle.Closed-system stability and steady error are analyzed for the existence of modeling errors.The simulations show that the proposed control systems have the better control performance than that by neural networks in the cases of the training samples with small size and noises.

Behavior Measurement Model Based on Prediction and Control of Trusted Network

In order to construct the trusted network and realize the trust of network behavior,a new multi-dimensional behavior measurement model based on prediction and control is presented.By using behavior predictive equation,individual similarity function,group similarity function,direct trust assessment function,and generalized predictive control,this model can guarantee the trust of an end user and users in its network.Compared with traditional measurement model,the model considers different characteristics of various networks.The trusted measurement policies established according to different network environments have better adaptability.By constructing trusted group,the threats to trusted group will be reduced greatly.Utilizing trusted group to restrict individuals in network can ensure the fault tolerance of trustworthiness of trusted individuals and group.The simulation shows that this scheme can support behavior measurement more efficiently than traditional ones and the model resists viruses and Trojans more efficiently than older ones.

Risk-based water quality decision-making under small data using Bayesian network

A knowledge-based network for Section Yidong Bridge,Dongyang River,one tributary of Qiantang River,Zhejiang Province,China,is established in order to model water quality in areas under small data.Then,based on normal transformation of variables with routine monitoring data and normal assumption of variables without routine monitoring data,a conditional linear Gaussian Bayesian network is constructed.A "two-constraint selection" procedure is proposed to estimate potential parameter values under small data.Among all potential parameter values,the ones that are most probable are selected as the "representatives".Finally,the risks of pollutant concentration exceeding national water quality standards are calculated and pollution reduction decisions for decision-making reference are proposed.The final results show that conditional linear Gaussian Bayesian network and "two-constraint selection" procedure are very useful in evaluating risks when there is limited data and can help managers to make sound decisions under small data.