Search

An ultra-long phase-sensitive optical time domain reflectometry(Φ -OTDR) that can achieve sensing distance 205.46km single fiber with spatial resolution of 15m is presented to solve the problem of long-distance fiber optic cable monitoring, which use dual-coherent optic source and bidirectional high-order Raman amplification. To the best of our knowledge, this is the record sensing distance using a commercial Raman-line Φ -OTDR system.

With modern rapid and emerging development trend of Internet technology, great number of data are accumulated in the society, which leads to a larger data scale. With the target to discover the potential value of general data, it has been usually necessary to flexibly apply various data mining algorithms based on the actual situation and scenarios. The implementation of data mining algorithm is faced with great challenges in terms of execution efficiency, algorithm parallelization and platform usability. Different from horizontal text, multi-directional text and curved text, the surface text of surface object has some geometric features. In the detection phase, the text box label of surface text needs more vertex coordinates to express its spatial information, which makes it difficult to label large-scale samples. In the recognition stage, it is more difficult to recognize the surface text of the surface object than the regular text on the plane. At present, several universities ignore the importance to the construction of political education micro platform, and fail to develop a scientific micro platform according to the actual needs and interests of students. Based on this situation, this paper applies the data mining models and text recognition algorithm to study innovative smart political education platform.

Digital analysis of the current situation of network ideological education propaganda in colleges based on remote data modeling is conducted in this manuscript. The introduction of the neural network aims to quickly and effectively remove auxiliary variables that are not sensitive to the target variable, improve the model training effect, and avoid the long training time caused by the selection of too many variables and the risk of overfitting. Data packets may need to be forwarded by the multiple intermediate nodes in the communication network to reach the destination node. Based on the innovations on the proposed model, it has been applied into the education applications. The experiment has provided efficient results for further analysis.

Model uncertainties and external disturbances can inevitably affect vehicle dynamic control accuracy and even cause the vehicle system to be unstable and unsafe. Therefore, vehicle dynamic controller must be able to suppress the influence of model uncertainties and external disturbances on vehicle dynamic control performance. To this aim, we design a matrix inequalities (both bilinear matrix inequalities (BMIs) and linear matrix inequalities (LMIs) are involved) based robust model predictive controller for vehicle dynamic control. Robust positive invariant (RPI) set is used to guarantee the controller is robust and to construct the matrix inequality equations. We test the usefulness of the proposed controller via a numerical example.

We developed a deep learning prototype for the purpose of rainfall prediction for the city of Beijing China. It predicts rainfall dynamics in the next two hours with spatial reso-lution of 1km and temporal resolution of 6 minutes. That is a magnitude better than traditional weather forecasting. The computation time is less than 10 seconds, several magnitudes of less computation time than traditional forecasting. This deep learning network combines the strengths of known structures such as U-Net, ResNet, Sqeeze-and-Excitation, and the spatial Attention mod-ule. We rely solely on full convolutional layers instead of RNN layers as used in other weather prediction ef-forts. Meteorologically assessment metrics suggest better performance by the FCN approach. The prototype is now up and running in anticipation for the summer rain season of 2019. This might be the first such service employed by a capital weather service.

A 28nm 16Kb nonvolatile ternary content addressable memory (nvTCAM) test chip based on logic-compatible resistive memory (ReRAM) is demonstrated with high density and low power for packet routing in network applications. The crossbar array exhibits a smallest cell size of 16.3F2 by using 2-diode-2-ReRAM (2D2R) nvTCAM cell, useful for further 3-dimension (3D) stacking. K-means clustering is employed to allocate the storage of routing table entry for given bank count. Then the search of destination IP address can be classified and confined to a specific bank, reducing active banks for power saving. Evaluations show >3X improvement in cell density and >70% reduction in search energy with limited overhead in silicon area for bank count of four.

Converter driven motors may cause energy accumulation in the DC-link capacitor of the converter when it is in the braking mode, and without a way to transmit this energy, this part of the energy will threaten the voltage balance. To reduce the impact as well as make use of the braking energy, a three-port cascaded multi-level hexverter (TCMH) based on partial energy input is proposed in this paper. Made up by six bridge arms of which only three have power inputs, it is able to drive two high power motors at the same time. By a loop current control also proposed in this paper, energy interaction exists among the six bridge arms. When the two motors are in asynchronous modes, the energy of the braking motor converter can be fed directly to the motoring mode motor, which solves the power balance problem as well as increases the power utilization rate. Finally, simulations are conducted to verify the effectiveness of the proposed hexverter and its loop current control.

Data collection is a fundamental task in Wireless Sensor Networks (WSNs). Recent research shows that significant benefits can be achieved for data collection in WSNs by introducing mobile sink (MS). However, the slow movement speed of MS usually incur unbearable latency. The rendezvous-based data collection scheme provides an efficient approach to tackle this problem. In this paper, we propose an rendezvous algorithm based on k-hop domination set to construct the rendezvous nodes (RNs). We first give the problem definition and then we propose an efficient distributed algorithm to construct k-hop domination sets. After that, we apply algorithm of solving Traveling Salesman Problem (TSP) to generate the traveling trajectory of MS, along which the MS collects the cached data when it arrives the communication range of RNs. Finally, the effectiveness of our algorithm is validated by performance simulations.

Based on the analysis of the construction process of collaborative development capability federation for ordnance equipment, the matching model and process of capability requirement and capability are constructed, and a capability matching algorithm is proposed. The matching algorithm includes three matching stages: the preliminary matching stage that matches the capability category and state; the primary matching stage that matches the basic information and functional information of capability; and the improvement matching stage that matches the capability performance information. Finally, the validity and practicability of the matching method is verified by the left cover of a military vehicle planetary transmission mechanism in the capability-based federal collaborative development platform of ordnance equipment.

Hybrid DC circuit breaker, a combination of disconnector and multiple power electronic devices, has gained popularity for its low-loss at normal operation and its quick and effective circuit breaking during fault protection. However, its bulkiness threatens to limit its application in multi-terminal DC system, where the terminal is usually connected to more than one transmission lines and hence the circuit breaker tends to appear in clusters. In this context, a multi-port DC circuit breaker topology is proposed here. The circuit breakers at one terminal is integrated into one. Through this, it is able to cut down by half the number of power electronic devices needed, greatly reducing the size and cost for circuit breaking. The effectiveness of the proposed topology is verified by simulation results.

In order to protect solid-state circuit breaker (SSCB) from transient impacts under short circuit fault, an effective and economical snubber design is quite important. Traditional RCD snubbers are widely used because of their good voltage-clamping ability and short fault clearing time. Meanwhile, metal oxide varistor (MOV) also emerges as a suitable solution due to its non-linear U-I characteristic and simplicity. To optimize snubber design, it is imperative to analyze and compare the performance of these two snubber candidates. Therefore, a set of measurements is selected to focus on the performance analysis of RCD snubber and MOV, which contains five indicators: peak SSCB voltage, peak bus current, fault clearing time, snubber reliability and cost. A ±200 V Low-voltage DC (LVDC) system experiment is conducted for comparison. The results show that improved RCD snubber, Discharge-Suppressing Type I has smaller overvoltage but its cost is much higher than that of MOV. MOV stands out for shorter fault clearing time, however, its transient fluctuation is larger than that of RCD.

To utilize the regenerated energy directly among motor loads in industrial field, a hexagonal cascaded multilevel converter topology and its control strategy are proposed to drive two motors in asynchronous operation. An AC power supply is added to the traditional hexagonal configuration forming a tri-end system with one input port and two output ports, which can drive two asynchronous-operation motors by sending different control instructions. According to the analysis of power flow, the six arm-modules equal to two groups of coupling power supply, in which energy recycling is realized by supplying regenerated energy from one motor to another directly. Which improves the efficiency of energy conservation and suppresses the bump-up of arm capacitors' DC voltage. Simulation in MATLAB/Simulink and experiment have both verified the effectiveness of proposed topology and control strategy.

Gradient-reference-current (GRC) control strategy is proposed to achieve better dynamic response for tri-state buck converter over a wide load variety without additional hardware. By designing the logical current comparing element without load current detection, the conduction mode of tri-state buck converter under GRC control could be switched between discontinuous conduction mode (DCM) and pseudo continuous conduction mode (PCCM) adaptively to avoid large ripple voltage when converter operates over wide load range. As it operates in load condition within a comparatively small range, the converter stays operating in PCCM and the reference current could be changed in the same trend with load current among the pre-set values to improve the efficiency of tri-state buck converter in light load condition and the dynamic performance of it in heavy load condition. Simulation results verify the effectiveness of GRC control strategy.

When a bus fault occurs in a low-voltage DC microgrid, solid-state circuit breakers (SSCB) on either end of the DC bus assume the responsibility of the isolation of the faulted section, while the freewheeling path attached between SSCB and the bus takes on the task of fault energy absorption and fault current damping. However, during fault interruption, the snubber attached in parallel with SSCB for overvoltage suppression will influence the fault energy absorption performance of freewheeling path, while the resistance of the freewheeling path will also influence the overvoltage suppression capability of the SSCB snubber. Hence, this paper analyzed the interaction between the two and proposed a combined design so as to optimize the fault interruption performance for minimum SSCB overvoltage, and shortest fault clearing time, with less cost and size. Preliminary experiments are conducted on a ±200V/DC prototype, and the experimental results demonstrated the effectiveness of the design.

Considering multiple operation condition in DC microgrid, a bus voltage control strategy with terminal coordination is proposed in this paper, which could stabilize the bus voltage in extreme operation condition through receiving the operation information and issuing commands to every terminal. The equivalent circuit of DC microgrid is established and analysed, which can provide instruction on voltage control. During control strategy designing, the operation modes and switching criteria under multiple conditions are made. Accordingly, terminals coordinate control and flow chart of selection are introduced in this paper. Finally, DC microgrid model with loop topology with four terminal ports is built in MATLAB/Simulink to validate this strategy, and experiment results will be added in final paper.

An improved RCD snubber for solid-state circuit breaker (SSCB) is proposed in this paper in a low-voltage DC microgrid to protect SSCB against overvoltage caused by its tripping at the occurance of a DC-bus fault. Instead of relying mostly on the snubber capacitor to absorb the fault energy like commonly-used RCD snubbers, the proposed RCD snubber uses the snubber diode and an additional snubber resistor to form a separate path, which allows most of the energy to exhaust itself away from the snubber capacitor. In this way, the overvoltage across SSCB can be suppressed without increasing value and capacity of snubber capacitor, while the fault energy can also be quickly spent. Snubber process is discussed detailedly at a line-to-line fault to prove its effectiveness. Also, a design method is presented because former studies on parameter design of RCD snubber usually aim for converter appliance and cannot be directly applied on the proposed RCD snubber, which must be able to operate under fault current and thereby lays more stress in the peak value damping of overvoltage instead of quick suppression. Finally, simulations are conducted to prove its effectiveness in overvoltage suppression and fault current absorption.

Characterized by grid impedance variation and background grid voltage harmonics, weak grid may give rise to injected grid current distortion and cause stability issues in a grid-connected AC/DC converter. To lessen the effect, a composite compensation method is proposed in this paper based on output impedance model, which can increase control bandwidth and improve system's robustness against grid impedance variation and background harmonics. Simulations and preliminary tests are both conducted to verify that the proposed method can retain a good current harmonic suppression and system stability when the grid impedance varies or grid voltage harmonics are injected.

We present an Yb-doped fiber frequency comb based on a tapered single mode fiber by locking the repetition frequency and the carrier-envelop-offset frequency simultaneously to the microwave frequency reference outside.

Ni-P alloy thin films are prepared by electroless plating as embedded thin-film resistor (ETFR) materials. The micro-structure, electrical, thermal, mechanical, and corrosion-resistant properties of Ni-P alloy thin films are investigated to optimize the electroless plating conditions. When the phosphorus content was greater than 9 %, Ni-P alloy thin films are amorphous. Ni-P alloy thin films' sheet resistance, corrosion resistance and hardness increase with the increase of P content.

330V high-voltage direct current (HVDC) power system is a preferred option of the magnetic train power supply system. The power supply for magnetic levitation control system is used to provide stable operation voltage to the suspension control circuits, which is the key part of magnetic train. The suspension control power supply is typically the converter with high input voltage and low multiple output voltages. The traditional solutions have these problems, such as uncontrolled duty ratio, poor cross-regulation rate, and low reliability. In order to solve these limits, a novel two-stage solution employing the double resonant tank LLC-DCX is proposed and developed in this paper, which can overcome the problems in traditional design. The proposed solution not only can increase the overall conversion efficiency significantly because of the achieving soft-switching over the entire operation range, but also can achieve the low input current ripple and high reliability owing to the uniform thermal distribution. A 210-W, 220-380V input prototype with four outputs, is fabricated and tested in the lab, and the results are presented in this paper. The declared features are demonstrated well by the experimental results.

Independent spanning trees(ISTs) have increasing applications in fault-tolerance, reliable communication protocols and secure message distribution. In this paper, we propose an O(N) parallel algorithm to construct n BISTs rooted at any vertex on n-dimensional twisted cube, where BISTs denote the ISTs isomorphic to binomial-like tree and N =2n.

In this paper, a control strategy is applied on an LLCL filter based single-phase grid-connected voltage source inverter (VSI). This strategy includes splitting the capacitor or both capacitor and inductor on the L-C series branch of the LLCL filter into two parts and using the current flowing between the two parts as a feedback to stabilize the system. By this means, the system is simplified to a first-order one. The model of the LLCL filter based inverter is proposed and simulations of inverter with an LLCL filter based on this control strategy are conducted respectively to verify the effectiveness of the method.

Recently, mobility-enabled wireless sensor networks have been exploited to improve energy efficiency at the cost of increased latency. In this paper, we propose a Grid-based solution for Real-time Data Gathering application with mobile sinks (GRDG). Like many hierarchical structures, the flexible virtual grid partition scheme based on power control is designed to enhance network scalability, reduce routing hop counts, and ensure balanced energy consumption. Moreover, in order to decrease the transmission latency and cater for real-time application requirement, a local routing tree updating approach is presented. Although the quickly constructed routing tree is not optimal, communication overheads are decreased significantly.

Using the cone theory, the existence and uniqueness of solutions for a class of systems of abstract binary operator equations in Banach space are discussed, and the error estimations for the convergent iterative sequence are given. The results presented here generalize and unify some known results.

Using the cone theory, the existence of fixed point for a class of nonlinear operators in Banach space are discussed, and new results on these operators are obtained. An application to the Hammerstein integral equations is given. The results presented here improve and generalize some known results.

We report the first non-destructive, depth resolved determination of the full stress tensor in Cu through-silicon vias (TSVs), using synchrotron based micro-beam X-ray diffraction. Two adjacent Cu TSVs were studied; one deliberately capped with SiO2, the other without (uncapped). Both Cu TSVs were found to be in a state of tensile hydrostatic stress that fluctuated considerably with depth. The average hydrostatic stress across the capped and the uncapped Cu TSVs was found to be (99 MPa ± 13 MPa) and (118 MPa ± 18 MPa), respectively. This apparent disparity between the mean hydrostatic stresses is attributed to local differences in their microstructure, and not to the differences in capping.

A screw axis identification (SAI) method for serial robot calibration is applied in the service robot. The proposed method is newly introduced by Wang and can be operated easily. The self-designed service robot is a head-eye robot consisting of a binocular camera fixed on the head and two arms with four-degree of freedom. The binocular camera is used to calibrate the parameters of the arms joints. Then the kinematic model of the robot can be built by these parameters. Finally, by comparing the nominal angle values of the joints with the calculated values, a group of error values within 1eare achieved, which can meet the requirement of the further research of the robot about grabbing objects. Therefore, the SAI method which can automatically calibrate parameters has great efficiency and can be calibrated easily, which makes an identified contribution to the spreading of the robot application.

Internet of things is a new technology system that consists of a number of information technologies, and it's an important part of developing mordern sevice industry. The advent of Internet of things is changing the previous tradition networks. Based on Internet of things, smartphone of Android system is connected with service-oriented robots by wireless communication, where it is programmed based on TCP socket, thus, smartphone controlling service robot can be realized. The robot can be controlled by the system to sing, dance, and so on according to the command, and the characteristics are visual and portable, and the motion control of the robot can be realized well.

The drawbacks of Service-Oriented Analysis and Design (SOAD) methodologies are discussed and a new SOAD method based on Use Case (UC) approach is presented in this paper, therefore, a service model is designed to facilitate service identification from Use Case script. Moreover, a process to achieve high quality service identification is introduced as a guideline to an analyst work. Through the analysis of the method basic characteristics and its illustration, we can conclude that the method has several advantages compared to other methods which are based on business process modeling and enterprise architecture.

The TiAl based alloy have fine lamina structures after solution treatment (1370°C, 120min, 50°C water quench). With the addition of W element, the O content is increased form base inside to outside, and the oxidation layer is upward 6um. The oxidation layer is reduced from 6um to 3um after Si element addition (0.5%at). The oxidation layers are alternated between TiO 2 and Al 2 O 3 , and the outmost layer is TiO 2 . With the addition of Si element, the oxidation resistance is improved.

In order to compare upflow and downflow biological activated carbon(BAC) filter's advanced treatment effect, the changes of COD Mn and turbiditity during biofilm colonization process was discussed firstly.Investigation was made on influence of filtration velocity and empty bed contact time(EBCT) after successful biofilm colonizing. Result shows that for the same filtration velocity, the removal rate of COD Mn and turbidity by upflow BAC were 28.2% and 52%, respectively. The removal rate of COD Mn were 10% higher than downflow BAC's, but turbidity's removal effect was 20% lower. After upflow BAC continuous ran 30 days without backwash, effluent's turbidity and head loss were stable, which prolong backwash cycle effectively. Similar conclusion was obtained when upflow/downflow BAC have same EBCT.

a pure culture of Chlorella vulgaris or Prococentrum lima, commonly found in the ship ballast water, was exposed to low-pressure (LP) UV lamps. Irradiated pure culture suspension was subsequently incubated for 10 days under white light fluorescent lamps. During incubation, cell density, absorbance and relative fluorescence intensity were analyzed. The results show that the growth of C.vulgaris and P.lima were both inhibited by LPUV irradiation, and the inhibition effect increased with LPUV dose. C.vulgaris and P.lima were inhibited for 10 days and 3 days, respectively, after LPUV irradiation of 40, 100 and 300 mJ/cm2.

The purpose of this paper is to establish sample classification algorithms in consistent and inconsistent decision tables. First, according to the definition of inclusion degree and the idea of positive region in rough set, we give the definition of set value vector inclusion degree; then, according to the maximum inclusion degree principle, the sample classification algorithms are put forward with respect to consistent and inconsistent decision tables respectively; at last, the validity of the classification algorithms are accounted for through examples.

First, we classify the objects in continuous domain decision table according to fuzzy clustering; then, combining rough set theory with fuzzy set theory, an attribute reduct algorithm of decision table with continuous attributes is put forward; at last, a rule extraction algorithm is proposed and also the validity of this algorithm is accounted for through an example.

Derived from the thought of data cube, this paper presents a new structure model which organizes multimedia data in a multi-dimensional data cube. For the problem of semantic gap in retrieving multimedia information, the paper gives the definition and operation laws of matrix division method, which describes the high-level semantic structure for a lot of kinds of multimedia information. We retrieve multimedia information by using the high-level semantic structure to solve the problem of semantic gap. Based on the structure, the paper presents a partial priority algorithm by the results users locate multimedia information. The locating efficiency is improved by the algorithm based on the structure through modifying the high-level semantic branch weight by every result users locate multimedia information automatically.

The paper describe a detailed study into the object- oriented design and implementation of distributed and parallel computing environment for the forward kinematics analysis of parallel manipulator on desktop computers cluster using message passing while communication are needed among related computing nodes. The solution routines were blind to whether the objects were local or remote. Numerical tests were carried out and reasonable speed-up was achieved, particularly for direction solution methods. It is concluded that message passing provides a viable mechanism for computing nodes coordination while implementing distributed parallel

Minimizing energy dissipation and maximizing network lifetime are among the central concerns when designing applications and protocols for wireless sensor networks. In this paper we propose an energy-efficient clustering and routing scheme for wireless sensor networks (EECR). It mainly includes distributed nodes clustering, dynamic cluster head rotation and inter-cluster routing selection. In clustering stage, we introduce an uneven clustering mechanism. Cluster heads (CHs) which are closer to the base station (BS) have smaller cluster size than those farther from BS, thus they can preserve some energy for the purpose of inter-cluster data forwarding. For the dynamic cluster head rotation mechanism, the sensor nodes perform cluster head function in turn which balances energy consumption well among CHs. What is more, we propose an energy-efficient multi-hop routing protocol executed by BS for inter-cluster communication. Our strategy is making the most of BS's energy, so the node energy in the network is saved furthest. Simulation results show that EECR clearly prolongs the network lifetime over some related routing protocols.

First, the definition of similarity degree of objects in continuous domain decision table is given; then, according to fuzzy clustering, an attribute reduct and attribute significance algorithm of continuous domain decision table is put forward; thirdly, a classification algorithm is proposed according to the principle of maximum membership degree; at last, the validity of this classification algorithm is accounted for through an example.

In resent years many routing protocols have been proposed for wireless sensor networks, among which the adaptive routing protocols are very attractive because their low routing overhead. In this paper we propose an energy efficient distributed clustering and routing protocol (DCR). It mainly includes distributed nodes clustering and distributed multi-hop routing. In clustering stage, we introduce an uneven clustering mechanism. Cluster heads which are closer to the base station (BS) have smaller cluster size than those farther from BS, thus they can preserve some energy for the inter-cluster data forwarding. In distributed routing stage, each cluster head chooses its next-hop neighbor for inter-cluster communication independently according to the distance to BS and its residual energy. The simulation results demonstrate that DCR significantly minimizes energy consumption of nodes and extends the network lifetime compared with some existing clustering protocols.