Using PMD

先天性长 QT 综合征2型 hERG 基因 G604S 突变真核表达载体的构建和表达韩稳琦;霍建华;李国良;蒋永荣;武金娥;孙超峰【摘要】目的：构建长QT综合征（ LQTS ）2型hERG 基因G604 S突变的真核表达载体G604 S-hERG-pcDNA3、G604S-hERG-pEGFP。

方法：采用重叠延伸PCR 法在pEGM-hERG基础上构建G604S突变体PGEM-hERG-G604S，通过限制性内切酶法和基因重组技术将突变体插入到真核表达载体pcDNA3和pEGFP-C2中并测序验证，用脂质体转染法将G604 S-hERG-pEGFP转染至HEK293细胞并观察其荧光表达。

结果：构建的含有突变位点的真核表达载体经DNA测序均成功验证hERG基因1810位点碱基G突变为A，构建的G604 S-hERG-pEGFP在HEK293细胞中成功表达绿色荧光。

结论：成功构建hERG基因G604 S h-ERG-pcDNA3和G604 S-hERG-pEGFP表达载体。

%Aim:To construct eukaryotic expression vectors G 604S-hERG-pcDNA3 and G604S-hERG-pEGFP linked to LQT2.Methods:PGEM-hERG-G604S containingG604S fragment was constructed using pEGM-hERG as a template by overlap extension PCR and then validated by DNA sequencing , then hERG-G604 S sequence was subcloned into pcDNA 3 and pEGFP-C2 vectors using restriction enzymes and gene recombination technology , respectively.After sequencing, G604 S-hERG-pcDNA3 and G604 S-hERG-pEGFP expression vectors were transfected into HEK 293 cells to obtain heterolo-gous expression system .Results:G604 S-hERG-pcDNA3 and G604 S-hERG-pEGFP eukaryotic expression vectors were con -structed successfully ,a missense mutation in hERG 1 810 site nucleotide G mutantto A was identified using DNA sequen -cing,hERG mutant was correctly combined to eukaryotic expression vector pEGFP -C2and expressing green fluorescent pro-tein confusion mutant G604S in HEK293 cells.Conclusion: The protocol can be used to construct the eukaryotic expres-sion vector G604S-hERG-pcDNA3 and green fluorescent protein expression vector G 604S-hERG-pEGFP.【期刊名称】《郑州大学学报（医学版）》【年(卷),期】2015(000)006【总页数】5页(P753-756,757)【关键词】先天性长QT综合征;hERG基因;突变;真核表达载体【作者】韩稳琦;霍建华;李国良;蒋永荣;武金娥;孙超峰【作者单位】西安交通大学第一附属医院心内科西安710061;西安交通大学第一附属医院心内科西安710061;西安交通大学第一附属医院心内科西安710061;西安交通大学第一附属医院心内科西安710061;西安交通大学第一附属医院心内科西安710061;西安交通大学第一附属医院心内科西安710061【正文语种】中文【中图分类】R541先天性长QT综合征(congenital long QT syndromes,cLQTS)为一种心肌离子通道病，是由于编码心肌离子通道蛋白的基因突变导致心肌细胞复极时间延长而触发的一组临床综合征。

1、PMD 包含16 个规则集，涵盖了Java 的各种常见问题，其中一些规则要比其他规则更有争议：基本（rulesets/basic.xml）——规则的一个基本合集，可能大多数开发人员都不认同它：catch块不该为空，无论何时重写equals()，都要重写hashCode()，等等。

命名（rulesets/naming.xml）——对标准Java 命令规范的测试：变量名称不应太短；方法名称不应过长；类名称应当以小写字母开头；方法和字段名应当以小写字母开头，等等。

未使用的代码（rulesets/unusedcode.xml）——查找从未使用的私有字段和本地变量、执行不到的语句、从未调用的私有方法，等等。

设计（rulesets/design.xml）——检查各种设计良好的原则，例如：switch语句应当有default块，应当避免深度嵌套的if块，不应当给参数重新赋值，不应该对double 值进行相等比较。

导入语句（rulesets/imports.xml）——检查import 语句的问题，比如同一个类被导入两次或者被导入ng的类中。

JUnit 测试（rulesets/junit.xml）——查找测试用例和测试方法的特定问题，例如方法名称的正确拼写，以及suite()方法是不是static 和public。

字符串（rulesets/string.xml）——找出处理字符串时遇到的常见问题，例如重复的字符串标量，调用String构造函数，对String变量调用toString()方法。

括号（rulesets/braces.xml）——检查for、if、while和else语句是否使用了括号。

代码尺寸（rulesets/codesize.xml）——测试过长的方法、有太多方法的类以及重构方面的类似问题。

Jjavabean（rulesets/javabeans.xml）——查看JavaBean 组件是否违反JavaBean 编码规范，比如没有序列化的bean 类。

vmd pbc wrap用法-回复VMD PBC Wrap的用法VMD（Visual Molecular Dynamics）是一种用于分子动力学模拟的软件工具，可用于研究生物大分子结构和功能。

PBC（Periodic Boundary Conditions）是一种在分子动力学模拟中引入的条件，以模拟大量分子的自由运动。

而VMD PBC Wrap是VMD软件中的一个功能，用于处理周期边界条件下的原子坐标，以便正确显示分子在周期性盒子中的位置。

本文将向您介绍VMD PBC Wrap的用法，一步一步地解释如何使用这个功能来处理周期边界条件下的原子坐标以及与周期性盒子相关的其他操作。

第一步：打开VMD首先，您需要打开VMD软件。

这可以通过在计算机上双击VMD的应用程序图标来完成，或者通过终端窗口中输入“vmd”命令来启动VMD。

第二步：导入分子结构文件在VMD的主界面中，您会看到一个菜单栏和一个图形显示窗口。

首先，您需要导入您想要处理的分子结构文件。

可以从菜单栏中选择“File”->“New Molecule”来导入文件。

选择您的分子结构文件，并点击“Open”按钮导入文件。

第三步：创建周期性盒子一旦您的分子被导入，您需要为其创建一个周期性盒子。

这样可以模拟实验中的无限空间，使分子在模拟中正确地交互。

在菜单栏中选择“Extensions”->“Analysis”->“Periodic”，然后在弹出的对话框中选择“PBC Wrap”。

第四步：选择操作和参数设置一旦PBC Wrap插件被打开，您将看到一个新的窗口。

在这个窗口中，您可以选择要执行的操作以及相关的参数设置。

最常用的操作之一是“Wrap”，它将确保原子坐标被限制在周期性盒子内。

通过点击“Wrap”按钮，VMD将对所有原子应用这个操作。

此外，您还可以选择将作用于特定分子中的原子。

在VMD的主窗口中，选择要操作的分子，然后在PBC Wrap窗口中选择“Active Molecule”选项卡。

bpmn中inconing,outgoing标签-回复BPMN中的incoming和outgoing标签：解析与应用引言：随着业务过程管理（Business Process Management）的不断发展和优化，BPMN（Business Process Model and Notation，业务流程建模与标记）作为一种业界广泛使用的建模标准，被广泛应用于组织的流程图设计和管理中。

在BPMN中，incoming和outgoing标签是非常重要的元素，用于描述流程中各任务之间的依赖关系。

本文将详细解析这两个标签的特性和用途，并介绍它们在实际业务场景中的应用。

第一部分：incoming标签的解析与应用1. incoming标签的定义与作用：incoming标签用于定义某个任务或事件的输入，指明了数据流从哪里来。

它表示了一个流程元素可以接收的信息来源。

在BPMN图中，可以通过incoming标签，清晰地表示出每个任务或事件依赖的前置条件，从而实现流程的精确控制和顺序。

它充当了流程图中信息传递的路径。

2. incoming标签的属性：- id：标识每个incoming标签的唯一ID。

- sourceRef：指明了流程元素的来源。

它可以引用定义在同一个BPMN定义文件中的一个任务、事件或者网关等。

一个incoming标签可以连接到多个sourceRef。

3. incoming标签的示例应用场景：a. 顺序控制：通过incoming标签，可以定义一个任务仅在其前置任务已经完成的情况下才能执行。

例如，在一个订单处理流程中，某个任务必须在其前置任务“创建订单”完成后才能执行，通过定义incoming标签，可以实现顺序控制。

b. 并行控制：BPMN允许并行执行多个任务或事件。

通过incoming 标签，可以将多个并行流程分支汇总成一个汇聚点，从而实现并行执行的控制。

例如，在一个采购流程中，可以定义多个并行任务，每个任务负责不同的采购品类，通过incoming标签，将这些并行任务汇总成一个汇聚任务进行进一步处理。

High current measurements using a series connected resistor can be very accurate, provided the resistor is Kelvin connected. A Kelvin connection is a simple 4-terminal connection, usually made to a 2-terminal device, separating the current path through the resistor from the voltage drop across the resistor.The 4-wire connection improves the measurement accuracy by directly sensing the voltage drop across the resistor. This eliminates inaccuracies caused by voltage drops across resistor leads and printed circuit board (PCB) traces. The sensed voltage is then applied to high-impedance circuitry.Figure 1 illustrates the 4-wire Kelvin principle applied to a 2-terminal surface mount sense resistor.Figure 1. Kelvin SensingCurrent sense resistors are available from a number of manufacturers in two basic styles: open air and resistor chips. Open-air resistors are metal strips and are available in both leaded and surface mount packages. Resistor chips are surface mount packages and offer excellent thermal characteristics. Both styles are available in resistance ranges from <1 milliohm to 1 ohm. True 4-terminal sense resistors are available, but are generally more expensive. Unless extreme precision is required, the 2-terminal resistor is the economical choice with PCB traces added to provide the 4-terminal connection.Figure 2. Kelvin Sense Connections on the SMT4004Figure 2 illustrates the 4-wire principle as applied to the 2-terminal resistor using the SMT4004 power supply controller. For optimal performance, maintain equal length PCB sense traces and locate the sense resistor(s) as close to the SMT4004 as possible. Since the highest level of the four voltages the SMT4004 controls is used to power the device itself, this sense (VI1) trace must be capable of handling 3mA maximum without dropping additional voltage. The highest level input can be any one of the VIx pins. The remaining three sense resistors should have identical Kelvin sense trace widths and lengths. A typical layout example is shown in Figure 3.Recommended sense resistor vendor: IRC-International Resistive Co., Inc. (/) A list of other vendors is shown in Table 1.Figure 3. Typical Layout ExampleFigure 4. Current Sense Resistor TypesTable 1. Suggested Sense Resistors for Use with the SMT4004* Ceramic substrate - .714" W x .275' L x .118' H. Power rating good to 50 Amp. max** .2" W x .4" L, low ohms and good power rating.SELECTING SENSE RESISTOR VALUES AND POWER MOSFETSSelection of MOSFET switches for the SMT4004 Trakker is a compromise between load regulation, board area, and MOSFET cost.To obtain good load regulation with low supply voltages the MOSFET must have a very low ON resistance (R DS(ON)). For example, a 1.8V supply with a 10A maximum load current is equivalent to 180 milliohms load resistance. If the total resistance of the sense resistor plus MOSFET ON resistance is 9 milliohms, the load regulation is approximately 5% for a load change from 0A to 10A. Great care must be taken in choosing the MOSFET and ensuring the PCB trace resistance does not degrade circuit performance.If the circuit breaker trip voltage is programmed to 25mV on the SMT4004, and if the voltage drop across the MOSFET is kept below 25mV at maximum current, then the total drop of 50mV yields a load regulation of less than 3% with a 1.8V supply, and 1% with a 5V supply.Choosing a suitable MOSFET is simply a matter of applying Ohm's law once the supply voltage, load current, and load regulation requirements are known. Returning to the 1.8V supply example with a maximum current of 10A - first choose the current sense resistor, then set the trip current higher than the operating current. Choosing 12.5A yields 25% over-current and allows for the tolerances of the resistor and trip voltage. With a nominal trip voltage of 25mV and a trip current of 12.5A, the current sense resistor is 2 milliohms. Therefore, the MOSFET R DS(ON) must be below 7 milliohms. A list of recommended MOSFETs is shown in Table 2.Table 2. Suggested Low R DS(ON) N-Channel MOSFET Switches for Use withthe SMT4004* The V GS maximum for the ST Microelectronics devices is 15V. The VGATE output from the SMT4004 is typically 14V, but the data sheet maximum is 16V. In order to protect the MOSFET four 13V zener diodes (1N5243B) can be added from each VGATE pin to ground. Alternately, a single 15V zener diode (1N5245B) can be added from the VGG_CAP pin to ground to clamp all VGATE outputs.PARALLELING MOSFETS REDUCES VOLTAGE DROPS AND POWER DISSIPATIONWhen supply regulation is unacceptable due to high R DS(ON), two or more MOSFETs may be wired in parallel to lower the R DS(ON). For example, a 1V supply delivering 15A of load current will have its load regulation improved by using two or more MOSFETs in parallel (see Figure 5). The R DS(ON) is halved when two identical MOSFETs with identical gate resistors (RGx) are connected as in the Figure. The SMT4004 has been demonstrated to drive as many as 6 high current MOSFETs connected in parallel from any of the VGATE outputs.Figure 5. Paralleling MOSFETs to achieve lower R DS(ON)ADVANCED CURRENT-SENSING TECHNIQUES REDUCE LOSSES, IMPROVE RELIABILITYSMT4004: Typical Current-SensingThe SMT4004 provides over-current protection by sensing the voltage drop across an external resistor. Voltage trip Levels are user-programmable to 25mV or 50mV. The voltage dropped across the resistor and the MOSFET reduces the available output voltage (Figure 6). This drop becomes substantial when low input voltage high current supplies are employed.Figure 6. SMT4004 Typical Curent-Sensing Scheme.Output voltage regulation also suffers as the power supply voltage is regulated on the input-side of the circuit. The output voltage varies according to:Reducing the voltage drop across the sense resistor without compromising fault protection is desirable for improved efficiency and output voltage regulation.REDUCE THE IR DROP: ADD AN OFFSET VOLTAGEAdding a simple resistive divider significantly reduces the required sense resistor voltage needed to trip the SMT4004 Circuit Breaker while maintaining fault protection (Figure 7).Figure 7. Simple Divider Reduces Voltage DropThe resistor values are calculated according to the maxi-mum desired voltage drop across R S to trigger a fault:Where:V SET = SMT4004 OC Trip Point (25mV or 50mV)V TRIP = Desired Trip Voltage (across R S).Given:V IN = 2.5V, V SET = 25mV, V TRIP = 10mVUsing 1% resistors the closest values are:R1 = 56.2 milliohms , R2 = 9.31K milliohmsNote: Adding a small valued capacitor (C1) is suggested to prevent nuisance tripping. The value is chosen for a cutoff point (3dB) of 1/10th the switching frequency of V IN.Assuming f SW = 100kHz:Use 0.33uF.IMPROVED REGULATION: CORRECT VOLTAGE-SENSINGNearly all power supplies make available 'Sense' terminals for remote regulation of the output voltage. Figure 8 displays the correct Sense terminal connection.Figure 8. Improved Power Supply SensingThis technique can be used with power supplies that have Sense inputs. Remote sensing eliminates the effect of the current sense resistor voltage drop. With this arrangement (Figure 9) only the MOSFET RDS(ON) must be considered and a wider selection of devices can be used.Figure 9. Remote Sensing Using a Four Terminal DC/DC BrickThe Sense lead is applied to the output side of the current sense resistor, thereby eliminating this voltage drop as a factor in output voltage regulation.ULTIMATE REGULATION: INTELLIGENT VOLTAGE-SENSINGIdeally,the Sense connections should be made to the load side (Card-Side) to obtain the best possible voltage regulation. Unfortunately, such a connection could lead to a Bus-Side overvoltage during the time the MOSFET is being turned-on (or turned off). An intelligent voltage-sensing scheme overcomes this dilemma (Figure 10).Figure 10. Intelligent Power Supply SensingThe gate drive voltage (VGATE1) is used to turn on the 2N7000 small signal MOSFET thereby connecting the+Sense lead to the Card-Side after the power MOSFET is fully enhanced. The1N4148 is used to quickly shut-off the 2N7000 when a fault occurs. Without it, the 2N7000 may remain enhanced long enough for the power supply to cause an overvoltage on the Bus-Side.Note: This example of switching the +Sense lead was tested using a 5V supply. When used with lower voltage supplies,the resistor values must be altered to prevent the 2N7000 from turning on before the power MOSFET is fully enhanced. Increase the 1M milliohms resistor according to this ratio:For example, when using a 3.3V supply, the 1M milliohms is replaced with a 1.5M milliohms resistor.。

Abstract —Applications using artificial neural networks (ANNs) for optical performance monitoring (OPM) are proposed and demonstrated. Simultaneous identification of optical signal-to-noise-ratio (OSNR), chromatic dispersion (CD), andpolarization-mode-dispersion (PMD) from eye-diagramparameters is shown via simulation in both 40 Gb/s on-off keying (OOK) and differential phase-shift-keying (DPSK) systems. Experimental verification is performed to simultaneously identify OSNR and CD. We then extend this technique to simultaneously identify accumulated fiber nonlinearity, OSNR, CD, and PMD from eye-diagram and eye-histogram parameters in a 3-channel40 Gb/s DPSK wavelength-division multiplexing (WDM) system.Furthermore, we propose using this ANN approach to monitor impairment causing changes from a baseline. Simultaneous identification of accumulated fiber nonlinearity, OSNR, CD, and PMD causing changes from a baseline by use of the eye-diagram and eye-histogram parameters is obtained and high correlation coefficients are achieved with various baselines. Finally, the ANNsare also shown for simultaneous identification ofin-phase/quadrature (I/Q) data misalignment and data/carver misalignment in return-to-zero differential quadrature phase shift keying (RZ-DQPSK) transmitters. Index Terms —Optical fiber communication, neural networks, optical performance monitoring, phase modulation I. I NTRODUCTIONIGH-PERFORMANCE optical networks are susceptible to various degrading effects that can change over time. Knowledge of the data channel degradation can be used to diagnose the network, repair the damage, drive a compensator/equalizer, and/or reroute traffic around a non-optimal link [1-3]. Therefore, it is valuable to monitor the channels for many types of impairments, such as optical signal-to-noise-ratio (OSNR), chromatic dispersion (CD), polarization-mode-dispersion (PMD), and fiber nonlinearity, which can change with temperature, plant maintenance, andManuscript received January 15, 2009. This work was supported by theDARPA CORONET PARAGON Program, the U.S. Department of Commerce,and Cisco Systems, and is not subject to U.S. copyright. Xiaoxia Wu, and Alan Willner are with Department of ElectricalEngineering, University of Southern California, Los Angeles, CA 90089, USA.(phone: +1 213-740-0024; fax: +1 213-740-4679; e-mail: xiaoxia@ ).Jeffrey A. Jargon is with Department of Commerce, National Institute of Standards and Technology (NIST), Boulder, CO 80305, USA. Ronald A. Skoog is with Optical Networking Research, TelcordiaTechnologies, Inc., Red Bank, NJ 07701, USA.Loukas Paraschis is with cisco, 170 Tasman Dr., San Jose, CA 95134, USA.path reconfiguration. Key features of any optical performancemonitors are simplicity in implementation and the ability to accommodate different modulation formats and impairments. Recently, optical networks have been evolving from closed systems to open systems, in which the optical layer is designed to allow transmitter/receiver add and drop without affecting the current structure. This trend has been reflected in service provider requirements for “alien wavelengths” and in the standards -- most notably, ITU-T G.698.2 [4]. Associated with changes in the number of channels are the power transients in the surviving channels arising from cross saturation in optical amplifiers and the nonlinear interactions among channels. To maintain system performance, agile optical performance monitoring (OPM) and automatic system control become increasingly important. OPM can be performed by measuring changes to the data and determining “real-time” changes resulting from variousimpairments, such that a change in a particular effect will change a measured parameter. This can employ: (i) optical techniques to monitor changes in a radio frequency (RF) tone power or in the spectral channel power distribution [5], or (ii) electrical post-processing techniques in the specific case of coherent detection [6, 7].The optical approaches have been shown to be powerful forOPM. However, the electrical distortions that are crucial for thesignal quality at the decision point tend to be neglected in the optical approaches. Several techniques have been proposed for OPM using off-line digital signal processing of received electrical data signals [8-21]. Four of these methods [8-11] utilize amplitude histograms, power distributions or asynchronous sampling to estimate bit error rate (BER); four [12-15] employ delay-tap plots to distinguish among impairments; three [16-18] use pattern recognition techniques to identify multiple impairments; and the rest [19-21] use parameters derived from eye diagrams and histograms for thesame purpose. The latter approach is to probe the network uponinitialization and train each receiver to record a specific data eye-diagram pattern that corresponds to a specified range of potential physical parameters. These eye diagrams can be generated either from a synchronized sampler, or by a technique that regenerates such diagrams from asynchronous samples [11]. Once the network is fully operational, variations in the received eye diagram from the ideal formation can then be attributed to specific physical parameters derived from the Applications of Artificial Neural Networks inOptical Performance MonitoringXiaoxia Wu, Student Member, IEEE, Jeffrey A. Jargon, Senior Member, IEEE, Ronald A. Skoog, Member, IEEE , Loukas Paraschis, Senior Member, IEEE, and Alan Willner, Fellow, IEEEHCopyright © 2009 IEEE. Personal use of this material is permitted. However, permission to use this materialprior network/receiver training.Recently, we have made use of a neural network approach to “train” receivers in an optical network to distinguish between resultant shapes of the data channel’s eye diagrams and the degrading effects of OSNR, CD, PMD [19, 20]. The ANN approach has further been applied to monitor the accumulation of accumulated fiber nonlinearity in addition to OSNR, CD, PMD [21]. By use of this method, the coefficients of the neural network algorithm are iteratively derived prior to live traffic being sent through the network. A similar technique has also been used for time misalignment monitoring in return-to-zero differential quadrature phase shift keying (RZ-DQPSK) transmitters [22], which extends the applications of our ANN approach to a broader sense of OPM.In this paper, we show various applications of ANNs in OPM. In Section II, the concept and structure of ANNs are introduced. The popularly used multilayer perceptron (MLP) neural network and various steps involved in the development of neural network models are described. In Section III, simultaneous identification of OSNR/CD/PMD is demonstrated in 40 Gb/s on-off keying (OOK) and DPSK systems via simulation. Subsequent experimental verification is performed to simultaneously identify OSNR and CD. In Section IV, we add accumulated channel nonlinear effects to CD, PMD, and OSNR. We demonstrate this technique in a 3-channel 40 Gb/s RZ-DPSK WDM system. Furthermore, we propose using our ANN approach to monitor impairment causing changes from a baseline instead of the absolute values. Simultaneous identification of accumulated fiber nonlinearity, OSNR, CD, and PMD introducing changes from a baseline by use of the eye-diagram and eye-histogram parameters in a 3-channel 40 Gb/s DPSK WDM system is obtained with various baselines. In Sections V, ANNs are used for simultaneous identification of in-phase/quadrature (I/Q) data and data/carver misalignments in RZ-DQPSK transmitters, which indicates the applications of ANNs in a broader sense of OPM.II. A RTIFICIAL N EURAL N ETWORKSA. ANN ConceptsAs bit rates increase, it becomes more difficult to predict the data degradation mechanisms in optical networks. In order to enable robust and cost-effective “self-managed” operation, it would be desirable for the network itself to agilely monitor the physical impairments and the quality of the data signals, and automatically diagnose and feed back information to control the network. By incorporating trained receivers, a simple structure of a self-managed network is shown in Fig. 1 (a). Impairments are indentified by the trained receivers in the optical network element (ONE) and error signals are generated and sent to the routers. Further actions can be taken so that the network controller can agilely control and manage the network. To illustrate how the trained receivers work, we introduce the concepts of ANNs. ANNs are information-processing systems that learn from observations and generalize byabstraction [23, 24], which are attractive alternatives to conventional methods such as numerical modeling methods, analytical methods, or empirical modeling solutions. ANNs have the ability to model multi-dimensional nonlinear relationships and are simple to use. Furthermore, the neural network approach is generic (i.e., the same modeling technique can be re-used for passive/active devices/systems) and the response is fast. Due to these features, the ANN approach has gained much attention as a powerful tool in a number of areas such as pattern recognition, speech processing, control, and bio-medical engineering, and recently been applied in RF modeling, microwave design, and optical performance monitoring. Oftentimes, neural networks are first trained tomodel the electrical/optical behavior of passive and active components/circuits/systems. These trained neural networks can then be used in high-level simulation and design, providing fast answers to the task they have learned [25, 26].An ANN consists of multiple layers of processing elements called neurons. Each neuron is connected to other neurons in neighboring layers by varying coefficients that represent the strengths of these connections, as shown in Fig.1 (b). ANNs learn the relationships among sets of input-output data that are characteristics of the device or system under consideration.Fig. 1. Concepts of ANNs.After the input vectors are presented to the input neurons and output vectors are computed, the ANN outputs are compared to the desired outputs, and errors are calculated. Error derivatives are then calculated and summed for each weight until all of the training sets have been presented to the network. The error derivatives are used to update the weights for the neurons, and training continues until the errors reach prescribed low values. MLP is the basic and most frequently used structure. In the MLP neural network, the neurons are grouped into layers. The first and last layers are called input and output layers, respectively, and the remaining layers are called hidden layers. Typically, an MLP neural network consists of an input layer, one or more hidden layers, and an output layer. For example, an MLP neural network with an input layer, one hidden layer, and an output layer, is referred to as 3-layered MLP or MLP3, as shown in Fig 1. (c). The hidden layer allows complex models of input-output relationships. The mapping of these relationships is given by Y = g[W’•g(W•X)], where X is the input vector, Y is the output vector, and W and W’ are the weight matrices between the input and hidden layers and between the hidden and output layers, respectively. The function g(u) can be the smooth switch-type activation functions, such as sigmoid, arc-tangent, and hyperbolic-tangent, which are bounded, continuous, monotonic and continuously differentiable. In our analysis, a nonlinear sigmoidal activation function given by g(u)=1/[1+exp(-u)] is used, where u is the input to a hidden neuron or an output neuron.In addition to MLP, there are other ANN structures [27], such as radial basis function (RBF) networks, wavelet networks, and recurrent networks. The universal approximation theorem [28] states that there always exists a 3-layer MLP neural network that can approximate any arbitrary, nonlinear, continuous, multidimensional function to any desired accuracy. The number of hidden neurons depends upon the degree of nonlinearity of the function and the dimensionality of the model. Highly nonlinear systems require more neurons, while smoother systems require fewer neurons. In our work, the number of hidden neurons is optimized via adaptive processes, which add/delete neurons during training.B.ANN Training and TestingThe most important step in neural network model development is the training process. In this sub-section, we will explain the ANN training and testing processes in more details. The neural network weight parameters (w) are initialized so as to provide a good starting point for training. The widely used strategy for MLP weight initialization is to initialize the weights with small random values (e.g., in the range [-0.5, 0.5]). To improve the convergence of training, one can use a variety of distributions (e.g., Gaussian distribution), and/or different ranges and different variances for the random number generators used in initializing the ANN weights [29].The training data consists of sample pairs, {(x n, d n) , n∈T r }, where x n and d n are I- and K-vectors representing the inputs and the desired outputs of the neural network and T r represents the index set of the training data. In our work, the inputs are the parameters derived from eye diagrams or other sources, e.g. RF tone power, and asynchronous diagrams, and the outputs are the impairments, e.g. OSNR, CD, and PMD. We define the neural network training error as [30]:()∑∑∈=−=rrTnKkknnkT dwxywE12,21)( (1)where d kn is the k th element of d n and y k(x n,w) is the k th neural network output for input x n. The purpose of neural network training is to adjust w such that the error function E Tr(w) is minimized. The error between training data and ANN outputs is fed back to the ANN to guide the internal weight update of the network. Here,Δw = ηh is called the weight update, and η is a positive step size known as the learning rate. Gradient based iterative training techniques determine update direction h based on error information E Tr(w) and error derivative information ∂E Tr(w)/ ∂w. Step size η can be determined in one of the following ways: (1) small value, either fixed or adaptive during training; or (2) line minimization to find best value of η.The time needed for training depends on the amount of training data involved, the structure of the neural network, and also the training algorithm. There are several gradient-based iterative training algorithms, including back propagation, conjugate gradient and quasi-Newton. Back propagation is relatively slow in converging, so second-order training algorithms, such as conjugate gradient and quasi-Newton, are oftentimes preferred for their increased efficiency. The quasi-Newton approach is relatively fast due to its quadratic converge property, although more computer memory is required since it relies on the Hessian matrix whose inverse needs to be calculated. The conjugate gradient method is a nice compromise in terms of memory and implementation effort, since the descent direction runs along the conjugate direction, which can be determined without matrix computations.We use feed-forward computation in our work. Given the input vector X and the weight vector W, neural network feed-forward computation is a process used to compute the output vector Y. It is useful not only during neural network training but also during the usage of the trained neural model. The external inputs are first fed to the input neurons and the outputs from the input neurons are fed to the hidden neurons. Continuing this way, the outputs of one layer neurons are fed to the next layer neurons [30]. During feed-forward computation, neural network weights W remain fixed.After training, the ANN can be tested by use of other sets of data. The correlation coefficient, which represents how close the ANN model outputs to the testing data, can be used as the quality measurement factor.III.ANN S FOR CD/PMD/OSNR M ONITORINGA.CD/PMD/OSNRWith the increase of system capacity, optical networks will be highly susceptible to deleterious and data-degraded fiber-based impairments. CD, PMD, and OSNR are among a few of the most important impairments due to the broad spectraof high-rate signals. Therefore, the ability of the network to identify the amount of the impairments is quite important tomaintain system performance.Fig. 2 shows the simulated eye diagrams for a 40 Gb/s RZ-OOK signal at a few select combinations of OSNR, CD and first-order PMD (i.e., differential group delay (DGD)). The simulated DGD emulation assumes that the signal polarization principle states have worst-case alignments with 50:50 powerin the fast and slow axes. Visually, it is obvious that these impairments produce distinct features. To quantify these attributes, we can calculate various eye-diagram parameters. For this example, we choose four such parameters, including Q-factor, eye closure, root-mean-square (RMS) jitter, and crossing amplitude. Q-factor is defined as the difference of the mean upper and lower levels divided by the sum of the upper and lower level standard deviations; eye closure is the ratio of the outer eye height to the inner eye height; crossing amplitude is the point on the vertical scale where the rising and falling edges intersect; and RMS jitter is usually defined as the standard deviation of the time data calculated in a narrow window surrounding the crossing amplitude. These four inputs are chosen because they change significantly with varying impairment combinations.The ANN architecture used in this work is a feed-forward, three-layer perceptron structure. The ANN consists of four inputs (Q-factor, closure, jitter, and crossing-amplitude), three outputs (OSNR, CD, and DGD), and twelve hidden neurons. The ANN is trained by use of a software package developed by Zhang et al. [31]. We first verify the concept via simulation in 40 Gb/s RZ-OOK and RZ-DPSK systems. The conjugate gradient method is used for training. The training data are obtained from the eye diagrams by use of one set of 125 samples (OSNR = 32, 28, 24, 20, 16 dB; CD = 0, 15, 30, 45, 60 ps/nm; DGD = 0, 2.5, 5, 7.5, 10 ps). Another set of 64 samples (OSNR = 30, 26, 22, 18 dB; CD = 7.5, 22.5, 37.5, 52.5 ps/nm; DGD = 1.25, 3.75, 6.25, 8.75 ps) is used for testing.The simulated fiber channel includes a laser with a full width at half maximum (FWHM) line-width of 10 MHz; a 40 Gb/s logic source; a single-arm, Mach-Zehnder modulator (MZM) biased at the quadrature point with V π driving voltage for generating OOK and at minimum point with 2V π driving voltage for generating DPSK, where V π is the half-wave voltage of the MZM, followed by another MZM for RZ pulse carving. Impairments are added through emulators in the link and then the signals are detected by using a single photodiode for RZ-OOK and a balanced receiver following a delay lineinterferometer (DLI) for RZ-DPSK, where the eye diagrams are recorded and the eye diagram parameters are extracted.Fig. 3 (a) shows the training error versus the epochs. An epoch is defined as a stage of ANN training that involves presentation of all the samples in the training data set to the neural network once for the purpose of learning. The testing and ANN-modeled data are compared in Fig. 3 (b) and (c). The ANN reports a correlation coefficient of 0.97 and 0.96 for OOK and DPSK systems, respectively. The measured average errors for OSNR, CD and DGD are 0.57 dB, 4.68 ps/nm, and 1.53 ps, respectively for 40 Gb/s RZ-OOK, and are 0.77 dB, 4.74ps/nm, and 0.92 ps, respectively for 40 Gb/s RZ-DPSK. B. Experimental VerificationThe experimental setup is shown in Fig. 4. 40 Gb/s RZ-DPSK or RZ-OOK signals are generated using two cascaded MZMs. The signal then goes through a tunable dispersion compensating module (TDCM) with +/- 400 ps/nm tuning range and 10 ps/nm tuning resolution, which serves as the CD emulator. The output of the TDCM is sent to an erbium-doped fiber amplifier (EDFA) with a variable optical attenuator (VOA) in front to adjust the received OSNR. The noise-loaded signal is then filtered by a bandpass filter (BPF) with 1 nm bandwidth, and sent to a scope, where the eye diagram parameters are extracted.Fig. 2. The impact of degradation effects on eye diagrams of an RZ signal. Tx: transmitter; Rx: receiver; DGD: differential group delay (first-order PMD).Fig.3. Simulation results for OSNR/CD/PMD monitoring in 40 Gb/s OOK and DPSK systems.In our experiment, we vary OSNR and CD to get two sets of eye diagram parameters for 40 Gb/s RZ-DPSK and RZ-OOK signals, respectively, including extinction ratio, eye opening factor and signal-to-noise ratio. One set with 20 samples (OSNR = 32, 28, 24, 20, 16 dB; CD = 0, 10, 30, 50 ps/nm) is sent to the ANN model for training, and the other set with 12 samples (OSNR = 30, 26, 22, 18 dB; CD = 10, 20, 40 ps/nm) is used for testing. The final training errors for the OOK and DPSK data are ~0.03 and ~0.04, respectively. Fig. 5 shows testing results with the experimental data. For the RZ-DPSK signal, we use the eye of the destructive port of the DLI to extract parameters since we cannot estimate balanced eye diagrams with the scope. The ANN reports a correlation coefficient of 0.99 for both of the 40 Gb/s RZ-OOK and RZ-DPSK systems. Fig. 5 compares the testing and ANN-modeled data. The measured average errors for OSNR and CD are 0.58 dB, 2.53 ps/nm, respectively for 40 Gb/s RZ-OOK and are 1.85 dB, 3.18 ps/nm, respectively for 40 Gb/s RZ-DPSK.The OSNR considered in this experimental work is 16 ~ 32 dB for illustration purpose. In real optical systems, the OSNR can be lower, such as 10-12 dB in 40 Gb/s DPSK systems, which is validated via simulation is the next sub-section. C. Monitoring Low OSNROSNR values in real optical networks may degrade to levelsas low as 10-12 dB for 40 Gb/s DPSK systems. Here, we perform a simulation for 40 Gb/s RZ-DPSK using parameters similar to that in the experiment above. Only OSNR and CD are varied for illustration purposes. We use 49 samples (OSNR = 34, 30, 26, 22, 18, 14, 10 dB; CD = 0, 10, 20, 30, 40, 50, 60 ps/nm) for training and 36 samples (OSNR = 32, 28, 24, 20, 16, 12 dB; CD = 5, 15, 25, 35, 45, 55 ps/nm) for testing. The eye-diagram parameters include Q-factor, eye closure, and RMS jitter. Fig. 6 compares the testing and ANN-modeled data. The ANN reports a correlation coefficient of 0.99, which shows the effectiveness of using ANNs for identification of lower OSNRs. In this case, the measured average errors for OSNR and CD are 1.23 dB, and 4.56 ps/nm, respectively.IV. ANN S FOR CD/PMD/OSNR/A CCUMULATEDN ONLINEARITY M ONITORING A. CD/PMD/OSNR/ Accumulated NonlinearityOne parameter that has not been explored much in OPM has been the accumulation of nonlinear impairment on the data channels, which has typically been one of the most difficult parameters to monitor in an optical network. Adding accumulated nonlinearity is also a challenge in terms of the neural network approach, due to its specific signatures on the eye diagrams.Fig. 7 shows simulated eye diagrams for the middle channel of a 3-channel 40 Gb/s RZ-DPSK WDM system at a few select combinations of OSNR, CD, DGD and accumulated fiber nonlinearity. We can clearly see that different impairment combinations imprint different signatures on the eye diagrams. In this case, the four outputs are input optical power, OSNR, CD, and PMD, and the eight inputs include Q-factor, eye-closure, RMS jitter, ‘0’-level crossing amplitude, mean ofFig. 7. The impact of degradation effects on the eye diagrams of RZ-DPSK.Fig.6. Simulation results for OSNR/CD monitoring in a 40 Gb/s DPSK system.Fig.5. Experimental results for OSNR/CD monitoring in 40 Gb/s OOK and DPSK systems.Fig.4. Experimental setup. CW: continuous wave.‘1’s and ‘0’s, standard derivation (SD) of ‘1’s and ‘0’s.Fig. 8 shows the 3-channel WDM configuration used in the simulation. The 40 Gb/s RZ-DPSK signals are generated by two cascaded MZMs and then coupled together with a channel spacing of 0.8 nm. The channels are decorrelated by use of logic sources with different pseudo-random bit sequence (PRBS) orders. The WDM signals then pass through 2 km of highly nonlinear fiber (HNLF) with a nonlinear coefficient of 18 W -1·km -1, zero dispersion wavelength of λ0 (1550 nm), and dispersion slope of 0.05 ps/nm 2/km, following by a CD emulator and a PMD emulator. The output is sent to an EDFA with a variable optical attenuator in front to adjust the received OSNR. The signal is then filtered by a BPF with 0.64 nm bandwidth, and sent to an oscilloscope, where the eye diagram and eye histogram parameters are extracted. A 3-channel case is chosen to illustrate the concept, although this approach isalso applicable to WDM networks with more channels.The middle channel is chosen for the analysis because it experiences the strongest interchannel nonlinearity. The training data are obtained from the eye diagrams by use of a set of 135 samples (optical power = -5, -3, -1, 1, 3 dBm; OSNR = 36, 28, 20 dB; CD = 0, 20, 40 ps/nm; DGD = 0, 4, 8 ps). Note that a few training samples are used in this work. In practical networks, a much larger amount of data will be required for training. Fig. 9 shows the training error versus epochs. The final training error is ~0.1 in our case.Once the model is trained, we validate its accuracy by use of a different set of testing data that includes 32 samples (optical power = -4, -2, 0, 2 dBm; OSNR = 32, 24 dB; CD = 10, 30 ps/nm; DGD = 2, 4 ps). Again, the simulated DGD emulation assumes that the signal polarization principle states have worst-case alignments with 50:50 power in the fast and slow axes. The ANN reports a correlation coefficient of 0.97.Fig. 10 compares the testing and ANN-modeled data for optical power, OSNR, CD, and DGD. The measured average errors for optical power, OSNR, CD and DGD are 0.46 dB,1.45 dB, 3.98 ps/nm, and 0.65 ps, respectively. It is shown that the ANN models, trained with parameters derived from eye diagrams and eye histograms, can potentially be used to simultaneously identify accumulated fiber nonlinearity, OSNR, CD, and PMD in WDM channels.B. ANNs for Identification of Impairment Causing Changes from a BaselineNormally, when considering a system to be monitored, we assume the system is impairment-free; then differentimpairments, such as CD and PMD, are added for the purposeof testing the monitoring approaches. However, systems are not perfect, and inevitably contain a certain amount of impairments. Thus, starting from a baseline is more practical in terms of performance monitoring.From the analyses and demonstrations in the former sections, ANNs have been shown to be a potentially powerful tool for OPM. Continuing with this ANN approach, we make use of changes in optical power, OSNR, CD and PMD as the outputs of the neural network, rather than absolute values. Similarly, different impairment combinations imprint different signaturesFig. 9. Training error versus number of epochs for a 40 Gb/s RZ-DPSK channel (middle channel) in a 3-channel WDM system.Fig. 8. Simulation setup. λ0: the wavelength of the channel of interest. λ0=1550 nm.Fig.11. Block diagrams of using ANN for monitoring changes from a baseline.RMS: root-mean-square; SD: standard derivation.Fig. 10. Simulation results for comparison of testing and ANN-modeled data for a 40 Gb/s RZ-DPSK channel (middle channel) in a 3-channel system.on the obtained eye diagrams, where the input parameters are extracted. Again, the ANN used is an MLP3 with 12 hidden neurons. Fig. 11 shows a block diagram for the training and testing, where 135 samples are used for training (Δ optical power = -4, -2, 0, 2, 4 dB; ΔOSNR = -8, 0, 8 dB; ΔCD = -20, 0, 20 ps/nm; ΔDGD = -4, 0, 4 ps) and 32 samples are used for testing (Δ optical power = -3, -1, 1, 3 dB; ΔOSNR = -4, 4 dB; ΔCD = -10, 10 ps/nm; ΔDGD = -2, 2 ps). Similar to Fig. 8, Fig. 12 shows the simulation setup for monitoring impairment causing changes from a baseline. The initial system has a certain amount of impairments, which is added with a spool of optical fiber. The rest of the system staysthe same, as shown in Fig. 8.The baseline is set to optical power = -1 dBm, OSNR = 28 dB, CD = 20 ps/nm and DGD = 4 ps for initial training and testing. The final training error is ~0.1 in this case. The ANN reports a correlation coefficient of 0.93. Fig. 13 compares the testing and ANN-modeled data for the optical power, OSNR, CD, and DGD changes. It is shown that the ANN models, trained with parameters derived from eye diagrams and eye histograms, can potentially be used to simultaneously identify the accumulated fiber nonlinearity, OSNR, CD, and PMD causing changes in WDM channels.To consider other cases, we vary the baseline arbitrarily and repeat the training and testing. Fig. 14 shows a plot of correlation coefficients for various baselines. We can clearly see that high coefficients are achieved regardless of the baseline, which shows that the ANN approach is largely independent of the system reference. This technique could potentially be valuable for performance monitoring in optical systems with dynamic traffic.The inputs of ANNs so far are derived from eye-diagrams, which in general need clock recovery and are considered high cost. Recently we are able to derive parameters from the asynchronously generated delay-tap plots and train the ANNs to simultaneously identify OSNR, CD and DGD [32]. Moreover, the inputs of ANNs can be any other types of parameters that reflect the changes of impairments. In the following section, we extend the monitoring work to identify the time misalignments in RZ-DQPSK transmitters, in which case the inputs of ANNs are the RF tone/low frequency power levels.V. ANN S FOR T IME M ISALIGNMENT I DENTIFICATION INRZ-DQPSK T RANSMITTERS As the modulation format becomes more advanced, the transmitter tends to become more complex in terms of number of components and time synchronization among the components. Due to unavoidable optical/electronic device aging, imperfections and temperature variations, maintaining the correct timing within the transmitter is quite difficult and yet crucial to maintain system performance. Therefore, a laudable goal would be to monitor the time misalignment in order to provide a feedback signal and maintain proper synchronization. For an RZ-QPSK transmitter, the following are important: (i) I and Q data must be aligned with each other, and (ii) the RZ pulse carver must be synchronized to the data. There have been reports of measurements of time misalignment/synchronization for serial and parallel types of RZ-DQPSK transmitters [33]. In these techniques, a specific parameter is measured, such as power in an RF tone or power in one part of the spectrum. These parameters will either increase or decrease with a particular temporal misalignment. One could use a simple feedback loop that would either maximize or minimize these measured values. However, it would be more valuable if the transmitter could be “trained” to recognize and directly relate RF tone power or spectral power to a specific temporal misalignment cause and value.Since ANNs have the ability to learn the relationships among sets of input-output data that are characteristic of the device or system under consideration and then apply the relationship toFig. 14. Variation of the correlation coefficient with various baselines.Fig. 13. Simulation results for comparison of testing and ANN-modeled data for a 40 Gb/s RZ-DPSK channel (middle channel) in a 3-channel WDM system. (a)Optical power change; (b) OSNR change; (c) CD change; (d) First-order PMD change.Fig. 12. Simulation setup. λ0: the wavelength of the channel of interest. λ0= 1550 nm. DLI: delay-line interferometer.。

八年级上学期期中综合测评卷一、听力（共20小题, 计20分）第一节: 你将听到五个句子. 请在每小题所给的A、B、C三个选项中选出一个你所听到的单词或短语. 每个句子读两遍.（）1. A. monkey B. money C. honey（）2. A. tell B. call C. fall（）3. A. along B. around C. away（）4. A. mouth B. math C. month（）5. A. look after B. take after C. run after第二节: 听下面10段对话, 每段对话后有一个问题, 读两遍, 请根据每段对话的内容和后面的问题, 从所给的三个选项中选出最恰当的一项.（）6. A. In a park. B. In a bookstore. C. In a zoo.（）7. A. She likes collecting coins.B. She likes collecting stones.C. She likes collecting stamps.（）8. A. Because he has to have a guitar lesson first.B. Because he wants to buy a guitar first.C. Because he doesn't want to come to the party.（）9. A. Because her mother doesn't let her go on a trip.B. Because she doesn't have enough time for the trip.C. Because she doesn't have enough money for the trip.（）10. A. Brown. B. Green. C. Blue.（）11. A. Some bread. B. Some cake. C. Nothing.（）12. A. See his aunt. B. Have a meeting. C. See a film.（）13. A. He is talking on the phone.B. He is reading.C. He is sleeping.（）14. A. Once a day. B. Twice a day. C. Three times a day.（）15. A. A bird. B. A goldfish. C. We don't know.第三节: 听下面两段对话, 每段对话后有几道小题, 请根据每段对话的内容和后面的问题, 从所给的三个选项中选出最恰当的一项. 每段对话读两遍.听下面一段对话, 回答第16至17小题.（）16. Whom did Jane go to the movies with?A. Her father.B. Her friend.C. Her mother.（）17. Why did they get to the cinema late?A. Because the bus broke down.B. Because the traffic was heavy.C. Because they forgot the time.听下面一段对话, 回答第18至20小题.（）18. How did Jack and his classmates get to Beijing?A. They drove there.B. They flew there.C. They took a train.（）19. How long did they stay in Beijing?A. For one week.B. For two weeks.C. For three weeks.（）20. What did Jack buy for the gil?A. A model of the Great Wall.B. A model of the Summer Palace.C. A model of the Palace Museum.二、单项选择（共15小题, 计15分）（）21. —Mum, I'm going to see dentist this morning.—Don't forget to take umbrella. It's going to rain.A. a; aB. an; aC. the; anD. /; an（）22. —Which of the two films do you like better?—of them. I think they're interesting.A. EitherB. BothC. AllD. None（）23. Thanks a lot helping me my English.A. to; inB. to; withC. for; inD. for; with（）24. It's 10: 30 p. m. now. I have to leave, Kate.A. almostB. neverC. alwaysD. only（）25. —More and more foreigners are becoming interested in Beijing Opera. —That's true. It's an important part of Chinese .A. cultureB. competitionC. serviceD. information （）26. —Can you tell me about this accident（事故）?—Yes, it at 3: 00 p. m. Three people lost their lives.A. cameB. gaveC. happenedD. picked（）27. —do you go to see your parents?—Twice a week.A. How longB. How oftenC. How soonD. How far（）28. We must what he is doing and then we can take action.A. get upB. bring outC. wake upD. find out（）29. —The sandwich is really delicious.—Thank you. I it myself last night.A. makeB. madeC. will makeD. am making（）30. Jean felt much after she told her worries to her close friend.A. goodB. betterC. the betterD. the best（）31. Dashan is from Canada, he can speak Chinese very well.A. WhenB. AlthoughC. IfD. Because（）32. Li Dong is the second boy in our class and he's very good at playing basketball.A. tallB. tallerC. tallestD. the tallest（）33. Chen Feng is trying English because he plans to America.A. to learn; goB. to lean; to goC. learning; goingD. learn; going （）34. —Is Jim coming by train?—I'm not sure. He drive his car.A. mustB. needC. mayD. can（）35. —What do you think of the sports show?—.A. GreatB. I'm afraid notC. I watch it every dayD. I like sports三、完形填空（共10小题, 计10分）Archie was a boy and he had some dreams. All his dreams were full of 36 . They kept him happy. He often 37 the things which appeared in his dreams.One day, Archie 38 of building the best and the biggest park. But when he told his friends about it, they said, "Get a 39 job. You can never build such a park. "So Archie found a job as a policeman. It was a good job, but Archie didn't like it. It wasn't what he wanted. So Archie 40 his job.Archie thought about his dream again. 41 his friends didn't believe his dreams, he decided to have a thy. He thought that no matter how difficult it was, he should believe in himself.It was hard to build the park. But it was very 42 for Archie. When he 43 all the work, people came from everywhere. They loved all the fun things in the 44 . Archie was happy to see his dream come true. He learned that things could be much 45 than they sometimes seemed to be. And Archie's story tells us: if you stick to（坚持）your dream, you can do it too.（）36. A. happiness B. sadness C. darkness D. tiredness（）37. A. forgot B. visited C. built D. destroyed（）38. A. heard B. dreamed C. reminded D. felt（）39. A. good B. hard C. special D. bad（）40. A. went on B. put up C. gave up D. gave away （）41. A. Because B. Until C. But D. Although（）42. A. difficult B. boring C. interesting D. dangerous （）43. A. started B. finished C. lost D. caught（）44. A. factory B. stadium C. palace D. park（）45. A. worse B. less C. better D. fewer四、阅读理解（共20小题, 计40分）AAll of us want to be healthy. To keep healthy, more and more people are doing sports. Do you like sports? If yes, which is your favorite sport? Look at the ads, they may help you.（）46. What do people in Nancy's Gym want to do?A. To enjoy swimming.B. To learn to skate.C. To keep healthy.D. To eat delicious food.（）47. If Mr. Smith goes skating with his little son, they need to pay for two hours' skating.A. $50B. $60C. $80D. $100（）48. People can go to Water World at .A. 8: 00 amB. 1: 30 pmC. 9: 30pmD. 10: 00pm（）49. You can call if you want to swim.A. 455—6223B. 886—8252C. 215—1829D. 215—1892（）50. Which of the following is TRUE?A. People can go to Nancy's Gym on Sunday.B. Ice World is only for adults.C. There are four large pools for adults in Water World.D. A ten-year-old boy needn't pay to go swimming.BIn some Western counties, many children do chores（杂务）to get pocket money （零花钱）. They usually start to do this when they are ten years old.School students have to do homework and study for tests. They don't have much free time on weekdays. They often do chores on weekends.Young kids only do easy chores. So they don't get much money. But that's enough. Many of them only want to buy candy（糖果）. And candy is cheap! They often help do the dishes, sweep the floor, or feed the pet cat or dog.When they get older, they want to buy more and more things. They want things that are more expensive than candy. So they have to work harder! They often help their parents wash the family car, cut the grass or cook meals.Some jobs are good ways for kids to learn new things. For example, they can learn how to use a lawn mower（割草机）or how to cook. Of course, their parents help them at first.（）51. Many children in some Western counties to get pocket money. A. study hard B. do chores C. do part-time jobs D. sell their pets （）52. The children usually begin to do chores at the age of .A. 10B. 13C. 15D. 17（）53. Mary wants to get more pocket money to buy something expensive.She can .A. do the dishesB. feed the pet catC. cook mealsD. sweep the floor（）54. Which of the following is NOT true?A. Young kids do easy chores because they can get much money from their parents.B. School students often do chores on Saturdays and Sundays.C. If kids get older and want something more expensive, they have to work harder.D. Kids can learn how to cook with the help of their parents.（）55. The passage mainly tells us how children in some Western countries .A. find jobsB. get pocket moneyC. study at schoolD. do choresCWe all want our skin（皮肤）to look good. If your skin is clear and healthy, it makes you feel good about yourself. So how do you get great skin? Here are some ways.How to wash your skin★Use mild soap（温和的肥皂）and warm water to wash your body.★Wash the soap off.★Wash your face using cleansing cream（洁面霜）, and then wash with cool water. Twice a day is enough, because too much cleaning is bad for your skin.Protect（保护）your skin from the sunIf you go outside and be in the sun, you should put on sun cream in summer.★Wear a hat to protect your face. And it will protect your ears and ncck, too.★Stay under the tree whenever you can.★Use sun cream about half an hour before going outside.What other things you should do★Remember to drink lots of water—your skin loves it.★Some exercise is good for your skin.★Eat healthy food so that your skin gets things that it needs.（）56. What kind of water can you use to wash your body?A. Cool water.B. Cold water.C. Warm water.D. Hot water.（）57. Washing your face a day is enough.A. onceB. twiceC. three timesD. four times（）58. Which isn't the right way to protect your skin if you are outside?A. Wearing a hat.B. Staying under the tree.C. Using your sun cream.D. Staying in the sun.（）59. Whose behavior（行为）isn't right to protect the skin?A. Jack drinks lots of water every day.B. Linda likes doing some exercise.C. Lily always eats some fruit and vegetables.D. Lucy uses sun cream after going back.（）60. What's the best title of the passage?A. Caring for your skinB. How to wash your skinC. How to keep healthyD. Protecting your bodyDLife is not always perfect, so you should learn to enjoy your life. Try to stay happy most of the time, because it can help you live healthier and longer. Here are some ways to help you live your life more happily.Write down all the good things. These can make you happier. 61 Do not try to compare with others. They might have a bigger house or a more expensive car. Who cares?Stay more with your family and friends. 62 Your life may be very busy, and you don't have to do what others ask you to do each time. Remember to tell your family and friends that you love them.63 You can go hiking in the mountains to breathe（呼吸）fresh air. It can make you relaxed. You can also go to the beach to enjoy beach volleyball.Learn new things. 64 Read a book or learn more about computers, and that can open your mind to the world.65 It is also helpful for yourself and others. You make them happier, and you will be the happiest of all.If you remember to follow the ideas above, you will have a perfect life. Get ready to enjoy yourself from now on!根据材料内容, 从下面五个选项中选出能填入文中空缺处的最佳选项, 使短文意思通顺、内容完整.A. Give a helping hand to others.B. There's always something new to learn.C. And as for the bad ones, do not keep thinking of them.D. They are the most important parts in your life.E. Take more outdoor activities.61. 62. 63. 64. 65.五、补全对话（共5小题, 计10分）根据下面的对话情景, 在每个空白处填上一个适当的句子, 使对话的意义连贯、完整.A: Hi, Bob.B: Hi, Lucy!A:66. ? B: It was great. I had great fun.A: 67. ? B: I played basketball.B: Twice a week. I like it very much. How did you spend your last weekend?A: 69. . B: Sounds interesting. Do you often go to the mountains?A: Yes, once a week. Would you like to go to the mountains with me?B: 70. . But I'm afraid I can't go with you this weekend. Maybe next week is OK.六、词语运用（共10小题, 计10分）阅读短文, 从方框中选择适当的词并用其正确形式填空, 使短文通顺、意思完整. 每空限填一词, 每词限用一次.Teachers are important. But what role do they 71 ? They help children learn. 72 they don't just help children with schoolwork. They also help them become good and 73 . People all thank their teachers. World Teachers' Day is on October 5. But many 74 celebrate Teachers' Day on another day. For example, Chinese people celebrate it on September 10.On Teachers' Day, students 75 their teachers. They may give them small gifts. Some students give flowers. Some students 76 go back to their primary school or high school. They visit 77 old teachers. If they can't visit them, they may send them cards. Some students stay friends with their old teachers for many years.What makes a good teacher? A good teacher 78 to teach! They love their subjects and love to share what they know with students. They find fun ways to do that, too. A good teacher's class is hardly boring! A good teacher also 79 about his or her students. They teach students classes, and they also teach them something 80 life.Do you know a good teacher? Thank him or her today!71. 72. 73. 74. 75.76. 77. 78. 79. 80.七、书面表达（共1题, 计15分）生命中总会有一些人给予过我们帮助, 我们也心存感激. 请以"The person （s）I want to thank most"为题, 写一篇英语短文, 可以感谢你的朋友、家人、老师甚至是陌生人.1. 要点: （1）最想感谢的人是谁;（2）感谢他/她（们）的原因;（3）想为他/她（们）做的事.2. 要求: （1）文中不得出现真实姓名和学校名称;（2）词数80左右.The person（s）I want to thank most参考答案一、听力1—5 BCACA6—10 BCACC11—15 CABBB16—20 CBCBA二、单项选择21. C【解析】考查冠词. see the dentist"看牙医", 为固定搭配; 第二空表泛指, 又因umbrella的发音以元音音素开头, 改用不定冠词an. 故答案为C.22. B【解析】either"两者中的任意一个"; both"两者都"; all"三者或三者以上都"; none"三者或三者以上都不". 结合问句中的"two filns"和答语中的"I think they're interesting"可知选B.23, D【解析】thanks for doing sth. 意为"谢谢做某事", help sb. with sth. 意为"在某方面帮助某人". 故选D24. A【解析】考查副词. almost"几乎", never"从不", always, 总是", only"仅仅". 结合句意可知选A.25. A【解析】考查名词词义辨析. culture"文化"; compelition"竞争"; service"服务"; information"消息". 句意: —一越来越多的外国人对京剧开始感兴趣. —确实如此. 它是中国文化的一个重要组成部分. 结合句意可知应用culture.26. C【解析】考查动词词义辨析. 根据语境可知, 此处表示"它发生在下午三点", 故选C.27. B【解析】考查疑问词组. how long"多久", 对一段时间提问; how often"多久一次", 对频率提问; how soon"多久以后", 对将来时间提问; how far"多远", 对距离提问. 根据答语"Twice a week"可知选B.28. D【解析】考查动词短语辨析. get up起床; bring out使显现; wake up醒来; find out查明. 此处表示"我们必须查明他在做什么, 然后才可以采取行动", 故选D.29. B【解析】考查动词的时态. 句意: —这个三明治真的很好吃. ——谢谢, 这是我昨天晚上自己做的. 做三明治这个动作发生在过去, 故用一般过去时.30. B【解析】考查比较级的用法. 结合设空后的"在她把她的忧虑告诉她亲密的朋友后"可知, 琼感到好多了. much可用来修饰比较级, good的比较级为better, 故选B31. B【解析】句意: 虽然大山来自加拿大, 但他中文说得很好. although"虽然"符合句意, 故选B32. C【解析】考查形容词的最高级. 句意: 李东是我们班第二高的男生, 并且他很擅长打篮球. 结合句意及题干中的"the second"可知选C33. B【解析】考查非谓语动词. try to do sth. 意为"尽力做某事", try cloing sth. 意为"尝试做某事", plan to do sth. 意为"计划做某事". 结合题意可知选B34. C【解析】考查情态动词. must"必须"; need"需要", may"可能"can"可以". 结合答语中的"I'm not sue"可知选C35, A【解析】上句句意: 你认为这个体育节目怎么样? 四个选项中只有A项符合语境.三、完形填空【短文大意】本文讲述了Archie把梦境变成现实的故事.36. A【解析】根据下文中的"They kept him happy"可知此处应选A37. C【解析】根据第二段中的"building the best and the biggest park"可知built符合文意.38. B【解析】dream of"梦见", 符合文意, 故选B.39. A【解析】根据下文中的"It was a good job"可知此处应选A.40. C【解析】根据上文中的"It wasn't what he wanted"可推断gave up符合文意.38. B【解析】dream of"梦见", 符合文意, 故选B39. A【解析】根据下文中的"It was a good job"可知此处应选A40. C【解析】根据上文中的"It wasn't what he wanted"可推断gave up符合文意.41. D【解析】结合文意可知此处表让步关系, 故用Although"虽然".42. C【解析】根据上文中的hard和But可推断interesting符合文意.43. B【解析】根据下文的描述可知此处是说, 当他完成了所有的工作时, 故用finished44. D【解析】根据上文的描述可知park符合文意.45. C【解析】此处是说, 有时事情可能比看上去的（情况）要好得多. 故选C四、阅读理解【A篇短文大意】本文是三则有关运动的广告.46. C【解析】根据第一个方框中的"Keep fit! Lose weight! "可知选C.47. D【解析】根据第二个方框中的内容可知答案.48. B【解析】根据第三个方框中的"Opening time: 9: 00 am—9: 00 pm every day"可知答案.49. C【解析】根据第三个方框中的内容可知选C50. A【解析】根据第一个方框中的"Opening time: Saturday and Sunday"可知答案. 【B篇短文大意】文章介绍了在一些西方国家, 孩子们是怎样获得零花钱的.51. B【解析】根据第一段的第一句话可知答案为B52. A【解析】由第一段内容可知孩子们从十岁就开始做家务了. 故选A.53. C【解析】通读倒数第二段内容可知, 如果他们想要买更贵的东西, 他们可以帮助父母洗车, 割草或者做饭. 故选C.54. A【解析】根据第三段中的"Young kids only do easy chores. So they don't get much money"可知A项的表述不正确.55. B【解析】文章主要介绍了在一些西方国家, 孩子们是怎样获得零花钱的. 故B项符合题意.【C篇短文大意】本文介绍了几种保护皮肤的方法.56, C【解析】从文中的"warm water to wash your body"可知答案.57, B【解析】从文中的"Tvice a day is enough"可知答案.58, D【解析】从Protect（保护）your skin from the sun下的内容可知D项的表述不正确.59. D【解析】从文中的"Use sun cream about half an hour before going outside", 可知D项符合题意.60. A【解析】通读全文可知, 本文主要介绍了几种保护皮肤: 的方法, 故选A 【D篇短文大意】本文主要讲了如何能更快乐地生活.61—65 CDEBA五、补全对话66. How was your last weekend67. What did you do（last weekend）68. How often do you play it/basketball69. I went to the mountains70. Yes. I'd like/love to六、词语运用71. play72. But73. successful74. countries75. thank76. even77. thein78. loves79. cares80. about七、书面表达One possible version:The person（s）I want to thank mostIn my life, my parents are the most important persons. They raise me. They give me not only food, clothes and the place to live in, but also care and love. When I argue with my friends, my parents tell me how to get on well with them. When I am upset about bad grades, they always encourage me.I want to thank my parents and do something for them. I will try my best to help them do the housework and take good care of them when they're old.听力材料第一节: 你将听到五个句子. 请在每小题所给的A, B, C三个选项中选出一个你所听到的单词或短语. 每个句子读两遍.1. The famous singer held a concert to raise money for the homeless people.2. The rocks on the hill are going to fall. Please be careful!3. He sang loudly as he walked along the lake4. It was very hot last month.5. The boy had to look after his father at home.第二节: 听下面10段对话, 每段对话后有一个问题, 读两遍, 根据每段对话的内容和后面的问题, 从所给的三个选项中选出最恰当的一项.6. M: What can I do for you, madam?W: Please show me the book on the shelf. Is that a book on computer science?Q: Where are they probably talking?7. M: Do you like collecting coins, Linda?W: No. I think it's boring. I like collecting stamps.Q: What does Linda like collecting?8. W: Would you like to come to my birthday party?M: I'd love to, but I'm afraid I will be a little late. I must have a guitar lesson first. Q: Why will the boy be late for the party?9. M: You look worried, Tina.w: Yes. I don't have enough money for the trip.Q: Why does Tina look worried?10. W: Excuse me, Jim. Is this your ruler?M: No, it isn't. Mine is blue, but this one is green.Q: What's the color of the boy's ruler?11. M: Would you like some bread or cake?W: Neither. Thanks. I have had enough.Q: What would the woman like to have?12. W: Are you going to have a meeting this afternoon?M: No, I'm going to see my aunt in the hospital.Q: What does the hoy plan to do this afternoon?13. M: Would you mind not talking so loudly?W: Oh, sorry. I didn't know you are reading.Q: What's the man doing?14. W: How often shall I take the medicine?M: Twice a day. I'm sure you will feel well soon.W: Thank you very much.Q: How often should the woman take the medicine?15. M: What do you think is the best pet?W: I think a bird is the best pet.M: I don't think so. Birds are too noisy. I think a goldfish is the best oneQ: What does the man think is the best pet?第三节: 听下面两段对话, 每段对话后有几道小题, 请根据每段对话的内容和后面的问题, 从所给的三个选项中选出最恰当的一项. 每段对话读两遍. 听下面一段对话, 回答第16至17小题.M: Where did you go last night, Jane?W: I vent to see a movie with my mother.M: How was the film?W: It was great! But we arrived a little late because of the heavy traffic.M: What a pity!听下面一段对话, 回答第18至20小题.W: Hi, Jack! Long time no see. Where did you go last month?M: I went to Beijing with my classmates.W: Great! How did you get there, by train or by air?M: We took a train.W: What did you do there?M: Many things. The most exciting one is visiting the Great Wall.W: Sounds good! How long did you stay there?M: For two weeks. Oh, I bought something special for you. W: Really? What is it?M: Look! A small model of the Great Wall.W: It's so nice. Thank you.M: You're welcome.。