Identification of the wave-particles interaction in space plasma using higher-order statistics

The interaction of wave-particles and wave-wave in the space plasmas are essentially non-linear or non-Gaussian processes. Using the higher-order statistical analyses methods (higher-order moments and bi-tri correlation or bi-tri spectrum), its physical properties can be described. The question addressed in this paper is that of the usefulness of higher-order statistical analysis for identification of the wave-particles interaction in space plasmas. The signals handled are from the ARCAD-3 ISOPROBE experiment on ELF frequency range, then strong electrostatic turbulence and electron density irregularities. Second and third order statistical analyses are applied: first, on time series associated with each type of measurement, then, on the two types. All results are presented for one typical case. Correlation functions estimated over the corresponding time intervals point out the existence of a, non-linear interaction between these fluctuations and electrostatic filed. Supported by National Natural Science Foundation of China Zhao Zhenyu; born in 1952, Ph. D.     

Identification of the wave-particles interaction in space plasma using higher-order statistics

The interaction of wave-particles and wave-wave in the space plasmas are essentially non-linear or non-Gaussian processes. Using the higher-order statistical analyses methods (higher-order moments and bi-tri correlation or bi-tri spectrum), its physical properties can be described. The question addressed in this paper is that of the usefulness of higher-order statistical analysis for identification of the wave-particles interaction in space plasmas. The signals handled are from the ARCAD-3 ISOPROBE experiment on ELF frequency range, then strong electrostatic turbulence and electron density irregularities. Second and third order statistical analyses are applied: first, on time series associated with each type of measurement, then, on the two types. All results are presented for one typical case. Correlation functions estimated over the corresponding time intervals point out the existence of a, non-linear interaction between these fluctuations and electrostatic filed. Supported by National Natural Science Foundation of China Zhao Zhenyu; born in 1952, Ph. D.     

隧道围岩变形的非线性自回归时间序列预测方法研究

针对传统时间序列预测模型的单一线性和忽略施工过程影响的静态局限性,提出非线性自回归(包括NARNN与NARXNN)时间序列预测模型.该模型通过引入动态施工影响因子作为附加的外部输入,同时结合模型本身的反馈结构和延迟单元,在结构和动态特性上更加符合实际系统,可以非线性动态地考虑隧道施工全过程.运用该模型对史家山2号隧道施工过程中的围岩水平收敛和地表变形进行预测.结果表明:1)非线性自回归预测模型比传统的ARMA预测模型的预测精度高、适应性好;2)通过多次预测并对结果取平均值,可以保证非线性自回归预测模型预测结果的预测精度和稳健性;3)通过优化动态施工影响因子的取值方法,可以进一步提高NARXNN时间序列预测模型的预测精度.     

Generation time and temporal scaling of bird population dynamics

Theoretical studies have shown that variation in density regulation strongly influences population dynamics, yet our understanding of factors influencing the strength of density dependence in natural populations still is limited. Consequently, few general hypotheses have been advanced to explain the large differences between species in the magnitude of population fluctuations. One reason for this is that the detection of density regulation in population time series is complicated by time lags induced by the life history of species that make it difficult to separate the relative contributions of intrinsic and extrinsic factors to the population dynamics. Here we use population time series for 23 bird species to estimate parameters of a stochastic density-dependent age-structured model. We show that both the strength of total density dependence in the life history and the magnitude of environmental stochasticity, including transient fluctuations in age structure, increase with generation time. These results indicate that the relationships between demographic and life-history traits in birds translate into distinct population dynamical patterns that are apparent only on a scale of generations.     

一种新的混合智能预测模型及其在故障诊断中的应用

针对机电设备运行状态受多因素影响且变化趋势复杂、难以用单一预测方法进行有效预测的问题，提出了一种新的基于经验模式分解、支持向量机和自适应线性神经网络的混合智能预测模型．首先，利用经验模式分解方法将非平稳时间序列按其内在的时间特征尺度自适应地分解为多个本征模式分量，然后根据这些分量各自趋势变化的剧烈程度选择合适的核函数，用支持向量机对其进行预测，最后通过自适应线性神经网络对这些预测分量进行自适应加权组合，得到原始序列的预测值．研究结果表明，对于标准算例和某机组振动趋势的预测，不论是单步预测还是多步预测，该模型的预测性能均好于单一的支持向量机预测方法。     

Nonlinear System Identification with Unknown Piecewise Time-Varying Delay

Identification of nonlinear systems with unknown piecewise time-varying delay is concerned in this paper.Multiple auto regressive exogenous(ARX) models are identified at different process operating points,and the complete dynamics of the nonlinear system is represented by using a combination of a normalized exponential function as the probability density function with each of the local models.The parameters of the local ARX models and the exponential functions as well as the unknown piecewise time-varying delays are estimated simultaneously under the framework of the expectation maximization(EM) algorithm.A simulation example is applied to demonstrating the proposed identification method.     

桂林市汽车销售量的时间序列预测模型

选取1999年1月到2007年3月桂林市各季度的汽车销售量原始数据,在SPSS系统下,运用滑动求和自回归(ARIMA)方法及非参数方法建立桂林市汽车销售量时间序列模型ARIMA(p,d,q),从模型识别、参数估计、适应性检验和实际拟合4个方面来确定模型的参数(p,d,q),并对模型的预测效果进行检验。结果表明,ARIMA(0,2,2)模型能够较好地包含桂林市汽车销售量的发展趋势,该模型对2007年2季度至2008年2季度汽车销售量的预测值与实际值的误差小,相对误差可以控制在3%以内。     

Effects of extremely low frequency magnetic fields on gap junctional intercellular communication and its mechanism

The study on biological effect of electromagnetic fields has been paid close attention in recent years. Gap junctional intercellular communication (GJIC) plays an important role in the maintenance of cell proliferation and differentiation, and in the multistage process of carcinogenesis. A series of researches showed that xtremely low frequency (ELF) magnetic fields not only enhance the inhibition of GJIC induced by 12-O-tetradecanoylphorbol-13-acetate, but also inhibit GJIC directly when the intensity is equal to or more than 0.4 　mT, and that the mechanisms of GJIC inhibition by ELF magnetic fields are due to hyperphosphorylation of connexin 43, which is mediated by protein kinase C-activated signal transduction, and the internalization of connexin 43 from plasma membrane to cytoplasm.     

基于波粒二象性的量子光通信系统的研究

结合光子的波粒二象性,提出利用孤子调制承载量子态的方案,着重从调制信道的角度,分别论证了孤子在NLSE光纤和SIT媒介中的传输演变,量子的压缩态和纠缠态可以在孤子的传输演变中产生,从而实现量子的高速调制.这种调制方案,融合了量子和孤子的特性,避免了量子长距离传送所带来的一系列问题,如量子中继等.     

基于遗传算法的双线性模型在大坝安全监测预报中的应用

文章提出了实用自相关系数图确定双线性模型的自回归项,应用了基于遗传算法的一套建模方法。通过实例表明,由于双线性模型实际拟合和预测过程中产生的残差信息进行反馈矫正,保证了模型高的拟合精度和稳健的预测性能,与门限自回归模型比较,双线性模型比门限自回归模型具有更好的预测精度。     

海杂波多重分形特性分析与仿真

多重分形消除趋势波动分析方法(MF-DFA)是描述非平稳时间序列复杂波动的有效方法,本文提出将其用于海杂波分析,发现海杂波具有多重分形特性。在此基础上,运用多重分形小波模型对海杂波进行建模与仿真。通过数学分析和MF-DFA方法对仿真结果进行检验,分别从概率密度分布特性和多重分形特性角度验证了该模型的有效性。     

二维流化床内水平射流横向脉动特性的研究

利用图像处理技术,得到了二维流化床内水平射流深度的时间序列,基于Ｒ／Ｓ分析方法,研究了流化数、射流速度和静态床层高度对水平射流横向脉动分类维的影响,结果表明,水平射流横向脉动脉随流化数的增加而加剧,随射流速度和静态床层高度的增加而减弱,探讨了水平射流横向脉动的原因,发现气泡与水平射流的合并是水平射流横向脉动的主要原因。     

基于粒子群-支持向量机的时间序列分类诊断模型

构建一种基于粒子群算法-支持向量机(PSO-SVM)的磁共振功能成像(fMRI)时间序列分类诊断模型,通过针对脑区多维时间序列数据的深层次分析实现病症患者和健康者的准确判断与区分,为面向fMRI时间序列数据的病症诊断和预测提供有效科学依据.该方法在以下4个方面不同于其他已有相关研究工作:(1)构建基于自回归模型的脑区多维时间序列数据特征表示;(2)构建基于支持向量机模型的脑区多维时间序列数据分类机制;(3)构建基于粒子群算法的分类学习参数寻优策略;(4)建立融合上述特征表示、优化分类与参数优选模式的fMRI时间序列数据分类诊断模型.通过以精神抑郁症作为实证分析的具体案例,所提出分类诊断模型已取得良好实验效果,展示出其有效性与合理性.     

极低频电磁场对小鼠脑组织脂质过氧化及白细胞的影响

在小鼠上观测了50Hz、4mT电磁场间断暴露后脑组织超氧化物歧化酶（SOD）活力和丙二醛（MDA）含量以及血液白细胞总数和分类的变化。结果表明：SOD活力在暴露10天和20天时下降显著;MDA含量在暴露1天和10天时明显下降,但在20天时显著升高。血液白细胞数在暴露1天和20天时均显著减少,但在10天时显著增多;嗜中性粒细胞百分率在暴露1天和20天时显著上升,与此同时淋巴百分率显著下降。提示在本实验条件下,50Hz电磁场可加剧脑组织的脂质过氧化,改变血液白细胞组分的计数;这些变化与磁场暴露的持续时间有关,类似于一种应激效应。     

调制白噪声激励下时滞非线性系统的瞬态响应

本文研究了调制白噪声激励下多自由度时滞非线性系统的近似瞬态响应概率密度．首先,由系统当前状态与时滞状态的关系,将原时滞系统近似等效为无时滞系统．然后,应用基于广义谐和函数的随机平均法,导出关于幅值瞬态概率密度的平均Fokker-Planck-Kolmogorov方程．该方程的解可通过级数式表示,基函数为幅值相关正交函数,系数为时间函数．应用Galerkin方法,系数可由一阶线性微分方程组解得,从而得出幅值响应的瞬态概率密度、状态空间概率密度及幅值统计矩的半解析表达式．最后,以调制白噪声激励下阻尼耦合的二自由度Duffing-vanderPol振子系统为例,验证其求解过程,并讨论不同时滞的影响．     

应用经验模态分解下的AR模型提取旋转机械故障特征

将时间序列的AR模型引入到旋转机械故障诊断中,采用了经验模态分解与AR模型相结合的方法提取旋转机械的故障特征。通过选取含有故障信息的固有模态函数进行功率谱分析,提取故障特征,分析故障原因。仿真和试验结果表明,此法能够有效地提取故障特征参数,为旋转机械的故障诊断提供了方法保障。     

利用样条函数建立季节性时间序列的预测模型

用B样条函数最小二乘法的非参数回归与时间序列相结合的方法建立了季节性时间序列预测模型. 利用滑动平均估计季节项,再利用B样条函数非参数回归估计长期项和周期波动,对于随机项建立ARMA模型,最后对某产品需求量进行了实例分析. 结果表明该方法有较高的预测精度.     

飞机自动增压系统仿真实验的设计与实现

飞机增压系统是飞机安全保障环节之一,也是民航院校实验教学的重要一课,利用虚拟仪器技术及其开发环境LabVIEW的强大功能,能够解决民航飞机增压系统在实验教学中的应用,实现民航B737-300型飞机增压控制系统在自动工作模式下的仿真实验。通过飞机增压系统基本原理的分析,结合实验教学的需要,模拟了飞机飞行全过程的自动增压变化,具有自动变换增压程序,实时测量数据显示,图形化实验数据输出等特点,并能够显示自动模式失效以及座舱高度警告等等。     

产品需求量非平稳时序的ANN-ARMA预测模型

针对基于非平稳时序的产品需求量预测方法存在的问题,研究了人工神经网络(ANN)与自回归滑动平均(ARMA)模型的集成建模与预测方法. 产品需求量的非平稳时序可分解为确定项和随机项两个部分,用人工神经网络模型拟合确定项,以表示非平稳的变化趋势;用自回归滑动平均模型拟合随机项,以表示平稳的随机成分. 将两个模型的预测值之和作为产品需求量的优化预测值. 仿真结果表明,集成模型的预测精度高于单一的人工神经网络模型.