水文序列分析中基于信息熵理论的消噪新方法

为准确有效地识别和分离水文时间序列中的噪声成分，应用信息熵理论并结合小波消噪的基本思路，建立了小波系数阈值优选熵准则和水文序列消噪新方法：即首先应用熵函数H值描述噪声成分的不确定度，并应用信息量系数ICF(information cost function)值描述主序列的复杂度；然后通过分析不同小波系数阈值对应的噪声成分H值和主序列ICF值的变化规律，可优选出合理的小波系数阈值；最后对小波系数进行阈值量化处理，即可实现水文序列消噪。通过对不同特性模拟序列和不同实测水文序列分别进行分析，并通过与常用小波消噪方法(FT、SURE、MAXMIN)的消噪结果对比，验证了该阈值优选熵准则的合理性和适用性。分析结果显示：水文序列中的噪声成分具有偏态特性，因此本文中应用偏态分布线型(P-III型分布)对噪声成分进行描述更为合理；且此阈值优选结果是基于信息熵理论而确定，因此是整体上最优值。     

岩溶地下河水文时间序列小波消噪处理

在对重庆青木关岩溶地下河进行水文地质调查和水文气象观测的基础上,针对岩溶地下河水文过程的复杂非线性特征及存在的观测噪声,利用小波消噪方法对岩溶地下河水文时间序列进行消噪。结果表明：对水文时间序列进行消噪是必要的;小波消噪后,曲线变得平滑而且自相似性更强,该方法是有效可行的;小波函数的选择、分解层数的选取、阈值的确定和处理,需进行更深入的研究。     

小波消噪在枯水期来水量预报中的应用

根据小波理论,将枯水期来水量序列进行小波分解,并用软阈值方法进行消噪处理。用消噪后的小波系数进行序列重构,对重构后的序列进行时间序列分析,构建时间序列AR（p）模型,用AR（p）模型进行来水量的预测。     

基于小波消噪的年径流预测SPA模型

集对分析(SPA)的年径流预测就是基于SPA原理从同、异、反3个方面刻画预测模型的误差分布情况,利用联系度描述水文预测模型的预测精度,从而建立预测模型。水文序列的多时间尺度和高度的非线性特性,使得建立的水文预测模型精度往往不高。应用小波消噪的特点,利用汾河水库坝下站1959—1983年的资料建立小波消噪的SPA模型,对1984—1989年的丰枯状态进行预测,将水文预测中的单一预测和综合预测结果分别与实测系列进行对比。结果表明,综合预测模型优于单一预测模型。     

基于格拉布斯准则的小波阈值去噪算法研究

在小波去噪理论的基础上,应用格拉布斯准则对各层分解系数的阈值进行选取,改进了小波系数阈值估计的模型,并结合实例进行了计算分析。     

基于小波分析的大坝监测数据处理

大坝监测系统所采集的监测数据资料存在异常值并受噪声的影响,小波分析具有多分辨率、时频分析的特点。小波分析方法能准确、迅速的定位监测数据序列中异常值,适合运用于数据较多的情况。利用点的小波变换模极大值和阈值法能有效的去除异常值,避免人工去除的繁琐过程。对去除异常值后的序列进行小波软阈值去噪,消除监测数据中噪声的影响,为后续监测资料分析评价提供能反映大坝真实性态的数据。同时该方法适应性较广,能运用到其他数据预处理中,但必须根据实际情况选择合适的小波函数以及阈值。     

基于小波消噪偏最小二乘回归模型的径流预测

小波消噪方法,可消除原始数据序列存在的噪声;偏最小二乘回归分析方法,可减弱自变量间多重相关性在系统建模中的不利影响。为此,引入基于小波消噪的偏最小二乘回归分析方法进行建模分析,对奴各沙水文站1960-2000年的径流序列进行拟合和预测。结果表明,该模型在径流的拟合和预测中表现较好,具有较高的精度和较好的稳定性,可作为径流预测的有效方法。     

基于小波变换的友谊农场年降水序列多时间尺度分析

采用Morlet复小波对友谊农场近53 a年降水序列进行多时间尺度分析,并对友谊农场旱、涝的时频变化特征进行了初步研究.年降水序列由于受到各种因素的影响而含有噪声,文章采用小波消噪原理对实测年降水资料进行消噪处理.结果表明友谊农场年降水具有3 a的主周期变化,并主导着年降水变化的特性.同时推测2008年后,该地区年降水...     

染污绝缘子安全区泄漏电流检测中去除信号干扰方法

针对染污绝缘子安全区泄漏电流信号包含大量噪声干扰、很难准确提取其有效特征量的问题，利用4种小波阈值去噪方法对不同信噪比的实测安全区泄漏电流信号进行去噪，提取去噪前后的泄漏电流波形、有效值以及3次谐波与基波幅值比这3个特征量，对比分析了其去噪效果，优选出最适合安全区泄漏电流特征量提取的小波去噪方法。通过分析得出在信噪比大于1.0时，对于安全区泄漏电流波形和有效值，自适应阈值是最佳的去噪方法;对于3次谐波与基波的幅值比， 4种阈值去噪法获得的比值对真实比值的逼近效果基本一样。综合比较实测信号提取的各个特征量去噪前后的效果可知，自适应阈值法是提取安全区泄漏电流特征量的最佳去噪方法。     

基于小波变换的降雨时间序列去噪方法研究

小波变换在降雨时间序列数据的去噪方面具有显著的优势，可有效提高降雨时间序列预测的准确性。为确定降雨时间序列小波去噪过程中小波基函数、分解尺度以及阈值估计方法的选择，实现最优去噪，以国家气象科学数据中心2008～2018年的日降雨时间序列为基础数据，以中国5个不同气候类型的省份为研究区域，基于复合指标T对57种小波基函数的去噪效果进行评价，并评价去噪过程中可能的分解尺度和常用阈值估计方法。结果表明：7～10阶的Daubechies小波是去噪效果最好的小波基函数组，最小T值在0.326 4～0.422 8之间，Symlets小波族的去噪效果最差；最优的分解尺度为3级，最小T值范围为0.184 4～0.252 6;混合阈值和Steins无偏风险估计阈值的去噪效果最好，最小T值在0.377 3～0.435 9之间。研究成果可为中国境内降雨时间序列和其他水文气象时间序列的去噪方法提供参考。     

A Practical Guide to Discrete Wavelet Decomposition of Hydrologic Time Series

Discrete wavelet transform (DWT) is commonly used for wavelet threshold de-noising, wavelet decomposition, wavelet aided hydrologic series simulation and prediction, as well as many other hydrologic time series analyses. However, its effectiveness in practice is influenced by many key factors. In this paper the ??reference energy function?? was firstly established by operating Monte-Carlo simulation to diverse noise types; then, energy function of hydrologic series was compared with the reference energy function, and four key issues on discrete wavelet decomposition were studied and the methods for solving them were proposed, namely wavelet choice, decomposition level choice, wavelet threshold de-noising and significance testing of DWT, based on which a step-by-step guide to discrete wavelet decomposition of hydrologic series was provided finally. The specific guide is described as: choose appropriate wavelet from the recommended wavelets and according to the statistical characters relations among original series, de-noised series and removed noise; choose proper decomposition levels by analyzing the difference between energy function of the analyzed series and reference energy function; then, use the chosen wavelet and decomposition level, estimate threshold according to series?? complexity and set the same threshold under each level, and use the mid-thresholding rule to remove noise; finally, conduct significance testing of DWT by comparing energy function of the de-noised series with the reference energy function. Analyses of both synthetic and observed series indicated the better performance and easier operability of the proposed guide compared with those methods used presently. Following the guide step by step, noise and different deterministic components in hydrologic series can be accurately separated, and uncertainty can also be quantitatively estimated, thus the discrete wavelet decomposition result of series can be improved.     

Improved Wavelet Modeling Framework for Hydrologic Time Series Forecasting

The combination of wavelet analysis with black-box models presently is a prevalent approach to conduct hydrologic time series forecasting, but the results are impacted by wavelet decomposition of series, and uncertainty cannot be evaluated. In this paper, the method for discrete wavelet decomposition of series was developed, and an improved wavelet modeling framework, WMF for short, was proposed for hydrologic time series forecasting. It is to first separate different deterministic components and remove noise in original series by discrete wavelet decomposition; then, forecast the former and quantitatively describe noise’s random characters; at last, add them up and obtain the final forecasting result. Forecasting of deterministic components is to obtain deterministic forecasting results, and noise analysis is to estimate uncertainty. Results of four hydrologic cases indicate the better performance of the proposed WMF compared with those black-box models without series decomposition. Because of having reliable hydrologic basis, showing high effectiveness in accuracy, eligible rate and forecasting period, and being capable of uncertainty evaluation, the proposed WMF can improve the results of hydrologic time series forecasting.     

小波方法在水资源趋势分析中的能力检验

通过Monte-Carlo模拟试验对小波多分辨分析法在水文时间序列趋势分析中的趋势识别能力进行了研究。结果表明,小波多分辨分析方法的趋势识别能力与时间序列成份组成有关,与序列的噪音分布类型无关,而分析中的分解尺度对于随机性影响大的序列,其值与变差系数成正相关。小波多分辨分析所具有的时频局部化特征,能够合理地反映出时间序列在时间域中的变化特点,其在频域中分析趋势比线性回归和多项式拟合等在时间域中分析趋势的方法更具有优势。另外,在趋势分析前应分析序列中合理的突变成分以及剔除序列中的周期性成分,否则小波趋势分析结果将不可靠。     

基于信息熵的洪水过程均匀度变异分析方法——以东江流域龙川站洪水过程为例

为考察洪水过程形态特征是否随环境变化而发生变异,本文提出了基于信息熵的洪水过程均匀度变异分析方法。该法首先借用信息熵构建洪水过程均匀度模型,然后采用水文变异诊断系统分析洪水过程均匀度序列的变异规律,最后分析造成变异的物理成因。以东江流域龙川站洪水过程为例,计算结果显示:龙川站洪水过程均匀度在1972年发生了跳跃强变异,变异后洪水过程均匀度的均值增大,说明变异后洪水过程较变异前更加均匀;通过成因分析,推测这种变异受气候变化的影响不大,主要是由于人类活动特别是枫树坝水库的修建运行造成的下垫面变化引起的。     

Application of Information Theory to Groundwater Quality Monitoring Networks

Using the criteria of maximizing information and minimizing cost,a methodology is developed for design of an optimal groundwater-monitoring network for water resources management. A monitoring system is essentially an information collection system. Therefore, its technical design requires a quantifiablemeasure of information which can be achieved through applicationof the information (or entropy) theory. The theory also providesinformation-based statistical measures to evaluate the efficiencyof the monitoring network. The methodology is applied to groundwater monitoring wells in a portion of Gaza Strip in Palestine.     

A Threshold Based Wavelet Denoising Method for Hydrological Data Modelling

This work developed a novel framework for considering wavelet denoising in linear perturbation models (LPMs) and simple linear models (SLMs). Rainfall and runoff time series data were decomposed using wavelet transforms to acquire approximate and detailed rainfall and runoff signals, respectively, at various resolution levels. At each resolution level, threshold quantifications were performed by setting the values of a detailed signal below a certain threshold to zero. The denoised rainfall and runoff time series data were obtained from the approximation at the final resolution level and processed detailed signals using threshold quantification at all resolution levels of rainfall and runoff, respectively, by wavelet reconstruction. The data were then applied to the SLM and regarded as the smooth seasonal mean employed in the LPM. The noise, i.e., original time series value minus denoised time series value, was employed as the perturbation term in the LPM. Moreover, a linear relationship between input and output noise was assumed. The denoised runoff and estimated noise of runoff were summed to estimate overall runoff in the LPM. To verify the accuracy of the proposed method, daily rainfall–runoff data were analyzed for an upstream area of the Kee-Lung River. The analytical results demonstrate that wavelet denoising enhances rainfall–runoff modelling precision for the LPM.     

Development of Nonlinear Model Based on Wavelet-ANFIS for Rainfall Forecasting at Klang Gates Dam

Rainfall is one of the most complicated effective hydrologic processes in runoff prediction and water management. Artificial neural networks (ANN) have been found efficient, particularly in problems where characteristics of the processes are stochastic and difficult to describe using explicit mathematical models. However, time series prediction based on ANN algorithms is fundamentally difficult and faces some other problems. For this purpose, one method that has been identified as a possible alternative for ANN in hydrology and water resources problems is the adaptive neuro-fuzzy inference system (ANFIS). Nevertheless, the data arising from the monitoring stations and experiment might be corrupted by noise signals owing to systematic and non-systematic errors. This noisy data often made the prediction task relatively difficult. Thus, in order to compensate for this augmented noise, the primary objective of this paper is to develop a technique that could enhance the accuracy of rainfall prediction. Therefore, the wavelet decomposition method is proposed to link to ANFIS and ANN models. In this paper, two scenarios are employed; in the first scenario, monthly rainfall value is imposed solely as an input in different time delays from the time (t) to the time (t-4) into ANN and ANFIS, second scenario uses the wavelet transform to eliminate the error and prepares sub-series as inputs in different time delays to the ANN and ANFIS. The four criteria as Root Mean Square Error (RMSE), Correlation Coefficient (R ²), Gamma coefficient (G), and Spearman Correlation Coefficient (ρ) are used to evaluate the proposed models. The results showed that the model based on wavelet decomposition conjoined with ANFIS could perform better than the ANN and ANFIS models individually.     

泄漏检测信号滤波技术比较

通过泄漏检测模型试验分析测量信号中的噪声来源,在对比研究传统小波去噪、改进神经网络去噪、最小二乘拟合去噪等方法在实测数据中去噪效果的基础上,借鉴神经网络反向传播学习算法的思路,提出了信号预滤波结合阈值自学习小波去噪的综合滤波方法。该方法通过对恒定状态下带噪压力信号阈值自学习使得重构信号与期望输出均方误差最小来获得单一工况下的最佳去噪阈值,再将此阈值用于同一工况下整个时间段的去噪,这样根据不同工况下得到的最佳阈值可以获得最优输出。数值计算结果比较表明该方法对噪声的抑制作用明显,比传统小波去噪、改进神经网络去噪等方法效果更好。     

径流序列变异诊断的近似熵与样本熵法对比研究——以新疆叶尔羌河为例

识别并诊断径流序列的变异点,是进行径流序列重构与一致性再建立的关键环节,对于认识水文机理、应对径流变化具有重要价值。选取1957-2016年叶尔羌河出山口卡群水文站实测日径流量数据,采用传统MannKendall突变检验法识别突变点,采用样本熵和近似熵法诊断序列的变异点并进行方法对比。结果表明:采用传统的突变检验方法 Mann-Kendall法进行突变点识别时,出现了多个虚假变异点,无法准确断定突变年份;采用近似熵法对径流量序列进行变异点诊断,发现1959、1987和2004年3个变异点,采用样本熵法对径流量序列进行变异点诊断,发现1987和2004年两个变异点。通过两种方法的对比分析发现,在同一维度下,样本熵较近似熵能更好地反映径流量序列的波动情况,并且样本熵具有更好的灵敏度。卡群站以上为高山分布区,人为干扰较小,自然因素成为影响研究区径流变化的主要因素。     

基于DEM和GIS的流域水文信息提取——以巴中市为例

为了探讨基于DEM和GIS的流域水文信息提取过程中阈值确定的有关问题,应用ArcGIS中的Hydrology水文分析工具,对巴中市水域的水文信息提取进行了研究。研究结果表明:1汇流累积量与河网密度、流域面积满足二阶导数关系,利用导数关系能够有效确定河网提取阈值。2阈值对河网信息提取具有较大的影响,阈值越小,河网越稠密。当阈值达到8500时,提取的河网密度和面积基本趋于稳定且与实际水系基本符合。3实际地形特征、原始DEM数据可能存在的误差以及其他人为因素等都会对水文提取结果产生影响。