A Kernel Time Structure Independent Component Analysis Method for Nonlinear Process Monitoring

Kernel independent component analysis(KICA) is a newly emerging nonlinear process monitoring method,which can extract mutually independent latent variables called independent components(ICs) from process variables. However, when more than one IC have Gaussian distribution, it cannot extract the IC feature effectively and thus its monitoring performance will be degraded drastically. To solve such a problem, a kernel time structure independent component analysis(KTSICA) method is proposed for monitoring nonlinear process in this paper. The original process data are mapped into a feature space nonlinearly and then the whitened data are calculated in the feature space by the kernel trick. Subsequently, a time structure independent component analysis algorithm, which has no requirement for the distribution of ICs, is proposed to extract the IC feature.Finally, two monitoring statistics are built to detect process faults. When some fault is detected, a nonlinear fault identification method is developed to identify fault variables based on sensitivity analysis. The proposed monitoring method is applied in the Tennessee Eastman benchmark process. Applications demonstrate the superiority of KTSICA over KICA.

Wavenumber-4 spectral component extracted from TIMED/SABER observations

The wavenumber spectral components WN4 at the mesosphere and low thermosphere(MLT)altitudes(70–10 km)and in the latitude range between±45°are obtained from temperature data(T)observed by the Sounding of the Atmosphere using Broadband Emission Radiometry(SABER)instruments on board the National Aeronautics and Space Administration(NASA)’s Thermosphere–Ionosphere–Mesosphere Energetics and Dynamics(TIMED)spacecraft during the 11-year solar period from 2002 to 2012.We analyze in detail these spectral components WNk and obtain the main properties of their vertical profiles and global structures.We report that all of the wavenumber spectral components WNk occur mainly around 100 km altitude,and that the most prominent component is the wavenumber spectral component WN4 structure.Comparing these long duration temperature data with results of previous investigations,we have found that the yearly variation of spectral component WN4 is similar to that of the eastward propagating non-migrating diurnal tide with zonal wavenumber 3(DE3)at the low latitudes,and to that of the semi-diurnal tide with zonal wavenumber 2(SE2)at the mid-latitudes:the amplitudes of the A4 are larger during boreal summer and autumn at the low-latitudes;at the mid-latitudes the amplitudes have a weak peak in March.In addition,the amplitudes of component WN4 undergo a remarkable short period variation:significant day-to-day variation of the spectral amplitudes A4 occurs primarily in July and September at the low-latitudes.In summary,we conclude that the non-migrating tides DE3 and SE2 are likely to be the origins,at the low-latitudes and the mid-latitudes in the MLT region,respectively,of the observed wavenumber spectral component WN4.

Effects of feeding time on complement component C7 expression in Pelteobagrus vachellii subject to bacterial challenge

Shifting the feeding time for fish from daytime to nighttime could alter their digestive behavior, disturb their metabolism, and may af fect immune-related genes. This study aimed to clone complement component C7 and analyze the different expression of C7 mRNA in fish fed during either the day or at night and then challenged with Aeromonas hydrophila infection. The P v-C7 cDNA of Pelteobagrus vachellii contained 2 647 bp with an open reading frame encoding a protein of 818 amino acids. Multiple sequences analysis indicated that P v-C7 included eight domains, which was similar to results for other species. Quantitative PCR analysis showed that P v-C7 was mainly expressed in the liver, spleen, intestine and head kidney tissues of healthy P. vachellii. Quantitative PCR analysis showed that C7 mRNA transcript in the liver, spleen and head kidney also increased significantly when the fish were fed at nighttime(20:00). In addition, the expression of Pv-C7 mRNA significantly increased with A. hydrophila challenge in the liver(48–96 h), spleen and head kidney(12–96 h) tissues of P. vachellii. Pv-C7 mRNA expression of the fish fed at nighttime showed significant higher than that in the fish fed at day time at 12–48 h in head kidney and 12–24 h in spleen. This study indicates that altering the feeding time from daytime to nighttime could increase P v-C7 mRNA expression, and feeding time may af fect the immune response involving C7.

Identification and classification of transient pulses observed in magnetometer array data by time-domain principal component analysis filtering

A method for identification of pulsations in time series of magnetic field data which are simultaneously present in multiple channels of data at one or more sensor locations is described. Candidate pulsations of interest are first identified in geomagnetic time series by inspection. Time series of these ＂training events＂ are represented in matrix form and transpose-multiplied to generate time- domain covariance matrices. The ranked eigenvectors of this matrix are stored as a feature of the pulsation. In the second stage of the algorithm, a sliding window （approxi- mately the width of the training event） is moved across the vector-valued time-series comprising the channels on which the training event was observed. At each window position, the data covariance matrix and associated eigen- vectors are calculated. We compare the orientation of the dominant eigenvectors of the training data to those from the windowed data and flag windows where the dominant eigenvectors directions are similar. This was successful in automatically identifying pulses which share polarization and appear to be from the same source process. We apply the method to a case study of continuously sampled （50 Hz） data from six observatories, each equipped with three- component induction coil magnetometers. We examine a 90-day interval of data associated with a cluster of four observatories located within 50 km of Napa, California, together with two remote reference stations-one 100 km to the north of the cluster and the other 350 km south. When the training data contains signals present in the remote reference observatories, we are reliably able to identify and extract global geomagnetic signals such as solar-generated noise. When training data contains pulsations only observed in the cluster of local observatories, we identify several types of non-plane wave signals having similar polarization.

Runtime environment for reflective real-time component

Reflective real-time component model is a special component model, which can identify timing constraint characteristics of component and support dynamic design-time amendment of real-time component according to users＇ requirements. The reflective real-time component runtime environment is a bearing space and reflective infrastructure for this special component model. It consists of three parts and manages the lifecycle and various relevant services of reflective real-time component. In this paper its mechanism and relevant key techniques in design and realization are formally specified with the communicating sequential processing （CSP） and the extended timed communicating sequential processing （TCSP）. Finally a prototype is established. Experimental study shows that this runtime environment can introduce a relevant reflective infrastructure guaranteeing dynamic and real-time features of software component.

Short-Term Financial Time Series Forecasting Integrating Principal Component Analysis and Independent Component Analysis with Support Vector Regression

Financial time series forecasting could be beneficial for individual as well as institutional investors. But, the high noise and complexity residing in the financial data make this job extremely challenging. Over the years, many researchers have used support vector regression (SVR) quite successfully to conquer this challenge. In this paper, an SVR based forecasting model is proposed which first uses the principal component analysis (PCA) to extract the low-dimensional and efficient feature information, and then uses the independent component analysis (ICA) to preprocess the extracted features to nullify the influence of noise in the features. Experiments were carried out based on 16 years’ historical data of three prominent stocks from three different sectors listed in Dhaka Stock Exchange (DSE), Bangladesh. The predictions were made for 1 to 4 days in advance targeting the short term prediction. For comparison, the integration of PCA with SVR (PCA-SVR), ICA with SVR (ICA-SVR) and single SVR approaches were applied to evaluate the prediction accuracy of the proposed approach. Experimental results show that the proposed model (PCA-ICA-SVR) outperforms the PCA-SVR, ICA-SVR and single SVR methods.

Testing for Deterministic Components in Vector Seasonal Time Series

Certain locally optimal tests for deterministic components in vector time series have associated sampling distributions determined by a linear combination of Beta variates. Such distributions are nonstandard and must be tabulated by Monte Carlo simulation. In this paper, we provide closed form expressions for the mean and variance of several multivariate test statistics, moments that can be used to approximate unknown distributions. In particular, we find that the two-moment Inverse Gaussian approximation provides a simple and fast method to compute accurate quantiles and p-values in small and asymptotic samples. To illustrate the scope of this approximation we review some standard tests for deterministic trends and/or seasonal patterns in VARIMA and structural time series models.

Online Batch Process Monitoring Based on Just-in-Time Learning and Independent Component Analysis

A new method was developed for batch process monitoring in this paper. In the devdopad method, just-in-time learning （ JITL ） and independent component analysis （ICA） were integrated to build JITL-ICA monitoring scheme. JITL was employed to tackle with the characteristics of batch process such as inherent time- varying dynamics, multiple operating phases, and especially the case of uneven length stage. According to new coming test data, the most correlated segmentation was obtained from batch-wise unfolded training data by JITL. Then, ICA served as the principal components extraction approach. Therefore, the non.Gaussian distributed data can also be addressed under this modeling framework. The effectiveness and superiority of JITL-ICA based monitoring method was demonstrated by fed-batch penicillin fermentation.

Effects of Different Sterilization Conditions on Active Components and Flavor of Apple Vinegar

[Objectives]To explore the effects of different sterilization conditions on nutrition and flavor of apple vinegar.[Methods]Five kinds of high temperature short time(HTST)sterilization conditions were selected to treat apple vinegar,and the volatile aroma components and the content of active components in apple vinegar before and after sterilization were analyzed.[Results]Compared with the control,the contents of total acid and malic acid in the samples after sterilization changed little,but the contents of citric acid increased significantly(P<0.01),and the contents of total phenols,ascorbic acid and total flavonoids decreased.Ethyl acetate,isopentyl acetate,ethyl caprylate,phenethyl acetate,1-pentanol,phenylethyl alcohol,acetic acid,and sec-butyl ether were the characteristic aroma components which contributed to the flavor of apple vinegar.As sterilization temperature increased,the content of esters decreased,while the content of acids,alcohols and aldehydes increased.The contents of nutrition,active components and volatile aroma components in apple vinegar under 100℃and 30 s sterilization conditions had little change compared with other sterilization conditions,so 100℃and 30 s were the optimal sterilization conditions.[Conclusions]Under different sterilization conditions,the content of flavor components in apple vinegar will change greatly,which will affect the quality of apple vinegar.

MECHANICAL RELAXATION TIME OF A TWO-COMPONENT EPOXY NETWORKLiClO_4 POLYMER ELECTROLYTE

The mechanical relaxation time of a two-component epoxy network-LiClO_4 system as a polymer electrolyte was investigated. The network is composed of diglycidyl ether of polyethylene glycol (DGEPEG) and triglycidyl ether of glycerol (TGEG), wherein LiCIO_4 was incorporated and acts as both the ionic carrier and the curing catalyst. As the relaxation time is informative to the segmental mobility, which is known to be essential for ionic conductivity, the average relaxation times of the specimens were determined through master curve construction. Experimental results showed that the salt concentration, molecular weight of PEG in DGEPEG and DGEPEG/TGEG ratio have profound effect on the relaxation time of the specimen. Among these factors , the former reinforces the network chains, leading to lengthen the relaxation time, whereas the latter two are in favour of the chain flexibility and show an opposite effect. The findings was rationalized in terms of the free volume concept.

荒漠草原不同植物群落蒸散组分特征及其环境因子影响分析

蒸散发是生态水文过程的关键环节,掌握蒸散组分的变化特征及其影响因子,对干旱半干旱地区的可持续发展至关重要。以荒漠草原多年生植物针茅群落和一年生植物猪毛蒿群落为研究对象,利用小型蒸渗仪开展了连续3年监测,分析了蒸散组分的日、月和年变化规律,探讨了影响蒸散组分的主要环境因子。结果表明:晴天时,多年生和一年生植物群落蒸散组分呈先增加后减小的抛物线型,夜间蒸散活动较弱,累积蒸散量较低,不足全天总累积蒸散量的20%;阴天时各蒸散组分无明显峰值,且日累积量均较小,一年生和多年生植物群落的蒸散量、蒸发量和蒸腾量无显著差异;10.64 mm/d及以上降雨对蒸散和蒸发的日变化具有明显影响,随着降雨量的增多,蒸散量和蒸发量也呈增大趋势,但蒸腾量则相对较小。从月动态来看,7—9月占全年蒸散量和蒸发量的一半左右,冬春季蒸散量和蒸发量维持在全年最低水平。年蒸散量与年降雨量接近,而蒸腾量占蒸散量的比例低于10%。总体来看,多年生植物群落蒸散量较一年生植物群落多。采用Mantel检验方法分析不同时间尺度影响蒸散组分的主要气候因素,在小时尺度上太阳辐射与蒸发量和蒸腾量显著性水平较高(P<0.01),但相关性较低(r<0.2);在日尺度上,蒸散量、蒸发量和蒸腾量与降雨量的显著性(P<0.01)和相关性(r≥0.4)均最高;而月尺度上,降雨量和降雨速率与蒸散量、蒸发量的相关性较高(0.2≤r<0.4),但显著性较低(P≥0.05)。因此,蒸散组分受环境因子的影响在不同时间尺度上具有较大差异,说明蒸散组分受多种因子共同影响。

基于特征图像组合与改进ResNet-18的电能质量扰动识别方法

针对传统电能质量扰动(power quality disturbance,PQD)识别体系中单一图像特征信息受限与算法识别能力不足等问题,依据特征融合的思想,提出一种基于特征图像组合与改进ResNet-18的PQD识别方法。首先,对PQD信号进行变分模态分解(variational mode decomposition,VMD)得到一系列固有模态函数(intrinsic mode functions,IMFs)与残差分量;其次,将IMFs、残差分量、原始扰动信号与Subtract分量纵向拼接成分量矩阵,利用信号-图像转化方法生成特征分量彩色图;再次,对原始扰动信号进行连续小波变换(continuous wavelet transform,CWT)生成小波时-频图;最后,将特征分量彩色图与小波时-频图组合输入改进的六通道ResNet-18中训练学习并完成扰动识别。通过仿真对PQD识别方法进行分析并将其与目前常用识别体系进行比较。结果表明,所提方法具有较好的抗噪性能并且能够更好地提取PQD特征信息,达到更高的识别准确率。

基于稠密连接的通道混合式PCANet的低分辨率有遮挡人脸识别

针对低分辨率有遮挡人脸识别问题提出了基于稠密连接的通道混合式主成分分析网络(DCH-PCANet)。现有的PCANet模型的卷积层只使用了通道无关式卷积(CIC)。通道无关式卷积由于未考虑特征图在通道方向上的相关性,可以更好地凸显单个特征图的局部纹理特征,对于补偿因低分辨率、遮挡等因素导致的特征损失具有重要意义,但也会强化遮挡区域的特征,从而放大坏特征的影响范围;而通道相关式卷积(CDC)由于充分考虑了各特征图在通道方向上的相关性,可以较好地抑制坏特征的作用,形成较为稀疏的特征图。在PCANet中添加了基于通道相关式卷积的特征图提取分支,形成了通道混合式PCANet;并且引入了稠密连接,以充分利用低阶特征提升有遮挡图像识别的鲁棒性。针对如下4种数据集进行了实验:受控环境、真实遮挡和模拟低分辨率的人脸数据集(AR人脸数据集),非受控环境、真实遮挡和模拟低分辨率的人脸数据集(MFR2和PKUMasked-Face),非受控环境、真实遮挡和真实低分辨率的人脸数据集(自建数据集)。实验结果表明,与现有方法相比,所提出的基于稠密连接的通道混合式PCANet具更好的遮挡鲁棒性和低分辨率鲁棒性,可以作为前沿方法的有效补充,提升其识别性能。

不同风格驾驶员眼动特性分析——夜间快速路分流区

为提高驾驶员在夜间快速路分流区换道的安全性,以不同风格的驾驶员为研究对象,分析不同风格驾驶员换道时眼动参数的变化规律。首先在前人研究基础上自编夜间快速路驾驶风格量表,对量表进行信度和Pearson相关系数法的效度检验;然后通过主成分分析法对问卷调查法得出的结果进行分析,得出驾驶风格综合得分量化模型,并使用K均值聚类分析法划分驾驶风格;最后基于眼动仪进行实车试验,获取三种风格驾驶员换道的眼动数据。试验结果表明:谨慎型、一般型、激进型驾驶员对于后视镜平均注视次数占比为48%、27%、25%;对左右两侧的平均累计注视时间占比是31%、25%、19%;平均瞳孔面积占比为56.3%、27.6%、16.1%。试验结果可以为提高驾驶员换道的行车安全性研究提供一定的数据支持。

基于时域卷积网络与Transformer的茶园蒸散量预测模型

在茶园水资源管理中,蒸散量(Evapotranspiration,ET)是评估作物水分需求的关键指标,由于茶园蒸散量预测具有时序性、不稳定性以及非线性耦合等特点,目前的茶园蒸散量预测模型存在预测精度较低的问题,针对此问题本文提出了一种新型的茶园蒸散量预测模型。首先使用互信息算法(Mutual information,MI)与主成分分析算法(Principal component analysis,PCA)相融合的数据处理算法(MIPCA),筛选强相关的特征并提取主成分;其次将时域卷积网络(Temporal convolutional network,TCN)与Transformer融合,利用灰狼算法(Grey wolf optimization,GWO)优化超参数,捕捉茶园数据的全局依赖关系;最后整合2个网络构建了MIPCA-TCN-GWO-Transformer模型,通过消融试验和对比试验验证了模型性能,并对模型在不同时间步长下的性能进行测试。结果表明,该模型平均绝对百分比误差(Mean absolute percentage error,MAPE)、均方根误差(Root mean square error,RMSE)和决定系数(Coefficient of determination,R^(2))3个评价指标分别为0.015 mm/d、0.312 mm/d和0.962,优于长短期记忆模型(Long short term memory,LSTM)等传统预测模型。在小时尺度、日尺度和月尺度下的R^(2)分别为0.986、0.978和0.946,在不同时间步长下展现了良好的适应性和准确性。本文构建的MIPCA-TCN-GWO-Transformer模型具有较高的预测精度和稳定性,可为茶园水资源优化管理和灌溉制度制定提供科学参考。

海上风电低频输电系统快速母线保护

海上风电低频输电系统母线故障后,换流器型电源等效电势与阻抗因控制的切换与调整过程,不再像传统电网可视为恒定。文中理论分析了母线电流差动保护适应性,明确故障控制策略对差动保护灵敏度的影响方式。利用母线区内外故障所对应的故障等效模型不同,提出基于时域全量模型识别的母线保护。该保护原理反映被保护元件自身拓扑的变化,理论上具有不受电源特性影响的优势,同样适用于其他换流器型电网。同时,所提保护在时域中实现,不存在相量提取问题,动作速度快。最后,在PSCAD/EMTDC中搭建了仿真模型,验证了所提母线保护的可靠性和快速性。

三维各向异性TI介质中的P/S波快速解耦技术及在弹性波逆时偏移中的应用

随着多分量采集技术的发展,弹性波逆时偏移技术在三维各向异性介质复杂地质构造成像中得到了广泛的应用.然而耦合的P波场和S波场,会在传播过程中产生串扰噪声,降低弹性波逆时偏移的成像精度.为了解决这一问题,本研究针对具有倾斜各向异性对称轴的三维横向各向同性(Transverse Isotropy,TI)介质,提出了一种矢量弹性波场快速解耦方法,可以有效提高偏移剖面的成像质量.该方法首先通过坐标转换,将观测系统坐标系的垂直轴旋转到TI介质的对称轴方向,在新坐标系下,根据具有垂直对称轴的三维横向各向同性(Vertical Transverse Isotropy,VTI)介质中的分解算子,推导出三维TI介质解耦算子表达式.接着引入一种在空间域快速计算分解波场的方法,来实现空间域矢量P波场和S波场分离,极大地提高了计算效率.最后,通过点积成像条件,将提出的P/S波分解方法引入到三维TI介质弹性波逆时偏移中,得到高精度的PP和PS成像.与以往的波场分解方法相比,本文方法具有数值稳定和计算效率高的特点.数值算例表明,应用上述三维TI分解算子得到的偏移剖面有效压制了噪声,提高了成像质量.

不同年份六堡茶中水溶性成分含量的变化

以不同年份六堡茶为原料,分析其水溶性成分含量的变化。选用冲泡法提取六堡茶中的水溶性成分,采用紫外分光光度法检测其含量。结果茶多酚含量在10.37~11.73之间,其含量随着茶样存放时间的增长而减少;咖啡碱含量在3.01~3.24之间,其含量随着茶样存放时间的增长而呈现出不同的改变趋势,但变化较小;茶褐素含量在65.89~81.24之间,随着陈化时间延长,茶褐素含量逐渐升高;游离氨基酸含量在5.21~11.54之间,随着陈化时间延长,氨基酸含量逐渐降低;可溶性糖的含量在3.39~3.67之间,放置陈化过程中无明显变化;黄酮类化合物含量在3.76~4.76之间,随着陈化时间延长,黄酮类化合物的含量逐渐升高。发现六堡茶存放时间和6种主要水溶性成分含量变化之间的关系,即茶多酚、游离氨基酸含量随着茶样存放时间的增长而减少,茶褐素、黄酮类化合物含量随着茶样存放时间的增长而增加,咖啡碱、可溶性糖的含量随着茶样存放时间的增长而无明显变化。

基于PCA-ShapeDTW-QWGRU的分布式光伏集群短期功率预测

针对分布式光伏短期功率预测建立基于主成分分析、改进的动态时间规整算法与量子加权门控循环单元(PCAShapeDTW-QWGRU)的集群功率预测模型。针对集群划分不够精细、光伏电站数据蕴含的信息难以捕捉的问题,提出基于主成分分析结合密度聚类算法(PCA-OPTICS)的集群划分方法;针对目前选取代表电站与集群相似性较低的问题,提出基于改进的动态时间规整算法(ShapeDTW)的代表电站的选取方法,利用ShapeDTW度量相似性距离,选取最小值作为代表电站,并利用基于均方根传播梯度下降法优化的量子加权门控循环单元(RMSprop-QWGRU)模型进行预测;为了解决代表电站与集群功率的变换系数转换差异较大的问题,采用实时变换系数对代表电站进行集群功率值预测计算。实验结果表明,所提方法能有效提升光伏集群功率预测的精度。

基于多分量LFM信号时频分析的水声多普勒和时延估计研究

在水声多普勒因子和时延估计研究实用化的进程中,利用多分量线性调频(LFM)信号实现估计的算法研究越来越普遍。针对多分量LFM信号时频域存有交叉项时各分量参数估计不准确的问题,提出基于非完全残差与脊线段匹配的自适应模态分解方法。该方法采用非完全残差函数保留了交叉点处的部分时频信息,利用脊线段匹配方法提供更精确的预设时频脊线,改进了各分量LFM信号调频斜率和起始频率的估计精度。联合两个估计量进一步给出了多普勒因子和时延估计的算法。仿真结果表示,较现有模态分解算法,所提改进方法有效解决了估计分量过程中交叉区间断裂带来的估计误差;水声多径的条件下,该方法的多普勒因子和时延估计精度优于对比的现有方法。