小样本下混合自注意力原型网络的风电齿轮箱故障诊断方法

针对部分风场因有标签故障样本数据稀少而导致风电齿轮箱故障诊断准确率不高的问题,提出了一种小样本下混合自注意力原型网络的故障诊断方法。首先,通过原型网络将振动信号映射到故障特征度量空间;然后采用位置自注意力机制和通道自注意力机制进行矩阵融合构建混合自注意力模块,建立原始振动信号的全局依赖关系,获取更具判别性的特征信息,学习风电齿轮箱各健康状态下的度量原型;最后通过度量分类器进行模式识别,实现小样本条件下风电齿轮箱的故障诊断。实验结果表明,所提出的混合自注意力原型网络故障诊断方法在不同小样本数据集上均能实现风电齿轮箱高精度故障诊断。     

基于等距特征映射和支持矢量机的转子故障诊断方法

针对振动信号的非线性特征,提出一种基于等距特征映射(Isometric feature mapping,ISOMAP)和支持矢量机(Support vector machine,SVM)的转子故障诊断方法。利用ISOMAP把数据从高维空间投影到低维空间而不改变数据内在属性的特点,对高维的故障振动信号降维并提取出低维的数据作为特征矢量,采用一种新核函数支持矢量机作为分类器进行故障诊断。将该方法应用于转子故障诊断,结果表明,ISOMAP-SVM方法不仅具有较高的故障诊断率,而且取得振动信号在低维空间的可视化表示。与其他核函数相比新核函数支持矢量机具有较好的诊断效果。     

基于协整分析的风力机多工况监测与故障诊断

为降低装备故障率、减小停机损失,提出一种基于协整分析的装备多工况监测与故障诊断方法。基于监控与数据采集系统采集的数据,利用随机森林特征选择算法提取与装备故障相关的关键特征变量;通过对关键特征序列的协整分析,计算协整系数,建立协整残差模型,获得反映装备状态变化的最优残差序列;采用概率图分析最优残差序列,完成了多工况状态的区间划分,确定了每种工况对应的残差预警阈值,实现了状态监测与故障预警。某型号直驱式风力机的研究结果表明：所提出的方法能有效分析非平稳时间序列,利用残差阈值可以监测电机故障、识别风力机工况,提高故障诊断的准确性。     

A data driven approach for condition monitoring of wind turbine blade using vibration signals through best-first tree algorithm and functional trees algorithm: A comparative study

Wind energy is one of the important renewable energy resources available in nature. It is one of the major resources for production of energy because of its dependability due to the development of the technology and relatively low cost. Wind energy is converted into electrical energy using rotating blades. Due to environmental conditions and large structure, the blades are subjected to various vibration forces that may cause damage to the blades. This leads to a liability in energy production and turbine shutdown. The downtime can be reduced when the blades are diagnosed continuously using structural health condition monitoring. These are considered as a pattern recognition problem which consists of three phases namely, feature extraction, feature selection, and feature classification. In this study, statistical features were extracted from vibration signals, feature selection was carried out using a J48 decision tree algorithm and feature classification was performed using best-first tree algorithm and functional trees algorithm. The better algorithm is suggested for fault diagnosis of wind turbine blade.     

变工况条件下的风电机组齿轮箱故障预警方法

针对风电机组运行工况复杂多变,难以实现故障特征提取和预警指标量化的特点,提出基于k邻近度异常检测技术的风电机组故障预警方法:首先利用阶比重采样技术将时域非平稳信号转换为角域的平稳或准平稳信号;其次构建出新量纲一幅域特征值,提取阶比重采样角域信号早期故障特征;最后将振动角域序列映射成多维特征向量,通过基于k邻近度的异常点检测技术挖掘机组潜在异常信息,实现机组的早期故障预警。试验仿真验证了该方法的有效性。     

基于云模型的风电机组输出功率特性分析

风电机组性能的优劣直接影响着风电场安全生产和经济效益。输出功率是风电机组最重要、最具代表性的性能指标之一,风功率曲线是机组发电能力最直观的表述。以输出功率和风速为数据源,应用云模型特征量研究风电机组输出功率的波动特性,有利于掌握风电机组性能状态。在对风电机组SCADA系统风速、功率数据筛选的基础上,描绘风电机组正常工作状态下的风功率散点图,采用比恩法建立风电机组实际风功率曲线;统计分析不同风速区间的输出功率,利用逆向云发生器建立不同风速下的输出功率云模型,得到不同机组的整体功率云;通过对比分析功率云的特征值,实现输出功率大小、波动范围和离散程度的量化分析;同时计算风速、功率相关系数反映和评价机组响应的灵敏度。云模型的应用,把机组状态从定性评价拓展到定量评价,从宏观综合评价深入到风速区间段精准评价,提高了风电机组性能分析的准确性和全面性。最后,应用实例验证了算法的有效性和可靠性。     

State variable technique islanding detection using time-frequency energy analysis for DFIG wind turbine in microgrid system

This paper introduces a new combined technique for wind turbine islanding detection using the trajectory of state variables and wavelet transform in microgrid system. The proposed relay is utilized of energy variation state of time-frequency transform coefficients of local signals in two-dimensional space. In order to improve of the proposed relay performance, a signal selection method based on the correlation concept between islanding and non-islanding signals. From of all patterns, the best of them with high correlation in islanding status is selected for the proposed relay learning. A neuro-fuzzy machine is used as learning of selected patterns to avoid threshold selection. The proposed technique is utilized to wind turbine islanding detection in a microgrid system contains various types of distributed generation including wind turbine, Combined Heat and Power (CHP) and photovoltaic system. Many islanding and non-islanding situation including motor starting, various load and capacitor switching in various operation conditions in the studied microgrid are simulated. Millstone characteristics of the proposed technique are included passive-based technique, negligible None Detection Zone (NDZ), suitable for microgrid application, performance increment in detection and fast detection time. Operation results of the proposed relay in various conditions confirm the performance of the proposed detection relay.     

基于动态特征矩阵的k近邻风电机组故障检测方法

受风的间歇性和随机性影响风电机组运行状态频繁切换,导致设备状态异常检测误报和漏报情况严重,风电企业运维成本居高不下。为此,提出了基于动态特征矩阵的k近邻故障检测方法,该方法采用基于互信息的动态特征矩阵描述风电机组的动态特性,通过加权k近邻同时考虑动态特征矩阵中的特征贡献率与累计互信息的影响,利用动态阈值计算降低运行状态突变造成的误报。分别以美国可再生能源实验室5 MW海上风机基准模型的常见传感器和执行器故障以及SCADA数据中记录的变桨系统故障为例,将所提方法的故障检测结果分别与PCA、KPCA、FD-kNN以及PC-kNN故障检测方法进行对比,结果表明所提方法能够准确进行故障信息的检测,所提方法优于其他对比故障检测方法。     

关联维数分析方法在液压泵多故障诊断中的应用研究

针对液压泵故障信号非线性和非平稳性特征,提出了利用相空间重构技术和分形理论相结合的特征关联维数提取方法。该方法将液压泵不同故障模式下获取的一维振动信号重构到高维相空间,进行信息深层挖掘;通过对相空间特征信号关联维数变化规律的分析,找出对故障反映敏感的关联维数,由此进行故障识别。通过实验验证,该方法提取的关联维数能有效反映液压泵的故障特征,为液压泵多故障诊断方法的研究提供可靠的特征信息,具有良好的应用前景。     

基于变分模态分解的风电齿轮箱轴承故障诊断方法

双馈异步风力发电机采用变转速变桨距的控制策略以保持风力最大功率捕获,风电齿轮箱时刻处于变速变载的恶劣工况,其关键部件极易受到损伤.针对齿轮箱轴承故障特征易受到风机变工况干扰的问题,提出了一种变分模态分解与瑞利熵相结合的特征分析方法,实现对风电齿轮箱高速轴轴承健康状态系数的估计.本文以双馈异步风机齿轮箱高速轴轴承作为研究...     

变风速条件下风力机轴承故障特征提取

针对风速变化条件下风力发电机轴承故障特征的检测问题,提出了一种基于灰狼优化( GWO )和双稳态杜芬振荡器的随机共振( SR )的故障特征提取方法.首先,根据风速估计故障特征信号的频率,通过合适的采样频率采集风力发电机的振动信号并对采集的信号做归一化处理.随后,根据风速尺度引入变换系数对频率 时间尺度进行变换.此外,利用灰狼算法方法将杜芬振子的阻尼比和系统参数调整到最优值.最后,通过杜芬系统和尺度恢复获得可识别信号.结果表明,所提出的方法能提取原始信号中的故障特征信号.     

基于LabVIEW的双馈风力发电机组在线谐波分析研究

针对双馈风力发电机组在并网运行时,其电力电子变流装置将向电网注入大量谐波电流的问题,将基于LabVIEW的在线谐波分析技术应用到双馈风力发电机组的谐波监测分析中。开展了针对双馈风力发电机组注入电网谐波的傅里叶分析和小波包分析,提出了将傅里叶变换与小波包变换联合使用的方法;在LabVIEW上搭建了仿真平台,对双馈风力发电机组可能存在的谐波进行了仿真分析试验。试验及研究结果表明,对于平稳信号与非平稳信号进行谐波分析时,将傅里叶变换与小波包变换联合使用,两者的效能得到明显的提高,优于单独使用傅里叶变换或小波包变换,且可更准确地获取基波和谐波的实时波形与频域特性,谐波定位更准确,为谐波补偿提供了更可靠、准确的信息。     

基于小波及非线性预测的轴承故障诊断方法

在非线性时间序列预测研究的基础上,提出非线性预测效果的特征提取方法.首先对采集到的足够长轴承数据采用小波变换进行消噪处理及边界延拓,使其满足预测需要的无限长、无噪声的条件,这样延迟时间取任意值均能重构原系统相空间;然后采用基于可预测性的选取嵌入维数的方法确定轴承各种状态信号的嵌入维数,进行相空间重构.应用实验结果表明:该方法提取的特征值能明显地区分轴承各种状态信号,且对数据分段长度的稳定性好,可以作为识别轴承故障的一种新途径.     

风力机叶片蒙皮初始裂纹的盲信号提取

针对风力机叶片初始裂纹特征难以提取的问题,提出了一种逐步提取并消减噪声源信号从而获得微弱裂纹故障特征的盲提取方法.首先基于卷积混合模型极小化改进代价函数推导自适应学习迭代算式,在仿真实验中确定非线性激励函数和滤波器的传输函数,根据输出信号的性能参数证明了改进算法对尖峰噪声的异常点更加敏感稳健.在风力机叶片疲劳实验台上模拟叶片蒙皮的初始横向裂纹,通过声发射信号采集系统获得观测信号,分析噪声源的特性并提取了初始裂纹的声发射信号特征,为风力机叶片状态监测和预警提供了依据.     

垂直轴风力机气动性能分析与数值模拟研究

先采用叶素理论对垂直轴风力机叶片受力进行静态的分析,然后采用多流管理论计算出叶尖速比与功率的关系曲线,最后采用滑移网格技术对风力机进行了数值模拟,得到的功率与用多流管理论计算的结果吻合较好.     

Automatic validation of vibration signals in wind farm distributed monitoring systems

A rapid growth of wind farms as a leading renewable energy sector has compelled a number of companies to develop dedicated distributed systems of monitoring and diagnosis (SM&D). Such systems are capable of early mechanical faults detection, which prevents from costly critical repairs. Fault detection of wind turbines is based on vibration and process signals analysis. Modern SM&D are usually advanced hardware and software technological products, which on the basis of collected data are capable of continuous execution of enormous number of analyses. However, the expansion of distributed systems has disclosed new challenges accompanying the transition from a singular signal analysis to automatic interpretation of large data sets. The paper deals with the assessment of data acquisition final products, namely the correctness of vibration signals recording process. The proposed validation, which is to be implemented a priori to data analysis, is required for significant calculation of any signal features as well as for legitimate storage of raw waveforms in the system database. The latter aspect is especially important in terms of gigabytes of disk space frequently wasted for corrupted data. The paper presents a step-by-step method for automatic signal validation, which can be implemented on a personal computer. Finally, the algorithm is evaluated on a real wind turbine database.     

Site calibration for the wind turbine performance evaluation

The accurate wind speed information at the hub height of a wind turbine is very essential to the exact estimation of the wind turbine power performance testing. Several methods on the site calibration, which is a technique to estimate the wind speed at the wind turbine’s hub height based on the measured wind data using a reference meteorological mast, are introduced. A site calibration result and the wind resource assessment for the TaeKwanRyung test site are presented using three-month wind data from a reference meteorological mast and the other mast temporarily installed at the site of wind turbine. Besides, an analysis on the uncertainty allocation for the wind speed correction using site calibration is performed.     

基于有限元结构动力学模型的塔架振动分析

为深入了解风机塔架的动态特性,并解决运行过程中的塔架振动问题,研究了塔架有限元动力学方程的建模方法,建立了塔架动力学方程; 依据动力学方程,研究了塔架模态的求解方法,得到了塔架基础刚度、塔架质量及塔架刚度对塔架模态的影响趋势,为风机塔架制造、安装过程提供理论参考依据。并依据动力学方程,研究了塔架动力学相应的求解过程,进而设计了基于塔架激励减缓的多目标变桨控制方法,仿真和风场实验结果表明,此方法可有效减小塔架振动告警,减小风机停机次数,从而提高发电量。     

基于GRU-XGBoost的风电场功率短期预测

随着风电技术的不断发展,更多的风电机组并入电网运行。考虑到电网的安全性与稳定性,精确的风电场发电短期预测技术越发重要。在利用自适应噪声的完备经验模态分解(CEEMDAN)风电原始序列信号的基础上,采用GRU-XGBoost模型对非线性、非平稳的功率序列进行建模和预测,以提高模型的预测能力和泛化能力。首先利用CEEMDAN将风电功率原始序列分解为一系列不同时间尺度的分量,将分解后的信号输入GRU神经网络输出预测信号,再输入XGBoost进行校正。通过与多种预测模型进行比较证明此模型拥有更好的预测精度。     

堆栈稀疏自编码器的风力机锥齿轮故障诊断

针对当前齿轮故障诊断存在着准确性不高、主观性强等问题,提出了一种基于堆栈稀疏自编码器(SSAE)和softmax分类器相结合的齿轮故障诊断方法。首先,运用时域分析以及样本熵方法对风力机锥齿轮振动信号进行特征提取,其次,将提取的特征输入到SSAE中进一步学习目标数据的深层本质特征,并进行特征降维,最后使用softmax分类器中进行分类识别。通过实验结果表明,和文中其他浅层学习模型相比,SSAE能够从齿轮振动信号中有效学习到所需的深层本质特征,拥有更高的识别准确率,因而证实了该方法优越性。