一种基于非线性偏最小二乘的风电机组齿轮箱状态监测方法北大核心CSCD

风电机组状态监测是提升机组运行水平和经济效益的重要手段。文章提出了一种基于非线性偏最小二乘(PLS)的风电机组齿轮箱状态监测方法,利用数据采集与监控系统(SCADA)数据对齿轮箱油温进行建模和监测。首先,基于无监督聚类对SCADA数据进行预处理,利用相关性分析选取与齿轮箱油温相关的输入变量;然后,构建用于表征非线性关系的输入变量,建立正常运行工况下齿轮箱油温的非线性PLS模型;最后,根据模型输出结果与齿轮箱油温的残差分布,设置合理阈值,用于齿轮箱状态监测。应用该模型对某大型风电机组齿轮箱进行状态监测。监测结果表明,相比于BP神经网络模型,该模型具有更高的拟合优度和预测精度。     

某风电场1.5MW风电机组齿轮箱点蚀及断齿故障分析

正齿轮箱是风电机组的重要部件,叶轮转速通过齿轮箱增速后使转速达到并网转速,有利于减少并网风电机组发电机的级数,缩小发电机的体积,提高机组效率。齿轮箱的损坏直接关系到风电场运行成本和收益。因受交变载荷的作用,在齿轮箱损坏的故障现象中,齿轮箱失效的主要形式是轮齿折断和齿面点蚀、剥落等,下面结合某风电场状况,对齿轮箱损坏的原因进行分析。     

多参数模型风电机组叶片结冰监测与预警研究

详细分析叶片结冰对风电机组运行性能和运行参数的影响,采用功率、叶轮转速和环境温度作为监测叶片结冰的变量.采用高斯过程回归分别建立功率模型和叶轮转速模型实现2个参数的实时监测.引入序贯概率比检验方法分析功率和叶轮转速模型的预测残差以发现2个参数在叶片结冰时的异常变化.当风电机组功率异常、叶轮转速异常且环境温度在0℃附近这...     

基于温度参数的风电机组异常识别

文章针对风电机组运行过程中机组早期的异常状态识别问题,提出一种考虑有功功率的基于机组温度参数变化特性的风电机组异常识别模型。首先,分析风电机组各系统与温度相关的参数。然后,利用相关性理论,确定了与有功功率相关的温度参数:齿轮箱高速轴轴承前端温度、齿轮箱高速轴轴承后端温度、齿轮箱油温、发电机驱动端轴承温度、发电机非驱动端轴承温度、发电机定子绕组温度,形成了异常检测的参数体系。再次,以正常状态下机组温度参数的偏度和峰度的最大区间作为阈值,建立风电机组异常识别模型。最后,采用滑动窗口对机组运行状态进行在线监测。通过实例研究发现,当机组发生异常状态时,温度参数的偏度或者峰度超过了阈值,比警报提前了15 d。该识别模型为风电机组的早期故障预警提供了参考。     

基于IPSO神经网络的风电机组主轴状态监测

风电机组主轴是叶轮和齿轮箱的连接部分,在机组传动链中具有传递转矩和能量的作用,因此对主轴进行状态监测关系到风电机组的稳定性。将改进粒子群算法(IPSO)与BP神经网络相结合构造主轴温度模型并进行预测。当主轴发生故障时,模型输入的观测向量发生异常变化,导致模型预测残差发生改变。为提高主轴异常预警的灵敏度和可靠性,文中采用基于莱依特准则的双滑动窗口对预测残差序列进行实时的统计分析,如果残差均值或标准差超出设定的故障报警阈值时,发出报警信息。     

风电齿轮箱油温高原因分析及解决方法

针对某1.5 MW风电齿轮箱普遍存在油温高的问题,通过对齿轮箱润滑油冷却系统进行理论计算,分析与对比同机型不同风电场热交换器的实际冷却效果,分析与研究同台机组不同运行状态下的实际冷却效果,认为引起该风电齿轮箱润滑油温度高的因素有:翅片结构选型不适、机舱密封不严等因素引起的翅片堵塞问题,润滑油设计流量偏低的问题,以及温控阀失效、油温传感器失效、发电机水冷缺水等元件问题。在排除元件问题的基础上,通过改变散热器翅片结构,增加热交换器换热面积和增加换热器冷却风扇流量等方法,彻底解决了齿轮箱油温高问题。工程实际应用表明,该方法可在该系列齿轮箱油温高技术改造上推广使用,同时也可为后续产品技术升级提供参考依据。     

基于SCADA数据的风电机组齿轮箱状态监测方法

为解决故障劣化渐变过程的长时间序列对齿轮箱状态监测模型的影响问题,提升其决策精度,提出一种基于数据采集与监控(SCADA)数据的组合建模方法。首先,采用主成分分析法(PCA)选取与齿轮箱温度密切相关的输入观测向量,并应用长短期记忆(LSTM)神经网络分别对齿轮箱正常工况和异常工况独立建立温度模型;其次,结合模型输出结果与SCADA数据提取残差分布特征向量,建立随机森林残差分布模型对机组齿轮箱运行状态进行监测;最后,对某大型风电场机组进行模型建立和仿真研究。结果表明,基于LSTM神经网络结合随机森林算法对风电机组齿轮箱状态监测有较强的实用性和较高的准确率,为后续开展齿轮箱健康度评价提供了新的方法和思路。     

风电齿轮箱油温高原因及解决方法

风电齿轮箱油温过高是一个常见的问题,可能导致油液老化和齿轮箱寿命的缩短。造成油温过高的主要原因有工作负载过大、过热的环境条件和油液循环不畅等。本文将主要研究风电齿轮箱油温高原因及解决方法,提出部分解决方案可以有效地降低齿轮箱油温,延长齿轮箱的寿命,提高风电设备的可靠性和效率。     

基于CAE与BiLSTM结合的风电机组齿轮箱故障预警方法研究

针对风电机组故障频发且早期故障监测难的问题,为实现风电机组智能监测,提出基于卷积自编码(CAE)与双向长短期神经网络(BiLSTM)的风电机组齿轮箱故障预警方法。首先对风电场数据采集与监视控制(SCADA)系统的数据进行预处理,选择能表征风电机组齿轮箱运行状态的监测量作为输出量,根据相关性分析选择与输出量相关性高的监测量作为输入参数;然后根据特征选择特性和参数非线性特性构建深度学习网络模型,对输出的预测值和残差进行统计分析,设置自适应阈值来监测风电机组异常状态的趋势变化。将CAE-BiLSTM模型应用于某风电场的算例分析中,并与其他模型的预测效果进行对比。结果表明：该方法解决了模型输入与结构冗余问题,提高了模型精度,能够有效预警风电机组齿轮箱早期故障。     

风电齿轮箱

风力发电机组中的齿轮箱是一个重要的机械部件，其主要功能是将风轮在风力作用下所产生的动力传递给发电机，使其得到相应的转速。通常风轮的转速很低，远达不到发电机发电所要求的转速，必须通过齿轮箱齿轮的增速作用来实现，故也将齿轮箱称之为增速箱。根据机组的总体布置要求，有时将与风轮轮毂直接相连的传动轴（俗称大轴）与齿轮箱合为一体，也有将大轴与齿轮箱分别布置，其间利用涨紧套装置或联轴节连接的结构。为了增加机组的制动能力，常常在齿轮箱的输入端或输出端设置刹车装置，配合叶尖制动（定浆距风轮）或变浆距制动装置共同对机组传动系统进行联合制动。     

基于网格搜索优化ERF模型的风电机组异常状态预警

提出一种基于网格搜索优化(GS)极端随机森林(ERF)模型的风电机组性能预测及异常状态预警方法.首先,采用离散度分析法清洗噪声和异常工况数据,以获取建模用正常运行状态数据.其次,通过分析风机运行与控制原理,选取与转速和功率具有较高相关度的特征参数作为模型输入,完成预测模型训练和验证,并对比ERF模型与其它几种模型的建模...     

Determination of the combined vibrational and acoustic emission signature of a wind turbine gearbox and generator shaft in service as a pre-requisite for effective condition monitoring

A review of current progress in Condition Monitoring (CM) of wind turbine gearboxes and generators is presented, as an input to the design of a new continuous CM system with automated warnings based on a combination of vibrational and Acoustic Emission (AE) analysis. For wind turbines, existing reportage on vibrational monitoring is restricted to a few case histories whilst data on AE is even scarcer. In contrast, this paper presents combined vibration and AE monitoring performed over a continuous period of 5 days on a wind turbine. The vibrational and AE signatures for a healthy wind turbine gearbox and generator were obtained as a function of wind speed and turbine power, for the full normal range of these operational variables. i.e. 5–25 m/s and 0–300 kW respectively. The signatures have been determined as a vital pre-requisite for the identification of abnormal signatures attributable to shaft and gearbox defects. Worst-case standard deviations have been calculated for the sensor data. These standard deviations determine the minimum defect signal that could be detected within the defined time interval without false alarms in an automated warning system.     

海上风电结构横风向气动力阻尼特性研究

基于横风向气动力阻尼理论计算模型,以NREL-5 MW海上风电机组为例,对其运行过程中横风向气动力阻尼进行计算,并采用FAST软件对计算结果进行验证。之后,研究转速、叶片桨距角和运行方式对横风向气动力阻尼的影响。研究结果表明：NREL-5 MW海上风电机组结构运行状态下的横风向气动力阻尼在0%~0.8%范围内变化,其随风电机组运行转速及叶片桨距角的增大而增大;此外,海上风电机组不同运行方式对其横风向气动力阻尼也会产生较大影响。     

Experimental identification of modal parameters of an industrial 2‐MW wind turbine

This contribution presents modal testing of a 2‐MW wind turbine on a 100‐m tubular tower with a 93‐m rotor developed by W2E Wind to Energy GmbH. This research is part of the DYNAWIND project of the University of Rostock and W2E. Beside classical modal analysis schemes, this contribution mainly focusses on the application of operational modal analysis techniques to a wind turbine. Specific problems are addressed, and hints for modal testing on wind turbines are given. Furthermore, an effective measurement setup is proposed for identification of the modal parameters of a wind turbine. The measurement campaign is divided in two parts. First, a measurement campaign using 8 sensor positions on a rotor blade was done while the rotor is lying on ground. Second, a detailed measurement campaign was done on the entire wind turbine with the rotor locked in Y position using 61 sensor positions on the tower, the mainframe, the gearbox, the generator, and the low‐voltage unit. While the rotor blade was tested by classical and operational modal analysis techniques, the entire wind turbine was tested by operational modal analysis techniques only. The mode shapes and eigenfrequencies of the wind turbine identified within the measurement campaigns are within the expected range of the design values of the wind turbine. But in contrast, the damping ratios differ strongly from those given in guidelines and literature. Furthermore, a strong influence of aerodynamic damping compared to structural damping is observed for the first tower mode even for a parked wind turbine.     

基于油液光谱监测的风电机组齿轮箱磨损状态的研究

风电机组齿轮箱的磨损微粒主要是铁颗粒,铁颗粒含量的增长趋势能直接反映出风电机组齿轮箱的磨损状态.以Spectro油液光谱分析仪监测风电机组齿轮箱在用齿轮油中的铁元素含量,通过一段时间内铁元素的增加量和风电机组可利用小时数,可计算得到单位可利用小时数下的铁元素增加量ΔQFe;引入可靠性理论研究了ΔQFe的分布规律,并以风...     

Monitoring wind turbine gearboxes

Concerns amongst wind turbine (WT) operators about gearbox reliability arise from complex repair procedures, high replacement costs and long downtimes leading to revenue losses. Therefore, reliable monitoring for the detection, diagnosis and prediction of such faults are of great concerns to the wind industry. Monitoring of WT gearboxes has gained importance as WTs become larger and move to more inaccessible locations. This paper summarizes typical WT gearbox failure modes and reviews supervisory control and data acquisition (SCADA) and condition monitoring system (CMS) approaches for monitoring them. It then presents two up‐to‐date monitoring case studies, from different manufacturers and types of WT, using SCADA and CMS signals. The first case study, applied to SCADA data, starts from basic laws of physics applied to the gearbox to derive robust relationships between temperature, efficiency, rotational speed and power output. The case study then applies an analysis, based on these simple principles, to working WTs using SCADA oil temperature rises to predict gearbox failure. The second case study focuses on CMS data and derives diagnostic information from gearbox vibration amplitudes and oil debris particle counts against energy production from working WTs. The results from the two case studies show how detection, diagnosis and prediction of incipient gearbox failures can be carried out using SCADA and CMS signals for monitoring although each technique has its particular strengths. It is proposed that in the future, the wind industry should consider integrating WT SCADA and CMS data to detect, diagnose and predict gearbox failures.Copyright © 2012 John Wiley & Sons, Ltd.     

基于深度卷积自编码器的风电机组故障预警方法研究

通过风电机组状态监测进行故障预警,可防止故障进一步发展,降低风场运维成本。为充分挖掘风电机组监控与数据采集（SCADA）各状态参数时序信息,以及不同参数之间的非线性关系,该文将深度学习中自动编码器（AE）与卷积神经网络（CNN）相结合,提出基于深度卷积自编码（DCAE）的风电机组状态监测故障预警方法。首先基于历史SCADA数据离线建立基于DCAE的机组正常运行状态模型,然后分析重构误差确定告警阈值,使用EMWA控制图对实时对机组状态监测并进行故障预警。以北方某风电场2 MW双馈型风电机组叶片故障为实例进行实验分析,结果表明该文提出DCAE状态监测故障预警方法,可有效对机组故障提前预警,且优于现有基于深度学习的风电机组故障预警方法,可显著提升重构精度、减少模型参数和训练时间。     

基于长短时记忆网络融合SCADA数据的风电齿轮箱状态监测

针对不具有时间记忆能力的机器学习方法融合风电机组数据采集与监控系统(SCADA)的时序数据而导致风电齿轮箱状态预测精度不高的问题,提出基于长短时记忆(LSTM)网络融合SCADA数据的风电齿轮箱状态预测模型。选择能表征风电齿轮箱运行状态的某个监测量作为模型的输出量,基于灰色关联度选择与该监测量关联密切的SCADA参数作为预测模型的输入量;使用正常状态下的SCADA数据训练LSTM预测模型,得出预测值和残差,通过3σ准则计算出上下预警阈值,用于风电齿轮箱状态监测和故障预警。某风电场风电齿轮箱的SCADA数据验证表明所提出的方法能有效预警风电齿轮箱故障。     

Effects of wind shear on wind turbine rotor loads and planetary bearing reliability

Recent studies suggest that wind shear and the resulting pitch moments increase bearing loads and thereby contribute to premature wind turbine gearbox failure. In this paper, we use momentum‐based modeling approaches to predict the pitch moments from wind shear. The non‐dimensionalized results, which have been validated against accepted aeroelastic results, can be used to determine thrust force, pitch moment and power of a general rotor as a function of the wind shear exponent. Even in extreme wind shear (m = 1), the actual thrust force and power for a typical turbine (R* < 0.5) were within 8% and 20% of the nominal values (those without wind shear), respectively. The mean pitch moment increased monotonically with turbine thrust, rotor radius and wind shear exponent. For extreme wind shear (m = 1) on a typical turbine (R* = 0.5), the mean pitch moment is ~25% the product of thrust force and rotor radius. Analysis of wind shear for a typical 750 kW turbine revealed that wind shear does not significantly affect bearing loads because it counteracts the effects of rotor weight. Furthermore, even though general pitch moments did significantly increase bearing loads, they were found to be unlikely to cause bearing fatigue. Analyses of more common low wind‐speed cases suggest that bearing under‐loading and wear are more likely to contribute to premature bearing failure than overloading and classical surface contact fatigue. Copyright © 2015 John Wiley & Sons, Ltd.