【作者】 刘强；

【导师】 高峰；

【作者基本信息】 华北电力大学(北京) ， 控制工程（专业学位）， 2018， 硕士

【摘要】 由于风电机组大多位于环境恶劣的地方,所以风机故障率居高不下,其故障的原因和特征也是各不相同的。风电机组的变桨系统以及叶片的各类故障在风电机组所有的故障中占比非常高,实际风场大多采用SCADA数据进行变桨系统故障诊断,这种方法易于在实际中实施,但故障阈值设定单一,准确性有待提高;而风电机组叶片损伤故障在实际机组故障诊断中没有较好的方法,大多采用观察法、图像处理法、声发射处理法等,这些方法存在一定的局限性,对于机组的运行状态以及环境条件有着较高的要求并且需要大量的人工作业,不利于应用推广。因此,对于风电机组变桨系统和叶片损伤故障诊断方法的研究有着重要的意义。本文对当前风电行业的发展状况以及现在风电机组变桨系统和叶片的损伤故障诊断方法进行了介绍,通过GH Bladed软件模拟实际机组运行状态,进而模拟出变桨系统和叶片的一些常见故障,并取得相关运行数据。对于变桨系统的各类故障,本文重点介绍了基于支持向量机的故障诊断方法,并采用粒子群算法对其进行了优化处理,优化后的诊断模型判别精度更高。对于叶片损伤故障,本文通过GH Bladed软件模拟叶片某单元损伤故障,然后首先对叶尖挠度进行小波包分解,得到各个频带的能量谱,基于能量谱的变化,初步对叶片损伤进行判别,在叶片发生损伤的情况下,利用模态应变能变化率法确定叶片具体损伤单元。实验结果表明上述两类研究方法可以准确、快速地对风电机组变桨系统故障和叶片损伤作出判别诊断。更多还原

【Abstract】 Since wind turbines are mostly located in harsh environments,the failure rate of wind turbines is high,and the causes and characteristics of the failures are also varied.The faults of the pitch system and the blade are very high of all the failure in the wind turbine.Most of the actual wind farms use SCADA data to diagnose pitch system faults.This method has many advantages in practical implementation and stability,but the fault threshold is set to be single and the accuracy needs to be improved.The failure of wind turbine blades in the actual unit fault diagnosis do not have good methods,most of the observation,image processing,acoustic emission treatment method,these methods have some limitations,for it needs good unit operating and environmental conditions and requires a lot of manual work,so it is not conducive to the application of promotion.Therefore,it is of great significance to study the pitch system and the blade failure diagnosis method.This paper introduces the current situation of the wind power industry and the present fault diagnosis methods for the wind turbines’ pitch systems and blades,and simulates the common faults of the pitch system and the blades by GH Bladed software to simulate actual unit operation.Obtain relevant operating data.For all kinds of faults of pitch system,this paper mainly introduces the fault diagnosis method based on support vector machine,and uses particle swarm optimization algorithm to optimize it.The optimized diagnosis model has higher discrimination accuracy.In this paper,the GH Bladed software is used to simulate the damage of a certain element of the blade.Then,the energy spectrum of each band is obtained by wavelet packet decomposition of tip deflection.Based on the change of energy spectrum,the damage of the blade is preliminarily determined,In the event of damage,the specific damage cell of the blade is determined using the modal strain energy change rate method.The experimental results show that the above two methods can accurately and quickly judge the failure of the wind turbine pitch system and blade damage.更多还原