摘要
针对风电机组塔架振动信号存在的非平稳特征,为了能依据塔架振动信号得到准确的塔架相对位移数据,提出了基于Ho⁃drick-Prescott算法的时域积分优化方法,并进行了基于风电机组设计仿真软件Bladed的数据分析和工程实际数据的验证。研究表明,基于傅里叶变换的频域积分不能提供准确相位,不适用于该非平稳工况;而相比采用多项式去趋势的时域积分,因其消除了高噪声产生的低频摆动趋势,其平方和误差指标可达3%~8%,适用于风电机组塔架结构系统的分析和诊断。
In view of the non-stationary working conditions of wind turbine,in order to obtain accurate relative displacement data according to vibration signals,an improved time domain integral method based on Hodrick-Prescott algorithm was proposed,and the simulation analysis based on wind turbine design simulation software bladed and the verification of field actual data were carried out.The research shows that the frequency domain integral based on Fourier transform can not provide correct phase and is not suitable for the non-stationary condition,while compared with the polynomial detrending time domain integral,it eliminates the low frequency swing trend caused by high noise,and its error evaluation index Ersq can reach 3%~8%,which is suitable for the analysis and diagnosis of wind turbine tower structure system.
作者
连而锦
Lian Erjin(Mingyang Smart Energy Group,Zhongshan,Guangdong 528400,China)
出处
《机电工程技术》
2021年第10期190-193,共4页
Mechanical & Electrical Engineering Technology
关键词
振动加速度
相对位移
非平稳
数字积分
时域积分
频域积分
vibration signal
relative displacement
non-stationary
digital integral
time domain integration
frequency domain integration
