海上风机结构工作模态识别的组合降噪方法  

A Compound De-Noising Method for Operational Modal Identification of Offshore Wind Turbine Structure

作　　者：董霄峰[1] 练继建[1] 杨敏[1] 王海军[1] 

机构地区：[1]天津大学水利工程仿真与安全国家重点实验室,天津300072

出　　处：《天津大学学报（自然科学与工程技术版）》2015年第3期203-208,共6页Journal of Tianjin University：Science and Technology

基　　金：国家高技术研究发展计划(863计划)资助项目(2012AA051702);国家创新研究群体科学基金资助项目(51021004);天津市应用基础及前沿技术研究计划(青年基金项目)资助项目(12JCQNJC04000)

摘　　要：海上风机结构现场原型观测所获取的振动加速度信号中,往往混有大量的电磁工频、机组转频及环境噪声等成分,这些干扰信号处理不当易导致模态识别失真与产生虚假模态等问题.针对这一问题提出了基于自适应滤波与集成经验模态分解(EEMD)法的组合降噪方法.通过对测试信号采取滤波、分解、降噪及重组等过程可以更有效地降低各种噪声干扰,以完整地保留结构自身的振动信息;再利用随机减量法(RDT)和特征系统实现法(ERA)完成模态信息的初步识别,并结合振型分析等手段剔除虚假模态,实现准确提取海上风机结构工作模态参数的目的.该方法相比传统的降噪方法在噪声统计特征不足情况下具有更好的降噪精度和降噪效率,对基于实测信号获取风机结构工作模态信息有很好的工程应用价值.To solve the problem in identifying modal parameters accurately and retraining false modes induced by severe disruption due to the electromagnetic frequency signals,rotational frequency signals and environmental noise usually mixed in the acceleration data observed from prototype test of offshore wind turbine structure,a compound de-noising method,combining self-adaptive filter and ensemble empirical mode decomposition,was put forward to achieve filtering,noise reduction and modal identification for the first time. By means of testing signal processing in the new way including filtering,decomposing,de-noising and refactoring,much noise interference was reduced and full information of the structural vibration was reserved. Then the modal information will be identified preliminar-ily using both random decrement technique and eigensystem realization algorithm,and the false modes were elimi-nated by vibration mode analysis in order to obtain the operation modal parameters of offshore wind power structure correctly. Besides,the method presented here not only shows better accuracy and higher efficiency in noise reduction compared to the classic approaches which are lack of the statistical characteristics of noise,but also provides opera-tional modal identification of wind turbine structure with a preferable applied value in engineering based on measured signals.

关 键 词：海上风电 模态识别 自适应滤波 集成经验模态分解 振型分析 

分 类 号：TK8[动力工程及工程热物理—流体机械及工程]