运行状态下海上风机支撑结构模态参数识别方法研究

Offshore wind turbine's support structure include the tower and foundation，whose state directly affects the operation of the wind turbine. Modal parameters is one of the most important parameters represent the health condition of the offshore wind turbine's support structure. Due to strong variation of modal parameters and complex excitation, the support structure does not meet the requirements of traditional modal analysis methods, in order to solve the problem, this project propose that extract the modal parameters of the support structure based on time-varying model. Oriented by solving the key problem, this project take time-varying subspace as the theoretical basis, adopt the reliable identification of the support structure as project goal, a novel modal parameters identification method for operating offshore wind turbine's support structure based on time-varying subspace is proposed. The research stratagem of this proposed method is as follows: 1) A novel method for free response extraction using various response components based blind source separation and hybrid feature recognition will be proposed, so as to achieve free response extraction of the structure. 2) Based on time-varying subspace with multiple types of data, a novel method will be proposed to achieve parameter identification under single data set of time-varying subspace. 3) A novel method of system's order setting based on excluding random pole and picking up clustering poles will be proposed to automatically set time-varying subspace's order. 4) Finally, the novel modal parameters identification method for operating offshore wind turbine's support structure will be developed and the method's performance will then be evaluated. These research have important theoretical and economic values not only for ensuring steady and reliable operation of offshore wind turbine, but also for the enrichment and development of the experimental modal analysis theory and technology.

海上风机支撑结构包括塔架和基础，其状态直接影响风机安全运行。模态参数是反映支撑结构健康状态的重要参数之一。支撑结构在风机运行状态下时变性强、所受激励复杂，不满足传统模态分析方法应用条件，项目针对该问题提出在时变模型基础上对支撑结构进行分析。以时变子空间为基础理论，以实现支撑结构模态参数可靠识别为目标，以解决关键问题为导向，研究运行状态下海上风机支撑结构的模态参数识别新方法。研究思路：提出基于多响应成分盲源分离和混合特征辨识的自由响应提取方法，实现结构自由响应提取；提出基于多类型数据时变子空间的模态参数识别方法，实现单组响应数据下时变子空间模态参数识别；提出随机极点剔除结合聚类极点提取的定阶方法，实现时变子空间自动定阶；最后集成实现运行状态下海上风机支撑结构模态参数识别方法进行系统评估。该方法对保证在役海上风机的稳定可靠运行，丰富和发展模态参数识别理论和技术，具有重要的理论和经济价值。

海上风机支撑结构是风机的基础结构，其损坏将直接影响风机正常运行。发展海上风机支撑结构健康监测技术对降低海上风电运营成本、延长海上风机使用寿命以及保证海上风电安全运营具有重要意义。模态参数是反映支撑结构健康状态的重要参数之一。海上风机支撑结构在风机运行状态下时变性强、所受激励复杂，不满足传统模态分析方法应用条件，项目提出了在时变子空间模型基础上对支撑结构进行模态参数识别。通过项目研究，完成以下工作：（1）提出了基于多响应成分盲源分离和混合特征辨识的自由响应提取方法，实现结构自由响应提取。利用盲源分离对海上风机支撑结构响应进行解耦，结合不同响应在统计特性、时频特性等形成混合特征指标提取运行状态下海上风机支撑结构自由响应；（2）提出了基于多类型数据时变子空间的模态参数识别方法，实现单组响应数据下时变子空间模态参数识别。研究不同类型响应在状态空间域的表征方式，利用不同类型响应的映射关系，基于时变子空间建立不同类型响应之间的映射模型，设计基于时变子空间的系统矩阵识别模型，实现时变模态参数识别；（3）提出随机极点剔除结合聚类极点提取的定阶方法，实现时变子空间自动定阶。利用物理模态与虚假模态的极点分布不同形成时变子空间虚假模态剔除方法，利用聚类分析方法实现物理极点自动拾取，实现自动定阶。该项目的研究对于丰富和发展海上风机支撑结构模态参数识别理论，保证海上风电稳定可靠运行，具有重要意义。