基于模态识别和数据驱动的海上风电健康监测方法研究

The health monitoring of offshore wind power structure is indispensable to achieve the large-scale commercial operation, and also the main restriction of the development of offshore wind turbines. The difficulties of the structural health monitoring includes the complexity of external load, the time variation of structure parameter, the difficulty of modeling, the limitation of sensing position, the diversity of operating conditions and the insufficiency of effective experience. The onshore wind turbine health monitoring method is not mature, not to solve the special problems of offshore wind turbines. But the sensor network technology provides a new technical means for the offshore wind turbine structure health monitoring. The data-driven based intelligent data analysis method offers a new idea for the real-time damage diagnosis and long-running data extraction. This project explores the intelligent detection method of offshore wind park which is taking the offshore wind parks as object and the sensor network as technical means and synthesizing the physical model and statistical model analysis. In this paper, we develop a new mode of operation analysis to solve the problem of modal identification of time-varying systems with externally excited non-Gaussian distributions of offshore wind turbines, research on the damage detection of offshore wind power which is based on limit learning machine and explore the intelligent optimization of virtual sensing technology and sensor layout.This study will provide the theoretical basis and technical support for the construction of offshore wind power structure health monitoring system.

海上风电结构健康监测是实现大规模商业化运行的必要保障，也是制约我国海上风电发展的关键问题。海上风机结构健康监测的难点在于外部载荷复杂、结构参数时变、模型难以建立、传感位置受限、运行工况多样和有效经验不足等，而陆地风机的健康监测方法无法解决海上风电所面临的特殊问题。传感网技术为海上风机结构健康监测提供了新的技术手段，基于数据驱动的智能化数据分析方法为实时损伤诊断和长期运行的数据中提取特征提供了新的思路。本项目以海上风电结构为对象，以传感器网络为技术手段，综合模态分析方法和基于数据驱动的方法，研究海上风电结构健康监测的智能方法。其中包括研究运行模态分析方法以解决海上风机这类具有外部激励非高斯分布的时变系统的模态识别问题；研究基于极限学习机的海上风电结构损伤检测方法；探索基于有限元模型的虚拟感知技术及传感器布局智能优化。本研究将为构建海上风电结构健康监测系统提供理论基础和技术支撑。

本研究以海上风电结构为对象，以传感器网络为技术手段，综合模态分析方法和基于数据驱动的方法，研究了海上风电结构健康监测的智能方法。其中包括研究了运行模态分析方法，解决了海上风机这类具有外部激励非高斯分布的时变系统的模态识别问题；研究了基于极限学习机的海上风电结构损伤检测方法；探索了基于有限元模型的虚拟感知技术及传感器布局智能优化方法。为构建海上风电结构健康监测系统提供理论基础和技术支撑。