风力机风轮结构振动及稳定性的概率分析方法研究

With the continuously developing of wind turbine towards large-scale and light weight, how to ensure the rotor with long soft blades having the well behaved dynamic characteristic is a vital challenge. It is a requirement to develop the refined dynamic design method for the advanced rotor structure design. This project is to research the probability analysis method for vibration and stability of wind turbine rotor, including the probability analysis method of vibration mode, the probability prediction method of resonance and the fuzzy probability prediction method considering the fuzzy of resonance criterion, and the probability analysis method of aeroelasitc stability, the probability prediction method of flutter and the fuzzy probability prediction method considering the fuzzy of flutter criterion for the wind turbine rotor. Based on the analysis of vibration and stability of the mistuned rotor, the vibration localization phenomenon of mistuned rotor is to be revealed and the influence of mistune to the aeroelastic stability for the rotor to be studied. Based on the substructure decomposition and the discipline decomposition, the multilevel surrogate model method is to be developed that can improve the accuracy and efficiency of the probability analysis of the multidiscipline complex structure. The research results in this project have the academic value to the development of the structural dynamic reliability design theory of complex structure and practical application prospect for the vibration and stability design of the wind turbine rotor.

随着风力机不断向大型化和轻量化的方向发展，对如何保证长柔叶片风轮具有良好的动态特性提出了严峻的挑战，需要不断发展更加精细化的动态设计方法，以满足日益先进的风轮结构设计的需要。本项目研究风轮结构振动及稳定性的概率设计分析方法，包括风轮结构振动模态概率分析、共振的概率预测和考虑共振判据模糊性的模糊概率预测方法；风轮气弹稳定性概率分析、颤振的概率预测和考虑颤振判据模糊性的模糊概率预测方法。通过失谐风轮结构振动及稳定性分析，揭示失谐风轮结构振动局部化现象，探索失谐对风轮气弹稳定性的影响。通过建立基于子结构分解和基于学科分解的多层次代理模型方法，显著提高多学科复杂结构概率分析的计算效率和精度。本项目研究成果将对发展复杂结构动态可靠性设计理论具有重要的学术意义，对风力机风轮结构的振动及稳定性设计具有现实的应用前景。

风电机组在不断大型化和轻量化过程中，长柔叶片面临共振和颤振问题，针对极端工况下风轮动态特性中存在的不确定性，本项目以代理模型概率分析方法为研究核心，综合考虑风轮结构振动及颤振的概率设计，发展了更加精细化的风力机风轮动态设计方法。考虑叶片截面翘曲、弯扭耦合、旋转效应的影响，提出一种同时适用于风轮共振和颤振分析的梁单元有限元模型，实现风轮共振和颤振分析高精度样本计算，为提高代理模型的精度提供基础；提出一种结合克里金代理模型的高效率动态采样方法，通过动态迭代采样，使得在极限状态附近生成更多的采样点，在保证代理模型精度同时最大限度减少采样点数目；考虑共振和颤振避开准则的模糊性，提出了基于相对熵的风轮结构共振和颤振的模糊概率预测方法，研究了颤振可靠度对随机参数的敏感性，为风轮动态设计提供方法和依据；考虑风轮参数的随机性，推导了叶片振动方程的一阶摄动格式，通过失谐风轮结构振动分析，揭示失谐风轮结构振动局部化现象；研究了台风工况下风轮载荷概率特性，基于子结构分解和基于学科分解的多层次代理模型，提出了风轮抗台风可靠度代理模型概率分析方法，显著提高多学科复杂结构概率分析的计算效率和精度。本项目在风轮共振和颤振代理模型概率分析方法研究方面取得了一定的成果，所研究的方法对复杂结构动态概率设计方法发展具有一定的学术意义，对大容量风力机风轮结构动态概率设计具有重要的实用价值。