基于多维核域谱的风机传动链系统微征兆健康状态识别方法

项目名称： 基于多维核域谱的风机传动链系统微征兆健康状态识别方法

项目编号： No.51505202

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 机械、仪表工业

项目作者： 刘文艺

作者单位： 江苏师范大学

项目金额： 21万元

中文摘要： 传动链系统是风力发电机组中的重要组成部分，其早期微征兆健康状态的识别直接影响风机的安全稳定运行和风电场电网的平稳输出供应。项目以风机传动链系统为主要研究对象，研究在复杂工况多维健康状态下风机传动链系统的微征兆健康状态识别方法。研究内容主要包括：1）探讨风机传动链系统干扰成份机理，提出交叉优化MHW的风机传动链系统微弱特征提取方法，解决强烈背景干扰下的微弱状态特征提取难题；2）风机传动链系统的微征兆健康状态识别方法：提出多维核域谱方法，解决样本数量较少、多维健康状态耦合时的早期微征兆健康状态识别难题。3）风机健康状态模拟试验台的完善：通过试验台验证项目方法的有效性，同时通过项目方法实现模拟试验台功能的扩展和可靠性的提升。项目研究可以维护风机的安全稳定运行和风电网供电系统的稳定，对整个风机行业都有着重要的理论意义和实践价值。

中文关键词： 风电机组；多维核域谱；传动链系统；健康状态识别；微征兆

英文摘要： Transmission Chain System (TCS) is the important component in wind turbine and its early micro-feature Healthy Condition Recognition (HCR) influences the turbine safety running and wind power smooth output directly. This project chooses the wind turbine TCS as the research object, and mainly study on the TCS micro-feature HCR method under complex condition and multi-dimension healthy conditions. The main research contents include three sections. 1) Discuss the interference components mechanism in wind turbine TCS, and propose the weak-feature extraction method of wind turbine TCS based on cross optimal Mexican-Hat Wavelet (MHW) method, in order to solve the weak-feature extraction problem of wind turbine TCS under strong background interference. 2) The micro-feature HCR method of wind turbine TCS. Propose the multi-dimension kernel domain spectrum method, in order to solve the micro-feature recognition problem of wind turbine TCS under small sample and multi-dimension healthy conditions coupling. 3) The improvement of the wind turbine healthy condition simulation testbed. To verify the effectiveness of the project methods using the simulation testbed, and at the same time to improve the testbed function and reliability using the project methods. This project research can not only maintain the wind turbine safety running and wind power smooth output, but also have important theoretical significance and practical value on the whole wind turbine industry.

英文关键词： Wind turbine;Multi-dimension kernel domain spectrum;Transmission chain system;Healthy condition recognition;Micro-feature