基于产品试验台数据的风电齿轮箱故障识别与诊断技术方法研究

项目名称： 基于产品试验台数据的风电齿轮箱故障识别与诊断技术方法研究

项目编号： No.51275453

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 机械、仪表工业

项目作者： 周晓军

作者单位： 浙江大学

项目金额： 80万元

中文摘要： 针对当前主流风电设备故障识别与诊断技术相对落后现状及存在的问题，提出通过在生产企业真实风电齿轮箱产品全功率模拟试验台基础上，利用风电齿轮箱生产企业试验台试验产品数据量大、故障信息数据可靠、针对性强和试验研究环境好等特点，解决目前风电齿轮箱故障识别与诊断专家系统研究需要大量原始故障信息要求和开展实验研究环境困难的制约，获得基于产品试验台数据的风电齿轮箱故障识别与诊断的新技术和新方法。研究内容包括：1）在传统齿轮箱故障振动信号处理方法基础上，研究针对风电设备齿轮箱各类故障信息传感器的优化布局和各类信号的获取方法；2）研究适用于变速度、变载荷工况下风电齿轮箱振动信号处理和故障特征提取的新方法；3）通过试验台具有模拟风电齿轮箱各种实际工况和主动设计真实产品故障模拟的特点，研究基于小波神经网络和专家系统的风电齿轮箱故障诊断新方法；4）研究建立基于产品试验台的风电齿轮箱故障诊断数据库系统。

中文关键词： 故障诊断；风电齿轮箱；角域振动信号；支持向量机；

英文摘要： In view of fault recognize and diagnosis technology of main wind turbines equipments is relatively laggard and has some defects, use of wind turbines gearbox manufacture corporations having plenty data, credible fault information, strong pertinence and friendly test environment characteristics through real wind turbines gearbox products maximal power simulation test platform, resolve a problem that now fault recognize and diagnosis expert system research needs huge original fault information of wind turbines gearbox products, and that test research is limited by environment, achieve a new technology and a new method of wind turbines gearbox fault recognize and diagnosis based on data of product test platform. The research contents include: 1）research on layout optimization of fault information sensors and method of obtaining all signals, aims at wind turbines equipments; 2）research on a new method of vibration signals processing and obtaining fault characteristics of wind turbines gearbox on speed variational and load variational circumstances. 3）in view of the platform can simulate all real work station and can design genuine product fault simulation of wind turbines gearbox, research a new method of fault diagnosis of wind turbines gearbox based on wavelet neural network and expert system. 4） research on build

英文关键词： Fault diagnosis；Wind turbine gearbox；Angular domain vibration signal；SVM；