摘要
针对实际工程中风力机传动系统未知输入难以完全解耦的问题,提出基于H_/H∞未知输入观测器(unknown input observer, UIO)故障诊断方法.首先,利用未知输入观测器的等式条件将未知输入分解为可解耦与不可解耦两部分,且将可解耦部分从估计误差中消除.其次,利用H_/H∞性能指标设计未知输入观测器的残差产生器,使残差对未知输入中不可解耦部分具有鲁棒性并对故障信号具有敏感性,并通过引入松弛矩阵解决观测器参数设计过程中的保守问题.考虑残差信号的随机特性,采用统计理论确定故障检测阈值.经仿真验证,该方法可对风力机传动系统传感器加性故障和乘性故障进行有效诊断.
In order to solve the problem that the unknown input of the wind turbine drive system is difficult to be fully decoupled, a fault diagnosis method based on H_/H∞ unknown input observer (UIO) is proposed. Firstly, the unknown input is decomposed into decouplable and uncoupled parts by using the equation condition of the unknown input observer, and the decouplable part is eliminated from the estimation error. Secondly, the residual generator of the unknown input observer is designed by using H_/H∞ performance index to make the residual robust to the uncoupled part of the unknown input and sensitive to the fault signal. The conservative problem in the parameter design of the observer is solved by introducing relaxation matrix. Considering the stochastic characteristics of residual signal, the fault detection threshold is determined by statistical theory. The simulation results show that the method can effectively diagnose the additive and multiplicative faults of sensors in wind turbine drive system.
作者
陈宣含
沈艳霞
吴定会
CHEN Xuan-han;SHEN Yan-xia;WU Ding-hui(Engineering Research Center of Application Internet of Things Technology, Jiangnan University, Wuxi Jiangsu 214000, China)
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2019年第10期1690-1697,共8页
Control Theory & Applications
基金
国家自然科学基金项目(61573167,61572237)资助~~
