基于改进多新息算法的风机液压变桨系统故障诊断  

Fault diagnosis of wind turbine hydraulic pitch system basedon improved multi-innovation algorithm

作　　者：胡昌选 文传博(指导)[1] HU Changxuan;WEN Chuanbo(School of Electrical Engineering,Shanghai Dianji University,Shanghai 201306,China)

机构地区：[1]上海电机学院电气学院,上海201306

出　　处：《上海电机学院学报》2020年第1期29-36,共8页Journal of Shanghai Dianji University

摘　　要：液压变桨系统故障是在风机运行时频发的一种故障,对其展开故障诊断具有重要意义。针对受到丢包和状态延时影响的风机变桨系统故障,提出一种基于改进多新息随机梯度(MISG)算法的故障诊断方法。将复杂的风机系统简化为状态空间模型,建立相应的辨识模型。利用强跟踪的思想,根据残差失配情况自适应调整梯度跟踪速率;同时,利用可变的新息长度,强化故障时系统状态参数发生变化时的状态跟踪。加入容错控制,使得系统发生故障时,仍然具有良好的性能,确保风机安全停机。结果表明:改进的算法比最小二乘算法具有动态响应性能更好、稳定性更高、测量精度更高的优点。The hydraulic pitch system failure frequently occurs with the running wind turbine.Therefore,it is of great significance to carry out the fault diagnosis.Aiming at the failure of the wind turbine hydraulic pitch system affected by the packet loss and the pitch angle control state delay,the present paper proposes a fault diagnosis method based on the improved multi-innovation stochastic gradient(MISG)algorithm.First,the complex system of the wind turbine is simplified into a state space model,and the corresponding identification model is established.With the idea of strong tracking,the gradient tracking rate is adaptively adjusted according to the residual mismatch.Next,the variable new interest length is used to enhance the state tracking when the system state parameter changes with the occurred fault.Finally,the fault-tolerant control is added so that when the system fails,it still has good performance and ensures safe shutdown of the wind turbine.The results show that the proposed algorithm has better dynamic response performance,stability and measurement accuracy than the least squares algorithm.

关 键 词：变桨距系统 系统辨识 自适应 状态跟踪 容错控制 

分 类 号：TM93[电气工程—电力电子与电力传动]