基于LSTM的塔架振动状态监测研究  

Research on monitoring of tower vibration condition based on LSTM

作　　者：苏连成[1] 朱娇娇 郭高鑫 李英伟[3] 姜浪朗 SU Liancheng;ZHU Jiaojiao;GUO Gaoxin;LI Yingwei;JIANG Langlang(School of Electrical Engineering Yanshan University,Qinhuangdao Hebei 066004 China;China Heavy Machinery Research Institute Co.Ltd,Xi'an Shaanxi 710018 China;School of Information Science and Engineering Yanshan University,Qinhuangdao Hebei 066004 China)

机构地区：[1]燕山大学电气工程学院,河北秦皇岛066004 [2]中国重型机械研究院股份公司,陕西西安710018 [3]燕山大学信息科学与工程学院,河北秦皇岛066004

出　　处：《燕山大学学报》2022年第5期437-445,共9页Journal of Yanshan University

基　　金：国防基础研究计划资助项目(JCKY2019407C002);河北省自然科学基金资助项目(F2021203054)。

摘　　要：针对风电机组塔架受到外部环境引起的复杂动载荷作用以及风机内部控制动作的激励而产生异常振动问题,本文提出了一种基于长短期记忆网络的塔架振动状态监测方法。首先,以风电机组监督控制和数据采集系统数据为基础,引入灰色关联度系数和最小角回归算法,分析各状态参量对塔架振动的影响;其次,采用长短期记忆网络建立塔架振动预测模型,通过对比分析预测值与实际值之间的残差判断风电机组的塔架振动状态;最后,利用某风电场的实际采集数据进行验证分析,结果表明所提方法能提前对风电机组的异常情况做出预警,可以有效避免机组因故障恶化导致的紧急停机,提高机组运行的可靠性。A tower vibration condition monitoring method based on long short-term memory network is proposed in this paper for solving the problem of tower abnormal vibration which is aroused by complex dynamic load caused by the external environment and the excitation of the internal control action of the wind turbine. Firstly, the grey correlation coefficient and the least angle regression algorithm are introduced to analyze the influence of various state parameters on the tower vibration based on the data from supervisory control and data acquisition system of wind turbines. Secondly, the long short-term memory network is used to establish a tower vibration prediction model, and the tower vibration state of the wind turbine is judged by comparing and analyzing the residual error between the predicted value and the actual value. Finally, the actual collected data of a wind farm is used for verification and analysis. The results show that the proposed method can give earlier warning of abnormal conditions of the wind turbine, which can effectively avoid the emergency shutdown of the turbine due to the deterioration of the fault and improve the reliability of the turbine operation.

关 键 词：风电机组 塔架振动 长短期记忆网络 监督控制和数据采集 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]