摘要
针对风机叶片故障引发的颤振与固有频率锁定现象导致的诊断性能下降问题,提出基于柯西蜂群算法的风机叶片故障诊断方法。分析风机叶片振动信号特性,通过计算叶片的振动频率和振动有效值,分析振动信号的变化特点,利用集合经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)算法分解叶片的振动信号,获取能够描述振动频率变化的模态分量;将分解结果输入随机森林模型中,该模型通过节点分裂和集成投票进行叶片故障分类诊断,并且引入柯西蜂群算法优化随机森林的分裂节点,随机森林依据获取的最佳分裂节点输出故障诊断结果,保证微小故障和早期故障的诊断效果。测试结果表明,所提方法通过监测风机叶片振动信号的频率变化,成功捕捉到振动幅值增加时频率的波动范围为0.22~0.78 Hz,能够依据此进行故障诊断。同时,所提方法对微小和早期故障的诊断应用性较好。
Addressing the issue of decreased diagnostic performance caused by flutter and natural frequency locking phenomena induced by wind turbine blade faults,a fault diagnosis method for wind turbine blades based on the Cauchy Bee Colony algorithm is proposed.Analyze the characteristics of the vibration signals of wind turbine blades,and calculate the vibration frequency and effective value of the blades’vibration.After analyzing the changing features of the vibration signals,the EEMD decomposition algorithm was utilized to decompose the vibration signals of the blades,obtaining modal components that could describe the changes in vibration frequency.The decomposition results were then input into a Random Forest model,which conducted blade fault classification and diagnosis through node splitting and ensemble voting.The Cauchy⁃based Artificial Bee Colony(ABC)algorithm was introduced to optimize the splitting nodes of the Random Forest.Based on the optimal splitting nodes obtained,the Random Forest output the fault diagnosis results,ensuring the diagnostic effectiveness for minor and early faults.The test results show that the proposed method successfully captures the frequency fluctuation range of 0.22 to 0.78 Hz when the vibration amplitude increases by monitoring the frequency changes of the fan blade vibration signal,which can be used for fault diagnosis.Meanwhile,the proposed method has good applicability in diagnosing small and early faults.
作者
杨兆辉
岳松
汪志伟
杨锦霖
常园礼
YANG Zhaohui;YUE Song;WANG Zhiwei;YANG Jinlin;CHANG Yuanli(Guotou Guangxi New Energy Development Co.,Ltd.,Qinzhou 535317,China)
出处
《电子设计工程》
2025年第23期94-98,共5页
Electronic Design Engineering
关键词
柯西蜂群算法
风机叶片
故障诊断
振动有效值
振动频率
模态分量
Cauchy Bee Colony algorithm
wind turbine blades
fault diagnosis
effective value of vibration
vibration frequency
modal component
