红外与激光工程
紅外與激光工程
홍외여격광공정
INFRARED AND LASER ENGINEERING
2014年
z1期
123-129
,共7页
模糊聚类%电动舵机%故障检测%舵机故障
模糊聚類%電動舵機%故障檢測%舵機故障
모호취류%전동타궤%고장검측%타궤고장
fuzzy cluster%electromechanical actuator%fault detection%electromechanical actuator fault
为了对电动舵机进行致命故障检测,根据模糊聚类方法建立了舵机标准状态样本,通过计算待测状态样本与标准状态样本之间的距离,将其归为最近的一类状态。首先,对舵机原始状态样本数据进行了归一化处理。接着,选用夹角余弦法建立了各样本间的模糊相似矩阵。然后,对初始聚类中心矩阵和隶属度矩阵进行迭代运算,设定最大迭代误差并结束迭代过程,得到了舵机标准状态样本。最后,搭建了舵机状态检测的试验平台,在拷机试验中舵机控制器实时运行状态检测程序,实时计算待测样本与标准状态样本的距离。实验结果表明:状态检测程序运行时间只需0.23 us,且检测结果全部正确。此状态检测方法满足了电动舵机故障检测准确性和实时性的要求。
為瞭對電動舵機進行緻命故障檢測,根據模糊聚類方法建立瞭舵機標準狀態樣本,通過計算待測狀態樣本與標準狀態樣本之間的距離,將其歸為最近的一類狀態。首先,對舵機原始狀態樣本數據進行瞭歸一化處理。接著,選用夾角餘絃法建立瞭各樣本間的模糊相似矩陣。然後,對初始聚類中心矩陣和隸屬度矩陣進行迭代運算,設定最大迭代誤差併結束迭代過程,得到瞭舵機標準狀態樣本。最後,搭建瞭舵機狀態檢測的試驗平檯,在拷機試驗中舵機控製器實時運行狀態檢測程序,實時計算待測樣本與標準狀態樣本的距離。實驗結果錶明:狀態檢測程序運行時間隻需0.23 us,且檢測結果全部正確。此狀態檢測方法滿足瞭電動舵機故障檢測準確性和實時性的要求。
위료대전동타궤진행치명고장검측,근거모호취류방법건립료타궤표준상태양본,통과계산대측상태양본여표준상태양본지간적거리,장기귀위최근적일류상태。수선,대타궤원시상태양본수거진행료귀일화처리。접착,선용협각여현법건립료각양본간적모호상사구진。연후,대초시취류중심구진화대속도구진진행질대운산,설정최대질대오차병결속질대과정,득도료타궤표준상태양본。최후,탑건료타궤상태검측적시험평태,재고궤시험중타궤공제기실시운행상태검측정서,실시계산대측양본여표준상태양본적거리。실험결과표명:상태검측정서운행시간지수0.23 us,차검측결과전부정학。차상태검측방법만족료전동타궤고장검측준학성화실시성적요구。
In order to detect electromechanical actuator fatal fault, an electromechanical actuator standard state sample was established according to fuzzy cluster method. Electromechanical actuator state can be classified to certain types by calculating least approximation distance between waiting test state sample and standard sample, so the state detection was realized. First, some original electromechanical actuator state sample data has been normalized. Then fuzzy similar matrix between each sample was established based on angle cosine law. After that iterative calculation of clustering center matrix and membership matrix was started. And iterative process was ended through setting maximum iterative error. So the standard state sample was obtained. Finally, test platform is established for state detection. The real- time state detection program runs in the electromechanical actuator controller in continuous operation test. And the program calculates distance of the sample under test and standard state sample. Experimental results indicate that runtime of electromechanical actuator state detection program just need 0.23us, and detection conclusion is completely right. This state detection method can satisfy the requirements of veracity and real- time for electromechanical actuator fault detection.