采用神经网络技术降低机电设备 BIT 虚警.PDFVIP

21 4 JOU RNAL OF NATIONAL UNIVERSITY OF DEFENSE T ECHNOLOGY Vol. 21 No. 4 1999 X BIT 徐永成　陶利民　温熙森　易晓山 ( 410073) BIT , BIT , BIT , , , , , , , , T N06 - Decreasing False Alarm of Mechantronics Equipment Built in Test Based on Neural Network Xu Yongcheng Tao Liming Wen Xisen Yi Xiaoshan ( Departmen tof M echantronics Engineering and Instrumenta tion, NU DT, Changsha, 410073) - ( ) . Abstract The significan tproblem of mechan tronics equipmen tbuil in tes M EBIT is i ts high false alarm ra te One of the importan tcauses is i ts sensor channels faul tof BIT system. The method of neural ne twork is adopted for the sensor channel faul tdiagnosis. The faul tm echanism and types of sensor channels of M EBIT in a large ship pow er engine and the faul tsam ples are obtained . After the training of neural network , the ac tual systems faults are diagnosed and i- dentified. The experimen tal results show tha he neural ne twork can diagnose the faults and identify their types. The method is com pact, effec tive and of prac tical value. - , , , Key words buil in tes faul tdiagnosis false alarm neural netw ork ( - , ) , Buil in Tes BIT BIT [ 1, 2] ( mechantronics equipmen tbuilt-in test, M EBIT ) BIT BIT BIT , BIT , MEBIT , , MEBIT , ( ) ( ) , , M EBIT BIT , , , , ,