Abstract:Static mechanical seal film thickness was measured directly and indirectly by using eddy current sensor and acoustic emission technique,separately.The film thickness variation characteristics of mechanical seal were studied under different rotational speed and pressure.Through advanced signal analysis and artificial intelligence method,the acoustic emission characteristics of mechanical seal were studied at different interval of film thickness,and the state recognition model of mechanical seal was established by artificial neural network so as to realize interval estimation of fluid film thickness of mechanical seal.The results show during working of mechanical seal,interval variation of film thickness is mainly influenced by pressure,while instant fluctuation of film thickness is by rotational speed.The average recognition rate of the film thickness of mechanical seal is reached 85% by the serial decision model constructed based on BP neural network,so as to implement transition of detection of film thickness of mechanical seal from destructive testing technique to a nondestructive one.