Condition Monitoring And Identification Method Research Of Hydraulic Reciprocating Seal

Reciprocating seal is very important for the hydraulic cylinder's normal work.Due to the poor working environment,reciprocating seal failure is one of the most common failure modes of hydraulic cylinder.If preventing measures are not taken in time,more serious consequence may happen.Nevertheless,it's really difficult to realize the precise monitoring and identification of hydraulic reciprocating seals' sealing condition.Traditional measures are usually taken by extracting the fluid leakage signal(such as pressure,flow signal,etc),but these methods' accuracy need to be improved,for leakage is only an indirect embodiment of sealing failure,and it is easy to be disturbed by other factors.Therefore,it will be very meaningful to find a timely and accurate method to monitor and identify the sealing condition of hydraulic reciprocating seal.In order to solve above problems,this paper takes the hydraulic cylinder piston seal as research object,focus on the following contents:(1)The theoretical background and sealing mechanism of hydraulic reciprocating seal are systematically expounded,and the key parameters that affect the state of reciprocating seal are found through analyzing the mathematical model of dynamic and static seal: the sealing surface contact pressure.(2)Through finite element analysis,the feasibility of reciprocating sealing condition monitoring scheme based on Fiber Bragg grating sensing is verified;and the strain signal monitoring of piston seal groove through the fiber Bragg grating sensor is realized by reasonable structure design.(3)The wear failure monitoring and identification test research of piston seal ring is carried out.The contact strain curves of the piston seal grooves in hydraulic cylinder are analyzed by the strain signal of the fiber Bragg grating sensor.(4)By methods of wavelet packet analysis,indexes of time domain analysis,spectrum analysis,power spectrum analysis,etc.8 kinds of fault feature parameters are put forward: Fiber Bragg grating signal's root mean square value,root mean square amplitude,margin index,steep degree index,the sideband amplitude entropy and 25 Hz power value;fluid pressure force signal's energy proportion of underlying frequency period,wavelet packet energy entropy,etc.It can provide a theoretical basis to achieve the condition identification of the hydraulic reciprocating seal.(5)Combined with the extracted fault feature parameters,the multi sensor signal fusion method based on support vector machine is selected to realize the state identification of hydraulic cylinder piston seal,and the recognition accuracy is 88.89%.