Research On Looseness Fault Detection Of Hydraulic Pipeline

As a medium to transport various liquids,the hydraulic pipeline has been widely used in various mechanical equipment;however the harsh environment in which the hydraulic pipeline has been employed for a long time certainly leads to all kinds of faults.The joint of hydraulic tubing is an important part for connecting pipe.It causes the looseness and even leakage of pipeline that the outer ring nut is expanded by variable hydraulic pressure and high temperature,and the wear of screw thread is generated by long-term vibration and bad installation.Moreover,the leakage of flammable,explosive or polluted liquids can give rise to serious consequences.Additionally,the clamp is one of the supporting accessories of pipeline,which also triggers the looseness of pipeline because of its long-term vibration.It is of significance to detect the health state of pipe joint and clamp for a hydraulic piping system.The clamps and joints distributed on the hydraulic pipeline are the research objects in this paper;and the experimental data of pipeline looseness fault would be analyzed in detail.The structure of the main research is as follows:(1)Aiming at the looseness fault of hydraulic pipe,this paper analyzes several different sorts of them and their mechanism.Then a specific platform and particular experiments are designed and realized for detection of looseness of hydraulic pipeline.Based on them,the different tightening states of the connecting nut are set as the distinct loosening degrees of the joint.And so does the screws for the clamp.All experiments provide supportive data for subsequent analysis on detection methods of looseness fault.(2)Aiming at the interference frequencies in the vibration signals,a signal reconstruction algorithm based on variational mode decomposition(VMD)is proposed.VMD can effectively extract various frequency components from the original signals.According to it,the fault characteristic can be amplified as much as possible and the interference frequency can be removed by selecting appropriate VMD components.Then,for declining the number of frequency components of fault signals,a configuration pattern amplitude(CPA),which is sensitive to the change of both signal frequency and amplitude,is proposed to distinguish the different components and extend them to multiple scales.The fault features of the reconstructed signal are extracted by the amplitude of the multi-scale configuration pattern,and the feasibility of this method is proved by experiments.(3)Aiming at the classification of hydraulic tubing joint loosening and clamp loosening,a loosening fault detection method based on quasi-binary variational mode decomposition(QBVMD)and dispersion entropy(DE)is proposed.Considering the multi-channel signals of three-axis accelerometer,the improved VMD method is used to decompose the selected two-dimensional acceleration data.Combined with DE,fault features are extracted easily and feature sets are reasonably constructed.Then the constraint score is able to reduce dimension of feature sets for computation efficiency.Finally,support vector machine(SVM)gives birth to the results of features classified.The experimental results show that this method can effectively figure out the difference of looseness in between hydraulic pipe clamp and pipe joint.