Health Diagnosis And Prediction Of Mine Hoisting Head Sheave

Multi-rope friction and multi-rope winding hoisting are widely used in deep mines over one kilometer.While for landing multi-rope and multi-rope winding hoisting system,head sheave is an essential component.The head sheave plays the role of supporting and guiding the wire rope in the hoisting system and is one of the main forced parts of the system.If the head sheave fails,it will induce a hoisting system failure,resulting in casualties and huge economic losses.In order to ensure the safe operation of it,it is necessary to analyze and study the fault of the head sheave and put forward reasonable and applicable health diagnosis and prediction methods.As such,carrying out the research on the health diagnosis and prediction of the head sheave has theoretical significance and practical application value for ensuring the safe operation of the hoisting system.First,the fault tree of head sheave system was established and the fault mechanism of the system was analyzed.According to the qualitative analysis results of the fault tree of the system,the characteristic parameter of health diagnosis modeling of the head sheave was determined.And the finite element software ANSYS Workbench was used to implement the modal analysis of the head sheave.The influence rule of the head sheave fault on the vibration amplitude variation was obtained through the modal analysis results and the vibration characteristics of the head sheave.The fault characteristic parameters are established for the head sheave.Secondly,the hardware and software systems for monitoring the running of head sheave were set up,including the selection of equipment for the hardware system and the selection of relevant monitoring software.The magnetic acceleration sensor and the non-contact displacement sensor were used to monitor the running state of the head sheave.The advantages of Fourier transform and Wavelet Transform(WT)on the vibration signal processing of the head sheave were analyzed.According to the characteristics of the monitored vibration signal of the head sheave,the wavelet threshold denoising method was adopted to filter and denoise the signal,which effectively reduced the noise of the vibration signal.Finally,based on the two-state head sheave health model and the vibration data of the operation state of head sheave on the hoisting test-bed,the multistate Hidden Markov Model(HMM)was built.The accuracy of HMM diagnosis was verified by relevant test data.The probabilistic reasoning,decoding and training problems of the life prediction model of the head sheave were discussed,and the implementation method and steps of the life prediction of the HMM were presented.