Research On Fault Diagnosis Of Gear Tooth Surface Of Ball Mill Based On Wavelet Analysis And EMD Decomposition

The transmission system is the key part of the ball mill operation with the large and small gears as the key components.As the running time increases,faults such as pitting,gluing,plastic deformation on the gear tooth surface will greatly affect the transmission efficiency and thus affect the equipment's production efficiency.Severe accidents may even occur due to the frequent failure.In this paper,research was performed on wavelet analysis and EMD decomposition,which are two of the signal processing method ideas.The feasibility of combining two methods research was concluded through calculation.A ball mill gear from a mining company was taken as the research object.The ball mill gear tooth surface is the main research object.The vibration detection was standardized by establishing the ideal experimental conditions and testing the experimental ball mill gear vibration data in groups based on the ball mill operation cycle.From this standard,five gears with different five failure levels were selected to be installed and run under this experimental condition.The gear tooth surface vibration data were collected.Based on the two signal processing methods,the vibration signal characteristics were extracted.A more accurate gear fault level standard was calibrated through analysis and calculation,which is used as the basis for quantitative and qualitative diagnosis of ball mill gear tooth surface faults.Finally,the signal processing method combining the two methods is compared with the fault diagnosis method for diagnostic effect through field fault cases,which proves that the new fault diagnosis scheme is effective.The specific work is as follows.(1)Research of the problems on the gear tooth surface of ball mill was performed.The significance of research and the development status was introduced.(2)Through analyzing the research ideas of two common signal processing methods and the research arithmetic examples of other related directions,the feasibility,effectiveness and shortcomings of the results of the two signal processing methods were pointed out.The fault judgment will be more accurate if the two methods are combined.(3)The experimental environment was constructed according to the equipment parameters and operating conditions of ball mills.Three experimental ball mills were measured several times and in multiple coordinates with respect to the operating time.The ball mill gear vibration standard suitable for field operation was standardized based on the national standard.And based on this standard,gears with five different fault levels were selected as experimental objects,operated under the same input conditions to collect vibration signals of different degrees.(4)Using two signal processing techniques,wavelet analysis and EMD decomposition,10 sets of fault simulation experiments were conducted using gears with 5 different degree of fault levels.The vibration signals during the operation of the collected ball mill were analyzed.The root mean square values of the primary high frequency wavelet coefficients,secondary high frequency wavelet coefficients and primary IMF functions of the vibration signals were obtained.The three presented sensitive feature pattern results to the gear tooth surface fault.The results were used to obtain the feature extraction results change rate curve.The conclusion of the comparative diagnostic effect that the first-level IMF function of vibration signal is better than the second-level HF wavelet coefficient,and the latter is better than the first-level HF wavelet coefficient was obtained.The diagnosis method combining the two methods is substituted into the fault case for fault analysis,and the former diagnosis is found to be more accurate and better by comparison.