Study On The Prediction Method Of Mine Fan Vibration Trend

Mine fan is one of the most important equipment in large mine equipment. It plays a key role in ensuring the safety of coal mine operation. Mine fan is the respiratory system of mine. Whether it can maintain normal operation directly related to the air quality of the working environment of the mine.Where there is mechanical equipment there is vibration. Vibration signal contains a wealth of information about the running status of the mechanical equipment.By analysising all kinds of vibration signals, extracting the key information from the vibration process. Using the appropriate forecasting method to forecast the future trend of mechanical equipment, people can grasp the running status of the mechanical equipment more accurately. For this,this paper carries on the research of the forcasting method of Mine fan vibration trend.In this paper, equipment condition prediction methods of time series, dynamic neural network, grey forecasting and Empirical Mode Decomposition are discussed. The principle and realizations of time series forecasting and dynamic neural network are introduced in detail.The vibration intensity of Kailuan Group's mine fan is regarded as the data sample, using single time series forecasting method and NAR neural network forecasting method to forcast the data sample. According to the excellent characteristics of Empirical Mode Decomposition, time series prediction mothod and NAR neural network prediction method based on Empirical Mode Decomposition are proposed. A comparative study of the prediction results is carried out, The validity of the method that proposed by this paper is demonstrated with respect to a single prediction method. Finally, the superiority of time series and NAR neural network forecasting mothod in predicting different datas are analyzed.Combining with Empirical Mode Decomposition,the final prediction Method of Mine Fan Vibration Trend is put forward: Time series- NAR neural network forecasting method based on Empirical Mode Decomposition. A comparative study of the prediction results is carried out and the accuracy and validity of the method are verified. The whole analysis and forecasting process is realized by MATLAB software.