The Key Technology Of Fault Prediction Based On Vector Spectrum

As science and technology and the rapid development of industrial technology, a variety of increasingly high degree of automation equipment, its function has become even more perfect, which increasing the complexity in the monitoring equipment, increasingly high demands on the accuracy of testing and fault diagnosis of signals. Practice has shown that the traditional single-channel signal analysis techniques cannot meet the requirements of high accuracy and high reliability. For the defects of incomplete information of single source vibration signals of rotating machinery, full information data processing and fault diagnosis technology came into being which based on data fusion. The technique of vector spectrum which based on the data level fusion can fully express vibration strength and frequency structure, and its expression pattern is compatible with the traditional method of spectrum analysis, has an unparalleled advantage.In the field of equipment status forecast, for the defects of incomplete information of single source vibration signals of rotating machinery, based on the method of data fusion, combing the vector spectrum technology with the time sequence autoregressive AR(n) model, parameter estimation based on least squares linear regression model, simple sliding smoothing forecasting model, exponential smoothing forecasting models combining prediction, proposed the time sequence autoregressive AR(n) model forecast method based on vector spectrum technology, the linear fitting forecast method based on vector spectrum technology, the simple moving average forecast method based on vector spectrum technology, the exponential smoothing forecast method based on vector spectrum technology, respectively give computational method and parameter optimization algorithms and to compare them. The results show that the spectrum prediction method based on vector spectrum technology has higher prediction accuracy in each frequency bands which can comprehensively and accurately reflect the state of the development trend of mechanical operation, is an accurate, effective, practical prediction method of rotating machinery. This article based on the prediction theory of vector spectrum technology, to the goal of building an equipment of inspection forecasting system, in the Windows operating system platform, using the Visual C++ programs development tools and SQL Server 2000 database tools to build an equipment condition forecasting system based on vector spectrum technology which realize the historical data, waveform, spectrum display, and prediction analysis. This work has a very important significance to the prediction of the equipment state and early prediction of the realization in engineering practice and full information forecasting techniques in practical application of engineering.