Time series and prognostication of vibration level for hydrotreaters
Roman Kvyetnyy, Liudmyla MelnykThis article is devoted to the problem of vibration level prognostication of hydrotreaters using the mathematical regression models that describe the time series with long or double long memory. The next problem investigation (using real data from Sayano-shushenskiy GES) shows that the processes, which describe the vibration level of hydrotreater’s knot, might be represented effectively with mathematical ARCH and GARCH models. Authors developed the prognostication algorithm for vibration level of hydrotreater’s knot that allows detect a long memory in time series with usage of Sugeno method, analyze a time series type, select time interval for prognostigation and using these results make a decision about working state of hydrotreater’s knot