Turbine Unit Vibration Condition Monitoring And Forecasting Based On Chaos Theory

Normalized algorithm improved correlation dimension in this paper,the feature of the trouble is extracted and the fault type is determined,the problems of consistency and irrelativeness in current correlation dimension calculation, a high-volume data as well as weak in noise are solved. According to the use of the phase space reconstruction method of chaotic theory, the paper introduces the lyapunov exponent and the correlation dimension as the standard to identify systemic running state. The basic prediction methods of chaos time series are summarized systematically and the Maximum lyapunov exponent and adaptive wavelet transform-Volterra prediction method is put forward, the forecast of steam turbine generator unit, digital simulation to prove their forecast accuracy, realtimeness are all improved better by this way.