Chaotic Time Series In Building Deformation Applied Research In The Forecast

Chaos is a seemingly irregular and random phenomenon in the deterministic system. In fact, there are a lot of phenomenon which were seen as random and irregular, they have some certainty rules. Chaos reveals the unity of the order and disorder, the unity of the uncertainty and randomness. The prediction of chaotic time series not only can make a model of the power system but also can detect and identify the chaotic.Chaos theory's application in the construction's deformation forecasting was studied in this paper. The paper revealed the deformation time series of the high building's monitoring point of Chang'an University had chaotic characteristic property based on the phase space reconstruction method, and used the weight local prediction model to predict the deformation monitoring. The main chievements are as follows:(1) By maximal Lyapunov expoment which was calculated by a method, namely Rosenstein method and G-P saturation correlation dimension method, the deformation time series of the high building's monitoring point of Chang'an University was considered a chaotic series, which was produced in nonlinear definite system.(2) Time delay of the deformation time series of the high building's monitoring point of Chang'an University was chosen by using autocorrelation function method.(3) Reconstruct dimension of the deformation time series of the high building's monitoring point of Chang'an University was obtained by charts of saturation correlation dimension method and false nearest neighbor percentage method.(4) The paper used local prediction model, the weight local prediction model and maximal Lyapunov expoment prediction model to predict the deformation time series of the high building's monitoring point of Chang'an University, and the conclusion was: the assessment precision of weight local prediction model was superior to local prediction model; the assessment precision of maximal Lyapunov prediction model was lower than the other two models. At the same time, the paper analysised the results of a multi-step weight local prediction model and concluded that chaos time series used on the one-step prediction of the construction's deformation time series was available, but it was disavailable on the multi-step prediction.