Study On Analysis Method Of Deformation Signal Based On Wavelet Entropy

The development of modern deformation monitoring technology make the deformation signal become more complicated. There is a wealth of information in the mass of the deformation data. Wavelet analysis is an effective tool for time-frequency localization analysis. It has a strong ability to capture the non-stationary signal characteristics. The information entropy can realize quantitative description of information, then contrast and estimate it. This thesis combined wavelet analysis theory and information entropy theory, and then used for deformation signal analysis, provided a kind of effective method of modern deformation analysis. The main research contents are as follows:Use coarse-grained statistical analysis method to deal with the wavelet coefficients which come form wavelet analysis, combined with the information entropy theory, to set up a new measure:wavelet time entropy. Then explored its parameters optimization and the principle of selection. Wavelet time entropy contains all kinds of singular points of signal singularity detection capability verified by simulation, and used for the singular values of the GPS deformation signal detection. The results showed:Wavelet time entropy locked the singular point in the window wide range, can sensitive detection point but anti-noise ability is limited. In the singular point, the entropy was proportional to the change rate of the amplitude and frequency, but can't quantitative identification regional amplitude mutations, and can be used in real-time dynamic deformation signal singularity of automated testing.Take wavelet energy as the information entropy processing object. Constructed wavelet energy entropy and discussed the implementation process of the parameters of the choice method. Structured containing different error simulation signal interference, verified the characteristics of wavelet energy entropy information detection ability, and it can be used in gross error detection for GPS deformation signals. Tests showed that the measure had good anti-noise ability, can achieve high noise background in nonlinear deformation signal feature extraction. The measure does not need prior data and parameter estimation, can detect regional states, discrete states and isolated state gross error, can real-time deformation signal gross error detection.Put the singular value decomposition theory into the multi-resolution analysis and information entropy theory to establish a wavelet singular entropy. Through transformed expression of the deformation signal space to complete an accurate extraction of the internal features of the signal information. Constructed different information content of the simulated signal to verify the complexity of the analysis, Used wavelet singular entropy to stability analysis of GPS stations. Compared the wavelet singular entropy of the deformation signal from different GPS continuous operation observatory. Provided certain reference for the error correction model for the GPS of evaluation and selection of the parameters for the model.