Manifold learning method for denoising dam safety monitoring data
The noise pollution cannot be avoided for dam deformation,seepage,stress-strain,and other safety monitoring data.It is difficult to achieve excellent denoising effect for the traditional linear noise reduction method.On the basis of reconstructing the phase space of dam safety monitoring data time sequences,with the cross application of local tangent space alignment method,maximum likelihood estimate,adaptive neighborhood and other methods,using the reconstructed phase space as a bridge,and through the extraction of deep information of the monitoring data sequences,the denoised data sequences of dam safety monitoring is obtained.Based on the application of prototype test data,it is evident that the noise reduction effect of the method proposed in this paper is superior to that of the wavelet soft-thresholding method and the fixed neighborhood local tangent space alignment(LTSA)method,demonstrating certain value for engineering applications.
dam safetymeasured datanoise reductionmanifold learningphase space reconstruction
万方数据
维普
