Research on sparse representation algorithm for reconstruction prediction of dam deformation
The original dam monitoring sequences collected in the sensors inevitably have external and human-caused noise,which brings challenges to the accurate prediction of dam deformation,and to solve this problem,the noise reduction and reconstruction training set method is proposed to predict the dam deformation.Aiming at the traditional noise reduction method affected by redundant basis functions,K-SVD method is introduced to sparsely represent the original dam monitoring sequences,and adaptively update the effective information of the atomically improved reconstructed dam monitoring sequences.Taking the real dam as an example,to verify the effectiveness of this paper's method,with different prediction ability of the machine learning model for the test,the test shows that this paper's training set noise reduction and reconstruction algorithm can improve the prediction accuracy of dam deformation,effectively show the characteristics of the dam deformation sequences,in the face of the traditional noise reduction algorithm has a better performance has a certain degree of robustness and reliability.
NETL
NSTL
万方数据
