Denoising Analysis Of Dam Deformation Data Based On EEMD-PCA Model

In the field of dam deformation monitoring,due to the complexity of the actual project and the uncertainty of the external environment,there are always some errors in the observed data.The existence of errors directly leads to the increase of fluctuation of observation data,which to a certain extent conceals the actual deformation of the project;in serious cases,because the information obtained from the observation data is totally inconsistent with the actual deformation,it will lead to errors in judgment and decision-making of the project safety,and consequently cause significant personal and property damage.Various denoising methods have also been applied in practice to eliminate data errors,but the traditional denoising models can not meet the current engineering needs because of their limitations,and most of the traditional models need to adjust model parameters manually according to the engineering situation,lacking self-adaptability.In view of this situation,a new denoising method based on Ensemble Empirical Mode Decomposition(EEMD)and Principal Component Analysis(PCA)is proposed.On this basis,an EEMD-PCA-ARIMA-based denoising method is proposed in combination with Autoregressive Integrated Moving Average Model(ARIMA).The prediction model of dam deformation data is analyzed experimentally by using simulation data and measured data.Meanwhile,the theory of multi-index comprehensive evaluation is introduced into the evaluation of denoising results.The results show that:(1)EEMD-PCA denoising method is effective and adaptive for the denoising of dam deformation data;(2)EEMD-PCA-ARIMA dam deformation data prediction model can better obtain the actual deformation curve of the dam,thus achieving better prediction accuracy;(3)The introduction of multi-index comprehensive evaluation theory is more intuitive and clear than the traditional denoising evaluation index.It is worth further research and promotion.The specific research contents and achievements of this paper are as follows:1.A denoising model based on EEMD-PCA is proposed.Sample matrix is constructed by empirical mode decomposition(EMD)of the observed data set,and then the corresponding principal components are selected to construct the mapping matrix.Then the mapping matrix is used to map the sample matrix constructed from the original data,so as to achieve the purpose of denoising.2.In order to reflect the effect of denoising more intuitively,the multi-index comprehensive evaluation theory is further introduced.By empowering the traditional multi-index,weighted average the results of the original evaluation index,a new evaluation index is obtained,and then the advantages and disadvantages of the denoising results are reflected intuitively.3.Using the model proposed in this paper to denoise the simulation data.Compared with wavelet denoising method and direct EEMD decomposition denoising method,the denoising method of removing high frequency components directly is discussed.In the process of wavelet denoising,sym6,Haar and db3 wavelet bases are used to denoise 1-10 layers of wavelet,and the multi-index comprehensive evaluation theory is used to evaluate.4.The denoising model proposed in this paper can be used to denoise the measured deformation value of the dam.The denoising effect is adaptive to a certain extent,and has practical significance.It is an effective data processing method for the actual engineering data.Furthermore,a dam deformation data prediction model based on EEMD-PCA-ARIMA is proposed by combining ARIMA model.The prediction accuracy is verified by comparing with several existing prediction models.