Research On Noise Suppression Method Of Simultaneous Source Data Based On PCA Optimal Rank

The simultaneous source acquisition technology adopts multiple gun excitation method,the number of covering is more in the same time and data acquisition is the way of simultaneous excitation and uninterrupted reception.Compared with the traditional single gun excitation acquisition technology,it has the advantages of high seismic data quality,high efficiency and low-cost.Therefore,the technology is widely used in oil and gas resources exploration and other fields.Because of the wide range and large time span of the simultaneous source acquisition area,the data collected are easily disturbed by a variety of noise,mainly including power-frequency noise,random noise and crosstalk noise,etc.In this paper,only for the first two types of noise,because there is a spectrum overlap between the two types of noise and the useful signal,which makes the signal to noise ratio improved a little by the conventional noise suppressing.Therefore,it is very important to suppress the noise of simultaneous source data effectively.Financed by ‘Comprehensive geological model and three-dimensional display of field experimental and demonstration bases' and ‘Deep field detection equipment and field test methods and comparative study Deep field detection equipment and field test methods and comparative study',subproject of the Sino Probe(China's largest national collaborative,multidisciplinary Earth science research project),the research of optimal rank of PCA in attenuating simultaneous source's noises is studied in this paper.Author analysis the properties of exploration database's noises by simultaneous source.The method of attenuate simultaneous source's industrial frequency noise based on optimal rank of PCA and the method of attenuate simultaneous source's random noise based on optimal rank of PCA-EMD are put forward in this paper.Finally,effectiveness and superiority of method presented in this paper is verified by experiment results of simulated and real data.Main contents of this paper are as follows:First,we analyzed the characteristics of signal and noise in simultaneous source data.In addition,simultaneous seismic acquisition process is described in detail.Based on above research,the signal-to-noise frequency aliasing phenomenon,the difference of correlation between signal-to-noise,frequency stability and Gaussian random noise characteristics are analyzed,respectively.we studied theory and calculation methods of the PCA,and expounds its geometric thought and denoising principle.In addition,the selection of PCA order number and was analyzed from the point of view of Hankel matrices.Second,we studied suppress power frequency noise of simultaneous source by the method of PCA.For the problem of signal-noise spectrum aliasing and power frequency instability,the relationship between the power frequency and the principal component order under the constraint criterion is analyzed,and a specific implementation method was given.Finally,we studied the method of suppressing simultaneous source strong random noise by PCA-EMD,and demonstrated the feasibility of suppressing random noise by PCA.Moreover,the mathematical theory and the suppressing noise principle of the EMD are also described.Based on above analysis and research,a combined noise suppression method of the optimal order number PCA-EMD under strong noise conditions is realized.