The Methods Research Of Microseismic Monitoring Data Signal Noise Separation Based On Blind Signal Theory

Microseismic monitoring technology is nearly two hundred and thirty years from the fracturing of water injection in low permeability reservoirs area gradually developed a kind of new technology, it through the vibrations caused by the human production activities to observe and analysis to evaluate the effects of production activities, effects, and the status of the underground. Due to the technology with the advantages of high efficiency and low cost, therefore is very extensive application in various fields. With the increase of oil consumption, the low permeability oil and gas, water injection in order to increase production by fracturing is a common method to increase. For fracturing require real-time monitoring in water injection from the formation breakdown of fracturing fracture azimuth, morphology and distribution, in line with the actual situation to the fracturing effect evaluation, so as to timely adjust well pattern arrangement and fracturing process. Microseismic monitoring technology is through the adjacent Wells or surface layout high sensitivity detector to real-time monitoring of fracturing Wells during the water injection fracturing rock failure produced by the seismic waves, seismic records by received inversion can be obtained when fracturing fracture geometry and spatial distribution of growth. Due to the short duration of microseismic events, amplitude is small, weak microseismic signal energy, high frequency, the actual collected seismic data generally contains noise interference. So when making microseismic monitoring is an important step in the effective signal before beginning to pick up the micro seismic data denoising processing, thereby eliminating noise because of seismic data inversion result of the interference of microseismic monitoring, achieve better guidance to the fracturing effect and process. This paper briefly introduces the commonly used several kinds of seismic data denoising method and the basic principle, and sums up some kind of noise in seismic data, and combining with the characteristics of seismic data and noise are introduced for seismic data denoising method.Seismic data denoising processing effect is better, but because it is in the light of the characteristics of seismic data in a certain respect to apply, so often some deficiencies in application, observed in the actual production of the ground microtremor data noise is complex, generally contains a large amount of random noise, so the seismic data is often a lower SNR. In view of this, this topic research and comparison of a series of suitable for the ground microtremor data denoising method, blind source separation technology was applied to seismic data denoising, the purpose is to be able to use this method on the microseismic monitoring data denoising a breakthrough and innovation.Fanaticism treatment usually includes fanaticism, separation (blind source separation), blind deconvolution, blind identify these three aspects of content, this paper mainly discusses the fanaticism, separation (blind source separation), the main principle and mathematical model of independent component analysis is introduced and several commonly used algorithms, and combine the characteristics of the seismic data denoising model is established and the objective function, and write program of synthetic seismic data and actual seismic data to deal with the noise. This article first through the analog signal to the independent component analysis of several common algorithm simulation test, and achieved good separation effect, and then to apply independent component analysis of synthetic seismic record and random noise mixed analog signal, and blind source separation denoising processing, separation, the results show that independent component analysis can be applied to solve the noise elimination of micro seismic data, so as to improve the signal-to-noise ratio of data. In applying this method to the actual seismic data denoising processing, found that the separation effect is not ideal, after mixing microseismic signal and the random noise model is compared with actual observation data in the signal component analysis, it is concluded that may exist the following problems:adjacent two-stage detector receive different noise in the seismic record, two stage to accept the effective signal may vary, time delay existing source to two level detector, downhole geophone inconsistencies in the Angle of each component and so on. The future research work should be put on these question discussion, considering the interference factors affect the separation efficiency as far as possible, so as to realize the method in the application of micro seismic monitoring data denoising processing.