Study On Automatic Time Picking For Microseismic Data Based On FCM Algorithm

The microseismic event is caused by dynamic underground stress field and rock stratum crack or fracture when hydraulic fracturing is used to monitor low permeability oil or gas reservoir.The most important means to deal with hydraulic fracturing is microseismic monitoring technology.Dynamic monitoring of fracture and oil or gas reservoir location is achieved by recording microseismic events continuously using three-component geophones which are placed around the treatment well.Microseismic monitoring technology has unique advantages in explaining and evaluating the effect of fracturing.Meanwhile,it provides theoretical basis and technical support for the application of hydraulic fracturing technology in modifying low permeability oil or gas fields.The purpose of explaining high density microseismic records in real-time is to locate microseismic hypocenter precisely.Automatic time picking is an important step in microseismic hypocenter location.On the basis of ensuring the accuracy of microseismic records,the design of real-time,fast,automatic and stable time picking algorithm is the key to improve the signal processing speed and field monitoring efficiency.The effective signal of the real microseismic record is of high frequency,weak energy and short duration.The low signal to noise ratio(SNR)of microseismic record is due to the interference of complex random noise.Manual picking is of low efficiency and can not meet the requirements of real-time processing.The results of traditional automatic microseismic time picking algorithms such as short and long time average(STA/LTA)algorithm and Akaike information criterion(AIC)algorithm are satisfactory in picking high SNR microseismic data.However,the results of time picking is not convincing when the SNR of data is low.In this paper,the types of microseismic noise,the frequency and energy characteristics of effective signals are discussed comprehensively and systematically based on the characteristic difference of effective signals and noise.A new automatic first arrival picking algorithm is proposed based on fuzzy C-means clustering(FCM)algorithm according to the difference of signals and noise.The membership degree matrix which represents the degree of similarity can be obtained by updating objective function constantly.Thesample point will be divided into the effective signal cluster when the value of the corresponding membership degree matrix is larger than the predetermined threshold.It is considered that the first arrival time of the microseismic data is the starting time of the effective signal cluster.Additionally,the numerical value and selection reason for the initial parameters of fuzzy index and membership matrix are given and its feasibility and validity are discussed in detail.The picking performance of FCM algorithm and the effect of processing simulation and actual microseismic data are analyzed integrally.The reliability and accuracy of the algorithm are verified by the receiver operating characteristic(ROC)curve and error statistics histogram.This proposed algorithm can pick up the first arrival precisely when the SNR is-8d B,and its accuracy ratio can reach 86%.The FCM algorithm is tested on both the synthetic and real microseismic record and the results are compared both with the STA/LTA algorithm and the AIC algorithm.The comparison results show that the accuracy of the proposed algorithm is better than the other two algorithms.And the approach resolves the problem which traditional time picking algorithms can not pick up the first arrival effectively when the SNR of microseismic data is low.