Research On Automatic Time Picking For Microseismic Data Based On LLE-ISPO Clustring And ABC-SVM Algorithm

With the great consumption of non-renewable oil and gas resources,the research on unconventional oil and gas reservoir exploitation has become a new hotspot.As an effective method to improve the development effect of unconventional oil and gas reservoirs,microseismic detection technology plays an important role in detecting and guiding the judgment of fracturing effect.The accuracy of first break picking has an important influence on event identification and source inversion.Therefore,first break picking has become an important part of microseismic technology.The purpose of picking up the first break of microseismic events is to find the first jumping point of useful microseismic signals.Time picking is of great concern in the processing of microseismic data.However,the traditional method based on time-domain-frequency can not pick the first arrival time accurately in low signal-to-noise ratio(SNR).Besides,the traditional time picking methods which based on clustering are sensitive to selecting the initial clustering centers and easy to converge to local optimal value.We propose a time picking method for microseismic data based on local linear embedding(LLE)and improved particle swarm optimization(IPSO)clustering algorithm.First,LLE algorithm can obtain the inherent characteristics and the rules hidden in high-dimensional data by calculating Euclidean distances and reconstruction weights between microseismic data points.The input is represented in a low-dimensional form.Then,the improved PSO clustering algorithm is used to select the optimal clustering centers from low-dimensional data through global search method.After that,the low-dimensional data can be classified into noise cluster and signal cluster by the K-means algorithm.Finally,the initial time of the signal cluster can be considered as the first arrival time of microseismic data.Local linear embedding and improved particle swarm optimization algorithm can separate microseismic signal from noise in unsupervised domain.In the field of supervised classification algorithm,this paper focuses on the time picking method based on artificialartificial bee colony optimization supportvector machine(ABC-SVM).We establish classification features based on the large differences in correlation and energy between signals and noises.Then,artificial bee colony(ABC)is used to optimize two important parameters of support vector machine(SVM),C and g.And the optimal classification surface is obtained.Next,the correlation and energy of signal and noise are trained by ABC-SVM and the ABC-SVM classification structure is formed.Finally,the signal and noise are divided into two clusters by using the trained structure.Experimental results of simulated microseismic records with very low signal-to-noise ratio and actual microseismic records can show: The proposed local linear embedding and improved particle swarm optimization algorithm can maintain more than 90% accuracy at the SNR of-10 dB.ABC-SVM algorithm can maintain more than 92% accuracy at-13 dB SNR..Compared with Akaike Information Criterion(AIC)and STA/LTA,the two methods have higher accuracy and stability.