Study On First Arrival Picking For Microseismic Data Based On AN Algorithm

Oil and gas resources are the cherished resources which people depend for their survival.However,due to the increasing consumption of oil and gas resources and the long-term limitation of exploration and production technologies,oil and gas resources have become especially important and precious.Unconventional oil and gas is a resource which cannot be obtained through traditional technologies.This part of the resource has a huge storage capacity.Its rational exploration and development are the solution to the problem of non-renewable energy consumption.Unconventional oil and gas exploration often uses hydraulic fracturing and microseismic detection techniques.Microseismic monitoring is usually divided into forward modeling,filtering denoising,first arrival picking and source inversion positioning.In this paper,the establishment of microseismic signal forward model and the related problems of the microseismic first arrival are studied in detail.The idea of establishing well-joint forward modeling and the approximation of the model are put forward through the principle analysis,experimental design and data processing.We proposed AN algorithm a new method to pick up first arrival of microseismic signal.The establishment of microseismic forward modeling is of great significance for the research of microseismic exploration technology.Microseismic exploration techniques are divided into two types: ground exploration and well exploration.There will be lots of noise in ground exploration,and the obtained data has a low signal-noise ratio and is ineffective.The microseismic exploration technology in wells is more effective,but it is expensive and difficult to operate.In this paper,a well-ground combination method is adopted to establish a two-dimensional forward modeling of microseismic signals.Based on the two-point ray tracing method,the exit angle is continuously adjusted and the optimal exit angle is obtained through iteration to simulate the direct wave,reflected wave,ascending wave and descending wave,get corresponding ray trace and multi-channel synthetic microseismic record,realize full-wave ray tracing.The law of seismic wave propagation can indicate that the forward model established can provide some help for signal analysis and processing,inversion of microseismic data and source localization.Accurate and dependable picking of the first arrival time for microseismic data is an important part in microseismic monitoring,which directly affects analysis and results of post-processing.This paper proposes a new method based on approximate negentropy(AN)theory for microseismic arrival time picking in condition of much lower signal-to-noise ratio(SNR).According to the differences in information characteristics between microseismic data and random noise,an appropriate approximation of negentropy function is selected to minimize the effect of SNR.And a weighted function of the differences between the maximum and minimum value of AN spectrum curve is designed to obtain a proper threshold function.In this way,the region of signal and noise is distinguished to pick the first arrival time accurately.To demonstrate the effectiveness of the AN method,we generate a series of pseudo-synthetic data with different SNR from-1 dB to-12 dB and test through comparing with the previously published Akaike information criterion(AIC)and the short and long time average ratio(STA/LTA).Through the experimental results,these three methods can achieve well picking effect when SNR is from-1dB to-8d B.However,when SNR is as low as-8dB to-12 dB,the proposed AN method is more accurate and stable as compared to AIC and STA/LTA methods.Furthermore,the application to the real three-component microseismic data also shows that the new method is superior to the other two methods in accuracy and stability.