Artificial neural network based autopicker for micro‐earthquake data

Geophysical Prospecting

Karrenbach³

2014

Self Cite

Microseismic monitoring is an increasingly common geophysical tool to monitor the changes in the subsurface. Autopicking involving phase arrival detection is a common element in microseismic data processing schemes and is necessary for accurate estimation of event locations as well as other workflows such as tomographic or moment tensor inversion, etc. The quality of first arrival picking is dependent on the actual seismic waveform, which in turn is related to the near surface and subsurface structure, source type, noise conditions, environmental factors, and monitoring array design, etc. We have developed a new hybrid autopicking workflow which makes use of multiple derived attributes from the seismic data and combines them within an artificial neural network framework. An evolutionary algorithm scheme is used as the network training algorithm. The autopicker has been tested and its applicability has been validated using a synthetically modelled seismic source, with promising results. In this work, we share the basic workflow and different attributes that have been tested with this algorithm for a synthetic data set to provide a framework for independent implementation, use and validation. We also compare the results obtained using the new neural network based autopicking routine with very robust contemporary autopicking algorithms in use within the industry.

Section: Introductionmentioning

confidence: 68%

Section: Introductionmentioning

confidence: 97%

Novel hybrid artificial neural network based autopicking workflow for passive seismic data

Geophysical Prospecting

Karrenbach³

2014

Self Cite

“…Based on the available dataset and initial analysis of data characteristics, a modified autopicker based on an ANN based workflow (Aminzadeh et al, 2011) was used to generate necessary first break arrival data (both p and s wave) and a catalog was generated. Hypoinverse (Klein, 2003) and SimulPS (Thurber, 1993) were used to invert for initial hypocentral estimates and velocity model refinement as well as relocation.…”

Section: Methodsmentioning

confidence: 99%

Framework for time lapse fracture characterization using seismic, microseismic & well log data

SEG Technical Program Expanded Abstracts 2012

2012

Self Cite

Extensive work has been done in the recent years involving use of conventional and passive seismic data for fracture characterization. This is particularly the case with unconventional reservoirs such as shale gas, shale oil and geothermal fields. The purpose of our study is to combine the benefits of conventional seismic data that provides relatively higher resolution reservoir characteristics with the relatively low resolution property estimates available from inversion of micro-earthquake data. Given the lower cost of the latter, we propose a cost effective dynamic reservoir characterization approach using a self-sustaining evaluation framework. The resulting time lapse fracture characterization technique is most suitable for those developments which involve the use of low cost passive seismic data acquisition arrays for reservoir monitoring. Our proposed method should allow for optimal use of microseismic data generated as part of passive seismic arrays common in unconventional field developments and thereby provide time lapse reservoir property predictions without having to carry out relatively expensive 4D seismic surveys.

“…The triggered data is run through an advanced ANN based autopicking workflow (Aminzadeh et al 2011) and the final picks obtained are used to detect phase arrivals for use with inversion algorithms for both location and velocity. Due to the nature of the data as well as limitations with the monitoring array design, only ~6% of the detected events are used in further analysis.…”

Section: Microseismic Data Analysismentioning

confidence: 99%

Fracture Characterization in Unconventional Reservoirs using Active and Passive Seismic Data with Uncertainty Analysis through Geostatistical Simulation

SPE Annual Technical Conference and Exhibition

2013

This study discusses a new workflow for fracture characterization and modeling using geophysical (microseismic and 3D surface seismic) data along with independent reservoir information (such as well logs). The framework is ideally suited for unconventional environments such as shale and tight reservoirs where modern technologies such as the use of hydraulic fracturing and passive seismic monitoring allow application of the proposed workflow.The workflow involves generating geomechanical property estimates (including stress and weakness estimates) as derived from passive seismic data analysis and relevant seismic attributes derived from 3D seismic data combined using ANN based reservoir property modeling framework. The training information for the networks is generated based on a-priori information through image logs. Resolution of passive seismic derived velocity models is improved by using sequential Gaussian co-simulation by combining low resolution velocity maps high resolution seismic impedance data for phase velocity estimation. Uncertainty estimates are quantified by adequate number of realizations and associated probability density functions for fracture properties within study volume.In this paper, different properties estimated through ANN modeling have been shared. New fracture identifier (FZI) properties have been defined and the models have been used to characterize fracture zones and major discontinuities for a representative unconventional reservoir (geothermal setting) used in our study. We also share uncertainty estimates for the identified fracture zones for improved characterization. Finally fracture property estimates for the study area (derived using FZI and other properties) have been generated for future reservoir simulation studies.The proposed method allows for improved understanding of shale and other unconventional reservoirs through fracture mapping and provides a workflow for improved volumetrics of the reservoir by making use of identified properties for fracture modeling. This work validates the potential for using relatively low resolution passive seismic data for improved reservoir characterization using Geostatistical tools. It also provides a valuable framework for pseudo 4D characterization where a single 3D seismic survey can be used as the basis to characterize the reservoir in a time lapse fashion using new information collected in time through passive seismic arrays as well as new well logs being obtained within the area of interest.