«UNIVERSITY OF OKLAHOMA GRADUATE COLLEGE SEISMIC DATA CONDITIONING FOR QUANTITATIVE INTERPRETATION OF UNCONVENTIONAL RESERVOIRS A DISSERTATION ...»
UNIVERSITY OF OKLAHOMA
SEISMIC DATA CONDITIONING FOR QUANTITATIVE INTERPRETATION OF
SUBMITTED TO THE GRADUATE FACULTYin partial fulfillment of the requirements for the Degree of
DOCTOR OF PHILOSOPHYBy SUMIT VERMA Norman, Oklahoma
SEISMIC DATA CONDITIONING FOR QUANTITATIVE INTERPRETATION OF
A DISSERTATION APPROVED FOR THE
CONOCOPHILLIPS SCHOOL OF GEOLOGY AND GEOPHYSICSBY ______________________________
Dr. Kurt J. Marfurt, Chair ______________________________
Dr. Deepak Devegowda ______________________________
Dr. Matthew J. Pranter ______________________________
Dr. Xiaowei Chen ______________________________
Dr. Steve Roche © Copyright by SUMIT VERMA 2015 All Rights Reserved.
To my family and friends like family
I received an enormous amount of guidance, knowledge, encouragement and care from my supervisor, Dr. Kurt Marfurt. In the beginning of my PhD, I was not sure what I wanted to do. Dr. Marfurt (Kurt Ji) has been patient with me and provided me with scope to think and figure out what I wished to do. I wanted to learn something new during my PhD. Coming into the program I had almost no knowledge of seismic data processing.
Dr. Marfurt wanted me to be exposed to processing as if it were a viral flu. I was meant to be developing an immunity to said flu. I am glad that I learned processing, which I now understand as well as like. He also provided me with an opportunity to be Lab guru for his classes, which not only helped me in my PhD, but also inspired me to choose academia as my career. I feel proud to be a student of Dr. Marfurt. To me he is one of the most wonderful people I have ever met in my life.
I would also like to express my sincere thanks to Dr. Matthew J. Pranter. He provided me geological inputs for my Mississippi lime project. He also encouraged me to present at the Mississippi lime consortium, which provided me a platform to share my research with the industry experts.
I would like to thank Dr. Deepak Devegowda, who is one of my committee members, for being very easy going and friendly as well as for his help with geo-statistics in Barnett Shale TOC and Brittleness estimation project. I also want to express my gratitude to Dr. Vikram Jayram who motivated and excited me to work on the Support Vector Machine.
members. He provided me with feedback on my ground roll suppression work as well as help with data for my project. I would also like to thank Dr. Jamie Rich for all his help.
I would like to thank Dr. Xiaowei Chen for agreeing to become my committee member. Without his help, my PhD would have been significantly more challenging.
I would also like to express my sincere thanks to Late Dr. J. Tim Kwiatkowski.
He was a source of inspiration to me. He also helped whenever there were any technical issues in the lab, preventing many setbacks along the way. I think I speak for all the students who ever worked in the 10th floor lab when I say that we did not notice how invaluable he was until after he was gone.
I would also like to thank Dr. Roger Slatt for his important comments and feedbacks on my research. I learned a lot of reservoir characterization from his courses.
I would like to thank the entire staff of ConocoPhillips School of Geology and Geophysics. I cannot forget the care of Nancy Leonard and Adrianne Fox, as they looked the finances and reimbursements and made my life simple. I also want to thank Donna Mullins who looked after issues concerning both the graduate college and courses. She has vast amounts of knowledge of the inner workings of the graduate college. She saved me from time to time by providing me correct information about graduate college rules.
I would like to thank Teresa Hackney, who made me feel at home whenever I visited the geology office. I would like to thank Rebecca Fay for taking care of all the dissertation defense and graduate college issues. I would also like to thank Jocelyn Cook and Devon Harr. I would like to thank Brandy Gunter, my graduate college advisor, who has made it simpler for me to transfers credit hours and apply for removing and adding courses.
to say thanks to Sean O'Bleness and Grant Butler for helping me with any IT related problems.
I would like to thank my friends and co-authors on my different papers Shiguang Guo, Marcus Cahoj, Tao Zhao, Yoryenys DelMoro, Fangyu Li, Onur Mutlu, Thang Ha, Roderick Perez, and Manuel Aguilar. I would like to also thank Alfredo Fernandez, Ben Dowdell, and Mark Aisenberg for their help with my dissertation. I would like to mention the names of my other friends who helped me during my journey of four years: Bo, Tengfei, Jie, Gabriel, Bryce, Jyot, Sayantan, Mohsen, Oswaldo, Luis, Richard, Dania, Tobi, Joe, Trey Stearns, Dustin, Mike, Caleb, Araceli, Alyssa, Yuji, Xuan, Murphy, Aliya and Bunmi.
I would like to give special dhannobad to Atish Roy for all his help and support.
I would like to show my gratitude to Supratik Da and Nabanita Di, who acted as my local guardians in the USA. I would like to also thank my friends Avinash, Bijit, Sara, Dhanya (Madam), Mahesh and Siddhesh for their support. I would like to also thank India Student Association for their support.
I also want to thank the AASPI Consortium sponsors, and my fellow graduate students at the ConocoPhillips School of Geology and Geophysics for their help and support.
Last but not the least, I would like to thank my family and my friends in India for supporting me to continue with my PhD, especially my mother who has made great sacrifices for my studies.
: Interpretation Pitfalls
Pitfalls in Prestack Inversion of merged seismic surveys
Assumptions for Prestack Inversion
Modeled to measured data misfit
Offsets, fold, and prestack migration
Validation of our hypothesis
The Solution – inversion using shorter offsets
: Seismic modeling
Calibration of attribute anomalies through prestack seismic modeling
Case study 1: Seismic Modeling of Chicontepec Basin’s Tectonic Structure........ 36 Seismic modeling of a pop-up structure
Seismic modeling of a graben structure
Discussion of results for case study 1
Case study 2: Seismic modeling of impedance anomalies associated with faults in the Woodford Shale
The Fault Model :
The Fracture Model :
Discussion on results for case study 2
Case studty 3 : Modeling Sags - are they Karst Collapse or Gas Chimneys Pushdown?
Karst Collapse Model:
Gas Chimney model:
Discussion on results for case study 3
Case 4 : Identifying processing challenges with seismic modeling
Figures and Tables
: Pitfalls in seismic processing
Pitfalls in seismic processing: part 1 groundroll sourced acquisition footprint....... 74
: Seismic Data Conditioning
Highly aliased groundroll suppression using a 3D multiwindow KL filter: Application to a legacy Mississippi Lime survey
Exploration Objectives and Data Description
Validation with a Synthetic patch
: Estimation of TOC and Brittleness
Geology of the study area
Correlating Core to Wireline Measurements
TOC estimation using Passey’s equation
Brittleness estimation using Wang and Gale equation
Correlation to wireline logs
Volumetric estimation of TOC and brittleness
Volumetric TOC estimation
Volumetric Brittleness estimation
Correlation of TOC and BI to Relative EUR: AASPI proto type Cigar Probe 135 Limitations of TOC and BI correlation with Production using cigar probe..... 137 Results
Disscussion and Conclusions
Figures and Tables
Table 3.1 Survey geometry created for models.
Computation cost for the processes.
Table 6.1 TOC estimated with window of ±1ft has least validation error and highest correlation with 4 well logs.
BI estimated with window of ±2ft has least validation error and highest correlation with 4 well logs.
Validation and Training error using a window of ±8ms for volumetric TOC estimation.
Table 6.4 Validation and Training error using a window of ±12 ms for volumetric BI estimation.
Location map of Anadarko basin area on map of Oklahoma, and location of study area in Anadarko basin marked by green boundary (modified from Northcutt and Campbell, 1988).
Stratigraphy of Anadarko basin in Pennsylvanian and Mississippian age, here Red Fork Formation and two of the geologic formations that appear as strong reflectors on seismic are highlighted in pink. Hunton (highlighted in blue) and Woodford (highlighted with green) are also formation of interest for current operators in the area (Modified from Clement, 1991).
“Fold Map” of the reprocessed megamerged 3D seismic data volume.
Superficially, this gives the impression that the data are greater than 25 fold throughout the survey
Phantom horizon slices 80 ms below Oswego cutting the Red Fork incised channels through (a) the P-impedance volume, ZP, (b) the S-impedance volume, ZS, computed from 2-42 input migrated gathers. For both of the figures, white arrows indicate artifacts in the resulting image. Black dotted arrow indicates a circular artifact.
(a) Synthetic gather generated at a well, with angles ranging between 0-42.
(b) Synthetic gather generated at a well, with offset range 0-22, and padded with zero
(a) and (b).
Representative gathers and base map indicating their locations. Note that location A and D have moderate amplitudes while B and C have low amplitudes at the farther offsets. The small residual amplitudes beyond these ranges are due to migration swings from the longer offset surveys.
Mean-squared error map showing the difference between the measured and modeled seismic gathers for the 2°–42° inversion.
Horizon slices along the Oswego surface through offset-limited stacked amplitude volumes: (a) 0-1520 m (0-5000 ft) (b) 1520-2450 m (5000-8000 ft) (c) 2450m (8000-11,000 ft) (d) 3350-4250 m (11000-14000 ft) and (e) 4250-5200 m (14000-17100 ft). The Oswego Lime was interpreted as a strong peak in the stacked seismic volume. Amplitude changes in c may be valid AVO effects. Often, inaccurate velocities (including anisotropic effects) result in misaligned gathers giving rise to zero crossings and troughs at far offsets. However, note how the amplitude approaches zero in the top right corner of the megamerged survey in (d) and (e) indicating that these large offsets were never recorded in these areas. White polygons in (c) indicate amplitude anomalies that will be used in subsequent quality control. White arrows indicate the major highways.
Phantom horizon slices 80 ms below the Oswego through (a) the P-impedance volume, ZP, (b) the S-impedance volume, ZS, computed from 2-22 input migrated gathers. Pink polygons correspond to an area of high error shown in Figure 2.10........ 27
modeled seismic gathers for the 2-22 inversion. To compare with Figure 6 the squared error was normalized with respect to the number of traces in each gather. White arrow corresponds to those drawn about amplitude anomalies shown in Figure 8c. Pink polygons encircle an area of high error that is posted on Figure 2.9.
Figure 3.1 Flowchart for generation of synthetics.
Figure 3.2 Horizon slice along the top Jurassic through co-rendered coherence, mostpositive curvature, and most-negative curvature.