Importance of Low Frequency Model in Model Based Inversion

Deterministic Acoustic Inversion = Developing the concept of inversion Reflection Coefficient R = Describes how much wave energy is reflected at interfaces Full stack seismic data = Assumed to represent zero-offset traces Recursive Inversion = Also known as Bandlimited Inversion

Reflection Coefficient R = Ratio of reflected wave amplitude to incident wave amplitude Acoustic Inversion = Estimating acoustic impedance by inverting seismic data Explicit Deterministic Methods = Taking into account well control availability and uncertainty quantification Recursive Inversion = Earliest form of inversion method

Inversion Objective Function = Discussing different terms related to inversion objectives Reflection Amplitude = Depends on the impedance difference between elastic half-spaces Acoustic Impedance Estimation = Result of inverting full stack seismic data Fine-scale Detail Requirement = Consideration in choice of inversion method

<p>Benefits of Inversion for Interpretation = Reviewing examples of benefits from inversion Deterministic Inversion Methods Choice = Based on well control availability, uncertainty quantification, and detail requirement Bandlimited Inversion = Simplest form of inversion method Impedance Difference Influence = Affects the reflection amplitude at interfaces</p> Signup and view all the answers

<p>Reflectivity = Approximation of the earth's properties based on seismic trace Impedance = Conversion of reflectivity trace obtained from seismic data Seismic trace = Representation of the earth's reflectivity Wavelet = Operator applied in frequency domain for inversion</p> Signup and view all the answers

<p>Colored inversion = Fast method using operator O in frequency domain for impedance transformation Sparse Spike Inversion = Identifies large spikes in reflectivity and generates synthetic trace for comparison Model Based Inversion = Assumes known seismic trace and wavelet to solve for reflectivity iteratively Bandlimited method = Inversion process considering missing low and high frequency ends of reflectivity spectrum</p> Signup and view all the answers

<p>Fast method suitable for initial look inversion = Colored inversion Less dependent on initial guess model = Sparse Spike Inversion Creates broad-band, high frequency model = Sparse Spike Inversion Solves iteratively to minimize 'misfit' = Model Based Inversion</p> Signup and view all the answers

<p>Low Frequency Model = Incorporating well data or seismic RMS/migration velocity volumes High frequency end of reflectivity spectrum = Missing in seismic trace due to bandlimitation Phase errors and noise contamination = Challenges in processing seismic trace Sparse spike embedded in a background of small spikes = Basis for Sparse Spike Inversion method</p> Signup and view all the answers

<p>Low-frequency information loss = Impacts impedance log resolution and velocity structure Bandlimited seismic signal spectrum = Narrower frequency spectrum characterized by missing low and high frequencies Convolution model in frequency domain = Equivalent to multiplication of two amplitude spectrums Full frequency spectrum from 0 Hz to Nyquist = P-impedance log shown in Model Based Inversion</p> Signup and view all the answers

<p>Simplicity Term (Fcontrast) = Aims to create an inverted impedance log with few reflection coefficients Seismic Term (Fseismic) = Controls the seismic residuals or misfit between real seismic and synthetic modelled traces Trend Term (Ftrend) = Stabilizes low frequencies and controls low frequency model contribution Spatial Term (Fspatial) = Controls the smoothness of the solution</p> Signup and view all the answers

<p>Delta IP = Elastic parameter change in the vertical (time) direction, normalized to low pass filtered trends S divided by N = Parameter to control the weight of Fseismic Delta P low pass = Difference between current low pass filtered P-impedance trace and low pass filtered trend u = Parameter to control the weight of Fcontrast</p> Signup and view all the answers

<p>Increasing uncertainty leads to smaller constraint u. = Simplicity Term (Fcontrast) Higher S/N ratio leads to more influence of seismic fit in inversion solution. = Seismic Term (Fseismic) Small residuals require a detailed model, i.e. a non-sparse solution. = Spatial Term (Fspatial) Low pass filtered trends are controlled by Trend Term Ftrend. = Trend Term (Ftrend)</p> Signup and view all the answers

<p>Inverse distance interpolation = Assigns values to unsampled locations based on the values at nearby sampled locations Kriging = Estimates the value of a point based on a weighted average of surrounding points Collocated co-kriging = Incorporates secondary data to improve the estimation of primary data External drift kriging = Uses a secondary variable to improve the estimation of the primary variable</p> Signup and view all the answers

<p>RMS velocities = Integrated as a primary component if well logs are missing Migration velocities = Integrated as a secondary property that guides the interpolation away from wells FWI velocities = Integrated as a primary component if well logs are missing Well logs = Record data which is both lower and higher frequency than seismic data</p> Signup and view all the answers

<p>Iteration process = Updates elastic attributes of the model to reduce the difference between synthetic and seismic data Cost function = Contains multiple misfit functions that need to be minimized during inversion Objective function = Determines the Optimum P-impedance by minimizing multiple misfit functions Misfit functions = Grouped into ensembles and used to evaluate the fit between seismic and synthetic data</p> Signup and view all the answers

<p>0 Iterations = Initial inverted solution is equal to the initial model with no contrasts or events in synthetic seismic 1 Iteration = Updated solution for acoustic impedance introduces contrasts and more detail in synthetic seismic 2 Iterations = Further updates improve inverted impedance leading to a reducing difference between real and synthetic seismic traces 3 Iterations = Each iteration further improves the model towards closely matching real seismic</p> Signup and view all the answers

<p>Minimization of Seismic-Synthetic misfit = Should be minimized during inversion process Inverted impedance close to well impedance = Captures contrasts and variations, influences two terms of cost function Control weights influence by testing of inversion = Project-dependent decision based on noise level in seismic data Project-dependent factor prioritization = May focus more on well control if seismic data are noisy</p> Signup and view all the answers

<p>Improved Horizon Interpretation = Simpler section due to attenuation of wavelet sidelobes Lateral Continuity = Analysis of interfaces between rock layers and elastic properties Rock Characterization = Analyzing changes in rock characteristics between well locations Prospect Identification = Using acoustic impedance for better prospect identification</p> Signup and view all the answers

<p>Stacked data represent zero-offset traces = Acoustic inversion assumptions All reflections in poststack data are primaries = Acoustic inversion assumptions The data are free of noise = Acoustic inversion assumptions AVO effects are not present = Acoustic inversion assumptions</p> Signup and view all the answers

<p>Model Based = Left side comparison method Sparse Spike = Middle comparison method Colored Inversion = Right side comparison method Inversion Results vs. Well Log Data = Comparison method at a well location</p> Signup and view all the answers

<p>Yellow Horizon Completion = Resulting in a much simpler section Lateral Continuity Analysis = Showing interfaces between rock layers and elastic properties Improved Prospect Identification = Using seismic tuning and acoustic impedance for better prospect identification Rock Characterization between Well Locations = Easier analysis of rock characteristic changes</p> Signup and view all the answers

<p>Interpretation of Southeast Asia Clastic Example = Close vertical positioning of contrasting layers within half of a wavelet length General Interpretation Challenges = Severe interference (tuning) and complication of the seismic section Yellow Reservoir Event Interpretation = Particularly difficult interpretation due to tuning effects Map Data Analysis Challenges = Lower sand appearing brighter due to seismic tuning, leading to more contrast</p> Signup and view all the answers

<p>Integration of all types of data for straightforward interpretation = Benefits summary provided by acoustic inversion Elimination of false stratigraphic-like effects = Wavelet side lobes attenuated for simpler interpretation Analysis of lateral continuity and interfaces between rock layers = Benefit related to seismic and elastic properties visualization Better characterization and prospect identification using different layers properties = Prospect identification benefits using acoustic impedance</p> Signup and view all the answers

<p>Analysis of rock characteristics changes between well locations = Detail provided by acoustic inversion on rock characterization Visualization showing interfaces between rock layers and elastic properties = Detail provided by acoustic inversion on lateral continuity analysis Identification of better prospects through seismic tuning and layer properties contrast = Detail provided by acoustic inversion on improved prospect identification Challenges in interpretation due to tuning effects and vertical layer positioning = Detail provided by acoustic inversion on interpretation challenges</p> Signup and view all the answers

<p>Changes in rock characteristic analysis between wells locations using P-impedance change visibility = Example illustrating benefit related to rock characterization analysis between wells locations Seismic data showing interfaces between rock layers and lateral continuity = Example illustrating benefit related to lateral continuity analysis Improved horizon interpretation due to attenuation of wavelet sidelobes = Example illustrating benefit related to improved horizon interpretation Prospect identification enhancement through seismic tuning and layer property contrast = Example illustrating benefit related to prospect identification enhancement</p> Signup and view all the answers

<p>Difficulties in interpreting Southeast Asia clastic example due to close vertical layer positioning = Interpretation challenge related to specific example in Southeast Asia clastic area Severe interference and complication in seismic section interpretations due to tuning effects = Interpretation challenge related to severe interference and complication Particularly difficult yellow reservoir event interpretation due to tuning effects = Interpretation challenge related to specific yellow reservoir event Challenges in map data analysis due to brightness differences caused by seismic tuning = Interpretation challenge related to map data brightness differences</p> Signup and view all the answers

<p>Integration of various data types for straightforward interpretation = Summary benefit related to integration of various data types for interpretation Elimination of false stratigraphic-like effects through wavelet sidelobes attenuation = Summary benefit related to elimination of false stratigraphic-like effects Analysis showcasing lateral continuity and interfaces between rock layers = Summary benefit related to visualization analysis on lateral continuity Better prospect identification through layer property contrast and seismic tuning = Summary benefit related to prospect identification enhancements</p> Signup and view all the answers