Based on the prestack P-S wave velocity ratio analogy method under drive waveform
Technical field
The invention belongs to oil-gas seismic exploration technical fields, more particularly, are related to a kind of prestack elastic parameter simulation side
Method.
Background technique
Inversion technique, which is developed so far, has nearly 40 years history, including a variety of algorithms of different.It can totally be divided into based on earthquake
Determinated back analysis and geostatistics stochastic inverse two major classes.
Determinated back analysis goes wavelet to seek reflection coefficient mainly from earthquake using convolution model, obtains opposite resistance
Anti-, resolution of inversion is substantially suitable with earthquake, and inversion result is determining.The Sparse Pulse seismic inversion side of such as Lvey proposition
Method, the inversion method are not had direct corresponding relationship by thin-layers interference function influence, earthquake information and thin reservoir, can not be to thin layer cross
It makes and being effectively predicted to variation, big additionally, due to changing between reservoir well, seismic resolution is low, and interwell reservoir forecasting is caused to exist very
It is big uncertain.
In order to make up the low defect of determinated back analysis resolution ratio, grown up statistics inversion method.This method be from
Well sets out, and stochastic model theory is combined with seismic inversion, using geostatistics thought, characterizes reservoir by variogram
Spatial Variability, carry out the stochastic simulation constrained by seismic impedance on this basis, estimate the high-resolution of lithology and reservoir
The regularity of distribution, as a result, one group of equiprobability stochastic simulation solution.Geostatistics are introduced the 1990s as Haas is equal to
Inverting, this method is distributed based on spatial domain sampling point, equally distributed to well more demanding.
Although there is high-resolution advantage to lead to its bullet due to the uncertainty of stochastic simulation solution for statistics inverting
Property parameter Vp/Vs spatial distribution characteristic and geologic rule and understanding it is usually inconsistent, therefore, phased statistics inversion method
Cause the attention of people.Many foreign well-known software companys all develop corresponding function, such as Jason, Petrel etc.,
The problem of being inherently the final inversion result of deposition phase control using artificial division, being primarily present three aspects: (1) sedimentary facies
The subjectivity of research leads to subjectivity since different people mutually has biggish difference using what data with different was divided in different time
Property it is very strong, system lack iteration correction function, can not judge the correctness and accuracy of subjective Division of Sedimentary Facies, will lead to simultaneously
The reasonability problem that phase boundray divides;(2) scale problem of facies analysis, sedimentary facies are divided into different scale scales, the form of expression
Although usually 2 d plane picture, what is substantially expressed is a three-dimensional concept, represents certain thickness.Different phase essences
The vertical scale of degree, expression is different, including sedimentary facies, parfacies, microfacies etc. respectively represent different stratigraphic units.Laterally dividing
On cloth, the sandy ground ratio of isopic zone, reservoir thickness do not change constantly yet, and therefore, actually deposition phasor General Expression is one group
There is the distribution relation of the stratum of genetic relationship spatially, it is difficult to establish accurate mathematical expression model;(3) facies analysis is fuzzy
Property, it is actually the concept of advantage phase as expressed by sedimentary facies, refers to after vertical a variety of depositional environments transition based on performance
The phase of environment is wanted, therefore inevitably leads to vertical facies tract fuzzy problem, influences the resolution ratio of Vp/Vs prediction, together
When also influence lithology, the resolution ratio of reservoir prediction.
Summary of the invention
Aiming at the problems existing in the prior art, it is an object of the invention to solve in deficiencies of the prior art
At least one of.For example, improving Vp/Vs (prestack it is an object of the invention to solve the problems, such as traditional facies control analysis manual intervention
P-S wave velocity ratio) simulation precision.
To achieve the goals above, the present invention provides a kind of prestack P-S wave velocity ratio simulations based under drive waveform
Method, the analogy method include: the input data body that the simulation of prestack P-S wave velocity ratio is calculated;To the input data
Body carries out phased sampling;According to the phased sampling as a result, simulating pseudo- well data;According to the pseudo- well data and fixed well number
According to establishing initial model, and the probability distribution of known well logging sample is counted as prior probability distribution;According to the initial model
With the input data body, likelihood function is calculated;According to the initial model and the prior probability distribution and it is described seemingly
Right function carries out Markov Chain-Monte Carlo stochastic simulation under Bayesian frame, obtains analog result.
In an exemplary embodiment of the present invention, can by prestack elastic parameter P-S wave velocity ratio invertomer with
Comprehensive wavelet convolution obtains the input data body.
In an exemplary embodiment of the present invention, the prestack elastic parameter can be obtained by prestack Simultaneous Inversion
P-S wave velocity ratio invertomer.
In an exemplary embodiment of the present invention, the prestack Simultaneous Inversion can according to Zuo Yipulizi equation or
Its approximate expression is sought with borehole restraint Sparse Pulse algorithm.
In an exemplary embodiment of the present invention, the synthesis can be extracted according to post-stack seismic data and well data
Wavelet.
In an exemplary embodiment of the present invention, the phased sampling can use seismic waveform similarity feature and
Maximum space distance.
In an exemplary embodiment of the present invention, the maximum space distance can be is sought by spatial function
Vector distance.
In an exemplary embodiment of the present invention, the simulation of the pseudo- well data may include:
The input data body well bypass road of N mouthfuls of fixed wells and the input data of the unknown well of a bite are calculated according to formula 1
Relation function between the road of body well bypass road.
Formula 1 are as follows: Snew=Σ wiSi, i=1,2,3..., N.
Wherein, wiThe relation function between the road of the input data body, SnewBy the input data body well of unknown well
Road, SiFor the input data body well bypass road of fixed well.
The relevant matches relation function of different wave parameter is sought according to formula 2.
Formula 2 are as follows: Pb=Σ λabPa。
Wherein, λabFor the relevant matches relation function of different wave parameter, PbFor the waveform parameter of the input data body,
PaFor the fixed well data waveform parameter.
The spatial relationship function between different wave parameter is sought according to formula 3.
Formula 3 are as follows:
Wherein,Spatial relationship function between different wave parameter, wiThe relationship between the road of the input data body
Function, λabFor the relevant matches relation function of the different wave parameter.
The pseudo- well data are sought according to formula 4.
Formula 4 are as follows:
Wherein, LnewFor the pseudo- well data,For the spatial relationship function between the different wave parameter, LiFor institute
State fixed well data.
In an exemplary embodiment of the present invention, the pseudo- well data can be Acoustic Impedance Data.
It in an exemplary embodiment of the present invention, can be by the initial model and input data body progress
Likelihood function is calculated with filtering.
Compared with prior art, the beneficial effect comprise that prestack elastic parameter analogy method under drive waveform
The cross directional variations of seismic waveform can be made full use of to reflect the phase change characteristics of reservoir space, and then analyze to avoid manual intervention
The vertical lithology combination high-frequency structure feature of reservoir, preferably embodies phased thought, is that a kind of real well shake combines high frequency
Analogy method makes inversion result from completely random to gradually determining, while not being strict with to the uniformity of well location distribution, mentions
High simulation precision, also increases the scope of application of lithology, reservoir and fluid prediction.
Detailed description of the invention
Fig. 1 show an exemplary embodiment of the present invention based on the prestack P-S wave velocity ratio mould under drive waveform
The modeling process chart of quasi- method.
Fig. 2 shows an exemplary embodiment of the present invention based on the prestack P-S wave velocity ratio mould under drive waveform
The phased sampling schematic diagram of quasi- method.
Fig. 3 show an exemplary embodiment of the present invention based on the prestack P-S wave velocity ratio mould under drive waveform
The pseudo- well digital simulation schematic diagram of quasi- method.
Fig. 4 show an exemplary embodiment of the present invention based on the prestack P-S wave velocity ratio mould under drive waveform
The simulation of quasi- method connects well profile.
Description of symbols:
A well, B well, C well, D well, E well, F well indicate drilling well, H8 indicate the bottom of 8 lower section of box, SX indicate 1 section of mountain bottom,
Vp/Vs indicates P-S wave velocity ratio.
Specific embodiment
Hereinafter, one kind that the present invention will be described in detail will be carried out in conjunction with exemplary embodiment and attached drawing to be based under drive waveform
Prestack P-S wave velocity ratio analogy method.
Fig. 1 show an exemplary embodiment of the present invention based on the prestack P-S wave velocity ratio mould under drive waveform
The modeling process chart of quasi- method.As shown in Figure 1, in one exemplary embodiment, it is of the invention based on folded under drive waveform
Preceding P-S wave velocity ratio analogy method can be achieved by the steps of:
Step (1) obtains prestack P-S wave velocity ratio invertomer according to prestack Simultaneous Inversion, and P-S wave velocity ratio is anti-
Body and comprehensive wavelet convolution are drilled, the input data body of prestack P-S wave velocity ratio simulation is obtained.
Step (2) carries out phased choosing to input data body according to seismic waveform similarity feature and maximum space distance
Sample.
Step (3), according to the phased sampling as a result, simulating pseudo- well impedance data using fixed well impedance data.Institute
The simulation for stating pseudo- well impedance data specifically includes:
The input data body well bypass road of N mouthfuls of fixed wells and the input data body well of the unknown well of a bite are calculated according to formula 1
Relation function between the road of bypass road.
Formula 1 are as follows: Snew=Σ wiSi, i=1,2,3..., N.
Wherein, wiThe relation function between the road of input data body, SnewFor the input data body well bypass road of unknown well, Si
For the input data body well bypass road of fixed well.
The relevant matches relation function of different wave parameter is sought according to formula 2.
Formula 2 are as follows: Pb=Σ λabPa。
Wherein, λabFor the relevant matches relation function of different wave parameter, PbFor the waveform parameter of the input data body,
PaFor the fixed well data waveform parameter.
The spatial relationship function between different wave parameter is sought according to formula 3.
Formula 3 are as follows:
Wherein,Spatial relationship function between different wave parameter, wiThe relationship between the road of the input data body
Function, λabFor the relevant matches relation function of the different wave parameter.
The pseudo- well data are sought according to formula 4.
Formula 4 are as follows:
Wherein, LnewFor the pseudo- well data,For the spatial relationship function between the different wave parameter, LiFor institute
State fixed well data.
Step (4) establishes initial model according to pseudo- well impedance data and fixed well impedance data, and counts known sample of logging well
This probability distribution is as prior probability distribution.
Initial model and input data body are carried out matched filtering, likelihood function are calculated by step (5).
Step (6) carries out Ma Erke according to initial model, prior probability distribution and likelihood function under Bayesian frame
Husband's chain-Monte Carlo stochastic simulation obtains meeting the impedance of earthquake intermediate frequency and well curvilinear structures characteristic simulation result.
Come below in conjunction with specific example to the prestack P-S wave velocity ratio simulation side of the invention based under drive waveform
Method is further described.
Example
Detailed description is based on the prestack P-S wave velocity ratio analogy method under drive waveform by taking the work area X as an example.
Step (1) carries out prestack Simultaneous Inversion to work area seismic data, obtains prestack elastic parameter Vp/Vs invertomer, leads to
The comprehensive wavelet of well shake combined extracting is crossed, comprehensive wavelet and Vp/Vs invertomer convolution are obtained into Vp/Vs generated data body, as folded
The input data body of preceding elastic parameter simulation.Prestack Simultaneous Inversion is according to Zoeppritz equation or its approximate expression, about with well
Beam Sparse Pulse algorithm can guarantee that the longitudinal wave obtained, S-wave impedance value are believed within the scope of seismic band with actual seismic in this way
Breath coincide, and it is final Vp/Vs mould that the inversion technique uniqueness is good, anti-noise ability is strong, small to well dependence, result stability is good
Quasi- result provides the certainty ingredient for respecting original earthquake data.
Step (2) carries out phase using seismic waveform similarity feature under chronostratigraphic architecture and the driving of input data body
Control sampling, preferably effective sample well out.Specific preferred method assumes a point to be predicted, according to waveform similarity as shown in Figure 2
Feature, for well preferably similar with the waveform as sample well, this method comprehensively considers two factors of waveform and distance, rather than
Only consider the well close from point to be predicted as sample well.In order to avoid the preferred error of the sample of big distance range, preferred process increases
Add the maximum space distance for considering to be sought by spatial function, it is excellent using waveform similarity and space length bivariate in fixed well
Choose the similar well of low frequency configuration as Spatial sample.
Step (3) simulates pseudo- well impedance curve using the impedance curve of preferred sample well, determines structure as shown in Figure 3
As pseudo- well impedance curve.Assuming that the seismic data in work area is per there is a corresponding well data together, wherein well number
It is constituted according to by fixed well data and unknown well puppet well data volume.The simulation of pseudo- well impedance curve specifically includes the following steps:
1., according to formula 1 calculate N mouthful fixed wells the input data body well bypass road and the unknown well of a bite the input number
According to relation function between the road of body well bypass road.
Formula 1 are as follows: Snew=Σ wiSi, i=1,2,3..., N.
Wherein, wiThe relation function between the road of the input data body, SnewBy the input data body well of unknown well
Road, SiFor the input data body well bypass road of fixed well.
2., seek according to formula 2 the relevant matches relation function of different wave parameter, i.e., according to fixed well different wave parameter
Pa、PbAnd its well bypass road Waveform Matching parametric function λab。
Formula 2 are as follows: Pb=Σ λabPa。
Wherein, λabFor the relevant matches relation function of different wave parameter, PbFor the waveform parameter of input data body, PaFor
Fixed well data waveform parameter.
3., seek according to formula 3 spatial relationship function between different wave parameter.Due to identical parameters well point earthquake and
Log data is all the reflection of identical geologic structure, therefore the two meets identical spatial function relationship.Different wave parameter it
Between spatial function relationshipIt may be different.
Formula 3 are as follows:
Wherein,Spatial relationship function between different wave parameter, wiThe relationship between the road of the input data body
Function, λabFor the relevant matches relation function of the different wave parameter.
4., pseudo- well data are sought according to formula 4.
Formula 4 are as follows:
Wherein, LnewFor pseudo- well data,Spatial relationship function between different wave parameter, LiFor fixed well data.
Step (4), under tectonic framework, comprehensive utilization fixed well impedance curve and pseudo- well impedance curve establish introductory die
Type, and the probability distribution of well logging sample is counted as prior probability distribution.
The Vp/Vs compound body of initial model and prestack Simultaneous Inversion is carried out matched filtering, is calculated seemingly by step (5)
Right function.According to the definition of likelihood function, some space is described here by the likelihood function that matched filtering is calculated
Position value is the probability of a certain particular value.Matched filtering is by certainty information (be originated from prestack Simultaneous Inversion) and gradually certainty
Information (being originated from the phased simulation of waveform) produces a new probability distribution space after being merged.
Step (6), Markov Chain-Monte Carlo stochastic simulation (Markov Chain Monte Carlo, MCMC).In
The radio-frequency component of initial model is simulated under Bayesian frame, analog result is made to meet the impedance of earthquake intermediate frequency and well curve knot
Structure feature (condition distribution probability).It is illustrated in figure 4 Vp/Vs simulation even well profile, ordinate indicates time, unit ms;H8
Indicate the bottom of 8 lower section of box, SX indicates the bottom on 1 section of mountain;Black portions are the sand body development area of simulation, are based on Rock physical analysis sand
The effective Vp/Vs threshold value of body is that can find out that the horizontal extension of sand body is preferable less than 1.7 from simulation drawing, meet sedimentation
Rule.8 lower section of box is braided stream deposit, and sandbody distribution is more in plane, and Shanxi group is meandering stream deposit, and is that thin interbed is heavy
Product, therefore the sand body of Shanxi group is less compared with 8 lower section of box, whole simulated section and well coincide preferably, simulate with a high credibility.
Although having been combined attached drawing above and exemplary embodiment describing the present invention, those of ordinary skill in the art
It will be apparent to the skilled artisan that can be carry out various modifications to above-described embodiment in the case where not departing from spirit and scope of the claims.