CN109298445A - A kind of inverse model update method based on Gaussian Profile M-H sampling - Google Patents
A kind of inverse model update method based on Gaussian Profile M-H sampling Download PDFInfo
- Publication number
- CN109298445A CN109298445A CN201811080984.3A CN201811080984A CN109298445A CN 109298445 A CN109298445 A CN 109298445A CN 201811080984 A CN201811080984 A CN 201811080984A CN 109298445 A CN109298445 A CN 109298445A
- Authority
- CN
- China
- Prior art keywords
- logarithm
- impedance
- indicate
- formula
- initial value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. analysis, for interpretation, for correction
- G01V1/282—Application of seismic models, synthetic seismograms
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. analysis, for interpretation, for correction
- G01V1/30—Analysis
- G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/62—Physical property of subsurface
- G01V2210/622—Velocity, density or impedance
- G01V2210/6226—Impedance
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/66—Subsurface modeling
- G01V2210/665—Subsurface modeling using geostatistical modeling
Abstract
The invention discloses a kind of inverse model update methods based on Gaussian Profile M-H sampling, are related to geophysical inversion technology field comprising step 1: carrying out initial model of the pretreatment acquisition to inverted parameters to earthquake information；Step 2: waiting for that the initial model of inverted parameters calculates impedance initial value logarithm L in the selection road t1；Step 3: calculating impedance initial value logarithm L1Variance, establish the state transition function of Gaussian distributed, update impedance initial value logarithm L repeatedly according to state transition function1；Step 4: the seismic channel whether t is greater than in earthquake information is judged, if so, terminating to update；If it is not, enabling t=t+1 skip to step 2 continues inverting；It solves the problems, such as that existing update inverse model uses based on equally distributed M-H sampling, lacks and lead to restrain that codomain is excessive, convergence rate is slow, falls into local optimum to actual data analysis, has achieved the effect that reduce convergence codomain range, quick optimizing.
Description
Technical field
The present invention relates to Geophysics Inversions and oil and gas reservoir to predict field, especially a kind of to be adopted based on Gaussian Profile M-H
The inverse model update method of sample.
Background technique
Seismic inversion is the important step of predicting oil/gas reservoir, its earthquake record data according to known to detector passes through
The mathematical model of earthquake record data and physical quantity to be asked establishes optimization problem, and by inversion method solve optimal
Change problem, to obtain the process of physical quantity optimal estimation to be asked.Based on the markovian seismic inversion in Monte Carlo-
It is the important method of seismic inversion, it updates inverse model by stochastical sampling, and is completed entirely by Markov Chain
The process of inverting.
M-H (Metropolis-Hastings) sampling, is the important method of stochastical sampling, and this method can guarantee sampling knot
Fruit has convergence；This method is to be sampled by introducing state transition function to model, then introduces decision function and determines
Whether sampling receives, to obtain the process of final sampled result by constantly sampling, this method is mentioned by Metropolis
Out, it is improved by Hastings, two people mathematically demonstrate the convergent reasonability of sampled result, and Lavielle will be based on uniform point
The M-H of cloth is sampled in Bayes's seismic inversion based on Gaussian Profile, it was demonstrated that M-H sampling in seismic inversion can
Row, but M-H sampling is uniformly distributed using being uniformly distributed using random value between 0-1 in the prior art, is as a result in appoint
What value distribution, causes codomain range excessive, and convergence rate is slow, while ram-jolt border data deficiency is analyzed over the ground, causes to fall into office
Portion is optimal；Therefore need a kind of inverse model update method that can overcome problem above.
Summary of the invention
It is an object of the invention to: the present invention provides one kind to sample inverse model update method based on Gaussian Profile M-H,
Solving existing update inverse model and using leads to convergency value to actual data analysis based on equally distributed M-H sampling, shortage
Domain is excessive, convergence rate is slow, falls into the problem of local optimum.
The technical solution adopted by the invention is as follows:
A kind of inverse model update method based on Gaussian Profile M-H sampling, includes the following steps:
Step 1: initial model of the pretreatment acquisition to inverted parameters is carried out to earthquake information；
Step 2: waiting for that the initial model of inverted parameters calculates impedance initial value logarithm L in the selection road t1, by impedance initial value
Logarithm L1As the initial model sampled based on Gaussian Profile M-H；
Step 3: calculating impedance initial value logarithm L1Variance, according to variance establish Gaussian distributed state shift letter
Number, according to the Gaussian Profile of state transition function to impedance initial value logarithm L1It is updated repeatedly and obtains updated t
Wait for inverted parameters in road；
Step 4: the seismic channel whether t is greater than in earthquake information is judged, if so, terminating to update；If it is not, t=t+1 is enabled to jump
Continue inverting to step 2.
Preferably, the step 1 includes the following steps:
Step 1.1: obtaining earthquake information, the seismic data includes seismic data and log data；
Step 1.2: according to earthquake data extraction layer position, log data being modeled to obtain using Kriging regression and is joined to inverting
Number initial model carries out well shake matching according to seismic data and log data and extracts wavelet.
Preferably, the step 2 includes the following steps:
Step 2.1: the selection road t waits for the initial model of inverted parameters, according to inverted parameters and wave impedance logarithm recursion
Relationship seeks impedance initial value logarithm L1, it is as shown in formula 1 to inverted parameters and wave impedance logarithm recurrence relation:
Lk=f (xk) formula (1)
Wherein, xkIt indicates to inverted parameters, LkIndicate wave impedance logarithm；
Step 2.2: by impedance initial value logarithm L1As the initial model sampled based on Gaussian Profile M-H.
Preferably, the step 3 includes the following steps:
Step 3.1: calculating impedance initial value logarithm L1Mean μk, according to mean μkCalculate its variance δk, calculation formula is such as
Shown in formula 2:
Wherein, n indicates vector L1Element number, L1iIndicate L1In i-th of element；
Step 3.2: according to variance δkThe state transition function of Gaussian distributed is established, is calculated as shown in formula 3:
f(L1,L2)∝N(0,δk) formula (3)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate updated wave impedance logarithm, N (0, δK) indicate that mean value is 0,
Variance is δkGaussian Profile；
Step 3.3: wave impedance logarithm being updated according to the Gaussian Profile of state transition function and obtains updated wave impedance pair
Number L2, it calculates as shown in formula 4:
L2=L1+ deltan formula (4)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate that updated wave impedance logarithm, n indicate have with wave impedance logarithm
There is the random vector of the Gaussian distributed of identical element, delta indicates step-length；
Step 3.4: computational discrimination function a (L1,L2), it calculates as shown in formula 5:
a(L1,L2)=min (1, P (L2)/P(L1)) formula (5)
Wherein, P (L1) indicate L1Posterior probability density function, P (L2) indicate L2Posterior probability density function；
Step 3.5: randomly selecting u by being uniformly distributed between 0-1, and judge whether u is less than a (L1,L2), if it is not, then will
Updated wave impedance logarithm L2As impedance initial value logarithm L1After skip to step 3.6；If so, without operation；
Step 3.6: judging whether the number of iterations is greater than the number of iterations setting value, repeated if it is not, then skipping to step 3.1
It updates；If so, waiting for inverted parameters according to the road updated wave impedance Logarithmic calculation updated t.
Preferably, posterior probability density function calculates as shown in formula 6 in the step 3.4:
Wherein, G indicates that the constant matrices obtained according to wavelet, L indicate the road the t wave impedance logarithm obtained according to layer position,
s0Indicate the road t original earthquake data, C1For constant term.
In conclusion by adopting the above-described technical solution, the beneficial effects of the present invention are:
1. being instead based on normal distribution M-H sampling the present invention is based on M-H sampling is uniformly distributed, Gaussian Profile passes through statistics well
The distribution of wave impedance logarithmic data after interpolation extends out obtains, and meets the distribution of real data, solves existing update inverting mould
Convergence codomain is excessive, convergence rate is slow, sunken using being caused based on equally distributed M-H sampling, shortage to actual data analysis for type
The problem of entering local optimum has achieved the effect that reduce convergence codomain range, quick optimizing；
2. the present invention, by calculating wave impedance logarithm variance, the state for establishing Gaussian distributed based on variance shifts letter
Number is updated by the M-H sampling based on Gaussian Profile, and Gaussian Profile samples last result still Gaussian distributed,
Codomain is constant, and avoiding and being uniformly distributed sampled result is that any codomain causes to restrain the excessive disadvantage of codomain, accelerates convergence.
Detailed description of the invention
Examples of the present invention will be described by way of reference to the accompanying drawings, in which:
Fig. 1 is method flow block diagram of the invention；
Fig. 2 is histogram distribution and the corresponding Gaussian Profile schematic diagram of single track well bypass road wave impedance of the invention；
Fig. 3 is the seismic profile of Noise of the invention；
Fig. 4 is multiple tracks wave impedance initial model schematic diagram of the invention；
Fig. 5 is that multiple tracks wave impedance of the invention updates result schematic diagram；
Fig. 6 is that single track well bypass road wave impedance of the invention updates result and contrast schematic diagram；
Fig. 7 is state transition function of the invention using the error comparison diagram being uniformly distributed with Gaussian Profile；
Fig. 8 is flow chart of the method for the present invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention, i.e., described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is logical
The component for the embodiment of the present invention being often described and illustrated herein in the accompanying drawings can be arranged and be designed with a variety of different configurations.
Therefore, the detailed description of the embodiment of the present invention provided in the accompanying drawings is not intended to limit below claimed
The scope of the present invention, but be merely representative of selected embodiment of the invention.Based on the embodiment of the present invention, those skilled in the art
Member's every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
It should be noted that the relational terms of term " first " and " second " or the like be used merely to an entity or
Operation is distinguished with another entity or operation, and without necessarily requiring or implying between these entities or operation, there are any
This actual relationship or sequence.Moreover, the terms "include", "comprise" or its any other variant be intended to it is non-exclusive
Property include so that include a series of elements process, method, article or equipment not only include those elements, but also
Further include other elements that are not explicitly listed, or further include for this process, method, article or equipment it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described
There is also other identical elements in the process, method, article or equipment of element.
It elaborates below with reference to Fig. 1-8 couples of present invention.
Present invention solves the technical problem that: solve existing update inverse model use based on equally distributed M-H sampling,
Lack leads to restrain that codomain is excessive, convergence rate is slow, falls into the problem of local optimum to actual data analysis
Technological means: a kind of inverse model update method based on Gaussian Profile M-H sampling includes the following steps:
Step 1: initial model of the pretreatment acquisition to inverted parameters is carried out to earthquake information；
Step 2: waiting for that the initial model of inverted parameters calculates impedance initial value logarithm L in the selection road t1, by impedance initial value
Logarithm L1As the initial model sampled based on Gaussian Profile M-H；
Step 3: calculating impedance initial value logarithm L1Variance, according to variance establish Gaussian distributed state shift letter
Number, according to the Gaussian Profile of state transition function to impedance initial value logarithm L1It is updated repeatedly and obtains updated t
Wait for inverted parameters in road；
Step 4: the seismic channel whether t is greater than in earthquake information is judged, if so, terminating to update；If it is not, t=t+1 is enabled to jump
Continue inverting to step 2.
Step 1 includes the following steps:
Step 1.1: obtaining earthquake information, the seismic data includes seismic data and log data；
Step 1.2: according to earthquake data extraction layer position, log data being modeled to obtain using Kriging regression and is joined to inverting
Number initial model carries out well shake matching according to seismic data and log data and extracts wavelet.
Step 2 includes the following steps:
Step 2.1: the selection road t waits for the initial model of inverted parameters, according to inverted parameters and wave impedance logarithm recursion
Relationship seeks impedance initial value logarithm L1, it is as shown in formula 1 to inverted parameters and wave impedance logarithm recurrence relation:
Lk=f (xk) formula (1)
Wherein, xkIt indicates to inverted parameters, LkIndicate wave impedance logarithm；
Step 2.2: by impedance initial value logarithm L1As the initial model sampled based on Gaussian Profile M-H.
Step 3 includes the following steps:
Step 3.1: calculating impedance initial value logarithm L1Mean μk, according to mean μkCalculate its variance δk, calculation formula is such as
Shown in formula 2:
Wherein, n indicates vector L1Element number, L1iIndicate L1In i-th of element；
Step 3.2: according to variance δkThe state transition function of Gaussian distributed is established, is calculated as shown in formula 3:
f(L1,L2)∝N(0,δk) formula (3)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate updated wave impedance logarithm, N (0, δK) indicate that mean value is 0,
Variance is δkGaussian Profile；
Step 3.3: wave impedance logarithm being updated according to the Gaussian Profile of state transition function and obtains updated wave impedance pair
Number L2, it calculates as shown in formula 4:
L2=L1+ deltan formula (4)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate that updated wave impedance logarithm, n indicate have with wave impedance logarithm
There is the random vector of the Gaussian distributed of identical element, delta indicates step-length；
Step 3.4: computational discrimination function a (L1,L2), it calculates as shown in formula 5:
a(L1,L2)=min (1, P (L2)/P(L1)) formula (5)
Wherein, P (L1) indicate L1Posterior probability density function, P (L2) indicate L2Posterior probability density function；
Step 3.5: randomly selecting u by being uniformly distributed between 0-1, and judge whether u is less than a (L1,L2), if it is not, then will
Updated wave impedance logarithm L2As impedance initial value logarithm L1After skip to step 3.6；If so, without operation；
Step 3.6: judging whether the number of iterations is greater than the number of iterations setting value, repeated if it is not, then skipping to step 3.1
It updates；If so, waiting for inverted parameters according to the road updated wave impedance Logarithmic calculation updated t.
Posterior probability density function calculates as shown in formula 6 in step 3.4:
Wherein, G indicates that the constant matrices obtained according to wavelet, L indicate the road the t wave impedance logarithm obtained according to layer position,
s0Indicate the road t original earthquake data, C1For constant term；Technical effect: the present invention is based on be uniformly distributed M-H sampling to be instead based on
Normal distribution M-H sampling, the distribution of wave impedance logarithmic data after Gaussian Profile is extended out by statistics well interpolation obtain, and meet reality
The distribution of border data solves existing update inverse model and uses based on equally distributed M-H sampling, lacks to real data point
Analysis leads to restrain that codomain is excessive, convergence rate is slow, falls into the problem of local optimum, has reached diminution convergence codomain range, quickly
The effect of optimizing；By calculating wave impedance logarithm variance, the state transition function of Gaussian distributed is established based on variance, is passed through
M-H sampling based on Gaussian Profile is updated, and Gaussian Profile samples last result still Gaussian distributed, and codomain is not
Become, avoiding and being uniformly distributed sampled result is that any codomain causes to restrain the excessive disadvantage of codomain, accelerates convergence.
Embodiment 1
As shown in figures 1-8, inverting is as follows when inverted parameters are wave impedance:
Step 1.1: input earthquake information uses Kriging regression according to log data according to earthquake data extraction layer position
Modeling obtains wave impedance initial model, extracts wavelet using well shake matching by earthquake information；
Step 2.1: the wave impedance AI of the selection road t modelk, asked just by the recurrence relation of wave impedance and wave impedance logarithm
Beginning wave impedance logarithm, wave impedance and the recurrence relation of wave impedance logarithm are as shown in formula 1:
Lk=ln (AIk) formula (1)
Wherein, AIkIndicate wave impedance, LkIndicate wave impedance logarithm；
Step 2.2: by impedance initial value logarithm L1As update initial model；
Step 3.1: calculating wave impedance logarithm LkMean μk, according to mean μkCalculate variance δk, calculation formula such as formula 2
It is shown:
Wherein, n indicates vector L1Element number, L1iIndicate L1In i-th of element；
Step 3.2: the state transition function of Gaussian distributed is established, is calculated as shown in formula 3:
f(L1,L2)∝N(0,δk) formula (3)
Wherein, L2Indicate updated wave impedance logarithm, N (0, δK) indicate that mean value is 0, variance δkGaussian Profile；
Step 3.3: wave impedance logarithm being updated according to the Gaussian Profile of state transition function and obtains updated wave impedance pair
Number L2, it calculates as shown in formula 4:
L2=L1+ deltan formula (4)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate that updated wave impedance logarithm, n indicate have with wave impedance logarithm
There is the random vector of the Gaussian distributed of identical element, delta indicates step-length；
Step 3.4: computational discrimination function a (L1,L2), it calculates as shown in formula 5:
a(L1,L2)=min (1, P (L2)/P(L1)) formula (5)
Wherein, P (L1) indicate L1Posterior probability density function, P (L2) indicate L2Posterior probability density function；
Posterior probability density function calculates as shown in formula 6:
Wherein, G indicates that the constant matrices obtained according to wavelet, L indicate the road the t wave impedance logarithm obtained according to layer position,
s0Indicate the road t original earthquake data, C1For constant term；
Step 3.5: randomly selecting u by being uniformly distributed between 0-1, and judge whether u is less than a (L1,L2), if it is not, then will
Updated wave impedance logarithm L2As impedance initial value logarithm L1After skip to step 3.6；If so, without operation；
Step 3.6: judging whether the number of iterations is greater than the number of iterations setting value, repeated if it is not, then skipping to step 3.1
It updates；If so, waiting for inverted parameters according to the road updated wave impedance Logarithmic calculation updated t.
Step 4 includes the following steps:
Step 4.1: the seismic channel data whether t is greater than in input data is judged, if so, terminating to update obtained update
Afterwards to inverted parameters section；If it is not, enabling t=t+1 skip to step 2 continues inverting.
Effect analysis: as shown in Fig. 2, the wave impedance logarithm that Gaussian Profile can very well after simulation interpolation, using Gauss
Distribution simulation M-H sampling has reasonability；As in Figure 3-5, the result of updated wave impedance section is cutd open compared to wave impedance
Face initial model can preferably obey the trend of earthquake record, react the correctness of the application, initial as shown in Figure 6
Indicate wave impedance initial model, inversion indicates updated wave impedance as a result, well indicates true wave impedance, update
As a result than initial model closer to truthful data, the correctness of the application has further been reacted；U as shown in Figure 7 is indicated using equal
The relationship of convergence number and error under even distribution M-H sampling, N indicate convergence number and error under Gaussian Profile M-H sampling
Relationship, as shown in the figure: the present processes reduce convergence codomain, can faster optimizing, the present invention passes through analysis source first
In the statistical property for the model data that log data modeling obtains, letter is shifted in conjunction with the state in Gaussian Profile building M-H sampling
Number sample using the Gaussian Profile that state transition function among the above is formed and passes through discriminant function more to initial model
Newly, by iterating, optimal inversion result is exported, the present invention solves existing seismic inversion use and is uniformly distributed as shape
State transfer function leads to restrain that codomain is excessive, convergence rate is slow, falls into the problem of local optimum, has reached diminution convergence codomain model
It encloses, the effect of quick optimizing.
Claims (5)
1. a kind of inverse model update method based on Gaussian Profile M-H sampling, characterized by the following steps:
Step 1: initial model of the pretreatment acquisition to inverted parameters is carried out to earthquake information；
Step 2: waiting for that the initial model of inverted parameters calculates impedance initial value logarithm L in the selection road t1, by impedance initial value logarithm L1
As the initial model sampled based on Gaussian Profile M-H；
Step 3: calculating impedance initial value logarithm L1Variance, the state transition function of Gaussian distributed, root are established according to variance
According to the Gaussian Profile of state transition function to impedance initial value logarithm L1It is updated repeatedly and obtains the road updated t and waited for instead
Drill parameter；
Step 4: the seismic channel whether t is greater than in earthquake information is judged, if so, terminating to update；If it is not, t=t+1 is enabled to skip to step
Rapid 2 continue inverting.
2. a kind of inverse model update method based on Gaussian Profile M-H sampling according to claim 1, feature exist
In: the step 1 includes the following steps:
Step 1.1: obtaining earthquake information, the seismic data includes seismic data and log data；
Step 1.2: according to earthquake data extraction layer position, log data being modeled to obtain at the beginning of inverted parameters using Kriging regression
Beginning model carries out well shake matching according to seismic data and log data and extracts wavelet.
3. a kind of inverse model update method based on Gaussian Profile M-H sampling according to claim 1 or 2, special
Sign is: the step 2 includes the following steps:
Step 2.1: the selection road t waits for the initial model of inverted parameters, according to inverted parameters and wave impedance logarithm recurrence relation
Seek impedance initial value logarithm L1, it is as shown in formula 1 to inverted parameters and wave impedance logarithm recurrence relation:
Lk=f (xk) formula (1)
Wherein, xkIt indicates to inverted parameters, LkIndicate wave impedance logarithm；
Step 2.2: by impedance initial value logarithm L1As the initial model sampled based on Gaussian Profile M-H.
4. a kind of inverse model update method based on Gaussian Profile M-H sampling according to claim 3, feature exist
In: the step 3 includes the following steps:
Step 3.1: calculating impedance initial value logarithm L1Mean μk, according to mean μkCalculate its variance δk, calculation formula such as formula 2
It is shown:
Wherein, n indicates vector L1Element number, L1iIndicate L1In i-th of element；
Step 3.2: according to variance δkThe state transition function of Gaussian distributed is established, is calculated as shown in formula 3:
f(L1,L2)∝N(0,δk) formula (3)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate updated wave impedance logarithm, N (0, δK) indicate that mean value is 0, variance
For δkGaussian Profile；
Step 3.3: wave impedance logarithm being updated according to the Gaussian Profile of state transition function and obtains updated wave impedance logarithm L2,
It calculates as shown in formula 4:
L2=L1+ deltan formula (4)
Wherein, L1Indicate impedance initial value logarithm, L2Indicate that updated wave impedance logarithm, n indicate there is phase with wave impedance logarithm
With the random vector of the Gaussian distributed of element, delta indicates step-length；
Step 3.4: computational discrimination function a (L1,L2), it calculates as shown in formula 5:
a(L1,L2)=min (1, P (L2)/P(L1)) formula (5)
Wherein, P (L1) indicate L1Posterior probability density function, P (L2) indicate L2Posterior probability density function；
Step 3.5: randomly selecting u by being uniformly distributed between 0-1, and judge whether u is less than a (L1,L2), if it is not, will then update
Wave impedance logarithm L afterwards2As impedance initial value logarithm L1After skip to step 3.6；If so, without operation；
Step 3.6: judging whether the number of iterations is greater than the number of iterations setting value, repeat more if it is not, then skipping to step 3.1
Newly；If so, waiting for inverted parameters according to the road updated wave impedance Logarithmic calculation updated t.
5. a kind of inverse model update method based on Gaussian Profile M-H sampling according to claim 4, feature exist
In: posterior probability density function calculates as shown in formula 6 in the step 3.4:
Wherein, G indicates that the constant matrices obtained according to wavelet, L indicate the road the t wave impedance logarithm obtained according to layer position, s0It indicates
The road t original earthquake data, C1For constant term.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811080984.3A CN109298445A (en) | 2018-09-17 | 2018-09-17 | A kind of inverse model update method based on Gaussian Profile M-H sampling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811080984.3A CN109298445A (en) | 2018-09-17 | 2018-09-17 | A kind of inverse model update method based on Gaussian Profile M-H sampling |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109298445A true CN109298445A (en) | 2019-02-01 |
Family
ID=65163338
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811080984.3A Pending CN109298445A (en) | 2018-09-17 | 2018-09-17 | A kind of inverse model update method based on Gaussian Profile M-H sampling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109298445A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113640871A (en) * | 2021-08-10 | 2021-11-12 | 成都理工大学 | Seismic wave impedance inversion method based on heavily-weighted L1 norm sparse constraint |
| CN113960655A (en) * | 2020-07-20 | 2022-01-21 | 中国石油天然气股份有限公司 | Seismic data sample updating method and system |
| CN113960655B (en) * | 2020-07-20 | 2024-04-30 | 中国石油天然气股份有限公司 | Method and system for updating seismic data samples |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152014A (en) * | 2013-01-30 | 2013-06-12 | 中国人民解放军理工大学 | Implementation method of Metropolis-Hastings variation particle swarm resampling particle filter |
| CN103149587A (en) * | 2013-02-19 | 2013-06-12 | 中国石油天然气股份有限公司 | Random-coupling four-dimensional-seismic-inversion monitoring method and device for oil reservoirs based on grid points |
| WO2016012780A1 (en) * | 2014-07-23 | 2016-01-28 | Iruiz Technologies Ltd | Improvements related to forecasting systems |
| CN105353407A (en) * | 2015-10-28 | 2016-02-24 | 中国石油化工股份有限公司 | Post-stack earthquake wave impedance inversion method |
| CN107203005A (en) * | 2016-03-18 | 2017-09-26 | 中国石油化工股份有限公司 | A kind of method that quantification calculates crack characterising parameter |
-
2018
- 2018-09-17 CN CN201811080984.3A patent/CN109298445A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152014A (en) * | 2013-01-30 | 2013-06-12 | 中国人民解放军理工大学 | Implementation method of Metropolis-Hastings variation particle swarm resampling particle filter |
| CN103149587A (en) * | 2013-02-19 | 2013-06-12 | 中国石油天然气股份有限公司 | Random-coupling four-dimensional-seismic-inversion monitoring method and device for oil reservoirs based on grid points |
| WO2016012780A1 (en) * | 2014-07-23 | 2016-01-28 | Iruiz Technologies Ltd | Improvements related to forecasting systems |
| CN105353407A (en) * | 2015-10-28 | 2016-02-24 | 中国石油化工股份有限公司 | Post-stack earthquake wave impedance inversion method |
| CN107203005A (en) * | 2016-03-18 | 2017-09-26 | 中国石油化工股份有限公司 | A kind of method that quantification calculates crack characterising parameter |
Non-Patent Citations (3)
| Title |
|---|
| 张广智 等: ""利用MCMC方法估算地震参数"", 《石油地球物理勘探》 * |
| 张建新: ""基于贝叶斯方法的有限元模型修正研究"", 《中国优秀硕士学位论文全文数据库-工程科技II辑》 * |
| 张繁昌 等: ""地震数据约束下的贝叶斯随机反演"", 《石油地球物理勘探》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113960655A (en) * | 2020-07-20 | 2022-01-21 | 中国石油天然气股份有限公司 | Seismic data sample updating method and system |
| CN113960655B (en) * | 2020-07-20 | 2024-04-30 | 中国石油天然气股份有限公司 | Method and system for updating seismic data samples |
| CN113640871A (en) * | 2021-08-10 | 2021-11-12 | 成都理工大学 | Seismic wave impedance inversion method based on heavily-weighted L1 norm sparse constraint |
| CN113640871B (en) * | 2021-08-10 | 2023-09-01 | 成都理工大学 | Seismic wave impedance inversion method based on re-weighted L1 norm sparse constraint |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dralle et al. | Event-scale power law recession analysis: quantifying methodological uncertainty | |
| Christie et al. | Uncertainty quantification for porous media flows | |
| CN108663711B (en) | A kind of Bayes's seismic inversion method based on τ distribution | |
| Makkonen et al. | An improved method of extreme value analysis | |
| CA2997608A1 (en) | History matching of hydrocarbon production from heterogenous reservoirs | |
| Sarma et al. | Efficient and robust uncertainty quantification in reservoir simulation with polynomial chaos expansions and non-intrusive spectral projection | |
| Nasri et al. | Non-stationary hydrologic frequency analysis using B-spline quantile regression | |
| Smith et al. | Comparing methods for calculating channel steepness index | |
| Lombardi et al. | On-line bayesian estimation of signals in symmetric/spl alpha/-stable noise | |
| La Boissonière et al. | Statistics of grain growth: experiment versus the Phase-Field-Crystal and Mullins models | |
| CN109298445A (en) | A kind of inverse model update method based on Gaussian Profile M-H sampling | |
| Shaochuan | A Bayesian multiple changepoint model for marked poisson processes with applications to deep earthquakes | |
| Socco et al. | Robust static estimation from surface wave data | |
| Mindham et al. | Extended State Dependent Parameter modelling with a Data-Based Mechanistic approach to nonlinear model structure identification | |
| Cai et al. | Influence of partially known parameter on flaw characterization in Eddy Current Testing by using a random walk MCMC method based on metamodeling | |
| Jahanbakhshi et al. | Impact of initial ensembles on posterior distribution of ensemble-based assimilation methods | |
| Jahanbakhshi et al. | Joint estimation of facies boundaries and petrophysical properties in multi-facies channelized reservoirs through ensemble-based Kalman filter and level set parametrization | |
| US20210149077A1 (en) | Optimized methodology for automatic history matching of a petroleum reservoir model with ensemble kalman filter (enkf) | |
| Mitchell et al. | Bedrock river erosion through dipping layered rocks: quantifying erodibility through kinematic wave speed | |
| Gańcza et al. | Regularized local basis function approach to identification of nonstationary processes | |
| Ortiz et al. | Integrating multiple-point statistics into sequential simulation algorithms | |
| CN112016956A (en) | BP neural network-based ore grade estimation method and device | |
| Chechkin et al. | Fractional Brownian motion approximation based on fractional integration of a white noise | |
| Zhao et al. | Characterization of the measurement error in time-lapse seismic data and production data with an EM algorithm | |
| Bardsley | A finite mixture approach to univariate data simulation with moment matching |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190201 |
|
| RJ01 | Rejection of invention patent application after publication |