CN107436452A - Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm - Google Patents
Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm Download PDFInfo
- Publication number
- CN107436452A CN107436452A CN201610375298.3A CN201610375298A CN107436452A CN 107436452 A CN107436452 A CN 107436452A CN 201610375298 A CN201610375298 A CN 201610375298A CN 107436452 A CN107436452 A CN 107436452A
- Authority
- CN
- China
- Prior art keywords
- toc
- hydrocarbon source
- source rock
- seismic
- prediction
- 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
- 239000011435 rock Substances 0.000 title claims abstract description 142
- 239000004215 Carbon black (E152) Substances 0.000 title claims abstract description 124
- 229930195733 hydrocarbon Natural products 0.000 title claims abstract description 124
- 150000002430 hydrocarbons Chemical class 0.000 title claims abstract description 124
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000011160 research Methods 0.000 claims abstract description 51
- 238000005553 drilling Methods 0.000 claims abstract description 50
- 238000012360 testing method Methods 0.000 claims abstract description 9
- 230000007812 deficiency Effects 0.000 claims description 9
- 238000012937 correction Methods 0.000 claims description 5
- 230000001537 neural effect Effects 0.000 claims 1
- 238000012549 training Methods 0.000 description 12
- 239000005416 organic matter Substances 0.000 description 5
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 4
- 229910052799 carbon Inorganic materials 0.000 description 4
- 230000001419 dependent effect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 208000035126 Facies Diseases 0.000 description 2
- 239000004575 stone Substances 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000000153 supplemental effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/40—Seismology; Seismic or acoustic prospecting or detecting specially adapted for well-logging
-
- 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. for interpretation or for event detection
- 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
- G01V2210/61—Analysis by combining or comparing a seismic data set with other data
- G01V2210/616—Data from specific type of measurement
- G01V2210/6169—Data from specific type of measurement using well-logging
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Acoustics & Sound (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The present invention provides a kind of hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm, and wherein method includes：Analysis test is carried out to hydrocarbon source rock rock core, the chip sample of the research each drilling well in area of acquisition, to obtain the Toc values of sample spot；The Toc prediction curves for obtaining each drilling well hydrocarbon source rock interval longitudinal direction in research area are calculated using log；Toc prediction curves are corrected according to the Toc values of sample spot, to obtain Toc curves；According to the seismic properties of Toc curves and seismic trace near well, seismic properties are trained using probabilistic neural network algorithm, to obtain preferred seismic properties, extract the preferred seismic properties of three dimensional seismic data, the Toc data volumes for obtaining distributed in three dimensions are calculated, to complete research area's hydrocarbon source rock prediction.Method provided by the invention, using the non-linear relation that can be preferably caught based on probabilistic neural network algorithm between seismic properties and Toc, make the prediction result to hydrocarbon source rock more accurate.
Description
Technical field
The present invention relates to petroleum exploration field, more particularly to a kind of hydrocarbon source rock prediction side based on probabilistic neural network algorithm
Method and device.
Background technology
Total content of organic carbon (total organic carbon, abbreviation Toc) is the main of evaluating hydrocarbon primary rock producing hydrocarbon ability
One of index, it is an important parameter of hydrocarbon research and resource assessment in oil-gas bearing basin.
In recent years, as logging technique develops, many scholars start to explore the Geochemical Parameters of hydrocarbon source rock and well logging is believed
Relation between breath, the response relation in hydrocarbon source rock between abundance of organic matter and different log parameters are also gradually realized, and
Evaluation of the identification of abundance, hydrocarbon source rock in organic matter, source-rock evolution maturity and hydrocarbon source rock etc. achieves many
Achievement, the development of Evaluation of source rocks technology is effectively promoted.As Zhu Guangyou thinks, due to organic matter, institute in hydrocarbon source rock be present
Have one in acoustic travel time logging, density log and resistivity logging containing less or without organic matter sedimentary rock than other with it
Fixed difference, and the number of the content of organic matter has certain correlation with this difference, therefore establish interval transit time, close
Spend with the model of resistivity to quantify the organic carbon content of calculating hydrocarbon source rock, and then predict longitudinal spread of hydrocarbon source rock.This set side
Method obviously compensate for the defects of rock sample geochemistry, but due to only depending on well-log information, can not be to the exhibition in hydrocarbon source rock plane
Cloth is predicted.And seismic data has the characteristics of cross direction profiles are wide, precision is high.Wang Zhihong is using individual well demarcation seismic facies inverting
Method, have studied the Distribution Characteristics of Song-liao basin north deep hydrocarbon source rock.In founding the state, two sections of Dongying Depression hole shop group of research is recognized
For undercompacted mud stone density is relatively low, in obvious low in the interval velocity section and wave impedance section after specially treated
It is worth abnormal belt.Using seismic data, by the analysis of seismic facies and Seismic Velocity Characteristics, potential hydrocarbon source rock is predicted.
But these are also simply to the qualitative forecasting of hydrocarbon source rock, rare to directly fixed with seismic data in domestic and foreign literature at present
Amount calculates the report of content of organic carbon of hydrocarbon source rock, and the geophysics evaluation degree of hydrocarbon source rock far lags behind evaluating reservoir.Therefore, base
In the quantitative forecast of the distribution of source rock of seismic data, it is of great significance for hydrocarbon amount and stock number estimation tool.
For a long time, geochemistry establish a set of ripe comment by doing geochemical analysis to rock sample or landwaste
Valency index and method.But this method has the shortcomings that obvious：First, can just enter only in the case of with source rock sample
OK；Secondly, the way of a whole set of thick-layer hydrocarbon source rock is replaced with limited analysis sample, masks the non-average of hydrocarbon source rock.At present
Due to being limited by drilling and coring delivery sample, continuous hydrocarbon source rock Toc measured values can not be typically obtained, typically with one spacing of interval
From Toc geometrical means evaluate hydrocarbon potentiality, cause sizable error.
The content of the invention
The present invention provides a kind of hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm, existing to solve
The larger technical problem of hydrocarbon source rock prediction error in technology.
One aspect of the present invention provides a kind of hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm, including：
Step 101, analysis test is carried out to hydrocarbon source rock rock core, the chip sample of the research each drilling well in area of acquisition, to obtain
The Toc values of sample spot；
Step 102, the Toc prediction curves for obtaining each drilling well hydrocarbon source rock interval longitudinal direction in research area are calculated using log；
Step 103, Toc prediction curves are corrected according to the Toc values of sample spot, to obtain Toc curves；
Step 104, according to the seismic properties of Toc curves and seismic trace near well, using probabilistic neural network algorithm to earthquake
Attribute is trained, to obtain preferred seismic properties, wherein it is preferred to which it is the seismic properties best with Toc correlations to shake attribute；
Step 105, the preferred seismic properties of three dimensional seismic data are extracted, calculate the Toc data volumes for obtaining distributed in three dimensions, with
Complete research area's hydrocarbon source rock prediction.
Further, step 102 specifically includes：
According to the cross plot of the log of research area's drilling well, response characteristics of the analysis hydrocarbon source rock Toc on log；
The log criterion of identification for being divided hydrocarbon source rock and country rock is obtained according to response characteristic, to pick out preferably
Log；
According to preferred log, the Toc prediction songs for obtaining each drilling well hydrocarbon source rock interval longitudinal direction are calculated using Δ LogR methods
Line.
Further, step 102 also includes：
Calculated in the low region of research area's degree of prospecting and obtain virtual log Toc prediction curves, the virtual log Toc predictions are bent
Line is in the case of drilling well deficiency, and the Toc prediction curves of acquisition are calculated by the way of virtual drilling well.
Further, in step 103 before the seismic properties of seismic trace near well are obtained, in addition to：
Fine stratum calibration is carried out to each drilling well in research area, determines the Seismic reflection character of hydrocarbon source rock top bottom reflecting interface；
The seismic interpretation that at least 10 × 10 grids are carried out to hydrocarbon source rock top bottom interface works, wherein, seismic interpretation saddlebag
Include the explanation of layer position and fault interpretation.
Further, preferably seismic properties include 3 to 7 seismic properties.
Another aspect of the present invention provides a kind of hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm, including：
Sample spot Toc value acquisition modules, hydrocarbon source rock rock core, chip sample for the research each drilling well in area to acquisition are carried out
Analysis test, to obtain the Toc values of sample spot；
Toc prediction curve acquisition modules, indulged for calculating the acquisition research each drilling well hydrocarbon source rock interval in area using log
To Toc prediction curves；
Toc prediction curve correction modules, Toc prediction curves are corrected for the Toc values according to sample spot, to obtain
Toc curves；
It is preferred that seismic properties acquisition module, for the seismic properties according to Toc curves and seismic trace near well, utilizes probability god
Seismic properties are trained through network algorithm, to obtain preferred seismic properties, wherein it is preferred to which it is related to Toc to shake attribute
The best seismic properties of property；
Toc data volume acquisition modules, for extracting the preferred seismic properties of three dimensional seismic data, calculate and obtain distributed in three dimensions
Toc data volumes, with complete research area's hydrocarbon source rock prediction.
Further, Toc prediction curves acquisition module includes：
Response characteristic acquisition submodule, for the cross plot of the log according to research area's drilling well, analysis hydrocarbon source rock Toc
Response characteristic on log；
Log criterion of identification acquisition submodule, hydrocarbon source rock and country rock are divided for being obtained according to response characteristic
Log criterion of identification, to pick out preferred log；
Toc prediction curve acquisition submodules, for according to preferred log, being calculated using Δ LogR methods and obtaining each drilling well
The Toc prediction curves of hydrocarbon source rock interval longitudinal direction.
Further, Toc prediction curves acquisition module also includes：
Virtual log Toc prediction curve acquisition submodules, obtained virtually for being calculated in the low region of research area's degree of prospecting
Well Toc prediction curves, the virtual log Toc prediction curves are in the case of drilling well deficiency, by the way of virtual drilling well
Calculate the Toc prediction curves obtained.
Further, preferably in seismic properties acquisition module before the seismic properties of seismic trace near well are obtained, in addition to：
Fine stratum calibration is carried out to each drilling well in research area, determines the Seismic reflection character of hydrocarbon source rock top bottom reflecting interface；
The seismic interpretation that at least 10 × 10 grids are carried out to hydrocarbon source rock top bottom interface works, wherein, seismic interpretation saddlebag
Include the explanation of layer position and fault interpretation.
Hydrocarbon source rock Forecasting Methodology and device provided by the invention based on probabilistic neural network algorithm, logged well with reference to hydrocarbon source rock
And seismic response features, (the Toc value)s of sample spot, Toc predictions are established using logging prediction Toc method from actual measurement Toc
Curve, and as dependent variable, choose seismic properties as independent variable, using based on probabilistic neural network algorithm establish them it
Between best fit relational expression, so as to obtain optimal seismic properties, then by being extracted optimally from 3-d seismic data set
Attribute is shaken, the Toc data volumes of distributed in three dimensions are calculated, so as to realize that hydrocarbon source rock is predicted.Calculated using based on probabilistic neural network
Method can preferably catch the non-linear relation between seismic properties and Toc, make the prediction result to hydrocarbon source rock more accurate.
Brief description of the drawings
The invention will be described in more detail below based on embodiments and refering to the accompanying drawings.Wherein：
Fig. 1 is that the flow for the hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm that the embodiment of the present invention one provides is shown
It is intended to；
Fig. 2 is that the flow for the hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm that the embodiment of the present invention two provides is shown
It is intended to；
Fig. 3 is that the structure for the hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm that the embodiment of the present invention three provides is shown
It is intended to；
Fig. 4 is that the structure for the hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm that the embodiment of the present invention four provides is shown
It is intended to.
In the accompanying drawings, identical part uses identical reference.Accompanying drawing is not drawn according to the ratio of reality.
Embodiment
Below in conjunction with accompanying drawing, the invention will be further described.
Embodiment one
Fig. 1 is to be shown according to the flow of the hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm of the embodiment of the present invention one
It is intended to；As shown in figure 1, the present embodiment provides a kind of hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm, including：
Step 101, analysis test is carried out to hydrocarbon source rock rock core, the chip sample of the research each drilling well in area of acquisition, to obtain
The Toc values of sample spot.
Specifically, collecting hydrocarbon source rock rock core, the chip sample of the research each drilling well in area, and analysis test is carried out, and then determined
The Toc values of sample spot.
Step 102, the Toc prediction curves for obtaining each drilling well hydrocarbon source rock interval longitudinal direction in research area are calculated using log.
Specifically, this step it is bent can to find the well logging such as Toc and density, interval transit time, resistivity by establishing logging module
Relation between line, the continuous Toc values of hydrocarbon source rock interval are obtained, that is, obtain the Toc of each drilling well hydrocarbon source rock interval longitudinal direction in research area
Prediction curve.
Step 103, Toc prediction curves are corrected according to the Toc values of sample spot, to obtain Toc curves.Utilize sample
The Toc values of point are corrected to Toc prediction curves, are Toc curves to the curve obtained after the correction of Toc prediction curves.
Step 104, according to the seismic properties of Toc curves and seismic trace near well, using probabilistic neural network algorithm to earthquake
Attribute is trained, to obtain preferred seismic properties, wherein it is preferred to which it is the seismic properties best with Toc correlations to shake attribute.
Specifically, using Toc prediction curves as aim curve, it is related to Toc to extract wave impedance, the amplitude of seismic trace near well etc.
Property preferable seismic properties, optimal nonlinear relation between them is established using probabilistic neural network algorithm, and choose with
The best seismic properties of Toc correlations preferably shake attribute.Further, training function is also obtained in this step, for
Used in subsequent step.
Probabilistic neural network algorithm is briefly described below, the data used in probabilistic neural network Algorithm for Training include
A series of training " example ", in analysis window, each earthquake sampling point has a bite drilling well to correspond to therewith.
{ A11, A21, A31, L1 }
{ A12, A22, A32, L2 }
{ A13, A23, A33, L3 }
…
{ A1n, A2n, A3n, Ln }
Here, there are n training sample and 3 attribute (the present embodiment is briefly described with 3 attribute).Li represents each
The realistic objective log value of example.
For given training data, probabilistic neural network algorithm assumes that each new output log value may be expressed as
The linear combination of training data log value.For the new data sample that property value is x={ A1j, A2j, A3j }, new log value
For
In formula,
D(x,xi) it is input point and each training points xiThe distance between.This distance is in the hyperspace by attribute
Vacuum metrics, and press σjScaling, therefore to every attribute, it may be different.
Two formulas describe the service condition of probabilistic neural network above.The training of probabilistic neural network is to determine optimal put down
Sliding parameter σj.The standard for judging these parameters is to export the verification error minimum of probabilistic neural network.Define m-th of target sample
Check results be：
When removing m-th of target sample from training data, the predicted value of m-th of sample can be drawn according to above formula.By
In known to sample value, it is possible to calculate the prediction error of the sample.This process is repeated to each training sample, then can be instruction
The total prediction error for practicing data is defined as：
It should be noted that the size of prediction error depends on parameter σjSelection.σjIt can be proposed by Master (1995) non-
Linear conjugate gradient algorithm minimizes.Derived probabilistic neural network has minimum verification error.
Probabilistic neural network has following several advantages：Fast convergence rate；Train function it is simple, and no matter classification problem
There is complexity how, as long as there are enough training samples, probabilistic neural network ensures that to be obtained under this criterion of Baily leaf
Take optimal solution；Sample supplemental capabilities are strong, and can tolerate error sample.
Step 105, the preferred seismic properties of three dimensional seismic data are extracted, calculate the Toc data volumes for obtaining distributed in three dimensions, with
Complete research area's hydrocarbon source rock prediction.
Specifically, preferred seismic properties are extracted from 3-d seismic data set, according to the training function obtained in step 106
The Toc data volumes for obtaining distributed in three dimensions are calculated, to complete research area's hydrocarbon source rock prediction.
The hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm that the present embodiment provides, with reference to hydrocarbon source rock well logging and ground
Shake response characteristic, from actual measurement Toc (the Toc value)s of sample spot, Toc prediction curves are established using logging prediction Toc method,
And as dependent variable, seismic properties are chosen as independent variable, are established using based on probabilistic neural network algorithm between them
Best fit relational expression, so as to obtain optimal seismic properties, then by extracting optimal earthquake category from 3-d seismic data set
Property, the Toc data volumes of distributed in three dimensions are calculated, so as to realize that hydrocarbon source rock is predicted.Using based on probabilistic neural network algorithm energy
The non-linear relation between seismic properties and Toc is preferably caught, makes the prediction result to hydrocarbon source rock more accurate.
Embodiment two
The present embodiment is the supplementary notes carried out on the basis of above-described embodiment.
Fig. 2 is that the flow for the hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm that the embodiment of the present invention two provides is shown
It is intended to；As shown in Fig. 2 the present embodiment provides a kind of hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm, including：
Step 201, analysis test is carried out to hydrocarbon source rock rock core, the chip sample of the research each drilling well in area of acquisition, to obtain
The Toc values of sample spot.
This step is consistent with the step 101 of embodiment one, will not be repeated here.
Step 2021, according to the cross plot of the log of research area's drilling well, hydrocarbon source rock Toc is on log for analysis
Response characteristic.
Specifically, making the cross plot of log according to research area's drilling well, hydrocarbon source rock Toc is in density, gamma, sound for analysis
Response characteristic on the logs such as the ripple time difference, resistivity, neutron.
Step 2022, the log criterion of identification for being divided hydrocarbon source rock and country rock is obtained according to response characteristic, with
Pick out preferred log.
Specifically, the log for being adapted to the hydrocarbon source rock in research area and country rock to divide is established according to response characteristic identifies mark
Standard, preferably log are the obvious log of response characteristic, and further, can selecting response characteristic, most significantly well logging is bent
Line is preferred log.
Step 2023, according to preferred log, calculated using Δ LogR methods and obtain each drilling well hydrocarbon source rock interval longitudinal direction
Toc prediction curves.
Each drilling well hydrocarbon source rock interval longitudinal direction in research area, continuous Toc prediction curves are obtained to calculate with Δ LogR methods, to protect
Demonstrate,prove the high precision on longitudinal direction.
Further, in addition to step 2024：
Calculated in the low region of research area's degree of prospecting and obtain virtual log Toc prediction curves, the virtual log Toc predictions are bent
Line is in the case of drilling well deficiency, and the Toc prediction curves of acquisition are calculated by the way of virtual drilling well.
Due to the situation of drilling well deficiency occurring in the low region of research area's degree of prospecting, thus it is bent calculating Toc predictions
Larger deviation occurs during line, therefore, assumes in the present embodiment with the presence of more mouthfuls of drilling wells, i.e., in the case of drilling well deficiency,
By the way of virtual drilling well, to calculate the Toc prediction curves of acquisition, to obtain virtual log Toc prediction curves.
Step 2024 is settable before step 2021, can also to be arranged on after step 2023 after step 201, walked
Before rapid 2031.
Step 203, Toc prediction curves are corrected according to sample spot Toc values, to obtain Toc curves；
Step 2041, fine stratum calibration is carried out to each drilling well in research area, determines the earthquake of hydrocarbon source rock top bottom reflecting interface
Reflectance signature.
Step 2042, the seismic interpretation that at least 10 × 10 grids are carried out to hydrocarbon source rock top bottom interface works, wherein, earthquake solution
Releasing work includes the explanation of layer position and fault interpretation.
Step 2041- steps 2042, it is mainly used in carrying out composite traces work to research area's drilling well, it is fine to carry out layer position
Demarcation, the Seismic reflection character of hydrocarbon source rock top bottom reflecting interface is determined, and at least 10*10 grids are carried out to hydrocarbon source rock top bottom interface
Seismic interpretation work, wherein, seismic interpretation include layer position explain and fault interpretation, then to its gridding handle, to connect down
The seismic properties come, which calculate, carries out data preparation.
Step 2043, according to the seismic properties of Toc curves and seismic trace near well, using probabilistic neural network algorithm to earthquake
Attribute is trained, to obtain preferred seismic properties, wherein it is preferred to which it is the seismic properties best with Toc correlations to shake attribute.
Further, preferably seismic properties include 3 to 7 seismic properties.
Step 205, the preferred seismic properties of three dimensional seismic data are extracted, calculate the Toc data volumes for obtaining distributed in three dimensions, with
Complete research area's hydrocarbon source rock prediction.
The hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm that the present embodiment provides, with reference to hydrocarbon source rock well logging and ground
Shake response characteristic, from actual measurement Toc (the Toc value)s of sample spot, Toc prediction curves are established using logging prediction Toc method,
And in the case of the low drilling well deficiency of research area's degree of prospecting, by the way of virtual drilling well, virtual log Toc is obtained to calculate
Prediction curve, and using Toc prediction curves or virtual log Toc prediction curves as dependent variable, choose seismic properties as independent variable, adopt
With the best fit relational expression established based on probabilistic neural network algorithm between them, so as to obtain optimal seismic properties, then
By extracting optimal seismic properties from 3-d seismic data set, the Toc data volumes of distributed in three dimensions are calculated, so as to realize hydrocarbon
Source rock is predicted.Using the non-linear relation that can be preferably caught based on probabilistic neural network algorithm between seismic properties and Toc, make
It is more accurate to the prediction result of hydrocarbon source rock.
Embodiment three
The present embodiment is device embodiment, for performing the method in above-described embodiment one.
Fig. 3 is that the structure for the hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm that the embodiment of the present invention three provides is shown
It is intended to；As shown in figure 3, the present embodiment provides a kind of hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm, including sample
Point Toc values acquisition module 301, Toc prediction curves acquisition module 302, preferably Toc prediction curves correction module 303, seismic properties
Acquisition module 304 and Toc data volumes acquisition module 305.
Wherein, sample spot Toc values acquisition module 301, hydrocarbon source rock rock core, landwaste for the research each drilling well in area to acquisition
Sample carries out analysis test, to obtain the Toc values of sample spot.
Toc prediction curves acquisition module 302, the research each drilling well hydrocarbon source rock interval in area is obtained for being calculated using log
The Toc prediction curves of longitudinal direction.
Toc prediction curves correction module 303, Toc prediction curves are corrected for the Toc values according to sample spot, with
Obtain Toc curves.It is preferred that seismic properties acquisition module 304, for the seismic properties according to Toc curves and seismic trace near well, profit
Seismic properties are trained with probabilistic neural network algorithm, to obtain preferred seismic properties, wherein it is preferred to shake attribute be with
The best seismic properties of Toc correlations.
Toc data volumes acquisition module 305, for extracting the preferred seismic properties of three dimensional seismic data, calculate and obtain three-dimensional
The Toc data volumes of distribution, to complete research area's hydrocarbon source rock prediction.
The present embodiment be with one corresponding device embodiment of embodiment of the method, for details, reference can be made to the description in embodiment one,
It will not be repeated here.
Example IV
The present embodiment is the supplementary notes carried out on the basis of embodiment three, for performing the side in above-described embodiment two
Method.
Fig. 4 is that the structure for the hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm that the embodiment of the present invention four provides is shown
It is intended to；As shown in figure 4, the present embodiment provides a kind of hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm, wherein Toc is pre-
Surveying curve acquisition module 302 includes response characteristic acquisition submodule 3021, the and of log criterion of identification acquisition submodule 3022
Toc prediction curves acquisition submodule 3023.
Wherein, response characteristic acquisition submodule 3021, for the cross plot of the log according to research area's drilling well, analysis
Response characteristics of the hydrocarbon source rock Toc on log；
Log criterion of identification acquisition submodule 3022, mud stone and sandstone are drawn for being obtained according to response characteristic
The log criterion of identification divided；
Toc prediction curves acquisition submodule 3023, for according to log criterion of identification, being calculated and being obtained using Δ LogR methods
Obtain the Toc prediction curves of each drilling well hydrocarbon source rock interval longitudinal direction.
Further, Toc prediction curves acquisition module 302 also includes：
Virtual log Toc prediction curves acquisition submodule 3024, obtained for being calculated in the low region of research area's degree of prospecting
Virtual log Toc prediction curves, the virtual log Toc prediction curves are in the case of drilling well deficiency, using virtual drilling well
Mode calculates the Toc prediction curves of acquisition.
Further, also wrapped before the seismic properties of seismic trace near well are obtained preferably in seismic properties acquisition module 304
Include：
Fine stratum calibration is carried out to each drilling well in research area, determines the Seismic reflection character of hydrocarbon source rock top bottom reflecting interface；
The seismic interpretation that at least 10 × 10 grids are carried out to hydrocarbon source rock top bottom interface works, wherein, seismic interpretation saddlebag
Include the explanation of layer position and fault interpretation.
The present embodiment be with two corresponding device embodiment of embodiment of the method, for details, reference can be made to the description in embodiment two,
It will not be repeated here.
Although by reference to preferred embodiment, invention has been described, is not departing from the situation of the scope of the present invention
Under, various improvement can be carried out to it and part therein can be replaced with equivalent.Especially, as long as being rushed in the absence of structure
Prominent, the every technical characteristic being previously mentioned in each embodiment can combine in any way.The invention is not limited in text
Disclosed in specific embodiment, but all technical schemes including falling within the scope of the appended claims.
Claims (9)
- A kind of 1. hydrocarbon source rock Forecasting Methodology based on probabilistic neural network algorithm, it is characterised in that including：Step 101, analysis test is carried out to hydrocarbon source rock rock core, the chip sample of the research each drilling well in area of acquisition, to obtain sample The Toc values of point；Step 102, the Toc prediction curves for obtaining each drilling well hydrocarbon source rock interval longitudinal direction in research area are calculated using log；Step 103, Toc prediction curves are corrected according to the Toc values of sample spot, to obtain Toc curves；Step 104, according to the seismic properties of Toc curves and seismic trace near well, using probabilistic neural network algorithm to seismic properties It is trained, to obtain preferred seismic properties, wherein it is preferred to which it is the seismic properties best with Toc correlations to shake attribute；Step 105, the preferred seismic properties of three dimensional seismic data are extracted, calculate the Toc data volumes for obtaining distributed in three dimensions, to complete Study the prediction of area's hydrocarbon source rock.
- 2. the hydrocarbon source rock Forecasting Methodology according to claim 1 based on probabilistic neural network algorithm, it is characterised in that step 102 specifically include：According to the cross plot of the log of research area's drilling well, response characteristics of the analysis hydrocarbon source rock Toc on log；The log criterion of identification for being divided hydrocarbon source rock and country rock is obtained according to response characteristic, to pick out preferred well logging Curve；According to preferred log, the Toc prediction curves for obtaining each drilling well hydrocarbon source rock interval longitudinal direction are calculated using Δ LogR methods.
- 3. the hydrocarbon source rock Forecasting Methodology according to claim 1 based on probabilistic neural network algorithm, it is characterised in that step 102 also include：Calculated in the low region of research area's degree of prospecting and obtain virtual log Toc prediction curves, the virtual log Toc prediction curves are In the case of drilling well deficiency, to calculate the Toc prediction curves of acquisition by the way of virtual drilling well.
- 4. the hydrocarbon source rock Forecasting Methodology according to claim 1 based on probabilistic neural network algorithm, it is characterised in that step In 103 before the seismic properties of seismic trace near well are obtained, in addition to：Fine stratum calibration is carried out to each drilling well in research area, determines the Seismic reflection character of hydrocarbon source rock top bottom reflecting interface；The seismic interpretation that at least 10 × 10 grids are carried out to hydrocarbon source rock top bottom interface works, wherein, seismic interpretation work includes layer Position is explained and fault interpretation.
- 5. the hydrocarbon source rock Forecasting Methodology according to claim 1 based on probabilistic neural network algorithm, it is characterised in that preferably Seismic properties include 3 to 7 seismic properties.
- A kind of 6. hydrocarbon source rock prediction meanss based on probabilistic neural network algorithm, it is characterised in that including：Sample spot Toc value acquisition modules, hydrocarbon source rock rock core, chip sample for each drilling well in research area to acquisition are analyzed Test, to obtain the Toc values of sample spot；Toc prediction curve acquisition modules, each drilling well hydrocarbon source rock interval longitudinal direction in research area is obtained for being calculated using log Toc prediction curves；Toc prediction curve correction modules, Toc prediction curves are corrected for the Toc values according to sample spot, to obtain Toc Curve；It is preferred that seismic properties acquisition module, for the seismic properties according to Toc curves and seismic trace near well, utilizes probabilistic neural net Network algorithm is trained to seismic properties, to obtain preferred seismic properties, wherein it is preferred to shake attribute be with Toc correlations most Good seismic properties；Toc data volume acquisition modules, for extracting the preferred seismic properties of three dimensional seismic data, calculate and obtain distributed in three dimensions Toc data volumes, to complete research area's hydrocarbon source rock prediction.
- 7. the hydrocarbon source rock prediction meanss according to claim 6 based on probabilistic neural network algorithm, it is characterised in that Toc Prediction curve acquisition module includes：Response characteristic acquisition submodule, for the cross plot of the log according to research area's drilling well, analysis hydrocarbon source rock Toc is being surveyed Response characteristic on well curve；Log criterion of identification acquisition submodule, for obtaining the survey for being divided hydrocarbon source rock and country rock according to response characteristic Well Curves Recognition standard, to pick out preferred log；Toc prediction curve acquisition submodules, for according to preferred log, being calculated using Δ LogR methods and obtaining each drilling well hydrocarbon source The Toc prediction curves of rock stratum section longitudinal direction.
- 8. the hydrocarbon source rock prediction meanss according to claim 6 based on probabilistic neural network algorithm, it is characterised in that Toc Prediction curve acquisition module also includes：Virtual log Toc prediction curve acquisition submodules, virtual log Toc is obtained for being calculated in the low region of research area's degree of prospecting Prediction curve, the virtual log Toc prediction curves are in the case of drilling well deficiency, are calculated and obtained by the way of virtual drilling well The Toc prediction curves obtained.
- 9. the hydrocarbon source rock prediction meanss according to claim 6 based on probabilistic neural network algorithm, it is characterised in that preferably In seismic properties acquisition module before the seismic properties of seismic trace near well are obtained, in addition to：Fine stratum calibration is carried out to each drilling well in research area, determines the Seismic reflection character of hydrocarbon source rock top bottom reflecting interface；The seismic interpretation that at least 10 × 10 grids are carried out to hydrocarbon source rock top bottom interface works, wherein, seismic interpretation work includes layer Position is explained and fault interpretation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610375298.3A CN107436452A (en) | 2016-05-27 | 2016-05-27 | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610375298.3A CN107436452A (en) | 2016-05-27 | 2016-05-27 | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107436452A true CN107436452A (en) | 2017-12-05 |
Family
ID=60458720
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610375298.3A Pending CN107436452A (en) | 2016-05-27 | 2016-05-27 | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107436452A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108376295A (en) * | 2018-01-31 | 2018-08-07 | 北京博达瑞恒科技有限公司 | A kind of oil gas dessert prediction technique and storage medium |
| CN108717211A (en) * | 2018-06-01 | 2018-10-30 | 北京师范大学 | A kind of prediction technique of the Effective source rocks abundance in few well area |
| CN109086773A (en) * | 2018-08-29 | 2018-12-25 | 电子科技大学 | Fault plane recognition methods based on full convolutional neural networks |
| CN109507733A (en) * | 2018-12-11 | 2019-03-22 | 中国石油天然气股份有限公司大港油田分公司 | A kind of method and device for predicting organic matter abundance in hydrocarbon source rock |
| CN109669210A (en) * | 2018-11-29 | 2019-04-23 | 中国石油大学(华东) | Favorable method based on a variety of seismic properties interpretational criterias |
| CN111812210A (en) * | 2020-07-08 | 2020-10-23 | 丁刚 | Method and device for positioning damage source of three-dimensional braided composite material |
| CN112444857A (en) * | 2019-09-02 | 2021-03-05 | 中国石油天然气股份有限公司 | Paste-salt earth physical prediction method based on post-stack seismic data |
| CN112703429A (en) * | 2018-09-14 | 2021-04-23 | Bp北美公司 | Machine learning based seismic attribute analysis |
| CN113837501A (en) * | 2021-11-25 | 2021-12-24 | 中国科学院地质与地球物理研究所 | Shale total organic carbon prediction method, system and equipment based on graph neural network |
| CN115796350A (en) * | 2022-11-23 | 2023-03-14 | 长江大学 | Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150039544A1 (en) * | 2013-07-31 | 2015-02-05 | Schlumberger Technology Corporation | Resource production forecasting |
| CN104977618A (en) * | 2014-04-09 | 2015-10-14 | 中国石油集团东方地球物理勘探有限责任公司 | Method for evaluating shale gas reservoir and finding dessert area |
-
2016
- 2016-05-27 CN CN201610375298.3A patent/CN107436452A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150039544A1 (en) * | 2013-07-31 | 2015-02-05 | Schlumberger Technology Corporation | Resource production forecasting |
| CN104977618A (en) * | 2014-04-09 | 2015-10-14 | 中国石油集团东方地球物理勘探有限责任公司 | Method for evaluating shale gas reservoir and finding dessert area |
Non-Patent Citations (4)
| Title |
|---|
| 王宏语: "《地震资料地质解释与应用》", 30 April 2016, 地震出版社 * |
| 窦玉坛 等: "黄土塬少井区采用果蝇优化算法构建虚拟井方法", 《地球物理学进展》 * |
| 谢锐杰 等: "《储层描述及油气预测方法与应用》", 31 March 2011, 湖北科学技术出版社 * |
| 金吉能 等: "地震多属性反演预测页岩总有机碳含量", 《石油天然气学报》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108376295A (en) * | 2018-01-31 | 2018-08-07 | 北京博达瑞恒科技有限公司 | A kind of oil gas dessert prediction technique and storage medium |
| CN108717211A (en) * | 2018-06-01 | 2018-10-30 | 北京师范大学 | A kind of prediction technique of the Effective source rocks abundance in few well area |
| CN108717211B (en) * | 2018-06-01 | 2019-05-24 | 北京师范大学 | A kind of prediction technique of the Effective source rocks abundance in few well area |
| CN109086773B (en) * | 2018-08-29 | 2022-03-04 | 电子科技大学 | Fault plane identification method based on full convolution neural network |
| CN109086773A (en) * | 2018-08-29 | 2018-12-25 | 电子科技大学 | Fault plane recognition methods based on full convolutional neural networks |
| CN112703429B (en) * | 2018-09-14 | 2024-02-13 | Bp北美公司 | Machine learning-based seismic attribute analysis |
| CN112703429A (en) * | 2018-09-14 | 2021-04-23 | Bp北美公司 | Machine learning based seismic attribute analysis |
| CN109669210A (en) * | 2018-11-29 | 2019-04-23 | 中国石油大学(华东) | Favorable method based on a variety of seismic properties interpretational criterias |
| CN109669210B (en) * | 2018-11-29 | 2020-05-01 | 中国石油大学(华东) | Favorable area prediction method based on multiple seismic attribute evaluation criteria |
| CN109507733A (en) * | 2018-12-11 | 2019-03-22 | 中国石油天然气股份有限公司大港油田分公司 | A kind of method and device for predicting organic matter abundance in hydrocarbon source rock |
| CN112444857A (en) * | 2019-09-02 | 2021-03-05 | 中国石油天然气股份有限公司 | Paste-salt earth physical prediction method based on post-stack seismic data |
| CN112444857B (en) * | 2019-09-02 | 2023-10-27 | 中国石油天然气股份有限公司 | Paste salt geophysical prediction method based on post-stack seismic data |
| CN111812210A (en) * | 2020-07-08 | 2020-10-23 | 丁刚 | Method and device for positioning damage source of three-dimensional braided composite material |
| CN113837501A (en) * | 2021-11-25 | 2021-12-24 | 中国科学院地质与地球物理研究所 | Shale total organic carbon prediction method, system and equipment based on graph neural network |
| CN113837501B (en) * | 2021-11-25 | 2022-03-08 | 中国科学院地质与地球物理研究所 | Shale total organic carbon prediction method, system and equipment based on graph neural network |
| US11880776B2 (en) | 2021-11-25 | 2024-01-23 | Institute Of Geology And Geophysics, Chinese Academy Of Sciences | Graph neural network (GNN)-based prediction system for total organic carbon (TOC) in shale |
| CN115796350A (en) * | 2022-11-23 | 2023-03-14 | 长江大学 | Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107436452A (en) | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm | |
| CN101930082B (en) | Method for discriminating reservoir fluid type by using resistivity data | |
| CN104453873B (en) | Evaluation method for shale oil and gas economic and effective layer section | |
| CN105221133B (en) | A kind of method and apparatus that content of organic carbon of hydrocarbon source rock is determined based on well logging multi-parameter | |
| CN104564041B (en) | Hyposmosis clastic reservoir rock efficiency evaluation method based on exploitation permeability limits | |
| CN104636588B (en) | Method and device for calculating content of organic carbon in source rock | |
| CN105675635B (en) | Method and device for determining relative content of components of compact rock and brittleness index | |
| CN103114840B (en) | Method and device for calculating organic carbon content of high-too-high mature shale | |
| CN105277982B (en) | A kind of mud shale total content of organic carbon earthquake prediction method | |
| US9244182B2 (en) | Method of assessing hydrocarbon source rock candidate | |
| CN105388525B (en) | Method for predicting reservoir and device | |
| CN102012526A (en) | Method for discriminating type of reservoir fluid by using resistivity data | |
| CN104727813B (en) | The porosity measurement method on one introduces a collection storage symbiotic type stratum | |
| CN105697002A (en) | Method for recognizing coal measure strata lithology | |
| CN105182424A (en) | Method and device for quantitatively predicting reservoir porosity based on patch saturation model | |
| CN106033127B (en) | Crustal stress azimuthal seismic Forecasting Methodology based on shear wave velocity rate of change | |
| Michael et al. | The South West Hub In-Situ Laboratory–a facility for CO2 injection testing and monitoring in a fault zone | |
| Chilton et al. | Uncovering the Controls on Fluvial Bedrock Erodibility and Knickpoint Expression: A High‐Resolution Comparison of Bedrock Properties Between Knickpoints and Non‐Knickpoint Reaches | |
| CN106285623A (en) | Determine the method and system of total content of organic carbon | |
| CN110118994B (en) | Continental facies hydrocarbon source rock quantitative prediction method based on seismic inversion and machine learning | |
| CN112528106A (en) | Volcanic lithology identification method | |
| Manda et al. | Influence of rock fabric on fracture attribute distribution and implications for groundwater flow in the Nashoba Terrane, eastern Massachusetts | |
| Rotimi et al. | Reservoir characterization and modeling of lateral heterogeneity using multivariate analysis | |
| Wallace | Use of 3-dimensional dynamic modeling of CO₂ injection for comparison to regional static capacity assessments of Miocene sandstone reservoirs in the Texas State Waters, Gulf of Mexico | |
| CN109598049B (en) | Method for drilling rock fracture development degree and regional rock fracture development rule |
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: 20171205 |
|
| RJ01 | Rejection of invention patent application after publication |