CN110502569A - A kind of standard well screen based on Discrete Choice Model selects visual analysis method - Google Patents
A kind of standard well screen based on Discrete Choice Model selects visual analysis method Download PDFInfo
- Publication number
- CN110502569A CN110502569A CN201910758272.0A CN201910758272A CN110502569A CN 110502569 A CN110502569 A CN 110502569A CN 201910758272 A CN201910758272 A CN 201910758272A CN 110502569 A CN110502569 A CN 110502569A
- Authority
- CN
- China
- Prior art keywords
- well
- standard well
- standard
- logging
- attribute
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 28
- 230000000007 visual effect Effects 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 55
- 238000005070 sampling Methods 0.000 claims abstract description 22
- 230000003044 adaptive effect Effects 0.000 claims abstract description 14
- 238000013461 design Methods 0.000 claims abstract description 14
- 230000002452 interceptive effect Effects 0.000 claims abstract description 9
- 238000012216 screening Methods 0.000 claims abstract description 9
- 230000008859 change Effects 0.000 claims abstract description 7
- 238000012800 visualization Methods 0.000 claims description 20
- 230000015572 biosynthetic process Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 4
- 238000012482 interaction analysis Methods 0.000 claims description 3
- 241001269238 Data Species 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 claims description 2
- 238000011002 quantification Methods 0.000 claims description 2
- 241000208340 Araliaceae Species 0.000 claims 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 1
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 1
- 235000008434 ginseng Nutrition 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 11
- 238000010586 diagram Methods 0.000 description 9
- 238000009394 selective breeding Methods 0.000 description 4
- 239000003245 coal Substances 0.000 description 3
- 239000012141 concentrate Substances 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000002708 enhancing effect Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 239000004744 fabric Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- RHZUVFJBSILHOK-UHFFFAOYSA-N anthracen-1-ylmethanolate Chemical compound C1=CC=C2C=C3C(C[O-])=CC=CC3=CC2=C1 RHZUVFJBSILHOK-UHFFFAOYSA-N 0.000 description 1
- 239000003830 anthracite Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013079 data visualisation Methods 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000007794 visualization technique Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/26—Visual data mining; Browsing structured data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mining & Mineral Resources (AREA)
- Economics (AREA)
- Animal Husbandry (AREA)
- Marine Sciences & Fisheries (AREA)
- Life Sciences & Earth Sciences (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Agronomy & Crop Science (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of standard well screen based on Discrete Choice Model and selects visual analysis method, comprising: keeps the overall space of standard well to be distributed using adaptive blue noise sampling model；The multiple dimensioned stratum Matching Model based on dynamic programming algorithm is designed, the similarity degree of well is measured from the differences in angle of different attribute；The priori knowledge of integrated expert user, designs the standard well screen choosing method based on Discrete Choice Model, maximizes the effectiveness between attributes similarity and standard well；The standard well screen choosing method of design iteration interactive mode supports user to change standard well according to priori knowledge, updates Discrete Choice Model, optimisation criteria well screening process and result iteratively.The present invention is on the basis of comprehensively considering well logging spatial distribution and multidimensional property information, that realizes visual analysis driving has the choosing of supervision standard well screen, standard well obtained can represent to the utmost around logs well, and is conducive to the promotion of subsequent stratum intelligent Matching precision and efficiency.
Description
Technical field
The present invention relates to a kind of, and the standard well screen based on Discrete Choice Model selects visual analysis method, belong to oil exploration,
Graphics and visualization technique field.
Background technique
3-D seismics wave number evidence and one-dimensional log data are two kinds of common data types of geologic structure interpretation field, largely
Research work around above two data type carry out geologic structure interpretation work.Seismic data is by task equipment
Record underground waveform signal simultaneously obtains after carrying out series of preprocessing to it, extracted by isochronous surface, volume, etc. analysis methods
User is helped to realize geologic structure interpretation.For example, the scholars such as Patel devise the seismic wave data analysis side based on two dimension slicing
Method (Patel D, Giertsen C, Thurmond J, et al.The seismic analyzer:interpreting and
illustrating 2D seismic data[J].IEEE transactions on visualization&computer
Graphics.2008,14 (6): 1571-1578.), user can carry out predecomposition to seismic wave data Layer bit architecture, and use
The rendering algorithms such as texture and Texture Transfer function enhancing display Geologic Structure Feature is deformed, to improve to architectonic explanation
Precision reduces the workload of geologists.However, factors all affect generation and the place of seismic data in the fact
Reason process, such as complicated orographic condition, limited appointed condition and inevitably calculating error, these all give geological structure solution
It releases and brings very big uncertainty (Zhou B, Hatherly P, Sun W.Enhancing the detection of
small coal structures by seismic diffraction imaging[J].International Journal
of Coal Geology,2017,178(6):1-12.)。
Different from secondhand seismic data, one-dimensional log data acquires acquisition directly from practical logging, a variety of
Closely related attribute is recorded in detail with geologic structure.For example, natural potential attribute can recorde between stratum and mud
The electrochemical action of generation and electrokinetics effect, for determining permeable formation and Water Flooding Layer (Li J, Liu D, Yao Y, et
al.Evaluation of the reservoir permeability of anthracite coals by
geophysical logging data[J].International Journal of Coal Geology,2011,87(2):
121-127.).Interval transit time attribute is to determine formation porosity in the spread speed on stratum by sound wave, to identify lithology
(Tang X M,Zheng Y,Patterson D.Processing array acoustic-logging data to image
near-borehole geologic structures[J].Geophysics,2007,72(2):87-97.).Based on more attributes
Log data carries out the important step that Strata Comparison is geologic structure interpretation, is more and more closed in geological exploration field
Note, such as based on the twin-well matching algorithm for intersecting optimization algorithm, rule-based expert system algorithm etc..
Therefore, representative standard well subset (Ren Z H, Zhang W is screened from the numerous well logging of original amount
C,Zhu X M,et al.Method to Determine the Collar Section Depth in Standard Well
Logging [J] .Petroleum Drilling Techniques, 2014 (6): 68-72.), carry out accurate and careful stratum
Comparison can quickly understand the geological conditions in subrange, and then effectively instruct the stratum matching of extensive original well logging, keep away
Exempt from a large amount of duplicate artificial labeling process, improves the efficiency of geological prospecting and exploitation to a certain extent.However, the sieve of standard well
Choosing relies on expert's priori knowledge, and especially traditional standard well selection course depend heavilys on handmarking, screening process
It is not only complicated and time-consuming, and tend not to consider the spatial distribution and multiattribute correlation of well logging, this gives subsequent ground texture
It makes explanation and brings very big uncertainty, it is difficult to promote the efficiency and precision of standard well selection.Therefore, in order to overcome artificial selection
The limitation of standard well comprehensively considers the spatial distribution of well logging and the correlation of multidimensional property, proposes a kind of based on discrete selection
Model has supervised standard well standard well screen to select visual analysis method.
Summary of the invention
The object of the present invention is to provide a kind of, and the standard well screen based on Discrete Choice Model selects visual analysis method, thus
During data visualization, realizes that the standard well screen for having supervised is selected, reduce error and uncertainty, improve log data
Visualize efficiency.
To achieve the above object, the technical scheme adopted by the invention is that: a kind of standard well based on Discrete Choice Model
Visual analysis method is screened, is specifically comprised the following steps:
(1) it is based on multidimensional log data, the stratum of the spatial distribution characteristic and multidimensional well log attributes that extract well logging respectively is closed
Connection relationship: under the premise of the standard well screen of determination selects ratio, meet local space using the acquisition of adaptive blue noise sampling algorithm
The Poisson disk of distribution；Meanwhile the multiple dimensioned stratum Matching Model based on dynamic programming algorithm is designed, from the angular amount of different attribute
Change measurement well log attributes similarity degree.
(2) standard well sample set is obtained according to expert's priori knowledge, is surveyed using described in Discrete Choice Model maximization steps 1
Effectiveness between well attribute similarity degree and standard well screen choosing traverses the pool that the adaptive blue noise sampling algorithm obtains
Loose disk according to the Discrete Choice Model screening criteria well, and obtains standard well filter information.
(3) it according to the standard well filter information, guides user's interactive mode specified by visualization analysis technique or changes
Standard well, standard well sample set information described in adaptive updates optimize Discrete Choice Model parameter, screen, met again
The standard well of user's current demand and experience.
Further, in step (1), described obtained using adaptive blue noise sampling algorithm meets local space distribution
The method of Poisson disk, specifically:
1) according to the spatial position of well logging, the Poisson disk for meeting local space distribution is obtained using blue noise sampling algorithm,
Only allow to sample a well logging inside each Poisson disk；
2) using the space distribution situation logged well in Poisson disk described in Density Estimator algorithm evaluation, the survey is calculated
The spacial distribution density value of well, and then the size of the Poisson disk radius is adaptively updated, keep the space point of the well logging
Cloth feature.
Further, in step (1), the side of the multiple dimensioned stratum Matching Model of the design based on dynamic programming algorithm
Method, specifically:
1) according to the multidimensional log data, using each well log attributes curve of median filter smoothness of image, Nonlinear Processing institute
It states multidimensional log data and is normalized between 0 and 1 according to discrete results；
2) according to the smooth log, the identification on stratigraphic horizon boundary is carried out to well logging using activity function and is drawn
Point；
3) the optimal layer position between two wells is solved by dynamic programming algorithm and matches sequence, by best on each well log attributes
Sum to obtain the quantified property difference between two wells with the corresponding formation thickness of sequence.
Further, described to utilize the similar journey of well log attributes described in Discrete Choice Model maximization steps 1 in step (2)
The method of effectiveness between degree and standard well screen choosing, specifically:
1) according to the Poisson disk for meeting local space distribution, the standard of handmarking is obtained using expert's priori knowledge
Well sample set；
2) fixed utility function is designed according to the multidimensional property similarity between logging well in the Poisson disk, using multinomial recurrence
Modeling random error distribution, both simultaneous obtain total utility function, to calculate different user select different well loggings as
The probability of standard well；
3) according to the standard well sample set, the phase in the total utility function is calculated automatically using Discrete Choice Model
It answers parameter and traverses the Poisson disk that the adaptive blue noise sampling algorithm obtains, and then select combinations of attributes in local space
Maximize the standard well of user utility.
Further, it guides user's interactive mode specified by visualization analysis technique described in step (3) or changes standard
The method of well, specifically:
1) a variety of visualization scheme collaborations are designed and show log data information: the real space position of well logging is mapped to ground
In figure, retain the visualization result of its spatial distribution characteristic；According to different attribute log data, design attributes curve graph, by office
The same attribute datas of different well loggings count in specified attribute space in portion space, using tone mapping technique by the standard
Well is distinguished with non-standard well；Quantify the difference between the log using variance, statistics obtains local space internal standard
Total attribute difference histogram of well, the contrast of reserved property feature.
2) according to the visualization result, design standard well screen selects visualization scheme, counts each attribute of each local space
Upper characteristic matching matches incidence matrix as a result, obtaining multidimensional property layer position；Screening annulus is set according to layer position matching result, on ground
It is intuitively compared on figure and enhances the attribute difference in display local space between well logging.
3) according to the comparison of the attribute difference, user's interaction analysis and replacement standard well, and then more new samples, root are supported
Optimize Discrete Choice Model parameter according to new sample set, recalculates and screen the standard well for meeting user demand.
Compared with prior art, the beneficial effects of the present invention are: not only comprehensively considering well logging during the choosing of standard well screen
Spatial distribution, the multidimensional property correlation of log data, go back the priori knowledge of quantitative evaluation domain expert, and then using visual
Analytical technology realizes that the standard well screen for having supervised is selected, and effectively reduces handmarking's workload and improves the choosing of standard well screen
Precision and efficiency.In conclusion the parameter tuning process in the method for the present invention Plays well preference pattern is simple and easy, well logging
Characteristic information in data obtains quick and intuitive displaying by the visual design of a variety of collaborations, and construct it is easily multi-functional can
Depending on analysis system, supports the dynamic of standard well to update and optimize, the analysis efficiency in standard well selection course can be significantly improved,
The precision of promotion standard well selection, can effectively meet the real-time application demand of user.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention；
Fig. 2 is blue noise sampling model schematic diagram, wherein (a) is initial schematic diagram, (b), (c) is selected for standard well
Journey exemplary diagram；
Fig. 3 is the multiple dimensioned stratum Matching Model schematic diagram based on dynamic programming algorithm；
Fig. 4 is system interface figure (geographic view), wherein (a) is information view, is (b) sampling panel, is (c) selection
Optimize panel, (d) be geographic view, (e) be multidimensional property view, (f) be stratum match views, (g) is stratum incidence matrix
View (h) is total attribute difference view；
Fig. 5 is the attribute evaluation result figure of selected standard well under given sample rate, wherein (a) is the ground of selected standard well
Figure projection result is (b) the well log attributes Drawing of Curve of standard well as a result, (c) being standard well in multidimensional property similarities,
It (d) is total attribute difference of standard well as a result, (e) being the user preference coefficient results being currently calculated, (f) between well logging
Actual formation correlation degree.
Fig. 6 is the comparative analysis schematic diagram of standard well after more new samples, wherein (a) is the user preference before Sample Refreshment
Coefficient results, (b) the stratum association results for the standard well before Sample Refreshment on multidimensional property are (c) weight after Sample Refreshment
The user preference coefficient results newly calculated, (d) the stratum association results for the standard well after Sample Refreshment on multidimensional property.
Specific embodiment
With reference to the accompanying drawing, the standard well visual analysis method to of the invention based on Discrete Choice Model is made further
Explanation.
It, should if Fig. 1 is the flow chart that a kind of standard well screen based on Discrete Choice Model of the present invention selects visual analysis method
Method specifically:
(1) it is based on multidimensional log data, the stratum of the spatial distribution characteristic and multidimensional well log attributes that extract well logging respectively is closed
Connection relationship: under the premise of the standard well screen of determination selects ratio, meet local space using the acquisition of adaptive blue noise sampling algorithm
The Poisson disk of distribution；Based on actual complex geological condition, the multiple dimensioned stratum Matching Model based on dynamic programming algorithm is designed,
Well log attributes similarity degree is measured from the angular quantification of different attribute.
If Fig. 2 is blue noise sampling model schematic diagram, described obtained using adaptive blue noise sampling algorithm meets part
The method of the Poisson disk of spatial distribution, specifically:
1) according to the spatial position of well logging, the Poisson disk for meeting local space distribution is obtained using blue noise sampling algorithm,
Only allow to sample a well logging inside each Poisson disk；
2) using the space distribution situation logged well in Poisson disk described in Density Estimator algorithm evaluation, the survey is calculated
The spacial distribution density value of well, and then the size of the Poisson disk radius is adaptively updated, keep the space point of the well logging
Cloth feature.
Firstly, wiIt indicates to log well flatly,Expression one is with wiCentered on, ri=r/f (wi) be radius sampling Poisson
Disk, parameter r indicate the sample rate defined by user's interaction, and function f (w) indicates to carry out estimation calculating according to well logging spatial distribution
Kernel density function, calculating process such as formula (1)；
In formula (1), KhIndicate that a bandwidth is the gaussian kernel function of h, f (wi) indicate well logging wiDensity Estimator value.
In the method, it is specified that each sampling disk only allows to select a well as standard well；It makes simultaneously as given a definition: detection
AnnulusIt indicates with wiCentered on, riFor internal diameter and di=2riFor the annular region of outer diameter, Q indicates the sampled well of artificial selection
Queue, SPQIndicate the set of all well loggings in Q queue in sampler tray corresponding to sampled well.
Secondly, all well loggings are collectively labeled as " enlivening ".One of well is randomly choosed as sample, and is added to
In queue Q.In iteration sampling process, by SPQIn well be changed to " inactive ".As shown in Fig. 2 (a), selected at random from queue Q
Select a bite sample well w0, determine that it samples diskWith detection annulusAs shown in Fig. 2 (b), fromMiddle random selection
" active " well w flatly2It is detected as possible sampled well, ifAny sampled well marked in queue Q is covered,
As shown in Fig. 2 (c), " active " well w2It will be changed to " inactive ", and will beIn reselect another " active " well w3
It continues to test, the well logging until finding the condition of satisfaction is marked as sampled well and is added in Q queue.If w0Surrounding does not have
There is effective sample well, then removes it from Q.During repeated sampling, until all well loggings all become an inactive state.
The method of the multiple dimensioned stratum Matching Model of the design based on dynamic programming algorithm, specifically:
1) corresponding to obtain using each well log attributes curve of median filter smoothness of image according to the multidimensional log data
Smooth log, multidimensional log data described in Nonlinear Processing simultaneously normalize between 0 and 1 according to discrete results；
2) according to the smooth log, the identification on stratigraphic horizon boundary is carried out to well logging using activity function and is drawn
Point；
3) the optimal layer position between two wells is solved by dynamic programming algorithm and matches sequence, by best on each well log attributes
Sum to obtain the quantified property difference between two wells with the corresponding formation thickness of sequence.
Detailed process such as Fig. 3 institute of the method for the multiple dimensioned stratum Matching Model of the design based on dynamic programming algorithm
Show, the first step, original well log attributes curve is smoothed, second step, is smoothly logged well song using activity function to each item
Line carries out the identification of layer position, shown in activity function equation such as formula (2):
Wherein, EiIndicate the active value at depth i.yjIndicate the well log attributes value in depth bounds [i-L, i+L], L is
Half window length.If the active value at well depth i is greater than given threshold value, corresponding depth i is identified ground as boundary
Layer.Third step finds out target using the two well matching process based on Dynamic Programming according to the ready-portioned stratigraphic horizon of target well
Optimum Matching path between well, to obtain the best match sequence between target well.Assuming that the given well A with m layers of stratum
With the well B with n-layer stratum, can be found between them according to global cost function most has coupling path, such as formula (3) institute
Show:
Wherein, C (Ai,Bj) indicate (A1,B1) arrive (Ai,Bj) coupling path on the sum of difference；d(Ai,Bj) it is two ground
The summation of all feature differences between layer (i-th layer of well A and well B jth layer), for measuring AiAnd BjSimilitude.And it is lacked in two wells
The difference of lost territory layer is expressed as g (Ai) or g (Bj).C (A is solved using dynamic programming methodi,Bj) minimum problems, obtain stratum
The optimal sequence of comparison.Finally, the formation thicknesses of the corresponding two target wells of best match sequence is summed, obtain target well it
Between multidimensional property difference.
(2) during traditional geologic interpretation, standard well, cannot always according to domain expert's priori knowledge artificial selection
The spatial distribution and multiattribute correlation for considering well logging, are easy to produce biggish error and uncertainty, seriously affect geology
The following explanations of construction.In the methods of the invention, a kind of Discrete Choice Model is introduced, proposes the standard based on Discrete Choice Model
Well supervision selection Visualization Framework, it can be considered that the correlation of well logging spatial distribution and a variety of attributes: being based on user preference and effect
With theoretical building utility function is maximized, it is arranged towards multidimensional property data and calculates personal preference coefficient, using discrete selection
Model is associated with the distribution of the regional area of well logging by multidimensional property well logging matching result.
Standard well sample set is obtained according to expert's priori knowledge, is logged well using described in Discrete Choice Model maximization steps 1
Effectiveness between attribute similarity degree and standard well screen choosing traverses the Poisson that the adaptive blue noise sampling algorithm obtains
Disk according to the Discrete Choice Model screening criteria well, and obtains standard well filter information.
It is described to select it using well log attributes similarity degree and the standard well screen described in Discrete Choice Model maximization steps 1
Between effectiveness method, specifically:
1) according to the Poisson disk for meeting local space distribution, the standard of handmarking is obtained using expert's priori knowledge
Well sample set；
2) fixed utility function is designed according to the multidimensional property similarity between logging well in the Poisson disk, using multinomial recurrence
Modeling random error distribution, both simultaneous obtain total utility function, to calculate different user select different well loggings as
The probability of standard well；
3) according to the standard well sample set, the phase in the total utility function is calculated automatically using Discrete Choice Model
It answers parameter and traverses the Poisson disk, and then select the standard well that combinations of attributes in local space maximizes user utility.
Firstly, the uniform spatial distribution in order to preferably guarantee standard well, is regarded as one alternatively for every mouthful of well logging
Well logging collection in each Poisson disk is regarded as a relatively independent standard Jing Beixuanfanganji, invites domain expert's base by scheme
Manual identification's standard well is carried out towards part well log collection in itself priori knowledge, obtains the standard well sample set of actual selection.
Then, it is based on Discrete Choice Model, constructs utility function, simulation meter using user preference and maximization of utility theory
Calculating different user selects different log well as standard well to be formed by total utility, as shown in formula (4):
Uqi=Vqi+εqi (4)
Wherein, UqiIndicate selection alternative " i " to user " q " possible total utility, VqiIndicate alternative
Expected utility value, εqiRandom residual is indicated, for quantifying the deviation of average utility value and actual utility value.
Since log data is multidimensional property data, it is assumed that fixed utility function V is Multivariable Linear function, it is such as public
Shown in formula 5:
Vqi=βTxqi (5)
Wherein, β=[β1,β2,...,βn]TIndicate user to the corresponding preference coefficient of n data attribute.X=xqi1,
xqi2,...,xqinIt then indicates total similarity degree of the alternative on each attribute, can be calculated according to step 2.Then, it uses
Multinomial Logit mode simulates the Density Distribution of random residual ε, as shown in Equation 6:
And then formula (5), formula (6) are substituted into total utility formula (4), simulation, which is calculated, concentrates user " q " in given scheme
The probability of selection scheme " i " that is to say that the effectiveness of alternative " i " is greater than given scheme and concentrates other all alternative effectiveness
Probability, as shown in equation 7:
Finally, resulting standard well sample set is substituted into formula (7), user preference system can be calculated automatically by returning
Number " β ", and updated in total utility function, help user to concentrate selection combinations of attributes to maximize its effectiveness in target data
Standard well.As procedure described above, the effectiveness between well log attributes similarity and standard well is maximized using Discrete Choice Model,
Realize the standard well screen choosing of supervision.
(3) it according to the standard well filter information, guides user's interactive mode specified by visualization analysis technique or changes
Standard well, standard well sample set information described in adaptive updates optimize Discrete Choice Model parameter, and then screen and expired again
The standard well of sufficient user's current demand and experience.
It is described to guide user's interactive mode specified by visualization analysis technique or change the method for standard well, specifically:
1) a variety of visualization scheme collaborations are designed and show log data information: in system main view shown in Fig. 4, will be logged well
Real space position be mapped in map, retain the visualization result of its spatial distribution characteristic；According to different attribute well logging number
According to the same attribute data of well loggings different in local space is counted in specified attribute space, used by design attributes curve graph
Tone mapping technique distinguishes the standard well with non-standard well；Quantify the difference between the log using variance,
Statistics obtains total attribute difference histogram of local space internal standard well, the contrast of reserved property feature；
2) according to the visualization result, design standard well screen selects visualization scheme: counting each attribute of each local space
Upper characteristic matching matches incidence matrix as a result, obtaining multidimensional property layer position；Screening annulus is set according to layer position matching result, on ground
It is intuitively compared on figure and enhances the attribute difference in display local space between well logging；
3) according to the comparison of the attribute difference, user's interaction analysis and replacement standard well, and then more new samples, root are supported
Optimize Discrete Choice Model parameter according to new sample set, recalculates and screen the standard well for meeting user demand, detailed process
As shown in Figure 6.
Fig. 4 illustrates complete system interface figure.Wherein, Fig. 4 (a) is information view, shows basic data set information,
It the well logging sum concentrated such as data, the standard well sum currently filtered out, the standard well serial number currently chosen and current divides
Regional area number；Fig. 4 (b) is to sample panel to determine sampled well to adjust sample rate；Fig. 4 (c) is selection optimization panel,
The more new function of sample set is provided, each attribute bias coefficient of current user is calculated and be shown；Fig. 4 (d) is geographic view, well logging
It is considered as dot, is projected on map according to its latitude and longitude value, the space distribution situation of well logging is intuitively presented on map；Fig. 4
(e) it is multidimensional property view, the curvilinear motion of different well log attributes is described according to raw log data；Fig. 4 (f) is stratum matching
Schematic diagram intuitively shows the stratum match condition between any two well；Fig. 4 (g) is stratum incidence matrix view, part is presented
Stratum matching result in range between more wells, each matrix respectively represent a kind of well log attributes, and each row or column indicates selected
A bite well logging in subrange is indicating the similar journey of attribute between a pair of of well logging with color mapping techniques in matrix unit lattice
Degree, the cell on diagonal line from the upper left corner to the lower right corner then indicate in subrange between every mouthful of well logging and other well loggings
Total similarity degree；Fig. 4 (h) is the total attribute difference histogram of standard well, is obtained by calculating variance of each attribute on log
It arrives, standard well and the total attribute difference of non-standard well is shown using different tone mapping schemes respectively.
Fig. 5 illustrates the attribute evaluation result figure of selected standard well under given sample rate, can help user's evaluation herein
The precision of standard well selected by illustration method.Wherein, Fig. 5 (a) is after user determines sample rate according to self-demand, in active user
Projection result and original well logging spatial distribution under preference coefficient, using the obtained standard well set of formula (1), on map
Maintain preferable consistency, illustrate that the local space distribution for the well logging that standard well can satisfy selected by this paper inventive method is special
Sign；Fig. 5 (b) is the well log attributes drawing result of given local space internal standard well and non-standard well, due to complex geological condition
The offset of layer position is caused, the similar attribute logged well in geographical location is also difficult to be fitted very well；Fig. 5 (c) is in given local space
Standard well and difference performance results of the non-standard well on multidimensional well log attributes, each fan-shaped region indicate that a kind of well logging belongs to
Property, each layer of annulus indicate the non-standard well of a bite, by the distribution of annular histogram, can be enhanced display standard well with it is non-standard
Similarity degree between well；Fig. 5 (d) is given local space internal standard well and difference of the non-standard well in each attribute dimensions
Summation, the variance by calculating well log attributes data obtain, can be belonged to by histogram quick sensing standard well in different well loggings
Similarities in property；Fig. 5 (e) is the Discrete Choice Model meter using the design of the method for the present invention shown in formula (4)-(7)
Obtained user preference coefficient results are shown, the positive negativity of coefficient are shown using different color mapping scheme enhancings, quickly
Active user is understood to the preference profile of different attribute；Fig. 5 (f) is to utilize the method for the present invention shown in formula (2), formula (3)
The actual formation matching result that designed stratum Matching Model is calculated, in given subrange, using matrix diagram
Form shows the association between any two well, and the actual formation correlation degree between two wells is presented using tone mapping technique, can
Similarity degree of the well logging on multidimensional property in local space is determined to show in detail and intuitively.
Fig. 6, which is illustrated, updates user preference coefficient front and back, the multidimensional property comparison in difference figure of standard well and non-standard well,
It is to illustrate to update user preference coefficient front and back, the attribute difference situation of non-standard well in standard well and given local space,
And preference performance of the standard well on multidimensional well log attributes.By visual analysis system as shown in Figure 4, user is available
Standard well filter information as shown in Figure 5, so as to the selection of the interactive change standard well of variation according to self-demand, and
This concentration of standard well sample is included in as new samples.Then it by clicking the Update button in Selection Floater, corrects again
And new user preference coefficient is generated, to meet the variation of the real-time application demand of user.As shown in Fig. 6 (a), user was originally more
Attribute of interest COND, AC less focuses on attribute SP, later the extremely attribute of interest SP according to the variation of self-demand.In user
After interactive mode change sample set, if Fig. 6 (c) new preference coefficient being calculated is the Demand perference for meeting user, illustrate this hair
Bright method can effectively reflect that more new standard well screen selects result to user to the real-time requirement preference of well log attributes, and accordingly.Meanwhile
Fig. 6 (b) and Fig. 6 (d) illustrate variation feelings of the standard well selection result provided by the method for the present invention in the incidence matrix of stratum
Condition, it is not difficult to find that selected standard well is in the positive negativity of the preference of similarities and user on each attribute on each attribute
Show stronger consistency.Comprehensive observing Fig. 6, it can be seen that the result images that the method for the present invention obtains, it can be intuitive and fast
Variance analysis between the realization standard well and non-standard well of speed, by visual analysis technical support user to standard well sample set
Be updated or optimize the standard well set new with quick obtaining, efficiently meet the real-time change of user's application demand, effectively
The efficiency and precision of raising standard well selection；
Compared with traditional artificial selection criteria well visualization process, the sharpest edges of the method for the present invention are proposed based on discrete
The standard well screen of preference pattern selects visual analysis method, has comprehensively considered spatial distribution characteristic, the multidimensional property phase of log data
Closing property and the priori knowledge of domain expert realize the standard well for having supervision by constructing convenient and fast multifunctional visible analysis system
The efficiency and precision of the selection of standard well is effectively promoted in screening.
Claims (5)
1. a kind of standard well screen based on Discrete Choice Model selects visual analysis method, which is characterized in that specifically include following step
It is rapid:
(1) it is based on multidimensional log data, the spatial distribution characteristic for extracting well logging respectively is associated with pass with the stratum of multidimensional well log attributes
System: under the premise of the standard well screen of determination selects ratio, meet local space using the acquisition of adaptive blue noise sampling algorithm and be distributed
Poisson disk；Meanwhile the multiple dimensioned stratum Matching Model based on dynamic programming algorithm is designed, from the angular quantification degree of different attribute
Measure well log attributes similarity degree.
(2) standard well sample set is obtained according to expert's priori knowledge, is belonged to using logging well described in Discrete Choice Model maximization steps 1
Property similarity degree and standard well screen choosing between effectiveness, traverse the Poisson that the adaptive blue noise sampling algorithm obtains
Disk according to the Discrete Choice Model screening criteria well, and obtains standard well filter information.
(3) it according to the standard well filter information, guides user's interactive mode specified by visualization analysis technique or changes standard
Well, standard well sample set information described in adaptive updates optimize Discrete Choice Model parameter, screen again, and acquisition meets user
The standard well of current demand and experience.
2. standard well screen selects visual analysis method according to claim 1, which is characterized in that described using certainly in step (1)
It adapts to blue noise sampling algorithm and obtains the method for meeting the Poisson disk of local space distribution, specifically:
1) according to the spatial position of well logging, the Poisson disk for meeting local space distribution is obtained using blue noise sampling algorithm, each
Only allow to sample a well logging inside Poisson disk；
2) using the space distribution situation logged well in Poisson disk described in Density Estimator algorithm evaluation, the well logging is calculated
Spacial distribution density value, and then the size of the Poisson disk radius is adaptively updated, keep the spatial distribution of the well logging special
Sign.
3. standard well screen selects visual analysis method according to claim 1, which is characterized in that in step (1), the design base
In the method for the multiple dimensioned stratum Matching Model of dynamic programming algorithm, specifically:
1) more described in Nonlinear Processing using each well log attributes curve of median filter smoothness of image according to the multidimensional log data
Dimension log data simultaneously normalizes between 0 and 1 according to discrete results；
2) according to the smooth log, the identification and division on stratigraphic horizon boundary are carried out to well logging using activity function；
3) the optimal layer position between two wells is solved by dynamic programming algorithm and matches sequence, by best match sequence on each well log attributes
It arranges corresponding formation thickness and sums to obtain quantified property difference between two wells.
4. standard well screen selects visual analysis method according to claim 1, which is characterized in that in step (2), it is described using from
Dissipate the method for the effectiveness between well log attributes similarity degree described in preference pattern maximization steps 1 and standard well screen choosing, tool
Body are as follows:
1) according to the Poisson disk for meeting local space distribution, the standard well sample of handmarking is obtained using expert's priori knowledge
This collection；
2) fixed utility function is designed according to the multidimensional property similarity between logging well in the Poisson disk, using multinomial regression model
Random error distribution is simulated, both simultaneous obtain total utility function, select different well loggings as standard to calculate different user
The probability of well；
3) according to the standard well sample set, the corresponding ginseng in the total utility function is calculated automatically using Discrete Choice Model
The Poisson disk that the adaptive blue noise sampling algorithm obtains is counted and traverses, and then it is maximum to select combinations of attributes in local space
Change the standard well of user utility.
5. standard well screen selects visual analysis method according to claim 1, which is characterized in that passing through described in step (3) can
The method of standard well is specified or changes depending on changing analytical technology guidance user's interactive mode, specifically:
1) a variety of visualization scheme collaborations are designed and show log data information: the real space position of well logging is mapped to map
In, retain the visualization result of its spatial distribution characteristic；According to different attribute log data, design attributes curve graph will be local
The same attribute datas of different well loggings count in specified attribute space in space, using tone mapping technique by the standard well
It is distinguished with non-standard well；Quantify the difference between the log using variance, statistics obtains local space internal standard well
Total attribute difference histogram, the contrast of reserved property feature.
2) according to the visualization result, design standard well screen selects visualization scheme, counts special on each attribute of each local space
Matching result is levied, multidimensional property layer position is obtained and matches incidence matrix；Screening annulus is set according to layer position matching result, on map
It intuitively compares and enhances the attribute difference in display local space between well logging.
3) according to the comparison of the attribute difference, user's interaction analysis and replacement standard well, and then more new samples are supported, according to new
Sample set optimize Discrete Choice Model parameter, recalculate and screen the standard well for meeting user demand.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910758272.0A CN110502569A (en) | 2019-08-16 | 2019-08-16 | A kind of standard well screen based on Discrete Choice Model selects visual analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910758272.0A CN110502569A (en) | 2019-08-16 | 2019-08-16 | A kind of standard well screen based on Discrete Choice Model selects visual analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110502569A true CN110502569A (en) | 2019-11-26 |
Family
ID=68588141
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910758272.0A Pending CN110502569A (en) | 2019-08-16 | 2019-08-16 | A kind of standard well screen based on Discrete Choice Model selects visual analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110502569A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017062009A1 (en) * | 2015-10-08 | 2017-04-13 | Halliburton Energy Services Inc. | Mud pulse telemetry preamble for sequence detection and channel estimation |
| CN106951581A (en) * | 2017-01-24 | 2017-07-14 | 同济大学 | Commercial complex simulator |
| CN108562950A (en) * | 2017-12-11 | 2018-09-21 | 中国石油天然气集团公司 | Method for intelligently dividing stratum horizon based on logging information |
| CN109740427A (en) * | 2018-11-26 | 2019-05-10 | 浙江财经大学 | Standard well screen based on blue noise sampling selects visual analysis method |
| CN109763814A (en) * | 2019-01-09 | 2019-05-17 | 浙江财经大学 | Stratum matching visual analysis method based on multi-dimensional logging data |
-
2019
- 2019-08-16 CN CN201910758272.0A patent/CN110502569A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017062009A1 (en) * | 2015-10-08 | 2017-04-13 | Halliburton Energy Services Inc. | Mud pulse telemetry preamble for sequence detection and channel estimation |
| CN106951581A (en) * | 2017-01-24 | 2017-07-14 | 同济大学 | Commercial complex simulator |
| CN108562950A (en) * | 2017-12-11 | 2018-09-21 | 中国石油天然气集团公司 | Method for intelligently dividing stratum horizon based on logging information |
| CN109740427A (en) * | 2018-11-26 | 2019-05-10 | 浙江财经大学 | Standard well screen based on blue noise sampling selects visual analysis method |
| CN109763814A (en) * | 2019-01-09 | 2019-05-17 | 浙江财经大学 | Stratum matching visual analysis method based on multi-dimensional logging data |
Non-Patent Citations (1)
| Title |
|---|
| 周志光 等: "基于蓝噪声采样的多维标准井筛选可视分析", 《计算机辅助设计与图形学学报》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11346970B2 (en) | Automatic quality control of seismic travel time | |
| Paasche et al. | Cooperative inversion of 2D geophysical data sets: A zonal approach based on fuzzy c-means cluster analysis | |
| CN101932954B (en) | Subsurface prediction method and system | |
| US20210319304A1 (en) | Identifying hydrocarbon reserves of a subterranean region using a reservoir earth model that models characteristics of the region | |
| US5724309A (en) | Method for geophysical processing and interpretation using instantaneous phase and its derivatives and their derivatives | |
| US6317384B1 (en) | Method for geophysical processing and interpretation using seismic trace difference for analysis and display | |
| US11486230B2 (en) | Allocating resources for implementing a well-planning process | |
| WO2020123084A1 (en) | Machine learning-augmented geophysical inversion | |
| CN111596978A (en) | Web page display method, module and system for lithofacies classification by artificial intelligence | |
| CN109763814B (en) | Stratum matching visual analysis method based on multi-dimensional logging data | |
| Wang et al. | Machine learning-based seismic spectral attribute analysis to delineate a tight-sand reservoir in the Sulige gas field of central Ordos Basin, western China | |
| US12060781B2 (en) | Method for generating initial models for least squares migration using deep neural networks | |
| US11221425B1 (en) | Generating a model for seismic velocities in a subsurface region using inversion with lateral variations | |
| US20220237891A1 (en) | Method and system for image-based reservoir property estimation using machine learning | |
| CN109522578A (en) | Complex Lithofacies prediction technique and system based on Bayes's fuzzy discrimination | |
| Wu et al. | Incremental correlation of multiple well logs following geologically optimal neighbors | |
| CN108663714A (en) | A kind of sedimentary micro depicting method | |
| Hussein et al. | Unsupervised machine learning using 3D seismic data applied to reservoir evaluation and rock type identification | |
| US20200271810A1 (en) | Hybrid Residual Moveout Error Estimation | |
| CN109283577A (en) | A kind of seismic layer labeling method | |
| CN110502569A (en) | A kind of standard well screen based on Discrete Choice Model selects visual analysis method | |
| CN109031423A (en) | Pre-stack seismic texture analysis method based on gradient co-occurrence matrix | |
| Nguyen et al. | Automated fault surfaces extraction from 3D fault imaging volume | |
| Hou et al. | Uncertainty analysis and visualization of geological subsurface and its application in metro station construction | |
| Lomask et al. | New tools for seismic stratigraphic interpretation: stratal convergence and instantaneous isochron attribute cubes derived from volumetric flattening of experimental strata |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191126 |