CN105425293B - Seismic properties clustering method and device - Google Patents
Seismic properties clustering method and device Download PDFInfo
- Publication number
- CN105425293B CN105425293B CN201510809921.7A CN201510809921A CN105425293B CN 105425293 B CN105425293 B CN 105425293B CN 201510809921 A CN201510809921 A CN 201510809921A CN 105425293 B CN105425293 B CN 105425293B
- Authority
- CN
- China
- Prior art keywords
- layer
- window
- sample point
- cluster
- time slip
- 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.)
- Active
Links
Landscapes
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses a kind of seismic properties clustering methods and device, this method to include：Determine the size of wave band cluster mode, clustering parameter, initial layers position, layer position offset and time slip-window, time slip-window be interval to be studied oscillogram in sample point number contained between two adjacent peaks or two adjacent troughs, the more than initial layers position layer digit of at least half time slip-window, the layer digit of at least half time slip-window below stop layer position；The seismic attributes data of interval to be studied is clustered using time slip-window, determines each layer position sample point generic；Unification is carried out to the cluster result of adjacent layer position according to space similarity, since next layer of position of initial layers position, the cluster result of each layer of position is adjusted according to the cluster result of adjacent last layer position.More hierarchical cluster attribute analyses are carried out using the band class information up and down of the point on layer plane, and unification is carried out to cluster result, it can be from the reservoir distribution in global angle represent layer section.
Description
Technical field
The present invention relates to technical field of geophysical exploration more particularly to a kind of seismic properties clustering methods and device.
Background technology
It, could be to exploration area after only having sufficient understanding to the geological condition of underground and be familiar in oil-gas exploration
The hydrocarbon storage situation in domain judges.An important means for obtaining geological information is exactly to analyze seismic data through mathematic(al) manipulation
The various seismic attributes datas obtained afterwards.Seismic attributes data is typically prestack or post-stack seismic data, in relation to the several of seismic wave
The parameters such as what form, kinematics character, dynamic characteristic.By the research to these parameters, survey area underground can be obtained
The features such as structure, lithology, the fluid of medium, and then infer the storage information of oil gas.Pass through from the seismic attributes data of acquisition
It is commonly referred to as seismic attributes analysis, a kind of side of most common of which that a series of analyses, which are inferred to this process of subterranean geology,
Method is exactly to cluster.
Since the 1980s, clustering method is gradually introduced in field of seismic exploration, especially more in earthquake
In terms of attributive analysis and seismic facies analysis, the reservoir in the case of greatly improving the predictive ability of subtle pool and being controlled without well
Predictive ability.
So-called cluster is exactly to be divided them into according to the size of difference between the seismic attributes data obtained at underground medium
Several classifications, it is larger per the data differences for differing smaller and different classes of between the data in one kind.Pass through the earthquake category to collection
Property is clustered, these seismic attributes datas can be divided into several big classifications, and then can be to the geology feelings of survey area
Condition is further analyzed.Such as geology Lithofacies dividing is carried out to target area, according to cluster result and result of log interpretation
Check analysis, to determine the facies tract corresponding to each classification.Especially during reservoir prediction, cluster analysis of seismic attributes
Be very important step, plays important effect.
Clustering technique has focused largely on designated layer plane at present, and reservoir is then integrated distribution in interval, is based on layer
The clustering method of plane cannot show the reservoir distribution in interval well.
Invention content
The present invention provides a kind of seismic properties clustering method and devices, existing based on layer plane at least to solve
Clustering technique cannot show the problem of reservoir distribution in interval well.
According to an aspect of the invention, there is provided a kind of seismic properties clustering method, including：Determine wave band cluster side
Formula, clustering parameter, initial layers position, layer position offset and time slip-window size, wherein the time slip-window is to wait grinding
Study carefully sample contained between sample point number contained between two adjacent peaks in the oscillogram of interval or two adjacent troughs
Point number, the layer digit of more than the initial layers position at least half time slip-window, at least half cunning below stop layer position
The layer digit of dynamic time window；The seismic attributes data of the interval to be studied is clustered using the time slip-window, really
Classification in the fixed interval to be studied belonging to the sample point of each layer of position；According to space similarity to the cluster knot of adjacent layer position
Fruit carry out unification, wherein since next layer of position of the initial layers position, by the cluster result of each layer of position according to its phase
The cluster result of an adjacent upper layer position is adjusted.
In one embodiment, the calculation formula of the time slip-window is：Wherein, Δ T tables
Show time slip-window, tstepIndicate sampling step length, t0Sample point described in wave band middle-range on expression layer position where sample point is nearest
Sampling time interval between trough or wave crest and the sample point.
In one embodiment, the wave band cluster mode includes：Wave band based on mean value is clustered mode, is spelled based on attribute
The wave band cluster mode of conjunction or the wave band based on difference sum cluster mode.
In one embodiment, if clustering mode using the wave band based on mean value, the time slip-window is utilized
The seismic attributes data of the interval to be studied is clustered, including：Wave where the sample point e on current layer position
Section, the attribute vector of the sample point e is determined according to following formula：
Wherein, the primitive attribute vector of the sample point e is (a1e, a2e... aie... ane), aiAjIndicate the sample point e's
J-th point of ith attribute value in point set A in time slip-window Δ T,Expression time slip-window is Δ
The average property value of the ith attribute of the sample point e, i=1,2 ..., n in the case of T；Work as to described according to above-mentioned formula
Other sample points on front layer position are calculated, and the attribute vector of each sample point on the current layer position is obtained；Successively along institute
Layer position where each sample point in point set A is stated, the attribute vector of each sample point is clustered.
In one embodiment, if mode is clustered using the wave band based on attribute split, when using the sliding
Between window the seismic attributes data of the interval to be studied is clustered, including：
By all properties of each sample point in the point set A in the time slip-window Δ T of the sample point e on current layer position
Split is carried out, the wave band attribute vector of the sample point e is obtained：(the full attributes of A1, the full attributes ... of A2, the full attributes ... of Aj),
In, j=1,2 ..., the full attribute of sample point is (a1, a2 ... ai ... an) in Δ T, the point set A；According to above-mentioned steps
Other sample points on the current layer position are calculated, obtain the wave band attribute of each sample point on the current layer position to
Amount；The wave band attribute vector of all sample points on the current layer position is clustered.
In one embodiment, if clustering mode using the wave band based on difference sum, the sliding time is utilized
Window clusters the seismic attributes data of the interval to be studied, including：It is calculated separately on current layer position according to following formula
Sample point e attribute vector and pre-stored multiple cluster centre distance D：Its
In, aiAjIndicate j-th point in the point set A in the time slip-window Δ T of sample point e of ith attribute value, aiCjIndicate poly-
J-th point of ith attribute value in point set C in the time slip-window Δ T at class center, the cluster centre and classification are one by one
It is corresponding；Determine the classification apart from the corresponding classification of nearest cluster centre for the sample point e with the sample point e.
In one embodiment, included uniformly to the cluster result of adjacent layer position according to space similarity：By described 5
Each sample of initial layers position is pressed classifies according to generic, and presses preset order into row label；Step A1, according to adjacent layer
The space similarity of position, under position is adjacent layer by layer layer by layer in position, determined with upper with it is described on position layer by layer corresponding region；Step
A2, according to the label of position layer by layer on described, will be described under the sample point in the region of position layer by layer label be adjusted to described on layer by layer
The label of corresponding region in position；Step A1 to A2 is repeated, the cluster of all layers of position until completing the interval to be studied
As a result it adjusts.
In one embodiment, the seismic attributes data of the interval to be studied is being carried out using the time slip-window
Before cluster, the method further includes：Following pretreatment is carried out to the seismic attributes data of the interval to be studied：To described
Abnormal data in shake attribute data is filtered processing；And filtered seismic attributes data is normalized.
According to another aspect of the present invention, a kind of seismic properties clustering apparatus is provided, including：Determination unit, with 5 in
Determine the size of wave band cluster mode, clustering parameter, initial layers position, layer position offset and time slip-window, wherein the cunning
Dynamic time window be interval to be studied oscillogram in sample point number or two adjacent troughs contained between two adjacent peaks
Between contained sample point number, the layer digit of more than the initial layers position at least half time slip-window, stop layer position with
Under at least half time slip-window layer digit；Cluster cell, for utilizing the time slip-window to the layer to be studied
The seismic attributes data of section is clustered, and determines the classification belonging to the sample point of each layer of position in the interval to be studied；Adjustment
Unit, for carrying out unification to the cluster result of adjacent layer position according to space similarity, wherein from the next of the initial layers position
A layer of position starts, and the cluster result of each layer of position is adjusted according to the cluster result of a layer position thereon.
Seismic properties clustering method through the invention and device are started with from the band class information up and down of the point on layer plane,
More hierarchical cluster attribute analyses are carried out using band class information, the earthquake volume data of specific interval is clustered by time slip-window,
And unification is carried out to the cluster result of adjacent layer position using space fit degree algorithm, it can be preferably from global angle represent layer section
Reservoir distribution.
Description of the drawings
Attached drawing described herein is used to provide further understanding of the present invention, and is constituted part of this application, this hair
Bright illustrative embodiments and their description do not constitute limitation of the invention for explaining the present invention.In the accompanying drawings：
Fig. 1 is the flow chart of the seismic properties clustering method of the embodiment of the present invention；
Fig. 2 is the structure diagram of the seismic properties clustering apparatus of the embodiment of the present invention；
Fig. 3 is the layer model figure of the embodiment of the present invention；
Fig. 4 is the schematic diagram of the wave band cluster mode of the embodiment of the present invention；
Fig. 5 is the layer position cluster centre vertical view of the embodiment of the present invention；
Fig. 6 A are the label distribution schematic diagrams of the upper position layer by layer of the embodiment of the present invention；
Fig. 6 B are the label distribution schematic diagrams of position layer by layer under the embodiment of the present invention；
Fig. 6 C are the label unified result schematic diagrames of position layer by layer under the embodiment of the present invention；
Fig. 7 is that the body cluster result of the embodiment of the present invention unifies schematic diagram；
Fig. 8 is the implementing procedure figure of the specific example of the seismic properties clustering method of the embodiment of the present invention.
Specific implementation mode
With reference to the attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete
Ground describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based on this
The embodiment of invention, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, belongs to protection scope of the present invention.
Most of researchs concentrate on layer plane in the more hierarchical cluster attributes put analysis at present, and any physical property of sedimentary formation is joined
Several variations are always reflected in the variation of seismic channel waveform shape, and the classification of seismic channel waveform shape represents the true of seismic signal
Real is laterally abnormal.In certain window scope, the wave character of same target zone should be consistent.For seismic properties number
According to classification, numerical values recited is not only considered, it is also contemplated that the variation of waveform.If only considering numerical values recited, earthquake is not accounted for
The waveform of signal, obtained classification results are undesirable.
An embodiment of the present invention provides a kind of seismic properties clustering method, Fig. 1 is that the seismic properties of the embodiment of the present invention are poly-
The flow chart of class method.As shown in Figure 1, this method comprises the following steps：
Step S101 determines wave band cluster mode, clustering parameter, initial layers position, layer position offset and time slip-window
Size, wherein time slip-window be interval to be studied oscillogram in sample point number contained between two adjacent peaks
Or contained sample point number between two adjacent troughs, the layer digit of more than initial layers position at least half time slip-window,
The layer digit of at least half time slip-window below stop layer position.
In specific implementation, the size of time slip-window can be determined according to the wave character of specific interval.Clustering parameter can
To include：Classification number.Preferably, the classification number for clustering setting can be 5 or 7.
Step S102 clusters the seismic attributes data of interval to be studied using time slip-window, determines to be studied
Classification in interval belonging to the sample point of each layer of position.
Step S103 carries out unification, wherein from initial layers position according to space similarity to the cluster result of adjacent layer position
Next layer of position starts, and the cluster result of each layer of position is adjusted according to the cluster result of a upper layer position adjacent thereto
It is whole.
Above-described embodiment through the invention is started with from the band class information up and down of the point on layer plane, using band class information into
The more hierarchical cluster attribute analyses of row, cluster the earthquake volume data of specific interval by time slip-window, and utilize space fit
Degree algorithm carries out unification to the cluster result of adjacent layer position, can be preferably from the reservoir distribution in global angle represent layer section.
In actual cluster process, preset classification number, time can be inputted at the interface of cluster calculation software
The seismic volume attributes data for being included between window size, designated layer position to be clustered and its offset, then output is in the interval
Every layer of sample point generic.Since there may be differences for adjacent layer position cluster classification in interval, so needing to clustering interval
Cluster result carries out unification.
The calculation formula of time slip-window can be：Wherein, Δ T indicates time slip-window, tstep
Indicate sampling step length, t0Wave band middle-range sample point on expression layer position where sample point nearest trough or wave crest and sample point it
Between sampling time interval.
Wave band clusters mode：Wave band cluster mode based on mean value, the wave band cluster side based on attribute split
Formula or wave band based on difference sum cluster mode.
Above-mentioned several wave band cluster modes are illustrated separately below.
(1) wave band based on mean value clusters mode
The seismic attributes data of interval to be studied is clustered using time slip-window, is included the following steps：
According to the wave band where the sample point e on current layer position, the attribute vector of sample point e is determined according to following formula：
Wherein, the primitive attribute vector of sample point e is (a1e, a2e... aie... ane), aiAjIndicate the sliding of sample point e
J-th point of ith attribute value in point set A in time window Δ T,Indicate that time slip-window is the feelings of Δ T
The average property value of the ith attribute of sample point e, i=1,2 ..., n under condition；
Other sample points on current layer position are calculated according to above-mentioned formula, obtain each sample point on current layer position
Attribute vector；
Successively along layer position where each sample point in point set A, the attribute vector of each sample point is clustered.
Clustering method in the prior art may be used in specific cluster process, for example, k-means clusters etc..Herein not
It repeats again.
(2) wave band based on attribute split clusters mode
The seismic attributes data of interval to be studied is clustered using time slip-window, is included the following steps：
By all properties of each sample point in the point set A in the time slip-window Δ T of the sample point e on current layer position
Split is carried out, the wave band attribute vector of sample point e is obtained：(the full attributes of A1, the full attributes ... of A2, the full attributes ... of Aj), wherein j
The full attribute of sample point is (a1, a2 ... ai ... an) in=1,2 ..., Δ T, point set A；
Other sample points on current layer position are calculated according to above-mentioned steps, obtain each sample point on current layer position
Wave band attribute vector；
The wave band attribute vector of all sample points on current layer position is clustered.Specific cluster process may be used
Clustering method in the prior art, for example, k-means clusters etc..Details are not described herein again.
(3) wave band based on difference sum clusters mode
The seismic attributes data of interval to be studied is clustered using time slip-window, is included the following steps：
The attribute vector of the sample point e on current layer position and pre-stored multiple clusters are calculated separately according to following formula
The distance D at center：Wherein, aiAjIndicate the point set in the time slip-window Δ T of sample point e
Close j-th point in A of ith attribute value, aiCjJ-th point is indicated in the point set C in the time slip-window Δ T of cluster centre
Ith attribute value, cluster centre and classification correspond；
Determine the classification apart from the corresponding classification of nearest cluster centre for sample point e with sample point e.
In one embodiment, unifying to the cluster result of adjacent layer position according to space similarity can be by following
Step is realized：Each sample of initial layers position is pressed and is classified according to generic, and presses preset order into row label；Step
A1, according to the space similarity of adjacent layer position, under position is adjacent layer by layer layer by layer in position, determination is corresponding with upper position layer by layer with upper
Region；Step A2, according to the label of upper position layer by layer, the label of the sample point in the region of position is adjusted in position layer by layer layer by layer by under
The label of corresponding region；Step A1 to A2 is repeated, the cluster result adjustment of all layers of position until completing interval to be studied.
In one embodiment, the seismic attributes data of interval to be studied is being carried out clustering it using time slip-window
Before, the above method can also include：Following pretreatment is carried out to the seismic attributes data of interval to be studied：To seismic attributes data
In abnormal data be filtered processing；And filtered seismic attributes data is normalized.It specifically, can be with
It is normalized using Min-max method or Z-Score methods.
Based on same inventive concept, a kind of seismic properties clustering apparatus is additionally provided in the embodiment of the present invention, can be used for
Method described in above-described embodiment is realized, as described in the following examples.It is solved the problems, such as due to seismic properties clustering apparatus
Principle is similar to seismic properties clustering method, therefore the implementation of the device may refer to the implementation of the above method, repeats place not
It repeats again.Used below, the software of predetermined function and/or the group of hardware may be implemented in term " unit " or " module "
It closes.Although system described in following embodiment is preferably realized with software, the combination of hardware or software and hardware
Realization be also that may and be contemplated.
Fig. 2 is the structure diagram of the seismic properties clustering apparatus of the embodiment of the present invention, as shown in Fig. 2, the seismic properties are poly-
Class device includes：Determination unit 21, cluster cell 22 and adjustment unit 23, are below specifically described the structure.
Determination unit 21, for determining wave band cluster mode, clustering parameter, initial layers position, layer position offset and sliding
The size of time window, wherein time slip-window be interval to be studied oscillogram in sample contained between two adjacent peaks
Contained sample point number between point number or two adjacent troughs, the layer of more than initial layers position at least half time slip-window
Digit, the layer digit of at least half time slip-window below stop layer position；
Cluster cell 22 is determined for being clustered to the seismic attributes data of interval to be studied using time slip-window
Classification in interval to be studied belonging to the sample point of each layer of position；
Adjustment unit 23, for carrying out unification to the cluster result of adjacent layer position according to space similarity, wherein from starting
Next layer of position of layer position starts, and the cluster result of each layer of position is adjusted according to the cluster result of a layer position thereon.
By above-mentioned apparatus, start with from the band class information up and down of the point on layer plane, more attributes are carried out using band class information
Clustering clusters the earthquake volume data of specific interval by time slip-window, and utilizes space fit degree algorithm pair
The cluster result of adjacent layer position carries out unification, can be preferably from the reservoir distribution in global angle represent layer section.
The calculation formula of time slip-window can be：Wherein, Δ T indicates time slip-window, tstep
Indicate sampling step length, t0Wave band middle-range sample point on expression layer position where sample point nearest trough or wave crest and sample point it
Between sampling time interval.
Wave band clusters mode：Wave band cluster mode based on mean value, the wave band cluster side based on attribute split
Formula or wave band based on difference sum cluster mode.
Above-mentioned several wave band cluster modes are illustrated separately below.
(1) wave band based on mean value clusters mode
Cluster cell 22 may include：
First determining module, for according to the wave band where the sample point e on current layer position, sample to be determined according to following formula
The attribute vector of this e：
Wherein, the primitive attribute vector of sample point e is (a1e, a2e... aie... ane), aiAjIndicate the sliding of sample point e
J-th point of ith attribute value in point set A in time window Δ T,Indicate that time slip-window is Δ T's
In the case of sample point e ith attribute average property value, i=1,2 ..., n；
First computing module is worked as being calculated other sample points on current layer position according to above-mentioned formula
The attribute vector of each sample point on front layer position；
First cluster module, for successively along layer position where each sample point in point set A, to the attribute of each sample point to
Amount is clustered.
(2) wave band based on attribute split clusters mode
Cluster cell 22 may include：
Die section is used for the various kinds in the point set A in the time slip-window Δ T of the sample point e on current layer position
The all properties of this point carry out split, obtain the wave band attribute vector of sample point e：(the full attributes of A1, the full attributes ... of A2, Aj belong to entirely
Property ...), wherein j=1,2 ..., Δ T, the full attribute of sample point is (a1, a2 ... ai ... an) in point set A；
Second computing module is worked as being calculated other sample points on current layer position according to above-mentioned steps
The wave band attribute vector of each sample point on front layer position；
Second cluster module is clustered for the wave band attribute vector to all sample points on current layer position.
(3) wave band based on difference sum clusters mode
Cluster cell 22 may include：
Third computing module, for calculated separately according to following formula the attribute vector of the sample point e on current layer position with
The distance D of pre-stored multiple cluster centres：Wherein, aiAjIndicate the cunning of sample point e
J-th point of ith attribute value, ai in point set A in dynamic time window Δ TCjIn the time slip-window Δ T for indicating cluster centre
Point set C in j-th point of ith attribute value, cluster centre corresponds with classification；
Second determining module, for determine with sample point e apart from the corresponding classification of nearest cluster centre for sample point e's
Classification.
In one embodiment, adjustment unit 23 includes：Label model, for pressing each sample of initial layers position according to institute
Belong to classification to classify, and presses preset order into row label；Area determination module, for similar according to the space of adjacent layer position
Degree under position is adjacent layer by layer layer by layer in position, is determining the corresponding region with upper position layer by layer with upper；Adjust module, for according to
The label of position layer by layer, the label of the sample point in the region of position is adjusted to the label of corresponding region in position layer by layer layer by layer by under.Weight
Area determination module and adjustment module is utilized to complete respective function again, the cluster of all layers of position until completing interval to be studied
As a result it adjusts.
In one embodiment, above-mentioned apparatus can also include：Pretreatment unit, for being treated using time slip-window
Before the seismic attributes data of research interval is clustered, following pretreatment is carried out to the seismic attributes data of interval to be studied：
Processing is filtered to the abnormal data in seismic attributes data；And place is normalized to filtered seismic attributes data
Reason.
Certainly, above-mentioned module divides a kind of only signal and divides, and the present invention is not limited thereto.As long as the present invention can be realized
Purpose module divide, be within the scope of protection of the invention.
In order to carry out apparent explanation to above-mentioned seismic properties clustering apparatus and device, implement with reference to specific
Example illustrates, however, it should be noted that the embodiment merely to the present invention is better described, is not constituted to this hair
It is bright improperly to limit.
By taking K-means clustering algorithms as an example, assumed as follows：
N kind seismic properties are selected, attribute-name is respectively a1, a2 ..., ai ..., wherein i=1,2 ..., n.
As shown in Figure 3, it is assumed that the sampling step length in the directions Time Slice is tstep, chooses layer bit line h1, and h1 increases downwards
Δ t obtains a layer bit line h2.Point e is taken on h1, body cluster is to carry out wave band cluster downwards along layer position, introduces sliding window here
Δ T, it is equal to the number of contained sampled point in wave band, and such as e points, Δ T is determined by the closest troughs of e or more, is based on e points here
Upper and lower waveform is symmetrical, if time of the lower wave trough away from e is different on e, takes smaller value, if this smaller value is t0.In the wave band of e points
T sampled point of shared Δ is arranged them by Time Slice ascending orders, composition set A.The calculation formula of Δ T：
Wherein：T0 indicates that e points or more sampling time interval of the closest trough away from e points, tstep indicate sampling step length.ΔT
Indicate sliding window.
Wave band cluster is the wave band cluster based on mean value, the wave based on attribute split respectively generally there are three types of processing mode
Section cluster, the wave band cluster based on difference sum, as shown in Figure 4.
(1) the wave band cluster based on mean value：
The primitive attribute vector of e：(a1e, a2e... aie... ane).The wave band for considering e points, takes attribute mean value as e's
Attribute vector：
Wherein：aiAjIndicate j-th point in the point set A in the sliding window Δ T of e points of ith attribute value,It indicates when sliding window is Δ T, the average property value of the ith attribute of e points, i=1,2 ..., n.
To other points carry out above-mentioned processing on layer position where e points, gathered successively along layer position where sampled point in set A
Class.
(2) the wave band cluster based on attribute split：
By all properties split of the sampled point in the corresponding set A of e points, wave band attribute vector is formed；
The primitive attribute vector of e：(a1e, a2e... aie... ane)。
The wave band attribute vector of e：(the full attributes of A1, the full attributes ... of A2, the full attributes ... of Aj), wherein j=1,2 ..., Δ T.
(3) the wave band cluster based on difference sum
It is clustered similar to K-means, first generates pre-set k cluster centre in layer plane, and it is different apart from calculating
It is clustered in common K-means.If C, which is layer plane h1, initializes the cluster centre generated when cluster centre (referring to Fig. 5)：
The distance of e to C calculates as follows：
Wherein：aiAjIndicate j-th point in the point set A in the sliding window Δ T of e points of ith attribute value, aiCjTable
Show that j-th point in the point set C in the sliding window Δ T of the cluster centre where e points of ith attribute value, D are apart from degree
Amount judges that the classification of e belongs to according to D.
Cluster labels be uniformly the cluster result based on adjacent layer plane similar to this it is assumed that corresponding to similarity using space
To unify the label of adjacent layer plane.Classify according to generic for example, each sample of initial layers position can be pressed, and presses
Descending label again, is denoted as 1,2 ..., according to upper label sequence, finds out down mark most in area of space in position layer by layer respectively
All labels of lower layer, are converted to the label in the upper layer region by label, then by corresponding upper label in lower layer be converted to by
The label that upper label is replaced has executed all layers of position, completes adjustment.
Referring to Fig. 6 A to 6C, Fig. 6 A show a layer cluster result of plane H, and solid line is layer plane H1 cluster knots in Fig. 6 B
Fruit, dotted line are the cluster result (it is assumed that layer plane H1 is adjacent with H, H1 is located at the lower section of H) of layer plane H, and Fig. 6 C show application
After cluster labels Unified Algorithm, the update result of layer plane H1 cluster labels.
Specifically, first, seismic volume attribute data is pre-processed, selects Min-max method to carry out in the present embodiment
Pretreatment.Then, setting program interface input parameter determines initial layers position (Top Horizon), layer position offset (Add
Time Slice), clustering algorithm parameter, half time window (Half Time).The present embodiment medium wave band clusters mode to be based on attribute
For the mode of split.After having executed body clustering algorithm, the unification of different layers position cluster result is carried out.Under initial layers position
One layer position starts, and each layer position adjusts its cluster result according to the cluster result of a layer position thereon.Fig. 7 is of the invention real
The body cluster result for applying example unifies schematic diagram, and in Fig. 7, classification number is 0~7.Fig. 8 is the seismic properties cluster of the embodiment of the present invention
The implementing procedure figure of the specific example of method, i.e. input parameter, are clustered, and cluster result (label) is unified, and storage is after reunification
Label, then visualized, user facilitated to check.
In conclusion the embodiment of the present invention is started with from the band class information up and down of the point on layer plane, using band class information into
The more hierarchical cluster attribute analyses of row, cluster the earthquake volume data of particular segment by time slip-window, improve Clustering Effect.It utilizes
Space fit degree algorithm carries out unification to the cluster result of adjacent layer position, can be preferably from the reservoir in global angle represent layer section
Distribution, this was applied in the past not available for the clustering algorithm of petroleum exploration domain.Practical application shows what this method obtained
Clustering Effect is substantially better than more hierarchical cluster attribute results a little.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes
It is one or more for realizing specific logical function or process the step of executable instruction code module, segment or portion
Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discuss suitable
Sequence, include according to involved function by it is basic simultaneously in the way of or in the opposite order, to execute function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
It should be appreciated that each section of the present invention can be realized with hardware, software, firmware or combination thereof.Above-mentioned
In embodiment, software that multiple steps or method can in memory and by suitable instruction execution system be executed with storage
Or firmware is realized.It, and in another embodiment, can be under well known in the art for example, if realized with hardware
Any one of row technology or their combination are realized：With the logic gates for realizing logic function to data-signal
Discrete logic, with suitable combinational logic gate circuit application-specific integrated circuit, programmable gate array (PGA), scene
Programmable gate array (FPGA) etc..
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
Centainly refer to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be any
One or more embodiments or example in can be combined in any suitable manner.
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical solution and advantageous effect
Describe in detail it is bright, it should be understood that the above is only a specific embodiment of the present invention, the guarantor being not intended to limit the present invention
Range is protected, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should be included in this
Within the protection domain of invention.
Claims (6)
1. a kind of seismic properties clustering method, which is characterized in that including：
Determine the size of wave band cluster mode, clustering parameter, initial layers position, layer position offset and time slip-window, wherein institute
It states sample point number contained between two adjacent peaks in the oscillogram that time slip-window is interval to be studied or two adjacent
Contained sample point number between trough, the layer digit of more than the initial layers position at least half time slip-window, stop layer
Or less at least half time slip-window layer digit；The calculation formula of the time slip-window is：
Wherein, Δ T indicates time slip-window, tstepIndicate sampling step length, t0Sample described in wave band middle-range on expression layer position where sample point
This puts the sampling time interval between nearest trough or wave crest and the sample point；The wave band clusters mode：It is based on
Wave band cluster mode, the wave band cluster mode based on attribute split or the wave band based on difference sum of mean value cluster mode；
The seismic attributes data of the interval to be studied is clustered using the time slip-window, determines the layer to be studied
Classification in section belonging to the sample point of each layer of position；
Unification is carried out to the cluster result of adjacent layer position according to space similarity, wherein next layer from the initial layers position
Position starts, and the cluster result of each layer of position is adjusted according to the cluster result of a upper layer position adjacent thereto；It is described to press
Included uniformly to the cluster result of adjacent layer position according to space similarity：
Each sample of the initial layers position is pressed and is classified according to generic, and presses preset order into row label；
Step A1, according to the space similarity of adjacent layer position, under position is adjacent layer by layer layer by layer in position, determined with upper with it is described on
The corresponding region of position layer by layer；
Step A2, according to the label of position layer by layer on described, will be described under the label of the sample point in the region of position layer by layer be adjusted to institute
State the label of corresponding region in position layer by layer；
Step A1 to A2 is repeated, the cluster result adjustment of all layers of position until completing the interval to be studied.
2. according to the method described in claim 1, it is characterized in that, if clustering mode using the wave band based on mean value,
The seismic attributes data of the interval to be studied is clustered using the time slip-window, including：
According to the wave band where the sample point e on current layer position, the attribute vector of the sample point e is determined according to following formula：
Wherein, the primitive attribute vector of the sample point e is (a1e, a2e... aie... ane), aiAjIndicate the sample point e's
J-th point of ith attribute value in point set A in time slip-window Δ T,Expression time slip-window is Δ T
In the case of the sample point e ith attribute average property value, i=1,2 ..., n；
Other sample points on the current layer position are calculated according to above-mentioned formula, obtain each sample on the current layer position
The attribute vector of point；
Successively along layer position where each sample point in the point set A, the attribute vector of each sample point is clustered.
3. according to the method described in claim 1, it is characterized in that, if using the wave band cluster side based on attribute split
Formula clusters the seismic attributes data of the interval to be studied using the time slip-window, including：
The all properties of each sample point in point set A in the time slip-window Δ T of sample point e on current layer position are carried out
Split obtains the wave band attribute vector of the sample point e：(the full attributes of A1, the full attributes ... of A2, the full attributes ... of Aj), wherein j
=1,2 ..., the full attribute of sample point is (a1, a2 ... ai ... an) in Δ T, the point set A；
Other sample points on the current layer position are calculated according to above-mentioned steps, obtain each sample on the current layer position
The wave band attribute vector of point；
The wave band attribute vector of all sample points on the current layer position is clustered.
4. according to the method described in claim 1, it is characterized in that, if using the wave band cluster side based on difference sum
Formula clusters the seismic attributes data of the interval to be studied using the time slip-window, including：
The attribute vector of the sample point e on current layer position and pre-stored multiple cluster centres are calculated separately according to following formula
Distance D：
Wherein, aiAjIndicate j-th point in the point set A in the time slip-window Δ T of sample point e of ith attribute value, aiCjTable
Show j-th point in the point set C in the time slip-window Δ T of cluster centre of ith attribute value, the cluster centre and classification
It corresponds；
Determine the classification apart from the corresponding classification of nearest cluster centre for the sample point e with the sample point e.
5. method according to claim 1 to 4, which is characterized in that in the utilization time slip-window to institute
State interval to be studied seismic attributes data clustered before, the method further includes：
Following pretreatment is carried out to the seismic attributes data of the interval to be studied：
Processing is filtered to the abnormal data in the seismic attributes data；And filtered seismic attributes data is carried out
Normalized.
6. a kind of seismic properties clustering apparatus, which is characterized in that including：
Determination unit, for determining wave band cluster mode, clustering parameter, initial layers position, layer position offset and time slip-window
Size, wherein the time slip-window be interval to be studied oscillogram in sample point contained between two adjacent peaks
Contained sample point number between number or two adjacent troughs, more than the initial layers position at least half time slip-window
Layer digit, the layer digit of at least half time slip-window below stop layer position；The calculation formula of the time slip-window is：Wherein, Δ T indicates time slip-window, tstepIndicate sampling step length, t0On expression layer position where sample point
Wave band middle-range described in sampling time interval between sample point nearest trough or wave crest and the sample point；The wave band is poly-
Class mode includes：Wave band cluster mode based on mean value, the wave band based on attribute split cluster mode or based on difference sum
Wave band clusters mode；
Cluster cell, for being clustered to the seismic attributes data of the interval to be studied using the time slip-window, really
Classification in the fixed interval to be studied belonging to the sample point of each layer of position；
Adjustment unit, for carrying out unification to the cluster result of adjacent layer position according to space similarity, wherein from the initial layers
Next layer of position of position starts, and the cluster result of each layer of position is adjusted according to the cluster result of a layer position thereon；Institute
It states and is included uniformly to the cluster result of adjacent layer position according to space similarity：
Each sample of the initial layers position is pressed and is classified according to generic, and presses preset order into row label；
Step A1, according to the space similarity of adjacent layer position, under position is adjacent layer by layer layer by layer in position, determined with upper with it is described on
The corresponding region of position layer by layer；
Step A2, according to the label of position layer by layer on described, will be described under the label of the sample point in the region of position layer by layer be adjusted to institute
State the label of corresponding region in position layer by layer；
Step A1 to A2 is repeated, the cluster result adjustment of all layers of position until completing the interval to be studied.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510809921.7A CN105425293B (en) | 2015-11-20 | 2015-11-20 | Seismic properties clustering method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510809921.7A CN105425293B (en) | 2015-11-20 | 2015-11-20 | Seismic properties clustering method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105425293A CN105425293A (en) | 2016-03-23 |
| CN105425293B true CN105425293B (en) | 2018-08-10 |
Family
ID=55503617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510809921.7A Active CN105425293B (en) | 2015-11-20 | 2015-11-20 | Seismic properties clustering method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105425293B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106443822A (en) * | 2016-08-16 | 2017-02-22 | 中国石油化工股份有限公司 | Geological integrated identification method and device based on gravity-magnetic-electric-seismic three-dimensional joint inversion |
| CN113589373A (en) * | 2020-04-30 | 2021-11-02 | 中国石油化工股份有限公司 | Well-seismic combined self-adaptive multi-parameter intelligent lithofacies identification method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102163224A (en) * | 2011-04-06 | 2011-08-24 | 中南大学 | Adaptive spatial clustering method |
| WO2014126650A1 (en) * | 2013-02-14 | 2014-08-21 | Exxonmobil Upstream Research Company | Detecting subsurface structures |
| CN104181597A (en) * | 2014-08-31 | 2014-12-03 | 电子科技大学 | Seismic facies analysis method based on prestack seismic data |
| WO2014197160A4 (en) * | 2013-06-06 | 2015-01-29 | Exxonmobil Upstream Research Comapny | Method for decomposing complex objects into simpler components |
| CN104570109A (en) * | 2013-10-29 | 2015-04-29 | 中国石油天然气集团公司 | Method for reservoir petroleum gas prediction |
| CN104914467A (en) * | 2015-05-22 | 2015-09-16 | 中国石油天然气股份有限公司 | Classification model channel extracting seismic facies clustering analysis method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7640582B2 (en) * | 2003-04-16 | 2009-12-29 | Silicon Graphics International | Clustered filesystem for mix of trusted and untrusted nodes |
-
2015
- 2015-11-20 CN CN201510809921.7A patent/CN105425293B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102163224A (en) * | 2011-04-06 | 2011-08-24 | 中南大学 | Adaptive spatial clustering method |
| WO2014126650A1 (en) * | 2013-02-14 | 2014-08-21 | Exxonmobil Upstream Research Company | Detecting subsurface structures |
| WO2014197160A4 (en) * | 2013-06-06 | 2015-01-29 | Exxonmobil Upstream Research Comapny | Method for decomposing complex objects into simpler components |
| CN104570109A (en) * | 2013-10-29 | 2015-04-29 | 中国石油天然气集团公司 | Method for reservoir petroleum gas prediction |
| CN104181597A (en) * | 2014-08-31 | 2014-12-03 | 电子科技大学 | Seismic facies analysis method based on prestack seismic data |
| CN104914467A (en) * | 2015-05-22 | 2015-09-16 | 中国石油天然气股份有限公司 | Classification model channel extracting seismic facies clustering analysis method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105425293A (en) | 2016-03-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Roy et al. | Generative topographic mapping for seismic facies estimation of a carbonate wash, Veracruz Basin, southern Mexico | |
| CN107976713B (en) | A kind of method and device of the lower removal sedimentation setting of higher-dimension seismic data input | |
| CN106842307B (en) | Reservoir stratum fine prediction method based on waveform classification and retrieval under forward constraint | |
| AU2013337322B2 (en) | Seismic waveform classification system and method | |
| CN107678059B (en) | A kind of method, apparatus and system of reservoir gas-bearing identification | |
| Wang | Reservoir characterization based on seismic spectral variations | |
| CN104101902B (en) | Seismic properties clustering method and device | |
| AU2009241409A1 (en) | System and method for interpretation of well data | |
| CN110837115B (en) | Seismic identification method and device for lithology of land-facies mixed rock compact reservoir | |
| CN105700016B (en) | The method of organic bank reservoir seismic prediction | |
| CN105277979B (en) | The optimization method and device of a kind of seismic properties | |
| CN105425293B (en) | Seismic properties clustering method and device | |
| Ha et al. | An in-depth analysis of logarithmic data transformation and per-class normalization in machine learning: Application to unsupervised classification of a turbidite system in the Canterbury Basin, New Zealand, and supervised classification of salt in the Eugene Island minibasin, Gulf of Mexico | |
| US10605940B2 (en) | Method for selecting horizon surfaces | |
| CN107991716B (en) | The deposition phasor of thin layer and the determination method and apparatus for depositing body thickness | |
| CN107102379A (en) | A kind of method that seat earth watery prediction is carried out based on many attribution inversions | |
| CN107688197B (en) | Thin bed prediction method and apparatus | |
| Mollajan et al. | Fuzzy classifier fusion: an application to reservoir facies identification | |
| Emelyanova et al. | Unsupervised identification of electrofacies employing machine learning | |
| CN112379435A (en) | Phase-controlled karst type seam hole aggregate carving method and device | |
| Jiang et al. | Saliency-map guided salt prediction by a multi-channel convolutional neural network | |
| Xu et al. | Application of seismic facies and attributes analysis on the identification of Permian igneous rock | |
| Al-Ali et al. | Lithofacies Classification and Distribution for Heterogeneous Channelized System Based on Neural Network Process: Case Study from Middle East Carbonate Reservoir | |
| CN114152979B (en) | Method and device for classifying seismic waveforms under time-varying window frame | |
| CN108267798B (en) | method and device for determining sedimentary phase control oil gas probability |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |