CN109143336A - A method of overcome the period in full waveform inversion to jump - Google Patents
A method of overcome the period in full waveform inversion to jump Download PDFInfo
- Publication number
- CN109143336A CN109143336A CN201810875786.XA CN201810875786A CN109143336A CN 109143336 A CN109143336 A CN 109143336A CN 201810875786 A CN201810875786 A CN 201810875786A CN 109143336 A CN109143336 A CN 109143336A
- Authority
- CN
- China
- Prior art keywords
- data
- waveform inversion
- full waveform
- prediction data
- arrival
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000007704 transition Effects 0.000 claims abstract description 36
- 238000005311 autocorrelation function Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 6
- 230000007423 decrease Effects 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 230000000875 corresponding effect Effects 0.000 description 19
- 238000010586 diagram Methods 0.000 description 12
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 244000145845 chattering Species 0.000 description 2
- 238000002939 conjugate gradient method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 206010042938 Systemic candida Diseases 0.000 description 1
- 230000002547 anomalous effect Effects 0.000 description 1
- 201000003984 candidiasis Diseases 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000002945 steepest descent method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/36—Effecting static or dynamic corrections on records, e.g. correcting spread; Correlating seismic signals; Eliminating effects of unwanted energy
- G01V1/364—Seismic filtering
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/20—Trace signal pre-filtering to select, remove or transform specific events or signal components, i.e. trace-in/trace-out
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The method that the period jumps in full waveform inversion is overcome the present invention relates to a kind of, which comprises the following steps: obtains the half period value of full waveform inversion;The first arrival of per pass observation data and per pass prediction data is picked up respectively;Obtain the time shift amount between the first arrival of per pass prediction data and the first arrival of corresponding observation data;Per pass prediction data is mobile to corresponding observation data, and corresponding generation middle transition data;Obtain the corresponding time histories sample of per pass prediction data;Objective function is established, by minimizing update of the objective function realization to model;Prediction data is regenerated according to updated model, and repeats step 2)~6), until the time shift amount between the first arrival of per pass observation data and the first arrival of new prediction data is respectively less than the half period of full waveform inversion.The new model for having been overcome the period to jump at this time, invention can be widely used in seismic data inversion technical field.
Description
Technical field
The method that the period jumps in full waveform inversion is overcome the present invention relates to a kind of.
Background technique
Full waveform inversion restores the model of earth interior by optimizing the objective function constrained by wave equation, including
P wave interval velocity, anisotropic parameters and density etc..Present full waveform inversion has been widely used in petroleum exploration domain,
However, the non-linear of objective function is a key factor for restricting the development of this technology.In addition to this, due to model and data
Dimension it is very big, the limitation of existing computer process ability can only use local minimum inversion algorithm, so if initially
The corresponding target function value of model is in objective function local minimum rather than the convex domain where global minimum, then Full wave shape
Inverting will converge to local minimum, so that the model and true model deviation that finally restore are larger, or even also than initial model
Difference.Full waveform inversion nonlinear one significant performance is exactly period jump, i.e., when prediction data deviates corresponding observation data
Distance more than half period will generating period jump.Therefore, period jump will lead to full waveform inversion and converge to an office
Portion's minimum is to the inverse model to make mistake.
In order to overcome the problems, such as that the period jumps, current most widely used method is multi-scale strategy, since frequency is lower, half
Period is bigger, therefore, can by maximum if the inverting since the low-limit frequency of observation data, is then stepped up inverting frequency
Energy the period is avoided to jump.On the other hand, it jumps for the period, scholars propose the objective function based on extension field to overcome
This problem.The opposite L2 norm based on surplus is traditional full waveform inversion of objective function, this kind of target based on extension field
The basic mechanism of functional based method is that convex domain more larger range of than traditional full waveform inversion is established around global minimum.No
It is same as data surplus, the L2 norm of seismic channel envelope has larger range of convex domain around global minimum, therefore, i.e.,
Make to be likely to converge to global minimum in the case of initial model error is larger.The objective function of this kind of extension field
It can be realized by weight estimation seismic channel and the cross-correlation function observed between seismic channel or Wiener filter.This mesh
Scalar functions can generate a very big convex domain to avoid period chattering near global minimum.Except mentioned above
Method except, there are also some objective functions to influence to mitigate period jump bring, such as inverting and adaptive full when walking entirely
Waveform inversion etc..
However, all there is some disadvantages: 1) multi-scale strategy for above-mentioned three classes method, this method needs to observe data
Including low-frequency information, if the low-limit frequency in data is there are still period jump, this method cannot overcome the period to jump
Jump converges to local minimum points, inverting failure so as to cause inverting.2) extension field objective function, in actual test, this side
The stability and robustness of method are very undesirable, cause the successful of inverting different with model with the data of inverting and change,
Basic reason needs further to be investigated.3) other methods, inverting and adaptive full waveform inversion when such as walking entirely, are similar to extension field
Objective function method, such methods the problem of there is also stability and robustness.
Summary of the invention
In view of the above-mentioned problems, the object of the present invention is to provide a kind of inverting can converge to global minima point and stability and
Robustness is good to overcome the method that the period jumps in full waveform inversion.
To achieve the above object, the present invention takes following technical scheme: a kind of to overcome period jump in full waveform inversion
Method, which comprises the following steps: step 1): according to the frequency and source wavelet of full waveform inversion, Full wave shape is obtained
The half period of inverting;Step 2): the first arrival of per pass observation data and per pass prediction data is picked up respectively;Step 3): it obtains every
Time shift amount between the first arrival of road prediction data and the first arrival of corresponding observation data;Step 4): according to obtained time shift amount and entirely
The half period value of waveform inversion, per pass prediction data is mobile to corresponding observation data, and corresponding generation middle transition data;
Step 5): according to the first arrival of per pass prediction data, the corresponding time histories sample of per pass prediction data is obtained;Step 6): according to per pass
Prediction data and its corresponding time histories sample and middle transition data, establish new objective function, and minimize this objective function
Model is updated;Step 7): prediction data is regenerated according to updated model, and repeats step 2)~6), until
Time shift amount between the first arrival of per pass observation data and the first arrival of new prediction data is respectively less than the half period of full waveform inversion, this
When overcome the period jump new model.
Further, the half of full waveform inversion is estimated according to the frequency and source wavelet of full waveform inversion in the step 1)
A period, detailed process are as follows: source wavelet is filtered, and retains the frequency of full waveform inversion;Focus is sub after calculating filtering
The auto-correlation function of wave;After measurement filtering in the auto-correlation function of source wavelet the extreme value nearest apart from zero-lag and zero-lag it
Between time shift amount, which is the length of full waveform inversion half period.
Further, according to the half period value of obtained time shift amount and full waveform inversion in the step 4), per pass is pre-
Measured data is mobile to corresponding observation data, and it is corresponding generate middle transition data, specifically: the first arrival of certain one of prediction data with
Time shift amount between the first arrival of corresponding observation data is Δ t0, which is observed into the mobile Δ t of data to correspondings, Δ tsIt is
Between time shift amount Δ t0Time shift amount between the half period value of full waveform inversion, the prediction data after movement are referred to as intermediate
Transit data.
Further, the time histories sample in the step 5) is Gaussian window or Hanning window, and the value of time histories sample is in correspondence
It is 1 at the first arrival of prediction data, the value further away from the first arrival then time histories sample of corresponding prediction data is smaller, and gradually decreases to
Zero.
Further, in the step 6) full waveform inversion objective function are as follows:
Wherein, W indicates that time histories sample, P indicate to extract the pickup operator of source wavefield in geophone station position, and u indicates prediction
Data, diIndicate middle transition data.
The invention adopts the above technical scheme, which has the following advantages: the present invention passes through setting middle transition data,
The middle transition data contain information of the prediction data relative to observation shortage of data, while not having with the time shift amount of prediction data
The half period value of full waveform inversion is had more than, so that inverting middle transition data can produce the correct of no period jump
Geophysical model, can jump problem efficiently against the period in full waveform inversion, restore background velocity, correctly to pass
System waveform inversion preferably restores intermediate frequency and high frequency velocity component provides basis, can be widely applied to seismic data inversion technology
Field.
Detailed description of the invention
Fig. 1 is the schematic diagram that middle transition data are generated in the present invention;
Fig. 2 is the comparison diagram of true velocity model and initial velocity model in the embodiment of the present invention, wherein Fig. 2 (a) is true
Real rate pattern, Fig. 2 (b) are initial velocity model;
Fig. 3 is the schematic diagram that the middle transition data generated in the embodiment of the present invention overcome the period to jump, wherein Fig. 3 (a)
For the big gun data generated by the true velocity model of the initial velocity model of Fig. 2 (b) and Fig. 2 (a), Fig. 3 (b) is by initial speed
The big gun prediction data and middle transition schematic diagram data that model generates are spent, Fig. 3 (c) is will be anti-using cosine square weighting function
The schematic diagram for being defined in preliminary wave is drilled, Fig. 3 (d) is the big gun prediction data generated by inversion speed model and observation data;
Fig. 4 is the gradient comparison schematic diagram of traditional full waveform inversion and the method for the present invention, wherein Fig. 4 (a) is traditional all-wave
The gradient schematic diagram of shape inverting, Fig. 4 (b) are the gradient of the first time iteration generated using the middle transition data of the method for the present invention
Schematic diagram;
Fig. 5 is the inversion result contrast schematic diagram of traditional full waveform inversion and the method for the present invention, wherein Fig. 5 (a) is to scheme
2 (b) the inversion result schematic diagram to carry out traditional full waveform inversion to initial velocity model, it is to first that Fig. 5 (b), which is with Fig. 2 (b),
Beginning rate pattern carries out the inversion result schematic diagram of the method for the present invention, and it is that initial velocity model is passed that Fig. 5 (c), which is with Fig. 5 (b),
The inversion result schematic diagram for full waveform inversion of uniting.
Specific embodiment
Come to carry out detailed description to the present invention below in conjunction with attached drawing.It should be appreciated, however, that attached drawing has been provided only more
Understand the present invention well, they should not be interpreted as limitation of the present invention.
The most widely used objective function of full waveform inversion is 2 norm squareds for observing the difference of data and prediction data,
That is:
It is constrained in:
Au=s (2)
Wherein, P indicates to extract the pickup operator of source wavefield in geophone station position;U indicates prediction data;D indicates observation
Data;A indicates wave equation operator;S indicates source wavelet.A and u is the function about model parameter m.Under normal conditions,
The element number of model parameter m can achieve million magnitudes, in addition, the calculation amount of numerical solution wave equation is also quite huge
's.Therefore, the method for minimizing the objective function of full waveform inversion is substantially the local inversion algorithm based on gradient, under steepest
Drop method and conjugate gradient method.However, objective function about model parameter m be it is nonlinear, when initial velocity model and true mould
When the error of type is larger, half period or more may be differed between prediction data and observation data, this phenomenon is known as the period
Jump, it may make the objective function of full waveform inversion converge to local minimum rather than global minimum.
For process cycle chattering, propose to generate the middle transition number between prediction data and observation data
According to as shown in Figure 1, thick solid curve indicates observation data, thin solid-line curve indicates prediction data, observes between data and prediction data
Time shift amount be Δ t0, this time shift amount is greater than half period, can be by prediction data to sight in order to overcome the period to jump
Measured data moves Δ ts, Δ tsIt is between time shift amount Δ t0Time shift amount between the half period value of full waveform inversion, is less than
Half period, the prediction data after movement is referred to as middle transition data, its observation data more closer than initial prediction data,
In figure, Δ tiIt indicates middle transition data and observes the time shift amount between data.Therefore, middle transition data have prediction data
Information when walking of opposite observation shortage of data, when replacing inverting to observe data with inverting middle transition data, Full wave shape is anti-
It drills and can produce correct update, and accumulate it on initial model, new prediction data will be closer to observation data, but
It is and the difference of observation data is still likely larger than half period.It is then possible to generate new middle transition data and carry out inverting.It is logical
This process of iteration is crossed, the position that initial model can be moved close to true model makes it that the period be overcome to jump.Later,
The inversion result of no period jump can be obtained using traditional full waveform inversion.
For earth's surface acquisition system, preliminary wave is direct wave near migration range, is refracted wave in remote offset distance.It is through
Wave only includes more new information of the model in earth's surface, and refracted wave includes information necessary to model Systemic candidiasis speed updates, because
This, using preliminary wave inverting can more new model background velocity, it is based on the above principles, provided by the invention to overcome Full wave shape anti-
The method for drilling middle period jump, comprising the following steps:
1) according to the frequency of full waveform inversion and source wavelet, the half period value i.e. length of 0.5T of full waveform inversion is estimated
Degree, specifically:
1.1) source wavelet is filtered, and retains frequency involved in full waveform inversion.
1.2) auto-correlation function of source wavelet after filtering is calculated, wherein calculating auto-correlation function value can use existing
Auto-correlation function calculation method disclosed in technology, detailed process do not repeat herein.
1.3) between extreme value and zero-lag nearest apart from zero-lag in the auto-correlation function of source wavelet after measurement filtering
Time shift amount, that is, length of delay, which is the length of full waveform inversion half period 0.5T.
2) first arrival of per pass observation data and per pass prediction data is picked up respectively, wherein observation data refer to using instrument
In the seismic data that field is recorded, it corresponds to the analogue data that true earth model generates.Prediction data refers to using meter
The seismic data that the simulation of calculation machine generates, it corresponds to the computation model (such as inverting start initial velocity model) in computer
The analogue data of generation.Meanwhile the method for any first break picking is used equally for picking up observation in geophysical exploration and seismology
The first arrival of data and prediction data.
3) the time shift amount Δ t between the first arrival of per pass prediction data and the first arrival of corresponding observation data is obtained0, and generate one
Time shift sequence, wherein the time shift sequence includes the time shift between the first arrival of per pass prediction data and the first arrival of corresponding observation data
Measure Δ t0, can be directly by the time shift amount Linear Mapping of the first arrival of prediction data and the first arrival of observation data to -0.5T~0.5T
Time shift amount within the scope of.
4) as shown in Figure 1, according to the half period value of the time shift sequence of generation and full waveform inversion, by per pass prediction data
To the corresponding mobile Δ t of observation datas, and corresponding generation middle transition data di, specifically:
Time shift amount in time shift sequence between the first arrival of certain one of prediction data and the first arrival of corresponding observation data is Δ t0,
Time shift amount Δ t0It, can be by the prediction data to sight in order to overcome the period to jump greater than the half period value of full waveform inversion
Measured data moves Δ ts, Δ tsIt is the time shift amount between time shift amount Δ tx and the half period of full waveform inversion, is less than half
A period, the prediction data after movement is referred to as middle transition data, its observation data more closer than initial prediction data.
5) according to the first arrival of per pass prediction data, the corresponding time histories sample W of per pass prediction data, time histories sample W are obtained
It can be Gaussian window or Hanning window, the value of time histories sample W is 1 at the first arrival of prediction data, further away from prediction data
The value of first arrival then time histories sample W is smaller, and gradually decreases to zero, wherein time histories sample W keeps Full wave shape anti-for selecting data
It drills and is limited in the range of preliminary wave.
6) according to per pass prediction data and its corresponding time histories sample W and middle transition data, it is anti-to establish new Full wave shape
The objective function drilled can indicate are as follows:
Wherein, u indicates prediction data, diIndicate middle transition data.By the target for minimizing above-mentioned full waveform inversion
Function realizes the update to "current" model.
Minimizing above-mentioned full waveform inversion objective function can be using optimization method disclosed in the prior art, such as steepest
Descent method or conjugate gradient method, detailed process do not repeat herein.
7) prediction data is regenerated according to updated model, and repeats step 2)~6), until per pass observes data
First arrival and new prediction data first arrival between time shift amount be respectively less than the half period of full waveform inversion.Overcome at this time
The new model of period jump.
Below by the specific embodiment method that the present invention will be described in detail overcomes period in full waveform inversion to jump.
The true model including a high-velocity anomaly body and a low-velocity zone is shown such as Fig. 2 (a), wherein background
Speed is 3000 meter per seconds, and the difference of anomalous body and background velocity is 1000 meter per seconds, and black dotted lines indicate focus arrangement, and white is empty
Line indicates geophone station distribution, using cross well survey system, excites 122 big guns altogether in 100 meters of depth, the distance between every big gun is 80
Rice, is distributed in 160 meters to 9840 meters of range.Source wavelet is the Ricker wavelet of a 10Hz, and 1001 wave detectors are consolidated
It is scheduled on 2900 meters of depth.If Fig. 2 (b) show the initial velocity model that background velocity is 2800 meter per seconds, with true velocity
There are greatest differences for model, as shown in Fig. 3 (a), in the comparison of big gun data it is not difficult to find out that, this species diversity cause prediction data with
Observe the phase difference that data are more than the half period.Period jump will be present using the full waveform inversion that the initial velocity model is implemented to ask
Topic.In order to illustrate this problem, full waveform inversion is carried out from 5Hz to 9Hz by interval of 1Hz, wherein a frequency is exactly one
Narrow frequency band, each frequency iteration 5 times, then to Whole frequency band iteration 15 times again, 40 iterative inversions in total.First time iteration
Shown in gradient such as Fig. 4 (a), it can be seen that there are the negative territories that the representation speed of large area should reduce, however in initial speed
It spends in model in addition to the speed in other external regions of low-velocity anomal should be increased.Therefore, period jump causes Full wave shape anti-
It drills and converges to wrong rate pattern as shown in Fig. 5 (a).
In order to overcome the period to jump, inverting should using one middle transition data of generation, and in each iteration by the present invention
The mode of middle transition data.Firstly, calculating the length of half period, the Ricker wavelet for 10Hz is about 48ms;Then,
It picks up prediction data and observes the first arrival of data, by two first arrival time shift amount Linear Mappings to -30ms~30ms time shift amount model
In enclosing, so that maximum time shift amount is less than half period;The time shift amount is applied in prediction data again to generate middle transition
Data, as Fig. 3 (b) show in a big gun data prediction data compared with middle transition data;Last inverting middle transition number
According to come replace inverting record data, surplus is weighted with the cosine square function as shown in Fig. 3 (c), in the inverting, using complete
Frequency band data.Fig. 4 (b) illustrates the gradient of first time iteration, and traditional full waveform inversion gradient difference such as Fig. 4 (a) shown in,
The gradient of middle transition data is dominated and nonnegative value by positive value, it means that inverting is attempting to increase speed, thus towards just
The development of true speed more new direction.After each iteration, middle transition data are regenerated.
Inversion result after showing 40 iteration such as Fig. 5 (b), by comparing it can be found that using middle transition data
Generated inversion result correctly has modified background velocity, has restored Velocity anomalies.The inverting of this high quality can also be with
It is verified by prediction data and preferably coincideing for observation data, as shown in Fig. 3 (d).It is generated using middle transition data anti-
Result is drilled as initial velocity model, implements 40 conventional full wave shape invertings and generates last inversion speed model such as Fig. 5 (c) institute
Show, by comparing discovery, that further enhances the structures of Velocity anomalies to converge to global minimum.
The various embodiments described above are merely to illustrate the present invention, wherein the structure of each component, connection type and manufacture craft etc. are all
It can be varied, all equivalents and improvement carried out based on the technical solution of the present invention should not exclude
Except protection scope of the present invention.
Claims (5)
1. a kind of overcome the method that the period jumps in full waveform inversion, which comprises the following steps:
Step 1): according to the frequency and source wavelet of full waveform inversion, the half period value of full waveform inversion is obtained;
Step 2): the first arrival of per pass observation data and per pass prediction data is picked up respectively;
Step 3): the time shift amount between the first arrival of per pass prediction data and the first arrival of corresponding observation data is obtained;
Step 4): according to the half period value of obtained time shift amount and full waveform inversion, per pass prediction data is observed to corresponding
Data are mobile, and corresponding generation middle transition data;
Step 5): according to the first arrival of per pass prediction data, the corresponding time histories sample of per pass prediction data is obtained;
Step 6): according to per pass prediction data and its corresponding time histories sample and middle transition data, new target letter is established
Number, and minimize this objective function and model is updated;
Step 7): prediction data is regenerated according to updated model, and repeats step 2)~6), until per pass observes data
First arrival and new prediction data first arrival between time shift amount be respectively less than the half period of full waveform inversion.Overcome at this time
The new model of period jump.
2. a kind of method for overcoming period jump in full waveform inversion as described in claim 1, which is characterized in that the step
1) according to the frequency of full waveform inversion and source wavelet in, the half period value of full waveform inversion, detailed process are estimated are as follows:
Source wavelet is filtered, and retains the frequency of full waveform inversion;
Calculate the auto-correlation function of source wavelet after filtering;
Time shift amount after measurement filtering between extreme value and zero-lag nearest apart from zero-lag in the auto-correlation function of source wavelet,
The time shift amount is the length of full waveform inversion half period.
3. a kind of method for overcoming period jump in full waveform inversion as described in claim 1, which is characterized in that the step
4) according to the half period value of obtained time shift amount and full waveform inversion in, per pass prediction data is moved to corresponding observation data
It is dynamic, and corresponding generation middle transition data, specifically:
Time shift amount between the first arrival of certain one of prediction data and the first arrival of corresponding observation data is Δ t0, by the prediction data to
The corresponding mobile Δ t of observation datas, Δ tsIt is between time shift amount Δ t0Time shift amount between the half period value of full waveform inversion,
The prediction data after mobile is referred to as middle transition data.
4. a kind of method for overcoming period jump in full waveform inversion as described in claim 1, which is characterized in that the step
5) time histories sample in is Gaussian window or Hanning window, and the value of time histories sample is 1 at the first arrival of corresponding prediction data, remoter
The value of first arrival from corresponding prediction data then time histories sample is smaller, and gradually decreases to zero.
5. a kind of method for overcoming period jump in full waveform inversion as described in claim 1, which is characterized in that the step
6) objective function of full waveform inversion in are as follows:
Wherein, W indicates that time histories sample, P indicate to extract the pickup operator of source wavefield in geophone station position, and u indicates prediction data,
diIndicate middle transition data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810875786.XA CN109143336B (en) | 2018-08-03 | 2018-08-03 | A method of overcome the period in full waveform inversion to jump |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810875786.XA CN109143336B (en) | 2018-08-03 | 2018-08-03 | A method of overcome the period in full waveform inversion to jump |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109143336A true CN109143336A (en) | 2019-01-04 |
CN109143336B CN109143336B (en) | 2019-09-13 |
Family
ID=64791491
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810875786.XA Active CN109143336B (en) | 2018-08-03 | 2018-08-03 | A method of overcome the period in full waveform inversion to jump |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109143336B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111239806A (en) * | 2020-03-09 | 2020-06-05 | 吉林大学 | Time domain full waveform inversion method based on amplitude increment coding |
CN111722287A (en) * | 2020-06-19 | 2020-09-29 | 南京大学 | Seismic phase characteristic identification waveform inversion method based on PDA strategy |
CN112630833A (en) * | 2020-11-19 | 2021-04-09 | 安徽理工大学 | Fast seismic first-arrival travel time joint inversion method based on logging curve |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103415786A (en) * | 2011-06-08 | 2013-11-27 | 雪佛龙美国公司 | System and method for seismic data inversion by non-linear model update |
GB2504158A (en) * | 2011-12-20 | 2014-01-22 | Nikhil Koolesh Shah | Checking the validity of a subsurface starting model for performing full waveform inversion on seismic data |
CN106054244A (en) * | 2016-06-16 | 2016-10-26 | 吉林大学 | Low-pass filter multi-scale full waveform inversion method of cut-off time window |
CN106908835A (en) * | 2017-03-01 | 2017-06-30 | 吉林大学 | Band limit Green's function filters multiple dimensioned full waveform inversion method |
CN107422379A (en) * | 2017-07-27 | 2017-12-01 | 中国海洋石油总公司 | Multiple dimensioned seismic full-field shape inversion method based on local auto-adaptive convexification method |
CN107505654A (en) * | 2017-06-23 | 2017-12-22 | 中国海洋大学 | Full waveform inversion method based on earthquake record integration |
CN107765302A (en) * | 2017-10-20 | 2018-03-06 | 吉林大学 | Inversion method when time-domain single-frequency waveform independent of source wavelet is walked |
CN108254731A (en) * | 2018-04-25 | 2018-07-06 | 吉林大学 | The multiple dimensioned staged layer stripping full waveform inversion method of Coherent Noise in GPR Record |
-
2018
- 2018-08-03 CN CN201810875786.XA patent/CN109143336B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103415786A (en) * | 2011-06-08 | 2013-11-27 | 雪佛龙美国公司 | System and method for seismic data inversion by non-linear model update |
GB2504158A (en) * | 2011-12-20 | 2014-01-22 | Nikhil Koolesh Shah | Checking the validity of a subsurface starting model for performing full waveform inversion on seismic data |
CN106054244A (en) * | 2016-06-16 | 2016-10-26 | 吉林大学 | Low-pass filter multi-scale full waveform inversion method of cut-off time window |
CN106908835A (en) * | 2017-03-01 | 2017-06-30 | 吉林大学 | Band limit Green's function filters multiple dimensioned full waveform inversion method |
CN107505654A (en) * | 2017-06-23 | 2017-12-22 | 中国海洋大学 | Full waveform inversion method based on earthquake record integration |
CN107422379A (en) * | 2017-07-27 | 2017-12-01 | 中国海洋石油总公司 | Multiple dimensioned seismic full-field shape inversion method based on local auto-adaptive convexification method |
CN107765302A (en) * | 2017-10-20 | 2018-03-06 | 吉林大学 | Inversion method when time-domain single-frequency waveform independent of source wavelet is walked |
CN108254731A (en) * | 2018-04-25 | 2018-07-06 | 吉林大学 | The multiple dimensioned staged layer stripping full waveform inversion method of Coherent Noise in GPR Record |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111239806A (en) * | 2020-03-09 | 2020-06-05 | 吉林大学 | Time domain full waveform inversion method based on amplitude increment coding |
CN111239806B (en) * | 2020-03-09 | 2021-06-22 | 吉林大学 | Time domain full waveform inversion method based on amplitude increment coding |
CN111722287A (en) * | 2020-06-19 | 2020-09-29 | 南京大学 | Seismic phase characteristic identification waveform inversion method based on PDA strategy |
CN111722287B (en) * | 2020-06-19 | 2021-10-08 | 南京大学 | Seismic phase characteristic identification waveform inversion method based on progressive data assimilation method |
CN112630833A (en) * | 2020-11-19 | 2021-04-09 | 安徽理工大学 | Fast seismic first-arrival travel time joint inversion method based on logging curve |
CN112630833B (en) * | 2020-11-19 | 2022-12-02 | 安徽理工大学 | Fast seismic first-arrival travel time joint inversion method based on logging curve |
Also Published As
Publication number | Publication date |
---|---|
CN109143336B (en) | 2019-09-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109143336B (en) | A method of overcome the period in full waveform inversion to jump | |
CN106405651B (en) | Full waveform inversion initial velocity model construction method based on logging matching | |
CN107894613B (en) | Elastic wave vector imaging method, device, storage medium and equipment | |
NO20101605A1 (en) | Consistent fall angle estimation for seismic imaging | |
AU2014203490B2 (en) | Survey techniques using streamers at different depths | |
AU2013402493A1 (en) | Automated horizon auto-picking on multiple volumes | |
AU2012208972B2 (en) | Quality-based steering methods and systems for 4D geophysical surveys | |
CN110187384B (en) | Bayes time-lapse seismic difference inversion method and device | |
CN109507726A (en) | The inversion method and system of time-domain elastic wave multi-parameter Full wave shape | |
CN110187382A (en) | A kind of diving Wave and back wave wave equation Travel Time Inversion method | |
WO2014195155A2 (en) | Method of constraining seismic inversion | |
CN110737018B (en) | Method for modeling anisotropy of VSP seismic data | |
CN113740906A (en) | Underwater vertical cable seismic wave interference imaging method and device | |
WO2015097557A2 (en) | Systems and methods for reducing noise in a seismic vibratory source | |
US20110182143A1 (en) | Technique and Apparatus for Seismic Data Quality Control Using Non-Linear Regression Analysis | |
CN107255832B (en) | A kind of inversion method of subsurface structure | |
KR101318994B1 (en) | Method and apparatus of estimating underground structure using a plurality of weighting value | |
Vigh et al. | Sparse-node long-offset velocity model building in the Gulf of Mexico | |
CN112305615B (en) | Seismic data angle domain common imaging point gather extraction method and system | |
CN109901221B (en) | Seismic data anisotropy modeling method based on dynamic correction velocity parameter | |
US10605938B2 (en) | 4D noise suppression | |
Zhang et al. | Residual moveout-based wave-equation migration velocity analysis | |
Zhang et al. | Autoencoded elastic wave-equation traveltime inversion: Toward reliable near-surface tomogram | |
Song et al. | The impact of local velocity anomaly on structure confirming and the varying velocity mapping strategy | |
Liu* et al. | The research of seismic data processing sequence in gas cloud zone-applied in Bohai P oilfield |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |