CN105510976B - A kind of many subwaves combination adaptive attenuation method - Google Patents

Abstract

A kind of many subwaves combination adaptive attenuation method, belong to earthquake multiple decay field, specifically comprise the following steps that first using protocol and many subwaves record of being obtained by Surface-related multiple Forecasting Methodology (SRMP) as original records, utilize lineups to optimize tracer technique and short time-window FK apparent velocity filter method extracts and multiple suppression lineups, thus obtain quasi-primary wave record and standard many subwaves record；Then the Wiener Filter Method utilizing extension eliminates the many subwaves of remnants in quasi-primary wave record further, and restores the primary wave information of former damage from accurate many subwaves record.The method fully utilize based on lineups optimization follow the trail of many subwaves elimination method and with extension the respective advantage of Wiener Filter Method, the common many subwaves of hyperbolic-type or the many subwaves of other type either all can effectively be suppressed by it, and can reduce the damage of primary wave to greatest extent.

Description

A kind of many subwaves combination adaptive attenuation method

Technical field

The invention belongs to earthquake multiple compacting field, more particularly to a kind of many based on lineups optimization tracer technique Subwave rejects the combination adaptive attenuation method combined with extension Wiener filtering.

Background technology

In recent years, such as wave field extrapolation method, Surface-related multiple damped system (Surface related Multiple Elimination, SRME) and inverse-scattering series method etc. obtain considerable based on many subwaves elimination method that wave equation is theoretical Development, the mainstream technology means of repeatedly wave pressure during wherein SRME method has become current production.Based on wave theory repeatedly Ripple elimination method generally includes the prediction of many subwaves and two steps of multiple attenuation, first needs wave theory to dope earthquake note Repeatedly wave component in record, then applies various repeatedly wave self-adaption damped systems by original according to the many subwaves record doped Many subwaves in earthquake record are rejected, and wherein to be an up wave equation class many in repeatedly wave self-adaption based on many subwaves record decay Subwave elimination method is for the key technology of many subwaves rejecting effect in complicated structure region.

The repeatedly wave self-adaption damped system being currently applied to wave equation class many subwaves elimination method mainly has two big classes, The first kind is multiple tracks Wiener Filter Method, and its principle is successfully to eliminate the free interface of prediction repeatedly according to least square criterion Ripple, so that SRME can be applied to reality veritably.

It is applied to the another kind of repeatedly wave self-adaption damped system of wave equation class many subwaves elimination method for based on various The matching attenuation method in mathematic(al) manipulation territory, as drawn the matching attenuation method in winter territory and primary wave based on bent wave zone and many based on parabola Subwave separation method etc..

In addition to above-mentioned two big class methods, people also proposed the matching attenuation method of other constraints, such as based on The self adaptation subtraction technique of pattern, adaptive attenuation based on independent component analysis, based on multidimensional prediction error filtering adaptive Answer subtraction technique, non-stationary to return self adaptation subtraction technique, unfilial son's wave zone is removed and is predicted the method for many subwaves and based on many The repeatedly wave self-adaption subtractive method etc. of road convolution signal blind separation.

Under normal circumstances, all kinds of many subwaves all can effectively be suppressed by multiple tracks Wiener filtering class method, but works as primary wave When intersecting with many subwaves lineups, it easily damages significant wave information；Comparatively speaking, the matching attenuation class method in mathematic(al) manipulation territory Preferably reject effect for occurring the many subwaves intersected to have with primary wave lineups, but it often assumes that many subwaves lineups accord with Close hyperbolic-type rule, thus it is for part non-double curve line style many subwaves rejecting poor effect in complicated structure region, in addition For when primary wave and many subwaves lineups apparent velocity difference are less (being generally present near migration range road), such method is also held Easily cause the damage of primary wave.It can be seen that the most conventional method can not take into account the extraction effect of significant wave and many subwaves Pressing result, it is therefore desirable to study a kind of new method so that it is be not only able to significantly improve the pressing result of many subwaves, Er Qieneng Enough it is prevented effectively from the damage to primary reflection signal during many subwaves are rejected, and the many subwaves being adapted to complicated structure region pick Remove.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind of many subwaves to combine adaptive attenuation method, the present invention based on Lineups optimize many subwaves of tracer technique and reject the combination adaptive attenuation method combined with extension Wiener filtering, first base In many subwaves information of prediction, utilize lineups to optimize tracer technique and short time-window FK apparent velocity filter method calculates, by former Record separation of beginning is as the criterion primary wave record and standard many subwaves record；Then the Wiener Filter Method recycling extension eliminates quasi-primary wave The many subwaves of remnants in record, and recover the primary wave information of damage in former standard many subwaves record simultaneously.

The present invention takes techniques below scheme:

A kind of many subwaves combination adaptive attenuation method, specifically includes following steps:

1) the many subwaves matching attenuation process followed the trail of based on lineups optimization, including the following steps:

(1) (x t), is carried out dynamic school to utilize Surface-related multiple Forecasting Methodology (SRMP) to obtain many subwaves record m Just and overlap-add procedure, obtaining many subwaves stack velocity spectrum, corresponding computing formula is

$E_m(v,\mathrm{\τ})=\left|\underset{n=1}{\overset{N}{\Σ}}m(x_n,t=\sqrt{{\mathrm{\τ}}^2+x_n^2/v^2})\right|$

In formula, n (1≤n≤N) is Taoist monastic name；x_nIt is the offset distance in the n-th road, to E_mAfter (v, τ) carries out appropriate smoothing, and m (x, t) Middle stack velocity value v₀, zero-offset time τ₀Hyperbola lineups, will spectrum E_m(v, τ) is upper to be formed with (v₀,τ₀Pole centered by) The bulk structural energy of value；

(2) application contour tracing method obtains spectrum E_mWith (v on (v, τ)₀,τ₀The bulk structural energy of extreme value centered by) Distribution, and search out its extreme point position, then according to the coordinate (v of this point₀,τ₀) simulate in time-space domain corresponding The whilst on tour t of lineups, i.e. its Suo Jingge road_n, it is represented by

$t_n=\sqrt{{{\mathrm{\τ}}_0}^2+{x_n}^2/{v_0}^2}$

Then with the extreme value place (v of former tracking₀,τ₀(x, local velocity's spectrum t), one to create protocol d centered by) Fixed speed and time range interscan then can obtain the accurate parameter v ' of these lineups₀With τ '₀, thus realize many subwaves homophase The optimization tracing process of axle；

(3) obtain each many subwaves lineups through the whilst on tour t of seismic channel_mAfter, utilize short time-window FK apparent velocity to filter Method from protocol d (x, t) and many subwaves record m (x, t) in eliminated, obtain the standard after removing many subwaves lineups once Ripple record d_p(x, t) with residue many subwaves record m_l(x, t), and the many subwaves lineups got rid of synthesize the repeatedly ripple record that is as the criterion d_m(x, t) with hyperbolic many subwaves record m_h(x,t)；

2) Wiener Filter Method pin alignment many subwaves record d of extension is utilized_m(x, t) with quasi-primary wave record d_p(x t) is carried out Adaptive attenuation again, to leach accurate many subwaves record d_m(x, t) in primary wave information, and eliminate further quasi-primary wave note Record d_p(x, t) in the many subwaves of remnants, corresponding processing procedure is expressed as

p₁(x, t)=ELSF [d_m(x,t),m_h(x,t)]

p₂(x, t)=ELSF [d_p(x,t),m_l(x,t)]

Wherein, ELSF represents extended least squares filter operator；p₁(x, t) in comprise be by record d_m(x, t) in leach Primary wave information；p₂(x, t) for eliminating record d_p(x, t) in the result of remaining many subwaves, the record after final MULTIPLE ATTENUATION It is shown below:

P (x, t)=p₁(x,t)+p₂(x,t)

In formula, (x, t) for the final result of combination adaptive attenuation for p.

Present invention beneficial effect compared with prior art:

This paper presents a kind of many subwaves combination adaptive attenuation method, the method optimizes tracer technique based on lineups Many subwaves reject the combination adaptive attenuation method combined with extension Wiener filtering, fully utilize and chase after based on lineups optimization The multiple elimination method of track and with the extension respective advantage of Wiener Filter Method, it is either to the common many subwaves of hyperbolic-type Or the many subwaves of other type all can effectively be suppressed, and can reduce the damage of primary wave to greatest extent.Theoretical model is real Test the process experimental result with field data and all show that the many subwaves of the method are rejected effect and are substantially better than the multiple tracks wiener filter of routine Wave method, its many subwaves rejecting being adaptive to complicated structure region processes.

Accompanying drawing explanation

The Grid Velocity model of Fig. 1 HORIZONTAL LAYERED MEDIUM WITH HIGH ACCURACY；

Fig. 2 many subwaves combination adaptive attenuation handling process；

Fig. 3 lineups optimize the 500th the CMP record that tracing process example (a) big gun collection record is corresponding；B () is based on freedom Interface many subwaves Forecasting Methodology obtains many subwaves record of the 500th CMP record；C () creates many subwaves based on many subwaves record Normal-moveout spectrum；D () follows the trail of the many subwaves lineups in the protocol obtained；

Fig. 4 optimizes many subwaves and primary wave separation process (a) the short time-window FK apparent velocity filtering of tracking method, and (b) is accurate Primary wave record example, (c) accurate many subwaves record example；

Fig. 5 (a) passes through multiple tracks Wiener filtering MULTIPLE ATTENUATION by the CMP record of many subwaves combination adaptive attenuation, (b) After CMP record, (c) many subwaves combination repeatedly wave component of being removed of adaptive attenuation method, (d) multiple tracks Wiener Filter Method removes Repeatedly wave component；

Fig. 6 original CMP record (a) and many subwaves record (b) example predicted；

Fig. 7 utilizes lineups to optimize the quasi-primary wave record example that tracking method obtains；

CMP record after Fig. 8 many subwaves combination adaptive attenuation method (a) are decayed repeatedly with multiple tracks Wiener Filter Method (b) shows Example；

Many subwaves information that Fig. 9 many subwaves combination adaptive attenuation method (a) and multiple tracks Wiener Filter Method (b) cut；

The reverse-time migration section of Figure 10 Pluto model protocol；

Figure 11 reverse-time migration section after many subwaves combination adaptive attenuation method multiple suppression；

Figure 12 reverse-time migration section after multiple tracks Wiener Filter Method MULTIPLE ATTENUATION.

Detailed description of the invention

Combine accompanying drawing below by embodiment technical scheme is further explained, but the protection of the present invention Scope is not by any pro forma restriction of embodiment.

Embodiment 1

A kind of many subwaves combination adaptive attenuation method, the dielectric model of detailed description of the invention includes HORIZONTAL LAYERED MEDIUM WITH HIGH ACCURACY mould Type and Pluto 1.5 model, wherein the speed of horizontal layer model is as shown in Figure 1.The present embodiment is as a example by horizontal layer model Detailed description is based on many subwaves combination adaptive attenuation method realizes flow process (flow chart is as shown in Figure 2), and its detailed process is such as Under:

1) Ricker wavelet using dominant frequency as 35Hz is as focus, uses the finite difference calculus analogy method of ACOUSTIC WAVE EQUATION to generate A set of earthquake record containing many subwaves.This earthquake record totally 500 big gun, every big gun contains 160 roads, and shot interval and road spacing are 10 Rice, smallest offset is away from being 0 (the 500th CMP that this big gun collection record is corresponding records as shown in Fig. 3 (a)).Shown in Fig. 3 (a) In CMP record, arrow point to for many subwaves lineups, both intersects with the primary wave lineups of same period, and serial number 2. many subwaves lineups almost overlap in position, near migration range road with primary wave lineups.

2) the many subwaves matching attenuation process followed the trail of based on lineups optimization, including the following steps:

(1) (x t), is carried out dynamic school to utilize Surface-related multiple Forecasting Methodology (SRMP) to obtain many subwaves record m Just and overlap-add procedure, obtaining many subwaves stack velocity spectrum, corresponding computing formula is

$E_m(v,\mathrm{\τ})=\left|\underset{n=1}{\overset{N}{\Σ}}m(x_n,t=\sqrt{{\mathrm{\τ}}^2+x_n^2/v^2})\right|$

In formula, n (1≤n≤N) is Taoist monastic name；x_nIt is the offset distance in the n-th road, to E_mAfter (v, τ) carries out appropriate smoothing, and m (x, t) Middle stack velocity value v₀, zero-offset time τ₀Hyperbola lineups, will spectrum E_m(v, τ) is upper to be formed with (v₀,τ₀Pole centered by) The bulk structural energy of value.

(2) application contour tracing method obtains spectrum E_mWith (v on (v, τ)₀,τ₀The bulk structural energy of extreme value centered by) Distribution, and search out its extreme point position, then according to the coordinate (v of this point₀,τ₀) simulate in time-space domain corresponding The whilst on tour t of lineups, i.e. its Suo Jingge road_n, it is represented by

$t_n=\sqrt{{{\mathrm{\τ}}_0}^2+{x_n}^2/{v_0}^2}$

Then with the extreme value place (v of former tracking₀,τ₀(x, local velocity's spectrum t), one to create protocol d centered by) Fixed speed and time range interscan then can obtain the accurate parameter v ' of these lineups₀With τ '₀, thus realize many subwaves homophase The optimization tracing process of axle.

For the horizontal layer model shown in Fig. 1, obtain many subwaves record (its based on Surface-related multiple Forecasting Methodology The 500th corresponding CMP records as shown in Fig. 3 (b)), and create repeatedly wave velocity spectrum (see Fig. 3 (c)) based on many subwaves record, The most on this basis, utilize contour tracing method to obtain the scope of many subwaves lineups stack power group, and utilize optimization Method for tracing determines many subwaves lineups in protocol (see the curve in Fig. 3 (d) 1.～2.).

(3) obtain each many subwaves lineups through the whilst on tour t of seismic channel_mAfter, utilize short time-window FK apparent velocity to filter Method from protocol d (x, t) and many subwaves record m (x, t) in eliminated, obtain the standard after removing many subwaves lineups once Ripple record d_p(x, t) with residue many subwaves record m_l(x, t), and the many subwaves lineups got rid of synthesize the repeatedly ripple record that is as the criterion d_m(x, t) with hyperbolic many subwaves record m_h(x,t)；

Shown in short time-window FK apparent velocity filtering such as Fig. 4 (a), the lineups in Fig. 3 are implemented short time-window the most respectively Obtain the quasi-primary wave record (as shown in Fig. 4 (b)) after removing tracking lineups after FK apparent velocity Filtering Processing and standard is repeatedly Ripple record (as shown in Fig. 4 (c)).In figure, the many subwaves lineups in quasi-primary wave record are completely eliminated, are substantially not present Remaining many subwaves information, but be 2. bordering on primary wave lineups due to many subwaves lineups and overlap, cause primary wave information to be subject to Slight damage (as shown in Fig. 4 (b) rectangular area), the information of its damage is present in accurate many subwaves record (such as Fig. 4 (c) rectangle Shown in region).

2) Wiener Filter Method pin alignment many subwaves record d of extension is utilized_m(x, t) with quasi-primary wave record d_p(x t) is carried out Adaptive attenuation again, to leach accurate many subwaves record d_m(x, t) in primary wave information, and eliminate further quasi-primary wave note Record d_p(x, t) in the many subwaves of remnants, corresponding processing procedure is expressed as

p₁(x, t)=ELSF [d_m(x,t),m_h(x,t)]

p₂(x, t)=ELSF [d_p(x,t),m_l(x,t)]

Wherein, ELSF represents extended least squares filter operator；p₁(x, t) in comprise be by record d_m(x, t) in leach Primary wave information；p₂(x, t) for eliminating record d_p(x, t) in the result of remaining many subwaves, the record after final MULTIPLE ATTENUATION It is shown below:

P (x, t)=p₁(x,t)+p₂(x,t)

In formula, (x, t) for the final result of combination adaptive attenuation for p.

For the model shown in Fig. 1, the 500th CMP based on many subwaves combination adaptive attenuation result records such as Fig. 5 Shown in (a).From Fig. 5 (a), after expanded Wiener filtering processes, the primary wave information of former damage has obtained preferable compensation. Fig. 5 (b) is the result by multiple tracks Wiener filtering MULTIPLE ATTENUATION, and it is adaptive that Fig. 5 (c) and (d) respectively illustrate the combination of many subwaves Answer the repeatedly wave component that damped method and multiple tracks Wiener Filter Method cut.Comparison diagram 5 (a) and Fig. 5 (b) and Fig. 5 (c) and Fig. 5 (d) Understanding, when many subwaves lineups and primary wave lineups are close, tradition multiple tracks Wiener Filter Method can frequently can lead to primary wave The damage of information, and many subwaves combination adaptive attenuation rule can preferably keep primary wave while effective multiple suppression Information.

Pluto 1.5 model be one for checking Surface-related multiple prediction and the master pattern of attenuating.Should Model seabed changes acutely from shallow to deep, and middle part exists salt dome structure at a high speed, causes major part wave impedance interface distortion, and shape Become and overlap tomography more, and the stratum of salt dome bottom is squeezed and acts on and become a series of fold building.Fig. 6 (a) is Pluto 1.5 mould 703rd original CMP record of type, Fig. 6 (b) is that many subwaves record of its correspondence is (pre-through Surface-related multiple Forecasting Methodology Record) CMP.It will be appreciated from fig. 6 that most many subwaves lineups remain as hyperbolic in Pluto 1.5 model protocol Form, but also grow and have the many subwaves of complexity not meeting hyperbola rule on a small quantity (relevant, such as Fig. 6 with salt brae, bottom interface In (a), (b) shown in arrow).

Many subwaves combination adaptive attenuation method is applied respectively protocol to be carried out repeatedly with tradition multiple tracks Wiener Filter Method Wave attenuation processes.Fig. 7 shows and optimizes the quasi-primary wave record that tracking method obtains, the original note with shown in Fig. 6 (a) based on lineups Picture recording ratio, the many subwaves of wherein most are all eliminated, and the primary wave information in record retains intact, the complexity that only arrow points to Many subwaves lineups (being bordering on linear event) are remained.

For the many subwaves of complexity that compacting further is remaining, this example pin alignment primary wave record has carried out extension Wiener filtering Method decays again, and shown in acquired results such as Fig. 8 (a), and Fig. 8 (b) is the record after multiple tracks Wiener filtering MULTIPLE ATTENUATION, Fig. 9 (a) then respectively illustrates many subwaves combination many subwaves of cutting with multiple tracks Wiener Filter Method of adaptive attenuation method and becomes with (b) Point.From Fig. 8 and Fig. 9, after multiple tracks Wiener filtering attenuation processing, the many subwaves of complexity in former quasi-primary wave record by Eliminate (as shown in Fig. 8 (a) arrow), and its primary wave information is almost without damage；And tradition multiple tracks Wiener Filter Method processes After earthquake record in wherein, the reflected amplitude in deep the most on the weak side, its part primary wave reflection line-ups is by serious Damage (as shown in Fig. 8 (b) arrow).

Earthquake record after original seismic data and MULTIPLE ATTENUATION is carried out reverse-time migration imaging, corresponding depth section See Figure 10～Figure 12 respectively.In the original offset section shown in Figure 10, the lineups for strong many subwaves imaging that arrow points to, It disturbs the interpretive analysis to effectively structure.As shown in Figure 11, the inverse time after many subwaves combination damped system processes is inclined Moving in section, strong many subwaves lineups that arrow points to are completely eliminated, and after being processed by conventional multiple tracks Wiener Filter Method Migrated section in still have remaining imaging (as shown in Figure 12 arrow), and interface, part salt brae reflection line-ups is distorted (as shown in Figure 12 ellipse scope), in addition contrast Figure 11 and Figure 12 understand, Figure 12 wherein, deep placement configurations imaging energy obvious Being weaker than Figure 11, these phenomenons all illustrate that multiple tracks Wiener Filter Method can major injury primary wave reflection letter while rejecting many subwaves Number, many subwaves combination damped system then can preferably keep primary wave information while effectively rejecting all kinds of many subwaves.

Claims (1)

1. the combination of subwave more than kind adaptive attenuation method, it is characterised in that it specifically includes following steps:

1) the many subwaves matching attenuation process followed the trail of based on lineups optimization, including the following steps:

(1) (x t), is carried out at dynamic(al) correction and superposition to utilize Surface-related multiple Forecasting Methodology to obtain many subwaves record m Reason, obtains many subwaves stack velocity spectrum, and corresponding computing formula is

$E_m(v,\mathrm{\τ})=\left|\underset{n=1}{\overset{N}{\Σ}}m(x_n,t=\sqrt{{\mathrm{\τ}}^2+x_n^2/v^2})\right|$

In formula, n is Taoist monastic name, 1≤n≤N；x_nIt is the offset distance in the n-th road, to E_m(v, τ) carry out appropriateness smooth after, m (x, t) in folded Accekeration v₀, zero-offset time τ₀Hyperbola lineups, will spectrum E_m(v, τ) is upper to be formed with (v₀,τ₀Extreme value centered by) Bulk structural energy；

(2) application contour tracing method obtains spectrum E_mWith (v on (v, τ)₀,τ₀Centered by), the bulk structural energy of extreme value divides Cloth scope, and search out its extreme point position, then according to the coordinate (v of this point₀,τ₀) simulate the corresponding homophase in time-space domain The whilst on tour t of axle, i.e. its Suo Jingge road_n, it is expressed as

$t_n=\sqrt{{{\mathrm{\τ}}_0}^2+{x_n}^2/{v_0}^2}$

Then with the extreme value place (v of former tracking₀,τ₀(x, local velocity's spectrum t), certain to create protocol d centered by) Speed and time range interscan then can obtain the accurate parameter v ' of these lineups₀With τ '₀, thus realize many subwaves lineups Optimize tracing process；

(3) obtain each many subwaves lineups through the whilst on tour t of seismic channel_mAfter, utilize short time-window FK apparent velocity filter method from former Begin record d (x, t) and many subwaves record m (x, t) in eliminated, obtain the quasi-primary wave record after removing many subwaves lineups d_p(x, t) with residue many subwaves record m_l(x, t), and the many subwaves lineups got rid of synthesize the repeatedly ripple record d that is as the criterion_m(x,t) With hyperbolic many subwaves record m_h(x,t)；

2) Wiener Filter Method pin alignment many subwaves record d of extension is utilized_m(x, t) with quasi-primary wave record d_p(x t) is carried out again Adaptive attenuation, to leach accurate many subwaves record d_m(x, t) in primary wave information, and eliminate quasi-primary wave record d further_p (x, t) in the many subwaves of remnants, corresponding processing procedure is expressed as

p₁(x, t)=ELSF [d_m(x,t),m_h(x,t)]

p₂(x, t)=ELSF [d_p(x,t),m_l(x,t)]

Wherein, ELSF represents extended least squares filter operator；p₁(x, t) in comprise be by record d_m(x, t) in leach one Subwave information；p₂(x, t) for eliminating record d_p(x, t) in the result of remaining many subwaves, the record after final MULTIPLE ATTENUATION is as follows Shown in formula:

P (x, t)=p₁(x,t)+p₂(x,t)

In formula, (x, t) for the final result of combination adaptive attenuation for p.