A geological constraint-based three-dimensional seismic velocity model fusion method and system

Abstract

The application provides a kind of geological constraint-based three-dimensional seismic velocity model fusion method and system, belong to the field of applied geophysics seismic exploration.The method is by establishing a regular grid model, respectively to the different velocity model to be fused is carried out spatial resampling, and the space is divided into different regions, then the constraint factor model is established to fuse and splice different velocity models, obtain the fused velocity model.The application can better fuse and splice two kinds of velocity models of different data sources, form a complete 3D fusion velocity model, the application considers the data source and reliability of velocity model, also considers the influencing factors such as geological structure.The velocity model to be fused and spliced can have missing or overlapping in transition zone data, by setting constraint factor C, using the weighted method to process the fusion area, finally obtain a transition smooth model.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of applied geophysical seismic exploration, and particularly relates to a three-dimensional seismic velocity model fusion method and system based on geological constraints. BACKGROUND

[0002] In the field of oil exploration, seismic velocity is an important prerequisite and basis for carrying out seismic forward modeling and seismic migration imaging. Accurate seismic velocity can greatly improve the seismic imaging effect and resolution, and thus improve the prediction accuracy of reservoirs. However, due to the limitations of geological and seismic exploration conditions, it is usually impossible to directly obtain an accurate seismic velocity model.

[0003] There are currently various methods for obtaining a seismic velocity model: obtaining a deep velocity model by a full-wave inversion (FWI) method; obtaining a near-surface velocity model by a tomographic inversion method; obtaining a surrounding formation reflection velocity by VSP logging; obtaining a layer velocity of a certain set of stable formations by rock physics or logging interpretation, etc. The scales and accuracies of the velocities obtained by different methods are different, and the applicable ranges are also different. For migration imaging and forward modeling, a complete set of formation velocities from the surface to the target layer in the subsurface is usually required. Therefore, the velocities obtained by different methods need to be fused and spliced to generate a reliable velocity field for seismic migration imaging or forward modeling research.

[0004] Chinese patent publication CN109975876A discloses a modeling method for well-seismic fusion velocity model based on structural horizon, which includes the following steps: S1, inputting horizon depth, logging velocity and prior velocity in prior velocity model; S2, performing outlier removal and smoothing processing on logging velocity in step S1; S3, calculating the ratio of logging velocity in step S2 and prior velocity in step S1 on a large scale, and correcting the logging velocity; S4, selecting a reference well, projecting logging velocities of other wells along the horizon to the reference well to form a depth-unified velocity curve; S5, based on Gaussian basis function, well distribution and depth-unified velocity curve, calculating the interpolation weight of each well; S6, based on the depth-unified velocity curve, performing weighted interpolation on the interpolation point to obtain the preliminary interpolation velocity of the point; S7, projecting the interpolation velocity from the reference well to the interpolation point to obtain the interpolation result; S8, performing weighted interpolation on each interpolation point, and the interpolation result constitutes a well velocity model; S9, fusing the well velocity model with the prior velocity model to form a top-down well-seismic fusion velocity model; Chinese patent publication CN109884700A discloses a multi-information fusion seismic velocity modeling method, which includes the following steps: step 1, inputting information required for modeling; step 2, smoothing the sonic logging velocity; step 3, setting the number of virtual wells and constructing virtual wells; step 4, filling the velocity model along the structural dip with each logging data, then based on Gaussian basis function, calculating the weighting coefficient of each logging data, weighted summing to obtain a logging velocity interpolation model; step 5, fusing the logging velocity interpolation model with the prior migration velocity, and calculating the ratio of the result with the prior migration velocity, adjusting the result based on the ratio to obtain a middle-deep layer velocity model; step 6, based on the ray coverage of the near-surface model, determining the fusion top surface and fusion area of the near-surface model; step 7, fusing the near-surface velocity and the middle-deep layer velocity in the fusion area to obtain a unified full velocity field as the final result output, and exporting the relative error of the model and the logging velocity; Chinese patent publication CN104570102A discloses a fusion method for near-surface velocity model and middle-deep layer velocity model, which includes: (1) inputting near-surface velocity model, observation system, velocity analysis reference surface and velocity function or seismic data obtained by velocity analysis; (2) calculating CMP preferred velocity analysis reference surface according to the near-surface velocity model and the observation system; (3) performing velocity analysis on the preferred velocity analysis reference surface obtained in step (2) to obtain a velocity function representing the middle-deep layer velocity model; or using the velocity function obtained by velocity analysis on a non-preferred velocity analysis reference surface, correcting the velocity function according to the difference between the velocity analysis reference surface and the preferred velocity analysis reference surface to obtain a velocity function representing the middle-deep layer velocity model; (4) fusing the near-surface velocity model and the middle-deep layer velocity model.Chinese patent publication CN104536043A discloses a deep domain overall velocity model fusion method and device, which comprises: picking up first arrival wave data in seismic data; obtaining a near-surface velocity model by travel time tomography method according to the picked up first arrival wave data; performing time domain velocity analysis on the seismic data to obtain a root mean square velocity model; converting the root mean square velocity model into a depth domain interval velocity model; taking the depth domain interval velocity model as an initial model, optimizing the initial model to obtain a middle-deep layer velocity model; fusing the near-surface velocity model and the middle-deep layer velocity model to obtain a deep domain overall velocity model; and performing pre-stack depth migration on the deep domain overall velocity model.

[0005] The common fusion or splicing methods at present include a fixed depth splicing method: a depth value is given, a tomographic velocity model above the depth value and a full-wave inversion velocity model below the depth value are directly selected, and then the two model velocities are spliced, which is simple and easy to implement, but has a disadvantage that the splicing may not conform to the actual geological law, and a false velocity mutation phenomenon may exist in the splicing part, affecting the application effect of the velocity model; another method is a splicing method using a transition zone depth time window: different velocity models are spliced by using a smoothing function in the time window, which effectively avoids velocity mutation and makes the velocity transition more natural. However, for complex structures, especially for mountain front zones and high and steep structure regions, this method also has limitations such as difficulty in designing a time window function. SUMMARY

[0006] The present application aims to solve the problems in the prior art, and provides a three-dimensional seismic velocity model fusion method and system based on geological constraints, which is used in the field of seismic exploration, and has certain advantages and limitations. Different velocity models are fused and spliced together according to the source, distribution range and reliability degree of the data itself by establishing a reasonable constraint model method, so as to meet the needs of velocity model splicing for complex structures, and to generate a reliable velocity model for seismic migration imaging or seismic forward simulation research.

[0007] The present application is implemented by the following technical solutions:

[0008] In a first aspect, the present application provides a three-dimensional seismic velocity model fusion method based on geological constraints, which comprises: establishing a regular grid model, spatially resampling different velocity models to be fused, dividing the space into different regions, and then establishing a constraint factor model to fuse and splice the different velocity models to obtain a fused velocity model.

[0009] The present application is further improved in that the method comprises:

[0010] First step, determine the top surface, bottom surface and boundary of the established fusion velocity model, and establish a regular grid model G;

[0011] Second step, using the regular grid model G to perform spatial resampling on different velocity models that need to be fused;

[0012] Third step, dividing the regular grid model G into A1 space, A2 space and to-be-fused splicing space A1UA2;

[0013] Fourth step, establishing a constraint factor model C;

[0014] Fifth step, using the constraint factor model C to obtain the velocity value V (i,j,k) of all grid cells in the to-be-fused splicing space A1UA2, and obtaining the fused velocity model;

[0015] Sixth step, outputting the fused velocity model.

[0016] Further improvement of the application is that the operation of determining the top surface, bottom surface and boundary of the established fusion velocity model in the first step comprises:

[0017] Determining the top surface, bottom surface and boundary of the established fusion velocity model according to geological information;

[0018] The geological information comprises the range of the study area, the surface elevation and the depth of the target layer.

[0019] Further improvement of the application is that the operation of establishing the regular grid model G in the first step comprises:

[0020] Establishing a regular cuboid grid model, and dividing the regular cuboid grid model into M, N, P grid cells in X, Y, Z directions respectively, wherein i, j, k represent the numbering of the grid cells in the three directions in space, i.e., the X, Y, Z direction grid cell coordinates;

[0021] The regular cuboid grid model is the regular grid model G;

[0022] The regular grid model G can enclose the boundary, top surface and bottom surface of the fusion velocity model.

[0023] Further improvement of the application is that the first step further comprises:

[0024] For the grid on the regular grid model G located outside the boundary of the fusion velocity model, the grid is set as an invalid grid, i.e., the value of each grid cell of the invalid grid is G (i,j,k) =0;

[0025] The remaining grids are valid grids, i.e., the value of each grid cell of the valid grid is G (i,j,k) =1.

[0026] Further improvement of the present application is that the operation of the second step comprises:

[0027] Comparing the regular grid model G with the first velocity model V1, if the grid cell size of the regular grid model G is greater than or equal to the grid cell size of V1, the arithmetic mean of the velocity values of all grid cells in V1 contained by the grid cell G(i,j,k) is calculated, and the calculated arithmetic mean is assigned to the grid cell G(i,j,k) as the resampled value of the regular grid G at this point, denoted as V1(i,j,k); if the grid cell size of the regular grid G is less than the grid cell size of V1, the velocity value of the V1 grid cell where the grid cell G(i,j,k) is located is directly assigned to G(i,j,k), or the arithmetic mean of the velocity values of all V1 grid cells intersected by the grid cell G(i,j,k) is assigned to G(i,j,k) as the resampled value of the regular grid model G at this point, denoted as V1(i,j,k);

[0028] Comparing the regular grid model G with the second velocity model V2, if the grid cell size of the regular grid model G is greater than or equal to the grid cell size of V2, the arithmetic mean of the velocity values of all grid cells in V2 contained by the grid cell G(i,j,k) is calculated, and the calculated arithmetic mean is assigned to the grid cell G(i,j,k) as the resampled value of the regular grid G at this point, denoted as V2(i,j,k); if the grid cell size of the regular grid G is less than the grid cell size of V2, the velocity value of the V2 grid cell where the grid cell G(i,j,k) is located is directly assigned to G(i,j,k), or the arithmetic mean of the velocity values of all V2 grid cells intersected by the grid cell G(i,j,k) is assigned to G(i,j,k) as the resampled value of the regular grid model G at this point, denoted as V2(i,j,k).

[0029] Further improvement of the present application is that the operation of the third step comprises:

[0030] According to the geological influencing factors and the distribution ranges of the first velocity model V1 and the second velocity model V2 to be fused, the boundaries of A1, A2 and A1UA2 are determined, and then the regular grid model G is divided into A1 space, A2 space and A1UA2 space to be fused according to the boundaries of A1, A2 and A1UA2;

[0031] In the A1 space, the velocity value V1 of the original first velocity model is adopted;

[0032] In the A2 space, the velocity value V2 of the original second velocity model is adopted;

[0033] The speed value of the to-be-fused splicing space A1UA2 is to be determined;

[0034] The geological influencing factors include a shallow structure model, a deep structure model, and a sedimentary facies model.

[0035] Further improvement of the present application is that the operation of the fourth step comprises:

[0036] The constraint factor model C is a grid model identical to the rule grid model G;

[0037] C (i,j,k) is the weight value of the (i, j, k) grid unit on the constraint factor model C, i, j, k represent the numbering of the grid unit in three directions of space, namely the X, Y and Z direction grid unit coordinates;

[0038] C (i,j,k) is in the range of 0-1;

[0039] When the (i, j, k) grid unit is located in the A1 space, C (i,j,k) = 1 is set;

[0040] When the (i, j, k) grid unit is located in the A2 space, C (i,j,k) = 0 is set;

[0041] When the (i, j, k) grid unit is located in the to-be-fused splicing space A1UA2, assuming that the distances of the grid unit to the nearest grid units of the A1 space and the A2 space are d1 and d2 respectively, C (i,j,k) = d2 / (d1+d2) is set.

[0042] Further improvement of the present application is that the operation of the fifth step comprises:

[0043] The following operation is performed on each grid unit in the to-be-fused splicing space A1UA2 to obtain the speed value V (i,j,k) of each grid unit:

[0044] (1) When only the V1(i, j, k) value exists in the grid unit, V (i,j,k) = V1 (i,j,k) ;

[0045] (2) When only the V2(i, j, k) value exists in the grid unit, V (i,j,k) = V2 (i,j,k) ;

[0046] (3) When the V1(i, j, k) and V2(i, j, k) values exist in the grid unit simultaneously,

[0047] V (i,j,k) = C (i,j,k) *V1(i,j,k) + (1 - C (i,j,k) ) * V2 (i,j,k) ;

[0048] (4) When the grid cell does not exist V1(i,j,k) and V2(i,j,k) values,

[0049] then V (i,j,k) = C (i,j,k) * V1 (最近点) + (1 - C (i,j,k) ) * V2 (最近点) ;

[0050] Wherein, V1 (最近点) and V2 (最近点) respectively represent the speed value of the grid cell in the V1 space closest to the grid cell, the grid cell in the V2 space;

[0051] After obtaining the speed value of all grid cells in the to-be-fused splicing space A1UA2, the fused speed model is obtained.

[0052] In a second aspect, the application provides a three-dimensional seismic velocity model fusion system based on geological constraints, which comprises a memory, a processor, and a computer program stored in the memory, wherein the computer program is executed by the processor to perform the following steps:

[0053] Firstly, the top surface, bottom surface and boundary of the established fusion speed model are determined, and a regular grid model G is established;

[0054] Secondly, the different speed models to be fused are spatially resampled by using the regular grid model G;

[0055] Thirdly, the regular grid model G is divided into A1 space, A2 space and to-be-fused splicing space A1UA2;

[0056] Fourthly, a constraint factor model C is established;

[0057] Fifthly, the speed value V (i,j,k) of all grid cells in the to-be-fused splicing space A1UA2 is obtained by using the constraint factor model C, so as to obtain the fused speed model;

[0058] Sixthly, the fused speed model is outputted.

[0059] Compared with the prior art, the present application has the advantages that the present application can better fuse and splice two velocity models from different data sources to form a complete 3D fused velocity model. The present application considers both the data sources and reliability of the velocity model and the influence factors such as geological structure. The velocity models to be fused and spliced can have missing data or overlapping data in the transition zone, and by setting a constraint factor C, the fusion region is processed by using a weighting method to finally obtain a transition smooth model. The present application can splice the models of complex structures, especially in the case of high and steep structures in the mountain front or on the surface, the region to be fused and processed can be arbitrarily selected, the difficulty of fusion and splicing is effectively reduced, and the reliability of the fused and spliced velocity is improved. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 Step block diagram of the method of the present application;

[0061] Figure 2 Regular grid model G, wherein G (i,j,k) is the value of the (i, j, k) grid cell;

[0062] Figure 3 Constraint grid model C, wherein C (i,j,k) is the weight value of the (i, j, k) grid cell;

[0063] Figure 4 Regular grid G spatial segmentation method, wherein A1, A2 are spaces determined by velocities V1, V2; A1UA2 is the region to be fused and spliced, which is a single isolated closed space;

[0064] Figure 5 Regular grid G spatial segmentation method, wherein A1, A2 are spaces determined by velocities V1, V2; A1UA2 is the region to be fused and spliced, which is a single isolated closed space;

[0065] Figure 6 Space A1 velocity distribution is V1, space A2 velocity distribution is V2, and space A1UA2 is the velocity distribution V to be fused;

[0066] Figure 7 Space A1 velocity distribution is V1, space A2 velocity distribution is V2, and space A1UA2 is the velocity distribution V to be fused; DETAILED DESCRIPTION

[0067] The present application will be further described in detail below with reference to the accompanying drawings:

[0068] In view of different data sources and spatial distribution of different velocity models, the application provides a three-dimensional velocity model fusion and splicing method.

[0069] Specifically, the space is divided into A1, A2 and A1UA2 three different space regions, wherein A1 space adopts the velocity value V1 of the original first velocity model, A2 space adopts the velocity value V2 of the original second velocity model, and the space A1UA2 to be fused and spliced is valued according to needs. A constraint factor model C is calculated, wherein C(i,j,k) is the weight value of the (i,j,k) grid unit of the constraint model, the fusion velocity V is calculated according to V1 and V2, and finally the fusion velocity model is obtained.

[0070] The implementation of the method of the application is as follows:

[0071]

Example 1

[0072] As shown in Figure 1 , the method of the application comprises:

[0073] Firstly, the top surface, bottom surface and boundary of the established fusion velocity model are determined according to the range of the research area, the ground elevation, the depth of the target layer and other information, and the top surface, bottom surface and boundary can be designed and adjusted according to actual needs.

[0074] Then, a regular grid model G is established, as shown in Figure 2 , which is divided into M, N, P grid units in X, Y, Z directions respectively, wherein i, j, k represent the numbering of the grid units in the three directions of space, i.e. the grid unit coordinates in X, Y, Z directions, and the size of the divided grid can be set according to the accuracy needs.

[0075] The regular grid model G is a regular cuboid grid model, and the spatial distribution range of the grid unit should be ensured to be greater than the spatial range of the research area model, i.e. the boundary, top surface and bottom surface of the fusion velocity model determined in the first step can be surrounded in the regular grid model G.

[0076] For the grid outside the boundary of the fusion velocity model on the regular grid model G, it is set as an invalid grid, i.e. the value of each grid unit of the invalid grid is set as zero. (i,j,k)= 0, no display and no operation, the rest of the grid is valid grid, that is, the value G of each grid unit of the valid grid (i,j,k) = 1, the fusion velocity model is contained in the valid grid. The grid outside the model boundary is invalid and does not participate in the operation.

[0077] Secondly, the different velocity models to be fused are spatially resampled by using the regular grid model G.

[0078] The spatial resampling is specifically as follows:

[0079] The regular grid model G is compared with the first velocity model V1. If the grid unit size of the regular grid model G is greater than or equal to the grid unit size of V1, the arithmetic mean of the velocity values of all the grid units (including the units intersecting the grid where G(i,j,k) is located) in V1 contained in the grid where G(i,j,k) is located is calculated (that is, the velocity values of these grid units are added and then divided by the number of the grid units), and the calculated arithmetic mean is assigned to the grid unit G(i,j,k) as the resampled value of the regular grid G at the point, denoted as V1(i,j,k). If the grid unit size of the regular grid G is less than the grid unit size of V1, the velocity value of the V1 grid unit where G(i,j,k) is located is directly assigned to G(i,j,k), or the arithmetic mean of the velocity values of all the V1 grid units intersecting G(i,j,k) is assigned to G(i,j,k) as the resampled value of the regular grid model G at the point, denoted as V1(i,j,k).

[0080] Similarly, the regular grid model G is compared with the second velocity model V2. If the grid unit size of the regular grid model G is greater than or equal to the grid unit size of V2, the arithmetic mean of the velocity values of all the grid units (including the units intersecting the grid where G(i,j,k) is located) in V2 contained in the grid where G(i,j,k) is located is calculated (that is, the velocity values of these units are added and then divided by the number of the units), and the calculated arithmetic mean is assigned to the unit G(i,j,k) as the resampled value of the regular grid G at the point, denoted as V2(i,j,k). If the grid unit size of the regular grid G is less than the grid unit size of V2, the velocity value of the V2 grid unit where G(i,j,k) is located is directly assigned to G(i,j,k), or the arithmetic mean of the velocity values of all the V2 grid units intersecting G(i,j,k) is assigned to G(i,j,k) as the resampled value of the regular grid model G at the point, denoted as V2(i,j,k).

[0081] Figure 1 V1 (i,j,k) , V2 (i,j,k) , V (i,j,k)These are the velocity values ​​of the (i,j,k)th grid cell after resampling according to the first velocity model rules, the velocity values ​​of the (i,j,k)th grid cell after resampling according to the second velocity model rules, and the velocity values ​​of the (i,j,k)th grid cell in the fused velocity model, respectively.

[0082] The third step involves determining the boundaries of A1, A2, and A1UA2 based on geological influencing factors such as shallow structural models, deep structural models, and sedimentary facies models, as well as the distribution range of the velocity models V1 and V2 to be merged. Then, the regular grid model G is divided into multiple spatial regions A1, A2, and A1UA2 based on the boundaries of A1, A2, and A1UA2.

[0083] Where A1 represents the distribution space of the first velocity model V1, A2 represents the distribution space of the second velocity model V2, and A1UA2 represents the spatial region that needs to be merged or spliced, serving as the transition zone between A1 and A2. A1UA2 can be an isolated, closed space, such as... Figure 4 As shown, it can also be multiple isolated enclosed spaces, such as Figure 5 As shown, the boundary of this spatial region is determined by seismic interpretation and geological understanding. It can be a separate closed continuous surface, or it can be connected to the model boundary to form a spatially closed continuous surface.

[0084] The fourth step is to establish a constraint factor model C based on the regular mesh model G, wherein the regular mesh model G (e.g., ...) Figure 3 The model shown is the same mesh model as the constraint factor model C. The difference is that G represents the newly established regular mesh. When G(i,j,k) = 0, it means that this mesh cell is invalid and no operation is performed. When G(i,j,k) = 1, it means that this mesh cell is valid and can be assigned values. Only when G(i,j,k) = 1 is it valid to assign values ​​to the constraint mesh C(i,j,k).

[0085] C represents the constrained mesh model, C (i,j,k) The weight value is used to constrain the (i,j,k)th grid cell in the model, where i,j,k represent its index in the three directions of the grid, i.e., the grid cell coordinates in the X, Y, and Z directions. (i,j,k) The value of C ranges from 0 to 1. (i,j,k) The value is determined based on the specific situation: when the (i,j,k) grid cell is located in space A1, C is set. (i,j,k) =1; When the (i,j,k) grid cell is located in space A2, set C (i,j,k) =0; When the (i,j,k) grid cell is located in space A1UA2, assuming the distances of this cell to the nearest points in space A1 and A2 are d1 and d2 respectively, then C (i,j,k) =d2 / (d1+d2).

[0086] In the fifth step, the following operation is performed on each grid cell in the A1UA2 range to obtain the velocity value V of each grid cell in the A1UA2 range (i,j,k) , to obtain the updated velocity model, i.e., the fused velocity model.

[0087] (1) When there is only a V1 value in the grid cell in the A1UA2 fusion range, i.e., V1(i,j,k)≠Null and V2(i,j,k)=Null, V (i,j,k) =V1 (i,j,k) ;

[0088] (2) When there is only a V2 value in the grid cell in the A1UA2 fusion range, i.e., V1(i,j,k)=Null and V2(i,j,k)≠Null, V (i,j,k) =V2 (i,j,k) ;

[0089] (3) When there are both V1(i,j,k) and V2(i,j,k) values in the grid cell in the A1UA2 fusion range (when the spatial distribution of the original velocity model of V1 and V2 is relatively large, there will be overlapping places in space, and at this time, after resampling, there may be both V1 values and V2 values in the same grid cell (i,j,k) in the A1UA2 fusion range, and from the geological factor, it is reasonable to have both, and it is not appropriate to delete any value), i.e., V1(i,j,k)≠Null and V2(i,j,k)≠Null,

[0090] V (i,j,k) =C (i,j,k) *V1 (i,j,k) +(1-C (i,j,k) )*V2 (i,j,k) ;

[0091] (4) When there are no V1(i,j,k) and V2(i,j,k) values in the grid cell in the A1UA2 fusion range (when the original spatial distribution of V1 and V2 has no contact, after resampling, some grid cells in space have neither V1(i,j,k) values nor V2(i,j,k) values, i.e., the regular grid cell G(i,j,k) has no velocity assignment), i.e., V1(i,j,k)=Null and V2(i,j,k)=Null,

[0092] V (i,j,k) =C (i,j,k) *V1 (最近点) +(1-C (i,j,k) )*V2 (最近点) ;

[0093] V1 (最近点)and V2 (最近点) respectively represent the velocity values of the grid cells in V1 space and V2 space closest to the current (i,j,k) grid cell.

[0094] Sixth step, output the fused velocity model.

[0095] The present application also provides a geological constraint-based three-dimensional seismic velocity model fusion system, and the implementation of the system is as follows:

[0096] [Example Two]

[0097] The system comprises a memory, a processor, and a computer program stored in the memory, and the computer program performs the following steps when executed by the processor:

[0098] First step, determine the top surface, bottom surface and boundary of the established fused velocity model, and establish a regular grid model G;

[0099] Second step, use the regular grid model G to perform spatial resampling on different velocity models that need to be fused;

[0100] Third step, divide the regular grid model G into A1 space, A2 space and to-be-fused splicing space A1UA2;

[0101] Fourth step, establish a constraint factor model C;

[0102] Fifth step, use the constraint factor model C to obtain the velocity values V of all grid cells in the to-be-fused splicing space A1UA2; (i,j,k) to obtain the fused velocity model;

[0103] Sixth step, output the fused velocity model.

[0104] The implementation of the application is as follows:

[0105] [Example Three]

[0106] Here, an example of establishing a whole velocity model from the ground surface to the deep layer is taken to illustrate. As shown in Figure 6 , the near-surface velocity model V1 and the deep-layer migration velocity model V2 need to be fused and spliced.

[0107] The near-surface velocity model V1 is obtained by seismic tomography inversion of the first arrival of the big gun, and contains the velocity information of the near-surface. Since the micro-logging data is used for constraint near the surface, the velocity near the surface is reliable, and the reliability of the velocity decreases with the increase of the depth. V2 is a deep-layer velocity model obtained by full-wave inversion, and the deep-layer velocity is relatively reliable, and the velocity near the near-surface is missing.

[0108] First, a rule grid model is established, and then the velocity models V1 and V2 are resampled, and then the models are divided into A1, A2 and A1UA2 three spatial parts according to the V1 and V2 distribution range and the stratum contact relationship, which correspond to the velocities V1, V2 and the to-be-fused velocity V respectively. Considering that the V1 and V2 velocity data of the to-be-fused space A1UA2 part are less reliable, the V1 and V2 grid values in the A1UA2 range are set to Null (i.e., step (1) in the fifth step above is completed), as shown in Fig. 1. Figure 6

[0109] Second, assuming that (i,j,k) is any grid cell in the space A1UA2, the distances of the grid cell to the nearest V1 and V2 are calculated, which are d1 and d2 respectively, and the constraint factor C is calculated (i,j,k) = d2 / (d1+d2). This operation is performed on all grids in the space A1UA2 to obtain the C values of all grid cells.

[0110] Third, the V values of all grids in the space A1UA2 are calculated, and the calculation formula is as follows:

[0111] V (i,j,k) = C (i,j,k) *V1 (最近点) +(1-C (i,j,k) )*V2 (最近点)

[0112] V1 (最近点) and V2 (最近点) represent the velocity values of the grid cells in the V1 space and the V2 space closest to the current (i,j,k) grid cell respectively.

[0113] Finally, the fused velocity model V is obtained, as shown in Fig. 2. Figure 7

[0114] The application belongs to the field of applied geophysical seismic exploration. It can be used to splice and fuse different velocity models to form a more complete velocity field model for seismic imaging migration or seismic forward simulation research.

[0115] Finally, it should be noted that the above technical solutions are only one embodiment of the application, and for those skilled in the art, on the basis of the application disclosed application method and principle, various types of improvements or modifications can be easily made, and are not limited to the methods described in the above embodiment, therefore, the above described method is only preferred, and does not have the meaning of limitation.​​

Claims

1. A method for fusing three-dimensional seismic velocity models based on geological constraints, characterized in that: The method involves establishing a regular mesh model, spatially resampling different velocity models to be fused, dividing the space into different regions, and then establishing a constraint factor model to fuse and stitch the different velocity models together to obtain a fused velocity model; including: The first step is to determine the top, bottom, and boundary of the fusion velocity model and establish a regular mesh model. G ; The second step is to utilize a regular grid model. G Spatial resampling is performed on the different velocity models that need to be fused, including: comparing the regular mesh model G with the first velocity model V1; if the mesh cell size of the regular mesh model G is greater than or equal to the mesh cell size of V1, then the mesh cell G is resampled. (i,j,k) The arithmetic mean of the velocity values ​​of all grid cells in V1 is calculated, and the calculated arithmetic mean is assigned to the grid cell G. (i,j,k) V1 is the value of the resampled regular grid G ​​at that point. (i,j,k) ,in i,j,k Represents the numbering of the grid cells in three spatial directions, i.e. X,Y,Z Orientation of grid cell coordinates; if the grid cell size of regular grid G ​​is smaller than the grid cell size of V1, then grid cell G... (i,j,k) The velocity value of the V1 mesh cell is directly assigned to G. (i,j,k) Or it will be with grid cell G (i,j,k) The arithmetic mean of the velocity values ​​of all intersecting V1 mesh cells is assigned to G. (i,j,k) V1 is the value of the resampled regular mesh model G at that point. (i,j,k) ; Compare the regular mesh model G with the second velocity model V2. If the mesh cell size of the regular mesh model G is greater than or equal to the mesh cell size of V2, then for mesh cell G... (i,j,k) The arithmetic mean of the velocity values ​​of all mesh elements in V2 is calculated, and the calculated arithmetic mean is assigned to mesh element G. (i,j,k) V2 is the value of the resampled regular grid G ​​at that point. (i,j,k) If the mesh cell size of the regular mesh G is smaller than the mesh cell size of V2, then mesh cell G will be... (i,j,k) The velocity value of the V2 mesh cell is directly assigned to G. (i,j,k) Or it will be with grid cell G (i,j,k) The arithmetic mean of the velocity values ​​of all intersecting V2 mesh elements is assigned to G. (i,j,k) V2 is the value of the regular mesh model G resampled at that point. (i,j,k) ; The third step involves dividing the regular grid model G into space A1, space A2, and a fusion space A1UA2. This includes determining the boundaries of A1, A2, and A1UA2 based on geological influencing factors and the distribution ranges of the first velocity model V1 and the second velocity model V2 to be fused. Then, the regular grid model G is divided into space A1, space A2, and fusion space A1UA2 based on these boundaries. The geological influencing factors include shallow structural models, deep structural models, and sedimentary facies models. The fourth step is to establish a constraint factor model. C ; Fifth step: Use constraint factor model C to obtain the space to be fused and stitched. A1UA2 velocity values ​​of all grid cells within V (i,j,k) The fused velocity model is obtained; The sixth step is to output the fused velocity model.

2. The method for fusing three-dimensional seismic velocity models based on geological constraints according to claim 1, characterized in that: The operation of determining the top, bottom, and boundary of the established fusion velocity model in the first step includes: The top, bottom, and boundaries of the established fusion velocity model are determined based on geological information; The geological information includes: the study area, surface elevation, and depth of the target layer.

3. The method for fusing three-dimensional seismic velocity models based on geological constraints according to claim 2, characterized in that: The operation of establishing the regular mesh model G in the first step includes: Establish a regular cuboid mesh model and divide it into M, N, and P mesh elements according to the X, Y, and Z directions, respectively; The rectangular grid model of this rule is the rule grid model G; The regular mesh model G can surround the boundary, top surface, and bottom surface of the fused velocity model.

4. The method for fusing three-dimensional seismic velocity models based on geological constraints according to claim 3, characterized in that: The first step further includes: For meshes on the regular mesh model G that lie outside the boundary of the fusion velocity model, they are designated as invalid meshes, i.e., the values ​​of each mesh element of the invalid mesh are... G (i,j,k) =0 ; The remaining grid cells are considered valid grid cells, meaning the values ​​of each cell in the valid grid cells are considered valid grid cells. G (i,j,k) =1 .

5. The method for fusing three-dimensional seismic velocity models based on geological constraints according to claim 4, characterized in that: Space A1 uses the velocity value V1 of the original first velocity model; Space A2 uses the velocity value V2 from the original second velocity model; The velocity value of the space to be merged and spliced, A1UA2, is yet to be determined.

6. The method for fusing three-dimensional seismic velocity models based on geological constraints according to claim 5, characterized in that: The fourth step includes the following operations: Establish constraint factor model C The constraint factor model C is a mesh model that is exactly the same as the regular mesh model G. C (i,j,k) For constraint factor model C (i,j,k) The weight values ​​of the grid cells, i,j,k Represents the numbering of the grid cells in three spatial directions, i.e. X,Y,Z Directional grid cell coordinates; C (i,j,k) The value of is between 0 and 1; when (i, j, k) Mesh cells are located A1 When in space, set C (i,j,k) =1 ; when (i, j, k) Mesh cells are located A2 When in space, set C (i,j,k) =0 ; when (i, j, k) The grid cells are located in the space to be merged and stitched together. A1UA2 At that time, assuming the distance of the grid cell is... A1 Space and A2 The distances to the nearest grid cell in space are respectively d1 , d2 ,but C (i,j,k) =d2 / (d1+d2) .

7. The method for fusing three-dimensional seismic velocity models based on geological constraints according to claim 6, characterized in that: The fifth step includes the following operations: Treating integrated splicing space A1UA2 The following operations are performed on each grid cell to obtain the velocity value of each grid cell. V (i,j,k) : (1) When only V1 exists in the mesh cell (i,j,k) When the value is [value], then V (i,j,k) = V1 (i,j,k) ; (2) When only V2 exists in the mesh cell (i,j,k) When the value is [value], then V (i,j,k) = V2 (i,j,k) ; (3) When the grid cell simultaneously has V1 (i,j,k) and V2 (i,j,k) hour, but V (i,j,k) = C (i,j,k) *V1 (i,j,k) + (1 - C (i,j,k) )*V2 (i,j,k) ; (4) When the grid cell does not have V1 (i,j,k) and V2 (i,j,k) When the value is, but V (i,j,k) = C (i,j,k) *V1 (最近点) + (1 - C (i,j,k) )*V2 (最近点) ; in, V1 (最近点) and V2 (最近点) These represent the velocity values ​​of the grid cells in space V1 and space V2 that are closest to this grid cell, respectively. Obtain the space to be merged and spliced A1UA2 The velocity model is obtained by taking the velocity values ​​of all grid cells within the model.

8. A three-dimensional seismic velocity model fusion system based on geological constraints, characterized in that: The system includes: a memory, a processor, and a computer program stored in the memory, the computer program being executed by the processor to perform the following steps: The first step is to determine the top, bottom, and boundary of the fusion velocity model and establish a regular mesh model G. The second step is to utilize a regular grid model. G Spatial resampling is performed on the different velocity models that need to be fused, including: Compare the regular mesh model G with the first velocity model V1. If the mesh cell size of the regular mesh model G is greater than or equal to the mesh cell size of V1, then for mesh cell G... (i,j,k) The arithmetic mean of the velocity values ​​of all grid cells in V1 is calculated, and the calculated arithmetic mean is assigned to the grid cell G. (i,j,k) V1 is the value of the resampled regular grid G ​​at that point. (i,j,k) If the mesh cell size of the regular mesh G is smaller than the mesh cell size of V1, then mesh cell G will be... (i,j,k) The velocity value of the V1 mesh cell is directly assigned to G. (i,j,k) Or it will be with grid cell G (i,j,k) The arithmetic mean of the velocity values ​​of all intersecting V1 mesh cells is assigned to G. (i,j,k) V1 is the value of the resampled regular mesh model G at that point. (i,j,k) ; Compare the regular mesh model G with the second velocity model V2. If the mesh cell size of the regular mesh model G is greater than or equal to the mesh cell size of V2, then for mesh cell G... (i,j,k) The arithmetic mean of the velocity values ​​of all mesh elements in V2 is calculated, and the calculated arithmetic mean is assigned to mesh element G. (i,j,k) V2 is the value of the resampled regular grid G ​​at that point. (i,j,k) If the mesh cell size of the regular mesh G is smaller than the mesh cell size of V2, then mesh cell G will be... (i,j,k) The velocity value of the V2 mesh cell is directly assigned to G. (i,j,k) Or it will be with grid cell G (i,j,k) The arithmetic mean of the velocity values ​​of all intersecting V2 mesh elements is assigned to G. (i,j,k) V2 is the value of the regular mesh model G resampled at that point. (i,j,k) ; The third step involves dividing the regular grid model G into space A1, space A2, and a fusion space A1UA2. This includes determining the boundaries of A1, A2, and A1UA2 based on geological influencing factors and the distribution ranges of the first velocity model V1 and the second velocity model V2 to be fused. Then, the regular grid model G is divided into space A1, space A2, and fusion space A1UA2 based on these boundaries. The geological influencing factors include shallow structural models, deep structural models, and sedimentary facies models. The fourth step is to establish a constraint factor model. C ; Fifth step: Use constraint factor model C to obtain the space to be fused and stitched. A1UA2 velocity values ​​of all grid cells within V (i,j,k) The fused velocity model is obtained; The sixth step is to output the fused velocity model.

Images