Intelligent radial interpolation seismic data processing method based on multi-objective ephemera algorithm

Abstract

The invention relates to the technical fields of tunnel engineering, engineering geophysical exploration, geophysics, seismic exploration and oil and gas exploration, and proposes an intelligent radial interpolation seismic data processing method based on a multi-objective mayfly algorithm, comprising: collecting well logging data and seismic data; when performing radial interpolation seismic gather processing, converting wide azimuth gathers into conventional gathers, and performing conventional gather optimization processing on each azimuth; initializing the parameters of the multi-objective mayfly algorithm, setting the stacking weight range of each azimuth; constructing the objective function of the multi-objective mayfly algorithm, and completing the initialization of the multi-objective mayfly algorithm; iterating using the multi-objective mayfly algorithm, and dynamically updating the stacking weight; extracting the mayfly position in the middle of the Pareto set, and using the stacking weight in the mayfly position as the optimal stacking weight for each azimuth. The invention adopts the multi-objective mayfly algorithm to determine the stacking weight of each azimuth of radial interpolation seismic gather data, and realizes high-fidelity processing of radial interpolation seismic gather data.

Description

Technical Field

[0001] The invention relates to the technical fields of tunnel engineering, engineering geophysical exploration, geophysics, seismic exploration and oil and gas exploration, and in particular to an intelligent radial interpolation seismic data processing method based on a multi-objective ephemera algorithm. Background Art

[0002] Seismic data preprocessing, also known as gather optimization, provides a good data foundation for subsequent inversion. High-quality seismic data preprocessing can maximize the effective signal of seismic data and improve the accuracy of subsequent inversion results. The accuracy of inversion results not only affects the reservoir identification rate in the oil and gas exploration industry, but also reduces the risk of water inrush and karst in tunnel engineering, engineering geophysical exploration, geophysics, seismic exploration, oil and gas exploration, and railway tunnel exploration.

[0003] Conventional seismic gather data only contains information in one azimuth, but it has met the needs of conventional exploration. In recent years, with the gradual improvement of seismic data acquisition technology, a variety of wide-azimuth seismic gather data have been generated, such as OVT seismic data and radial interpolation seismic gather data. Compared with conventional seismic gather data, wide-azimuth OVT gather data usually contains 4 to 6 azimuths, and each azimuth is an independent conventional seismic gather. In addition, wide-azimuth data also contains characteristics such as azimuth time difference and anisotropy, which are usually used to predict the strike, dip and dip of fractures, and can also realize multi-scale fracture prediction. However, the amplitude of OVT seismic data changes with azimuth. Radial interpolation seismic gather data are often generated by radial interpolation method. The number of azimuth acquisitions is generally lower than that of OVT data. Many azimuth gather data are realized by interpolation, which is suitable for scenes requiring high-precision spatial reconstruction, such as urban areas, complex geological structures or special exploration tasks. In general, wide-azimuth gather data contain more effective seismic signals.

[0004] At present, in conventional radial interpolation seismic gather data, it is necessary to convert the radial interpolation seismic gather data into conventional gathers (single azimuth), so the radial interpolation data needs to be stacked in different azimuths. Since the offset range of each azimuth at the same location of such data is different, it is not possible to stack based on different azimuths and offsets; since the number of coverage times of seismic data acquisition is fixed, it is necessary to convert the offset into angle gathers before stacking in different azimuths and angles. The traditional stacking operation is to directly add the weighted gather data of each azimuth angle. This commonly used stacking method is an effective means to convert wide azimuth gather data into conventional gather data. But in fact, the stacking weight of each azimuth is not uniform, because it is difficult to synchronize the gather data of each azimuth during the acquisition or processing of seismic data, which is bound to produce cumulative errors. Therefore, a more scientific method is needed to more intelligently determine the stacking weight of each azimuth.

[0005] With the introduction of intelligent optimization algorithms in this field, single-objective intelligent optimization algorithms can incorporate the superposition weight of each direction into the dimension of the population, and at the same time establish an objective function for solving in order to achieve the purpose of intelligently determining the superposition weight. Single-objective solution algorithms include: Genetic algorithm (GA), particle swarm optimization algorithm (PSO), whale optimization algorithm (WOA) and mayfly optimization algorithm (MA). In the update and improvement of the single-objective global nonlinear sneaker algorithm, relevant high-level scholars in the field of computer science have successively proposed multi-objective intelligent optimization algorithms, such as: multi-objective mayfly optimization algorithm (MMA), multi-objective genetic algorithm three generations (NSGA-III) and multi-objective African vultures optimization algorithm (AVOA).

[0006] In view of the particularity of determining the stacking coefficient of radial interpolation seismic gather data, it is necessary to establish multiple objective functions as constraints. On this basis, a multi-objective algorithm is used to intelligently determine the stacking weight of each azimuth, so that the stacked seismic angle gathers and the wellside synthetic seismic traces are more closely matched and more in line with the real laws of the strata, thereby ensuring the effective signal of the seismic data to the greatest extent and improving the resolution of the processed seismic data. Summary of the invention

[0007] The purpose of the present invention is to use a multi-objective ephemera algorithm to intelligently determine the stacking weight of each azimuth of radial interpolation seismic gather data, achieve high-fidelity processing of radial interpolation seismic gather data, provide strong data support for subsequent inversion, and provide an intelligent radial interpolation seismic data processing method based on a multi-objective ephemera algorithm.

[0008] In order to achieve the above-mentioned object of the invention, the embodiment of the present invention provides the following technical solutions:

[0009] The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm comprises the following steps:

[0010] Step 1, load the layer data and well location, and collect well logging data and seismic data;

[0011] Step 2, when performing radial interpolation seismic gather processing, convert the wide azimuth gather into a conventional gather, and perform conventional gather optimization processing on each azimuth;

[0012] Step 3, initialize the parameters of the multi-objective mayfly algorithm, set the superposition weight range of each direction; construct the objective function of the multi-objective mayfly algorithm, and complete the initialization of the multi-objective mayfly algorithm;

[0013] Step 4, use the multi-objective ephemera algorithm to iterate, dynamically update the superposition weights, gradually optimize, and obtain the Pareto set;

[0014] Step 5, extract the mayfly position in the middle of the final Pareto set, and use the superposition weight in the mayfly position as the optimal superposition weight for each orientation.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] The intelligent radial interpolation seismic gather data processing based on the multi-objective ephemera algorithm of the present invention has a significant advantage over the prior art in that it uses an intelligent optimization algorithm to determine multiple objective functions as constraints, and intelligently determines the stacking weight of each azimuth, thereby achieving efficient processing of wide-azimuth seismic gather data. Compared with the single-objective ephemera algorithm that can only optimize one objective function, the multi-objective ephemera algorithm can simultaneously optimize multiple objective functions, thereby improving the balance of solving the objective function. This method helps to maximize the effective signal of seismic data, improve the accuracy of subsequent inversion results and the resolution of processed seismic data. Compared with the traditional linear direct addition method, the present invention can more accurately determine the stacking weight of each azimuth, thereby reducing cumulative errors and improving the quality and accuracy of seismic data. This has a positive impact on the subsequent high-quality reservoir identification rate and the risk in exploration. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 It is a flow chart of a radial interpolation seismic gather data processing method in an embodiment of the present invention;

[0019] Figure 2 This is a spectrum analysis diagram of the 75° azimuth as conventional gather data before and after bandpass filtering in an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of Radon transformation of the 15° azimuth of the M1 well as conventional gather data in an embodiment of the present invention. Figure 3 (a) is the data waveform of the conventional gather data at 15° azimuth of M1 well before Radon transformation. Figure 3 (b) is the data waveform of the conventional gather data at 15° azimuth of M1 well after Radon transformation. Figure 3 (c) in the figure is the noise component;

[0021] Figure 4 This is a schematic diagram of the inverse Q filtering of the M2 well 45° azimuth as conventional gather data in an embodiment of the present invention. Figure 4 (a) is the data waveform of the conventional gather data at 45° azimuth of M2 well before inverse Q filtering. Figure 4 (b) is the data waveform of the conventional gather data at 45° azimuth of M2 well after inverse Q filtering;

[0022] Figure 5 This is a schematic diagram of deconvolution of CDP123 at 135° as conventional gather data in an embodiment of the present invention. Figure 5 (a) is the data waveform of CDP123 channel 135° azimuth conventional gather data before deconvolution. Figure 5 (b) is the data waveform after deconvolution of the conventional gather data of CDP123 channel 135° azimuth;

[0023] Figure 6 Schematic diagram of the Pareto set at the 50th iteration of the multi-objective ephemera algorithm in an embodiment of the present invention;

[0024] Figure 7 Schematic diagram of the Pareto set at the 100th iteration of the multi-objective ephemera algorithm in an embodiment of the present invention;

[0025] Figure 8Schematic diagram of the Pareto set at the 200th iteration of the multi-objective ephemera algorithm in an embodiment of the present invention;

[0026] Fig. 9 Schematic diagram of the Pareto set at the 300th iteration of the multi-objective ephemera algorithm in an embodiment of the present invention;

[0027] Fig.10 Schematic diagram of the convergence curve in the iteration of the single-target mayfly algorithm in the embodiment of the present invention;

[0028] Fig.11 It is a schematic diagram of seismic data slices along the horizon in an embodiment of the present invention, Fig.11 (a) is a schematic diagram of the slice along the horizon of the original observed seismic data. Fig.11 (b) is a schematic diagram of the slice along the horizon of the seismic data obtained by stacking with the initial stacking weight. Fig.11 (c) is a schematic diagram of slices along the stratigraphic level of seismic data obtained by stacking the stacking weights optimized by the multi-objective ephemera algorithm. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0030] It should be noted that similar numbers and letters represent similar items in the following figures, so once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance, or implying any such actual relationship or order between these entities or operations. In addition, the terms "connected", "connected", etc. can be directly connected between elements, or indirectly connected via other elements.

[0031] Embodiment 1:

[0032] The present invention is achieved through the following technical solutions: Figure 1As shown, the present invention proposes an intelligent radial interpolation seismic data processing method based on a multi-objective mayfly algorithm, which has obvious advantages over traditional technologies. The objective function of the multi-objective mayfly algorithm is determined as a constraint condition by an intelligent optimization algorithm, and the stacking weight of each azimuth is optimized by the multi-objective mayfly algorithm to achieve efficient processing of wide-azimuth seismic gather data. Compared with the optimization method of the single-objective mayfly algorithm, the multi-objective mayfly algorithm can optimize multiple objective functions at the same time, improve the balance of solving the objective function, and retain the effective signal of the seismic data to the greatest extent. Compared with the traditional linear direct addition method, the present invention can dynamically and accurately determine the stacking weight of each azimuth, reduce the cumulative error, improve the quality and accuracy of seismic data, have a positive impact on subsequent exploration work, and help to improve the management effect of reservoir identification rate and exploration risk. In summary, the method of the present invention has brought important improvements to improving the quality of seismic data processing, interpretation accuracy and exploration success rate, and is expected to bring more reliable data support and more effective exploration practices to the field of geological exploration.

[0033] The actual radial interpolation seismic data volume used in this embodiment has 6 azimuths (15°, 45°, 75°, 105°, 135° and 165°), the sampling interval is 0.002s, the time range is 2600-3000ms, and a total of 200 data points are sampled. The conventional angle gather interval range of each azimuth can be set to 1°-10° as a near angle, 11°-20° as a medium angle, and 21°-30° as a far angle.

[0034] The present invention is achieved through the following technical solutions: Figure 1 As shown, the method comprises the following steps:

[0035] Step 1: Load the layer data and well locations, and collect well logging data and seismic data.

[0036] The collected logging data include the longitudinal wave velocity logging data and density logging data; the collected seismic data include the radial interpolation seismic gather data and the layer data file.

[0037] The wells loaded in this embodiment are wells M1 to M5 in the XBOOS block. The P-wave velocity logging data and density logging data from wells M1 to M5 in the XBOOS block are collected for synthesizing wave impedance curves to calculate reflection coefficient sequences, and synthesizing seismic records through convolution operations for subsequent well-seismic calibration. Please refer to Table 1, which shows some logging data used for well-seismic calibration.

[0038] Table 1 Well logging data for establishing the initial model (partial)

[0039] Depth(m) P-wave velocity / (m / s) <![CDATA[Density / (g / cm 3 )]]> <![CDATA[Longitudinal wave impedance / (m / s) * (g / cm 3 )]]> 5647.12 4535.92 2.51 11365.8 5647.25 4524.6 2.48 11205.33 5647.38 4532.82 2.45 11115.13 5647.5 4546.35 2.45 11145.83 5647.62 4535.48 2.5 11320.34 5647.75 4477.1 2.52 11280.65 5647.88 4416.89 2.52 11125.9 5648 4319.63 2.5 10815.06 5648.12 4285.35 2.49 10672.33 5648.25 4274.64 2.48 10606.05 5648.38 4299.09 2.48 10658.84 5648.5 4344.52 2.48 10772.67

[0040] Step 2: When processing radial interpolation seismic gathers, the wide azimuth gathers are converted into conventional gathers, and conventional gather optimization processing is performed on each azimuth.

[0041] The radial interpolation seismic data volume of this embodiment includes six azimuths: 15°, 45°, 75°, 105°, 135° and 165°. First, each azimuth is treated as conventional gather data and bandpass filtered to filter out low-frequency surface wave interference in the original data. Taking the azimuth of 75° as an example, in the following example, Figure 2 In the spectrum diagram shown, it can be seen that the low-frequency part of the signal is successfully cut off after filtering.

[0042] Afterwards, each azimuth is treated as conventional gather data for optimization processing, including Radon transform, inverse Q filtering, deconvolution, gather flattening and gather conversion operations. Among them, Radon transform is mainly used to remove random noise and multiple wave noise in the data, such as Figure 3 (a) is the data waveform of the conventional gather data at 15° azimuth of M1 well before Radon transformation. Figure 3 (b) is the data waveform of the conventional gather data at 15° azimuth of M1 well after Radon transformation. Figure 3 (c) in the figure is the noise component. Inverse Q filtering is mainly used to compensate for the amplitude, such as Figure 4 (a) is the data waveform of the conventional gather data at 45° azimuth of M2 well before inverse Q filtering. Figure 4 (b) in the figure is the data waveform after the inverse Q filtering of the conventional gather data at 45° azimuth of M2 well. Deconvolution is mainly used to improve the vertical resolution of seismic data, such as Figure 5 (a) is the data waveform of CDP123 channel 135° azimuth conventional gather data before deconvolution. Figure 5 (b) is the data waveform after deconvolution of the conventional gather data of CDP123 channel 135° azimuth. The gather conversion operation means converting the offset gather into the angle gather.

[0043] After conventional gather optimization, stacking operations are performed in different directions and angles according to the initial stacking weights to obtain the initial stacking volume. For example, the first direction is stacked at three angles: near, medium, and far to obtain the initial stacking volume of the direction. The initial well-seismic calibration operation is then performed to establish an initial time-depth framework of well-seismic data to facilitate the subsequent construction of the objective function.

[0044] Step 3, initialize the parameters of the multi-objective mayfly algorithm, set the superposition weight range of each direction; construct the objective function of the multi-objective mayfly algorithm, and complete the initialization of the multi-objective mayfly algorithm.

[0045] Before iterating the multi-objective mayfly algorithm, it is necessary to initialize several key parameters of the multi-objective mayfly algorithm. Set the number of male mayflies q = 20, the number of female mayflies p = 20, the number of matings of the father generation mayflies Nc = 20, the number of mutations of the offspring generation mayflies Nm = 10, the gravity coefficient g = 0.8, the positive attraction constant a1 = 1, the positive attraction constant a2 = 1.5, the mutation coefficient Mu = 0.02, the corrected visibility coefficient Beta = 2, the random flight coefficient FL = 0.77, the wedding dance coefficient Dance = 0.77, the wedding dance coefficient reduction rate DD = 0.99, the random flight coefficient reduction rate FD = 0.99, the Pareto set capacity Np = 50, the maximum number of iterations t max =300.

[0046] Determine the size of the mayfly population and initialize the positions of male and female mayflies:

[0047] M={x i |x i =(x i1 , x i2 , …, x id ), i = 1, 2, ..., q}

[0048] F={y i |y i =(y i1 ,y i2 , ..., y id ), i=1, 2, ..., p}

[0049] Wherein, M represents the male mayfly population; F represents the female mayfly population; x i represents the position of the ith male mayfly; y i represents the position of the i-th female mayfly; q represents the number of male mayflies; p represents the number of female mayflies; d represents the total number of dimensions (d = 6); x id represents the position of the dth dimension of the ith male mayfly; y id Represents the position of the d-th dimension of the i-th female mayfly.

[0050] In the initialization stage of the multi-objective mayfly algorithm, the initial superposition weights need to be assigned to the weights of the first male mayfly's 1 to 6 dimensions. Table 2 shows the superposition weight boundary range settings for the mayfly population.

[0051] Table 2 Overlay weight boundary range setting of mayfly population

[0052] Dimensions Initial stacking weight Overlay weight lower limit Overlay weight limit 1 1 / 6 0.10 0.20 2 1 / 6 0.10 0.20 3 1 / 6 0.10 0.20 4 1 / 6 0.15 0.25 5 1 / 6 0.10 0.20

[0053] The objective function is established based on the stacking volume divided by azimuth and angle, which can ensure the balance of the final stacking weight and the convergence of the objective function. Based on this concept, this scheme constructs three objective functions, corresponding to the number of stacking volumes divided by angle, including near-angle stacking volume, medium-angle stacking volume and far-angle stacking volume:

[0054] Minimize the objective function J1:

[0055]

[0056] Minimize the objective function J2:

[0057]

[0058] Minimize the objective function J3:

[0059]

[0060] Where n represents the total number of wells, i = 1, 2, ..., n; corrcoef(.) represents the Pearson correlation coefficient of the two series; obs near(i) 、obs mid(i) 、obs far(i) Respectively represent the actual seismic records of the near-angle, medium-angle, and far-angle stacks of the ith well; syn near(i) ,syn mid(i) ,syn far(i) They represent the wellside synthetic seismic records of the near-angle, medium-angle, and far-angle stacks of the ith well, respectively.

[0061] If the sum of the stacking weights of the six directions is far greater than 1 or less than 1, it will inevitably affect the amplitude of the seismic record after the azimuth stacking. In order to achieve the purpose of "amplitude preservation", an additional objective function J4 is added to constrain the sum of the stacking weights:

[0062]

[0063] Among them, N azi Indicates the number of directions (N azi =6); represents the superposition weight of the j-th orientation.

[0064] According to the different optimization properties of the single-objective ephemera algorithm or the multi-objective ephemera algorithm, the objective functions corresponding to these two types of algorithms are established respectively. In the single-objective ephemera algorithm, the optimization process is completed by minimizing the objective function Singleobjective:

[0065] singleobjective=J1+J2+J3+J4

[0066] For example, the sub-item calculation values ​​in the single objective function Singleobjective based on the initial superposition weight are: J1 = 1.5434, J2 = 1.5989, J3 = 1.7240, J4 = 0, Singleobjective = 4.8663. Similarly, the calculation method of the single objective function of other mayfly populations is the same.

[0067] In the multi-objective mayfly algorithm, the optimization process is completed by minimizing the objective function Multiobjective:

[0068] Multiobjective=(J1,J2,J3,J4)

[0069] After updating the current position of the first male mayfly, the current objective function value is updated to (J1, J2, J3, J4) = (1.5434, 1.5989, 1.7240, 0). Afterwards, the current position of the second male mayfly in the initialization phase is calculated, and the current objective function value (J1, J2, J3, J4) = (1.5424, 1.5984, 1.7245, 0) is calculated according to the current position of the second male mayfly, and the historical position and historical objective function value of the second male mayfly are updated. The current positions and current objective function values ​​of all remaining male mayflies are calculated in the same way, and the historical optimal positions and historical optimal objective function values ​​of all male mayflies are updated.

[0070] Then, the current position of the first female mayfly is calculated, and the current objective function value (J1, J2, J3, J4) = (1.5424, J2 = 1.5981, 1.7264, 0) is calculated based on the current position of the first female mayfly, and the historical position and historical objective function value of the first female mayfly are updated. The current positions and current objective function values ​​of all remaining female mayflies are calculated in the same way, and the historical optimal positions and historical optimal objective function values ​​of all remaining female mayflies are updated.

[0071] Combine 20 male and 20 female mayflies, and perform a Pareto non-dominated sort on a total of 40 mayflies. Take the top 50 (Np=50) mayfly individuals as the Pareto set, and take all individuals that are less than 50 as the Pareto set.

[0072] The above completes the initialization of the multi-objective mayfly algorithm.

[0073] Step 4: Use the multi-objective ephemera algorithm to iterate, dynamically update the superposition weights, gradually optimize, and obtain the Pareto set.

[0074] Enter the main loop of the multi-objective mayfly algorithm, use the multi-objective mayfly algorithm to iterate, and update the position of the mayfly population (ie, the superposition weight).

[0075] Start the first iteration of the multi-objective mayfly algorithm (t=1):

[0076] Select the mayfly individual ranked first in the Pareto set as the leading mayfly;

[0077] The calculation starts from the female mayfly 1 in the initialization stage. At this time, the male mayfly 1 dominates the female mayfly 1. The Euclidean distance between the male mayfly 1 and the female mayfly 1 is calculated:

[0078]

[0079] Among them, r mf express and The Euclidean distance between represents the initial position of the jth dimension of male mayfly 1, Represents the initial position of the j-th dimension of female mayfly 1.

[0080] Afterwards, the current speed of the female mayfly 1 is updated, and the speed calculation results of each dimension of the female mayfly 1 are:

[0081]

[0082] in, represents the current speed of the jth dimension of the i-th female mayfly; g represents the gravity coefficient; a2 represents the positive attraction constant; Beta represents the corrected visibility coefficient; represents the position of the jth dimension of the i-th male mayfly at time t, represents the position of the jth dimension of the i-th female mayfly at time t; f() represents the objective function; FL is the random flight coefficient; r is a random number in [-1,1];

[0083] Then, update the current position of female mayfly 1:

[0084]

[0085] in, represents the position of the i-th female mayfly at the current moment; represents the current speed of the i-th female mayfly; represents the position of the i-th female mayfly at time t.

[0086] The current objective function value (J1, J2, J3, J4) is calculated according to the current position of the female mayfly 1, and the current position of the female mayfly 1 and the current objective function value are updated.

[0087] The same calculation is performed on the remaining 19 female mayflies in this manner to calculate the current positions and current objective functions of all the remaining female mayflies, and to update the historical optimal positions and historical optimal objective function values ​​of all the remaining female mayflies.

[0088] The calculation starts from the female mayfly 1 in the initialization stage. At this time, the female mayfly 1 dominates the leading mayfly, and the speed calculation results of the male mayfly 1 in each dimension are:

[0089]

[0090] in, represents the current speed of the jth dimension of the i-th male mayfly; represents the position of the jth dimension of the ith male mayfly at time t; a1 and a2 are positive attraction constants; pbest ij represents the current optimal position of the jth dimension of the ith male mayfly; Beta represents the corrected visibility coefficient; r p Represents x i and the current optimal position pbest i The Euclidean distance between g Represents x i and the Euclidean distance between the global optimal position gbest.

[0091] Then, update the current position of Mayfly 1:

[0092]

[0093] in, represents the position of the i-th male mayfly at the current moment; represents the current speed of the ith male mayfly; represents the position of the i-th male mayfly at time t.

[0094] The current objective function value (J1, J2, J3, J4) is calculated according to the current position of male mayfly 1. At this time, the historical optimal objective function value of male mayfly 1 dominates the current objective function value of male mayfly 1. A number in the range of [0,1] is randomly selected. The current selected number is 0.326 (less than 0.5). The historical optimal position of male mayfly 1 is updated, and the historical optimal objective function value of male mayfly 1 is updated.

[0095] The same calculation is performed on the remaining 19 male mayflies in this manner to calculate the current positions and current objective functions of all the remaining male mayflies, and to update the historical optimal positions and historical optimal objective function values ​​of all the remaining male mayflies.

[0096] If the current objective function value of mayfly 1 dominates the historical optimal objective function value of male mayfly 1, then the historical optimal position and historical optimal objective function value of male mayfly 1 are updated; if the historical optimal objective function value of male mayfly 1 dominates the current objective function value of male mayfly 1, then there is a 1 / 2 probability of updating the historical optimal position and historical optimal objective function value of male mayfly 1. For example, if male mayfly 12 does not meet the above conditions, then the historical optimal positions and historical optimal objective function values ​​of male mayfly 2 to male mayfly 11 and male mayfly 13 to male mayfly 20 are updated.

[0097] Then, every 2 mayfly individuals are randomly selected from the Pareto set for mating calculation, and half of the mayfly individuals are randomly selected from the Pareto set for mutation calculation; the mayfly population after mating and the mayfly population after mutation are combined, and 1 / 2 is taken from each and respectively incorporated into the current male mayfly population and female mayfly population.

[0098] Next, the male and female mayfly populations are sorted by Pareto non-dominated sorting, and the top q male mayfly individuals and the top p female mayfly individuals are selected as the new Pareto set. The objective function values ​​of all populations in the new Pareto set are taken to obtain the Pareto set under the first iteration.

[0099] For example, two mayfly individuals are randomly selected from the Pareto set and the first mating calculation is performed. The calculation results are shown in Table 3.

[0100] Table 3 Calculation results of the first mating of mayfly population

[0101]

[0102] The remaining 48 mayfly individuals in the Pareto set were subjected to 24 more mating calculations. After 15 mating calculations were completed, 10 mayfly individuals were randomly selected from the Pareto set for mutation calculation. Among them, the mutation calculation was performed on the first randomly selected mayfly individual, and the current objective function value of the mayfly individual after mutation was updated. This time, the fifth dimension was randomly selected for mutation, and the calculation results of the mutation dimension are shown in Table 4.

[0103] Table 4 Calculation results of the mutation dimension of the first mayfly individual

[0104] Mutation Dimension Before mutation After mutation 1 0.1459 0.1459 2 0.1127 0.1127 3 0.1415 0.1415 4 0.2351 0.2351 5 0.1256 0.1186 6 0.1696 0.1696

[0105] The combination of the mating mayfly population and the mutated mayfly population, a total of 30 mayfly individuals, is merged into one set, and then 1 / 2 of each is taken and incorporated into the male and female mayfly populations in the initialization stage. Afterwards, the male and female mayfly populations are Pareto non-dominated sorted, and the top 20 male and female mayfly individuals and the existing Pareto set combination are taken as new Pareto sets. The populations in the new Pareto set are Pareto non-dominated sorted. Finally, the objective function values ​​of the top 50 populations are taken as the Pareto set under the first iteration. If the number of populations is less than 50, the objective function values ​​of all populations are taken as the Pareto set under this iteration.

[0106] The parameter wedding dance coefficient in the multi-target mayfly algorithm is modified to Dance = 0.77*0.99, and the random flight coefficient is modified to FL = 0.77*0.99.

[0107] The Pareto set at the 50th iteration is shown in Figure 6 .

[0108] Repeat the same process for another 50 times to obtain the Pareto set at the 100th iteration. Figure 7 , repeat the same process for another 100 times to obtain the Pareto set at the 200th iteration. Figure 8 , iterate 100 times in the same way, and obtain the Pareto set at the 300th iteration as the final Pareto set of this optimization.

[0109] Step 5, extract the mayfly position in the middle of the final Pareto set, and use the superposition weight in the mayfly position as the optimal superposition weight for each orientation.

[0110] Extract the mayfly position in the middle of the final Pareto set, and use the superposition weight in the mayfly position as the optimal superposition weight in each direction, see Fig. 9 Similarly, in order to compare the two algorithms, the single-objective ephemera algorithm is used to optimize the objective function Singleobjective. The convergence curve can be found in Fig.10. Table 5 shows the comparison of the optimal stacking weights optimized by the single-objective mayfly algorithm and the multi-objective mayfly algorithm. Table 6 shows the improvement of the correlation coefficient between the synthetic seismic records of M1-M5 wells and the actual seismic records after optimization by the single-objective mayfly algorithm. Table 7 shows the improvement of the correlation coefficient between the synthetic seismic records of M1-M5 wells and the actual seismic records after optimization by the multi-objective algorithm. Compared with the stacking weights optimized by the single-objective algorithm, the number of cases where the correlation coefficient value decreased dropped from 7 to 5, and the percentage of decrease was also reduced, indicating that the multi-objective mayfly algorithm is better than the single-objective mayfly algorithm in optimizing such objective functions.

[0111] Table 5 Comparison of optimal superposition coefficients of single-target / multi-target mayfly algorithm optimization

[0112] position 15° 45° 75° 105° 135° 165° Initial stacking weight 1 / 6 1 / 6 1 / 6 1 / 6 1 / 6 1 / 6 Single target mayfly algorithm optimization 0.1843 0.1006 0.1993 0.25 0.1184 0.1474 Multi-objective Mayfly Algorithm Optimization 0.1441 0.1648 0.1742 0.2281 0.1263 0.1582

[0113] Table 6 Improvement of correlation coefficient after optimization using single-target mayfly algorithm

[0114] hashtag Close Angle promote medium angle promote Far Angle promote Initial stacking weight Superposition Thousandths (‰) Superposition Thousandths (‰) Superposition Thousandths (‰) M1 0.773 -3.398 0.765 -13.992 0.748 -3.675 M2 0.504 -7.822 0.522 -0.869 0.518 -0.539 M3 0.739 6.595 0.661 4.586 0.638 -2.382 M4 0.475 53.859 0.497 4.74 0.397 28.567 M5 0.776 5.967 0.678 1.951 0.608 2.233

[0115] Table 7 Correlation coefficient improvement after optimization using multi-objective ephemera algorithm

[0116] hashtag Close Angle promote medium angle promote Far Angle promote Initial stacking weight Superposition Thousandths (‰) Superposition Thousandths (‰) Superposition Thousandths (‰) M1 0.773 -2.911 0.767 -11.418 0.75 -1.693 M2 0.507 -1.225 0.524 2.829 0.52 3.444 M3 0.735 1.216 0.659 2.447 0.636 -5.397 M4 0.466 34.701 0.5 11.345 0.389 8.891 M5 0.775 4.345 0.679 2.35 0.608 3.344

[0117] In summary, compared with the optimization results of the single-objective mayfly algorithm, see Fig.10 The present invention establishes a set of intelligent stacking methods based on multi-objective ephemera algorithm for radial interpolation seismic gather data. The multi-objective ephemera algorithm is used to optimize and determine the optimal stacking weight for each azimuth, thereby improving the resolution of seismic data. Fig.11 , Fig.11 (a) in the figure is the slice along the horizon of the original observed seismic data. Fig.11 (b) in the figure is the slice along the horizon of the seismic data obtained by stacking with the initial stacking weight. Fig.11 (c) in the figure is a slice along the horizon of the seismic data obtained by stacking the stacking weights optimized by the multi-objective ephemera algorithm. Through comparison and visual observation, Fig.11 The result in (c) is more delicate and the resolution of the seismic data after stacking is higher; secondly, by comparing the correlation coefficients between the synthetic seismic records on the well and the actual seismic records, the results obtained by the intelligent stacking weight determination method established by the present invention are more consistent with the real seismic records and have a higher fidelity.

[0118] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent radial interpolation seismic data processing method based on a multi-objective ephemera algorithm is characterized by: The following steps are involved: Step 1, load the layer data and well location, and collect well logging data and seismic data; Step 2, when performing radial interpolation seismic gather processing, convert the wide azimuth gather into a conventional gather, and perform conventional gather optimization processing on each azimuth; Step 3, initialize the parameters of the multi-objective mayfly algorithm, set the superposition weight range of each direction; construct the objective function of the multi-objective mayfly algorithm, and complete the initialization of the multi-objective mayfly algorithm; Step 4, use the multi-objective ephemera algorithm to iterate, dynamically update the superposition weights, gradually optimize, and obtain the Pareto set; Step 5, extract the mayfly position in the middle of the final Pareto set, and use the superposition weight in the mayfly position as the optimal superposition weight for each orientation.

2. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 1 is characterized in that: In step 2, when performing radial interpolation seismic gather processing, the radial interpolation seismic data volume is divided into six azimuths of 15°, 45°, 75°, 105°, 135° and 165°; the angular gather interval range of each azimuth is an integer angle of 1° to 30°; in the angular gather interval range of each azimuth, 1° to 10° is a near angle, 11° to 20° is a medium angle, and 21° to 30° is a far angle.

3. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 2 is characterized in that: In step 2, the step of converting the wide azimuth gather into a conventional gather and performing conventional gather optimization processing on each azimuth includes: Each azimuth is treated as conventional gather data and bandpass filtered to filter out low-frequency surface wave interference in the original data; Each azimuth is treated as conventional gather data and optimized, including Radon transform, inverse Q filtering, deconvolution, gather flattening and gather conversion operations.

4. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 2 is characterized in that: In step 3, the step of initializing the parameters of the multi-objective mayfly algorithm includes: Set the number of male mayflies q = 20, the number of female mayflies p = 20, the number of mating of father mayflies Nc = 20, the number of mutations of offspring mayflies Nm = 10, the gravity coefficient g = 0.8, the positive attraction constant a1 = 1, the positive attraction constant a2 = 1.5, the mutation coefficient Mu = 0.02, the corrected visibility coefficient Beta = 2, the random flight coefficient FL = 0.77, the wedding dance coefficient Dance = 0.77, the wedding dance coefficient reduction rate DD = 0.99, the random flight coefficient reduction rate FD = 0.99, the Pareto set capacity Np = 50, the maximum number of iterations t max =300; Determine the size of the mayfly population and initialize the positions of male and female mayflies: M={x i |x i =(x i1 ,x i2 ,...,x id ),i=1,2,...,q} F={y i |y i =(y i1 ,y i2 ,...,y id ),i=1,2,...,p} Wherein, M represents the male mayfly population; F represents the female mayfly population; x i represents the position of the ith male mayfly; y i represents the position of the i-th female mayfly; q represents the number of male mayflies; p represents the number of female mayflies; d represents the total number of dimensions, each dimension corresponds to a direction, d = 6; x id represents the position of the dth dimension of the ith male mayfly; y id Represents the position of the d-th dimension of the i-th female mayfly.

5. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 2 or 4, characterized in that: In step 3, the step of setting the range of the superposition weights of each orientation includes: setting the initial superposition weight of each orientation to 1 / 6, setting the lower limit of the superposition weight to 0.1, and setting the upper limit of the superposition weight to 0.

2.

6. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 4 is characterized in that: In step 3, the objective function of the multi-objective mayfly algorithm is constructed to complete the steps of initializing the multi-objective mayfly algorithm, including: Construct three objective functions, corresponding to the number of superposition bodies at different angles, including near-angle superposition bodies, medium-angle superposition bodies, and far-angle superposition bodies: Minimize the objective function J1: Minimize the objective function J2: Minimize the objective function J3: Where n represents the total number of wells, i = 1, 2, ..., n; corrcoef(.) represents the Pearson correlation coefficient of the two series; obs near(i) 、obs mid(i) 、obs far(i) They represent the actual seismic records of the near-angle, medium-angle, and far-angle stacks of the ith well; syn near(i) ,syn mid(i) ,syn far(i) They represent the synthetic seismic records of the near-angle, medium-angle, and far-angle stacks of the ith well respectively; Construct the objective function J4 constraint superposition weight sum: Among them, N azi Indicates the number of directions (N azi =6); represents the superposition weight of the j-th orientation; In the multi-objective mayfly algorithm, the optimization process is completed by minimizing the objective function Multiobjective: Multiobjective=(J1,J2,J3,J4) After updating the current position of the first male mayfly, update the current objective function value to (J1, J2, J3, J4); start calculating the current position of the second male mayfly in the initialization phase, calculate the current objective function value (J1, J2, J3, J4) according to the current position of the second male mayfly, and update the historical position and historical objective function value of the second male mayfly; calculate the current position and current objective function value of all remaining male mayflies in the same way, and update the historical optimal position and historical optimal objective function value of all male mayflies; Then, the current position of the first female mayfly is calculated, and the current objective function value (J1, J2, J3, J4) is calculated according to the current position of the first female mayfly, and the historical position and historical objective function value of the first female mayfly are updated; the current positions and current objective function values ​​of all the remaining female mayflies are calculated in the same way, and the historical optimal positions and historical optimal objective function values ​​of all the remaining female mayflies are updated; Combine q male mayflies and p female mayflies, and perform a Pareto non-dominated sort on the total q+p mayflies. Take the first Np mayflies as the Pareto set, and take all the individuals that are less than Np as the Pareto set. The above completes the initialization of the multi-objective mayfly algorithm.

7. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 6 is characterized in that: The step 4 specifically comprises the following steps: Step 4-1, enter the main loop of the multi-objective mayfly algorithm, perform the t-th iteration of the multi-objective algorithm, calculate the current position and current objective function value of the mayfly individual in the Pareto set, and update the historical optimal position and historical optimal objective function value of all mayfly individuals; Step 4-2, randomly select every 2 mayfly individuals from the Pareto set for mating calculation, and then randomly select half of the mayfly individuals from the Pareto set for mutation calculation; combine the mating mayfly population and the mutated mayfly population, and take 1 / 2 of each to merge into the current male mayfly population and female mayfly population respectively; Step 4-3, perform Pareto non-dominated sorting on the male and female mayfly populations respectively, and take the top q male mayfly individuals and the top p female mayfly individuals as the new Pareto set; take the objective function values ​​of all populations in the new Pareto set to obtain the Pareto set under the tth iteration; Step 4-4, determine whether t is less than t max If yes, set t=t+1 and return to step 4-1; otherwise, obtain the tth max The Pareto set under iterations.

8. The intelligent radial interpolation seismic data processing method based on the multi-objective ephemera algorithm according to claim 7 is characterized in that: The step 4-1 specifically includes the following steps: Select the mayfly individual ranked first in the Pareto set as the leading mayfly; The calculation starts from the female mayfly 1 in the initialization stage. At this time, the male mayfly 1 dominates the female mayfly 1. The Euclidean distance between the male mayfly 1 and the female mayfly 1 is calculated: Among them, r mf express and The Euclidean distance between represents the initial position of the jth dimension of male mayfly 1, represents the initial position of the j-th dimension of female mayfly 1; Afterwards, the current speed of the female mayfly 1 is updated, and the speed calculation results of each dimension of the female mayfly 1 are: in, represents the current speed of the jth dimension of the i-th female mayfly; g represents the gravity coefficient; a2 represents the positive attraction constant; Beta represents the corrected visibility coefficient; represents the position of the jth dimension of the i-th male mayfly at time t, represents the position of the jth dimension of the i-th female mayfly at time t; f() represents the objective function; FL is the random flight coefficient; r is a random number in [-1,1]; Then, update the current position of female mayfly 1: in, represents the position of the i-th female mayfly at the current moment; represents the current speed of the i-th female mayfly; represents the position of the i-th female mayfly at time t; Calculate the current objective function value (J1, J2, J3, J4) according to the current position of the female mayfly 1, and update the current position of the female mayfly 1 and the current objective function value; Perform the same calculation on all remaining female mayflies in this manner, calculate the current positions and current objective functions of all remaining female mayflies, and update the historical optimal positions and historical optimal objective function values ​​of all remaining female mayflies; The calculation starts from the female mayfly 1 in the initialization stage. At this time, the female mayfly 1 dominates the leading mayfly, and the speed calculation results of the male mayfly 1 in each dimension are: in, represents the current speed of the jth dimension of the i-th male mayfly; represents the position of the jth dimension of the ith male mayfly at time t; a1 and a2 are positive attraction constants; pbest ij represents the current optimal position of the jth dimension of the ith male mayfly; Beta represents the corrected visibility coefficient; r p Represents x i and the current optimal position pbest i The Euclidean distance between g Represents x i The Euclidean distance between the global optimal position gbest; Then, update the current position of Mayfly 1: in, represents the position of the i-th male mayfly at the current moment; represents the current speed of the ith male mayfly; represents the position of the i-th male mayfly at time t; The current objective function value (J1, J2, J3, J4) is calculated according to the current position of male mayfly 1. At this time, the historical optimal objective function value of male mayfly 1 dominates the current objective function value of male mayfly 1. A number in the range of [0, 1] is randomly selected to update the historical optimal position of male mayfly 1 and the historical optimal objective function value of male mayfly 1. The same calculation is performed on all the remaining male mayflies in this manner to calculate the current positions and current objective functions of all the remaining male mayflies, and to update the historical optimal positions and historical optimal objective function values ​​of all the remaining male mayflies.

Images