Method and system for obtaining formation elastic parameters

Abstract

The application discloses a method and system for obtaining formation elastic parameters. The method comprises the following steps: establishing a rock physics model according to logging data, and establishing a sequence model according to seismic data; obtaining wavelets of different angle gathers according to the rock physics model and the sequence model; constructing a low-frequency background model according to the wavelets of different angle gathers; and performing prestack inversion to obtain formation elastic parameters according to the low-frequency background model, wherein in the process of each iteration calculation of the prestack inversion, the particles in a particle group previously constructed by randomization are evolved by using a Markov chain Monte Carlo method, and then a self-adaptive particle swarm optimization method is used to find global optimal values and local optimal values. The system comprises a processor and a memory, and the processor realizes the method when executing a computer program saved in the memory. According to the application, the problems of low efficiency and easy falling into local minimum values in the iteration calculation process of the existing method for obtaining formation elastic parameters based on prestack stochastic inversion can be solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of geophysical exploration, and more particularly, relates to a stratum elastic parameter acquisition method and system based on random adaptive particle swarm optimization. BACKGROUND

[0002] In the process of seismic wave propagating from the ground to the underground, the interface of stratum impedance difference (i.e. stratum interface) will be encountered, at which time the reflected seismic wave will be generated. The reflected seismic signal can be received by the geophone arranged on the ground. The seismic signal contains rich amplitude information, and the corresponding pretreatment of the seismic signal can obtain the amplitude information varying with the reflection angle. When the stratum contains fluid, such as water, oil or gas, the amplitude value will change significantly with the reflection angle, so we can use the change of the amplitude to identify the oil and gas bearing area in the stratum. In the prior art, the prestack wave impedance inversion technology can estimate the stratum elastic parameter by using the amplitude value, so as to achieve the purpose of oil and gas identification. The existing prestack wave impedance inversion technology includes deterministic inversion and random inversion, both of which can estimate the stratum elastic parameter. However, the stability of the deterministic inversion method is poor, and it is easily affected by the seismic noise, so the inversion result is not accurate. The random inversion method has good stability and noise resistance, but the calculation speed of the existing random inversion method is slow, the efficiency is low, and it is easy to fall into local minimum value in the calculation process and stop the iteration calculation. SUMMARY

[0003] The present application aims to solve the problems of low efficiency and easy to fall into local minimum value in the iteration calculation process of the existing stratum elastic parameter acquisition method based on prestack random inversion.

[0004] In order to achieve the above-mentioned purpose, the present application provides a stratum elastic parameter acquisition method and system based on random adaptive particle swarm optimization.

[0005] According to the first aspect of the present application, a stratum elastic parameter acquisition method based on random adaptive particle swarm optimization is provided, which comprises the following steps:

[0006] Establishing a rock physics model according to the logging data and a sequence model according to the seismic data;

[0007] Obtaining wavelets of different angle gathers according to the rock physics model and the sequence model;

[0008] Constructing a low-frequency background model according to the wavelets of different angle gathers;

[0009] According to the low-frequency background model, pre-stack inversion is performed to obtain elastic parameters of the stratum, wherein in each iteration calculation of the pre-stack inversion, firstly, particles in a particle group previously constructed are evolved by using a Markov chain Monte Carlo method, and then a self-adaptive particle swarm optimization method is used to find global optimal values and a plurality of local optimal values of a preset target function.

[0010] As preferred, the wavelet is expressed as w(θ, t), wherein θ is an incidence angle parameter, and t is a corresponding time depth on a seismic record.

[0011] As preferred, the construction of the low-frequency background model is implemented based on a Kriging interpolation method.

[0012] As preferred, before initial iteration calculation of the pre-stack inversion, the following steps are included:

[0013] A particle group is constructed by using a random sampling method;

[0014] A target function and an iteration number are set.

[0015] As preferred, the particle group is expressed as wherein i is a particle number, k is an iteration number, v p is a P-wave velocity, v s is a S-wave velocity, and ρ is a density.

[0016] The target function is expressed as wherein W1 and W2 are weight coefficients, D obs (θ, t) is observed seismic data, S(θ, t) is a synthesized seismic record, S(θ, t) = w(θ, t) * R(θ, t), R(θ, t) is a reflection coefficient calculated in each iteration, M is a low-frequency background model, X j is a sampling point number in single particle inversion.

[0017] As preferred, the self-adaptive particle swarm optimization method is expressed as x i = x i + v i , wherein v i = ω × v i + c1 × rand1 × (Pbest i - x i ) + c2 × rand2 × (Gbest - x i ), Gbest is an optimal value up to the current iteration, Pbest i is an optimal value in the current iteration, ω is a step coefficient, and v iis a step length, c1 is a weight of the optimal value of the previous iteration, c2 is a weight of the optimal value of the current iteration, c1 = c2 = 2, rand1 and rand2 are random numbers between 0 and 1.

[0018] Preferably, the expression of the step coefficient is

[0019] Preferably, the evolution expression of the Markov Chain Monte Carlo method is wherein P is an evolution matrix, is a set of all particles after the kth inversion of i particles.

[0020] Preferably, the formation elastic parameters are the shear wave velocity, the compressional wave velocity and the density contained in the global optimal solution of the pre-stack inversion.

[0021] According to the second aspect of the present application, there is provided a system for obtaining formation elastic parameters based on random adaptive particle swarm optimization, which comprises a processor and a memory, and the processor realizes any of the above-mentioned methods for obtaining formation elastic parameters based on random adaptive particle swarm optimization when executing a computer program stored in the memory.

[0022] The present application has the following advantages:

[0023] The method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to the present application firstly establishes a rock physics model according to well logging data and a sequence model according to seismic data; secondly obtains wavelets of different angle gathers according to the rock physics model and the sequence model; thirdly constructs a low-frequency background model according to the wavelets of different angle gathers; and finally performs pre-stack inversion to obtain formation elastic parameters according to the low-frequency background model, wherein in the process of each iteration calculation of the pre-stack inversion, the particles in a particle swarm constructed in advance are evolved by using a Markov Chain Monte Carlo method, and then a global optimal value and a local optimal value are found by using an adaptive particle swarm optimization method.

[0024] The method for obtaining formation elastic parameters based on random adaptive particle swarm optimization of the application calculates the elastic parameters of the formation to extract favorable information of oil and gas bearing layers according to the phenomenon that the amplitude of seismic waves will change in oil and gas bearing layers and in combination with the method of pre-stack elastic parameter inversion. The particle swarm is constructed by using the random sampling method, the adaptive attribute is defined according to the distance between the particles in the iteration process, and the particle swarm is evolved by using the Markov chain Monte Carlo method to avoid falling into local minimum value in the process of iterative calculation. The method for obtaining formation elastic parameters based on random adaptive particle swarm optimization of the application can accurately invert and calculate the elastic parameters of the formation, and can stably calculate the elastic parameters of the formation even when the signal-to-noise ratio is low, which is beneficial to improving the calculation efficiency of inversion and enhancing the stability of the inversion process.

[0025] The system for obtaining formation elastic parameters based on random adaptive particle swarm optimization of the application belongs to the same general inventive concept as the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization described above, and therefore has the same beneficial effects as the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization described above, which will not be described here again.

[0026] Other features and advantages of the application will be described in detail in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0027] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the figures, and wherein:

[0028] Figure 1 An implementation flowchart of the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to embodiment 1 of the application is shown;

[0029] Figure 2 A step flowchart of the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to embodiment 1 of the application is shown;

[0030] Figure 3 A synthetic seismic record made from the actual single-channel seismic record according to embodiment 1 of the application and the inversion result after 2000 iterations is shown;

[0031] Figure 4 A comparison between the elastic parameter inversion result after 2000 iterations according to embodiment 1 of the application and the actual logging data is shown;

[0032] Figure 5 A low-frequency background model established based on the actual work area data of the test according to embodiment 1 of the application is shown.

[0033] Figure 6 Fig. 6 shows the P-wave impedance obtained after inversion of the extracted single pick-up line according to the embodiment 1 of the present application;

[0034] Figure 7 Fig. 7 shows the S-wave impedance obtained after inversion of the extracted single pick-up line according to the embodiment 1 of the present application;

[0035] Figure 8 Fig. 8 shows the calculation results of the Poisson's ratio parameter of the extracted single pick-up line according to the embodiment 1 of the present application;

[0036] Figure 9 Fig. 9 shows the trend of the global minimum value with the number of iterations in the iteration process according to the embodiment 1 of the present application;

[0037] Figure 10 Fig. 10 shows the comparison results of the adaptive particle swarm optimization method according to the embodiment 1 of the present application and other optimization methods. DETAILED DESCRIPTION

[0038] Preferred embodiments of the present application will be described in more detail below. Although the following describes preferred embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present application is more thorough and complete, and the scope of the present application is fully conveyed to those skilled in the art.

[0039] Embodiment 1: Figure 1 Fig. 1 shows a flowchart of the implementation of the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to the embodiment of the present application, Figure 2 Fig. 2 shows a flowchart of the steps of the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to the embodiment of the present application. Referring to Figure 1 and Figure 2 the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to the embodiment of the present application includes the following steps:

[0040] S100, establishing a rock physics model according to logging data and a sequence model according to seismic data;

[0041] S200, obtaining wavelets of different angle gathers according to the rock physics model and the sequence model;

[0042] S300, constructing a low-frequency background model according to the wavelets of different angle gathers;

[0043] S400, performing pre-stack inversion to obtain the formation elastic parameters according to the low-frequency background model, wherein in the process of each iteration calculation of the pre-stack inversion, firstly, particles in a particle group constructed in advance are evolved by using a Markov Chain Monte Carlo method, and then a self-adaptive particle swarm optimization method is used to find a global optimal value and a plurality of local optimal values;

[0044] The formation elastic parameters include a shear wave velocity, a compressional wave velocity and a density.

[0045] Specifically, in step S100 of the embodiment of the present application, a relationship among the compressional wave velocity, the shear wave velocity and the density is established by using a rock physics model.

[0046] Further, in step S200 of the embodiment of the present application, the wavelet is represented as w(θ, t), wherein θ is an incident angle parameter, and t is a corresponding time depth on a seismic record.

[0047] Still further, in step S300 of the embodiment of the present application, the construction of the low-frequency background model is realized based on a Kriging interpolation method.

[0048] Still further, before initial iteration calculation of the pre-stack inversion, the embodiment of the present application includes:

[0049] constructing a particle group by using a random sampling method;

[0050] setting a target function and an iteration number, wherein the iteration number is a standard for iteration stop.

[0051] Still further, in the embodiment of the present application, an expression of the particle group constructed in a random manner is wherein i is a particle number, k is an iteration number, v p is the compressional wave velocity, v s is the shear wave velocity, and ρ is the density.

[0052] An expression of the target function is wherein W1 and W2 are weight coefficients, D obs (θ, t) is observed seismic data, S(θ, t) is a synthesized seismic record, S(θ, t) = w(θ, t) * R(θ, t), R(θ, t) is a reflection coefficient calculated in each iteration, M is a low-frequency background model, X j is a sampling point number in single particle inversion.

[0053] Still further, in the embodiment of the present application, an expression of the self-adaptive particle swarm optimization method is x i = x i + v i , wherein v i= ω x v i + c1 x rand1 x (Pbest i - x i + c2 x rand2 x (Gbest - x i ), Gbest is the optimal value up to the current iteration, Pbest i is the optimal value of the current iteration, ω is a step coefficient, v i is a step length, c1 is a weight of the optimal value up to the current iteration, c2 is a weight of the optimal value of the current iteration, c1 = c2 = 2, rand1 and rand2 are random numbers between (0, 1).

[0054] Further, in the embodiment of the present application, the expression of the step coefficient is ω (f) = ω0

[0055] Further, in the embodiment of the present application, the evolution expression of the Markov Chain Monte Carlo method is wherein, P is an evolution matrix, is a set of all particles after the kth inversion of i particles.

[0056] Further, in the embodiment of the present application, when the iteration stopping criterion is met, the iteration is stopped, and the global optimal solution is output, wherein the global optimal solution contains the shear wave velocity, the longitudinal wave velocity and the density. If the iteration stopping criterion is not met, the iteration process is continued until the iteration stopping criterion is met.

[0057] The beneficial effects of the method for obtaining formation elastic parameters based on the random adaptive particle swarm optimization in the embodiment of the present application are described in detail below based on two specific examples:

[0058] Figure 3 The actual single-channel seismic record and the synthetic seismic record made based on the inversion result after 2000 iterations are shown. The left graph is the actual angle gather data obtained after processing at 5°, 10° and 15°, and it can be found that the actual seismic data has much noise. The right graph is the synthetic seismic record made based on the inversion result after 2000 iterations, and it can be known from the comparison between the left graph and the right graph that the synthetic seismic record is very close to the actual seismic record, and the influence of noise is eliminated.

[0059] Figure 4 The comparison between the elastic parameter inversion result obtained after 2000 iterations and the actual logging data is shown. It can be found from the comparison that the inversion result is very close to the actual logging data.

[0060] Figure 5 The low-frequency background model established based on the actual work area data of the test is shown, Figure 6The longitudinal wave impedance obtained after inversion of the extracted single trace line is shown, Figure 7 The transverse wave impedance obtained after inversion of the extracted single trace line is shown.

[0061] Figure 8 The calculation result of the Poisson ratio parameter of the extracted single trace line is shown. The gas layer contains a lower Poisson ratio, and the distribution position and thickness of the gas layer can be seen from the figure.

[0062] Figure 9 The trend of the global minimum value with the number of iterations in the iteration process is shown. It can be seen from the trend that the search result of the global minimum value is continuously reduced with the increase of the number of iterations.

[0063] Figure 10 The comparison result of the adaptive particle swarm optimization method of the embodiment of the present application with other optimization methods is shown. According to the comparison result, the loss function value of the adaptive particle swarm optimization method of the embodiment of the present application is smaller, and the optimization effect is better. Figure 10 It can be seen that the loss function value of the adaptive particle swarm optimization method of the embodiment of the present application is smaller, and the optimization effect is better.

[0064] The embodiment of the present application proposes a new method specially used for calculating the elastic parameters of the stratum. According to the phenomenon that the amplitude of the seismic wave changes with the offset distance, and in combination with the characteristics of the random inversion, the prestack seismic signal obtained after processing is added to the input data, the low-frequency background model is established by using the logging data and the seismic data, and the elastic parameters of the stratum are calculated by using the unique random inversion method. When the stratum contains oil or gas, the amplitude of the seismic wave will change; when the oil saturation or the gas saturation increases, the change of the amplitude of the seismic wave is more violent. The method for calculating the elastic parameters proposed in the embodiment of the present application can better identify the characteristics of the oil and gas reservoir in the seismic profile, and is beneficial to the prediction and description of the oil and gas reservoir.

[0065] Embodiment 2: On the basis of the method for obtaining the elastic parameters of the stratum based on the random adaptive particle swarm optimization proposed in embodiment 1, the embodiment of the present application proposes a system for obtaining the elastic parameters of the stratum based on the random adaptive particle swarm optimization.

[0066] The system for obtaining the elastic parameters of the stratum based on the random adaptive particle swarm optimization of the embodiment of the present application comprises a processor and a memory. When the processor executes the computer program saved in the memory, the method for obtaining the elastic parameters of the stratum based on the random adaptive particle swarm optimization proposed in embodiment 1 is realized.

[0067] The above has described the embodiments of the present application, and the above description is exemplary, is not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for obtaining formation elastic parameters based on random adaptive particle swarm optimization, characterized in that: include: Build rock physics models based on well logging data and sequence models based on seismic data; Acquire wavelets of different angle gathers according to the rock physics model and the sequence model; constructing a low-frequency background model according to the wavelets of the different angle gathers; Based on the low-frequency background model, prestack inversion is performed to obtain formation elastic parameters. During each iterative calculation of the prestack inversion, the Markov chain Monte Carlo method is first used to evolve particles in a pre-randomly constructed particle swarm, and then the adaptive particle swarm optimization method is used to find the global optimal value and the local optimal value.

2. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 1, characterized in that: The wavelet is expressed as w(θ, t), where θ is the incident angle parameter and t is the corresponding time depth on the seismic record.

3. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 2, characterized in that: The construction of the low-frequency background model is achieved based on the Kriging interpolation method.

4. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 3 is characterized in that: Before the first iteration calculation of the prestack inversion, the following steps are included: The random sampling method is used to construct the particle swarm; Set the objective function and number of iterations.

5. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 4 is characterized in that: The expression of the particle swarm is: Among them, i is the number of particles, k is the number of iterations, v p is the longitudinal wave velocity, v s is the shear wave velocity, ρ is the density; The expression of the objective function is Among them, W1 and W2 are weight coefficients, D obs (θ, t) is the observed seismic data, S(θ, t) is the synthetic seismic record, S(θ, t) = w(θ, t) * R(θ, t), R(θ, t) is the reflection coefficient calculated in each iteration, M is the low-frequency background model, X j is the number of sampling points in the inversion of a single particle.

6. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 5, characterized in that: The expression of the adaptive particle swarm optimization method is x i =x i +v i , where v i =ω×v i +c1×rand1×(Pbest i -x i )+c2×rand2×(Gbest-x i ), Gbest is the optimal value up to the current iteration, Pbest i is the optimal value of the iteration, ω is the step coefficient, v i is the step length, c1 is the weight of the optimal value up to the iteration, c2 is the weight of the optimal value of the iteration, c1=c2=2, rand1 and rand2 are both random numbers between (0, 1).

7. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 6, characterized in that: The expression of the step coefficient is:

8. The formation elastic parameter acquisition method based on stochastic adaptive particle swarm optimization according to claim 7, characterized in that: The evolution expression of the Markov chain Monte Carlo method is: Among them, P is the evolution matrix, is the set of all particles after the kth inversion of particle i.

9. The method for obtaining formation elastic parameters based on stochastic adaptive particle swarm optimization according to claim 8, characterized in that: The formation elastic parameters are the shear wave velocity, compressional wave velocity and density contained in the global optimal solution of the prestack inversion.

10. A formation elastic parameter acquisition system based on random adaptive particle swarm optimization, characterized in that: The method comprises a processor and a memory, wherein when the processor executes the computer program stored in the memory, the method for obtaining formation elastic parameters based on random adaptive particle swarm optimization according to any one of claims 1 to 9 is implemented.

Images