High-resolution dipole bending wave velocity inversion method

Abstract

The invention discloses a high-resolution dipole bending wave velocity inversion method, and relates to the field of oil-gas exploration of unconventional reservoirs such as shale and tight sandstone, and the method comprises the steps: collecting conventional logging information and array acoustic logging information of a well, and carrying out the curve preprocessing of the conventional logging information and array acoustic logging information; processing waveform data of the monopole array and the dipole array obtained by preprocessing by adopting a correlation method to obtain low-resolution longitudinal wave velocity, longitudinal wave arrival time and low-resolution bending wave velocity; based on the obtained data, establishing a numerical model and performing forward modeling to obtain a dipole bending waveform and performing frequency-wavenumber domain analysis to obtain a high-order bending wave frequency range; through the range, wave form data of an actually measured dipole array is subjected to filtering processing and then is subjected to correlation analysis of bending wave arrival time constraint, so that the high-resolution dipole bending wave speed is obtained. According to the method, the longitudinal resolution of the bending wave velocity can be improved to 0.3 m, so that the logging interpretation of the thin interbed is more accurate, and the characteristics of the oil reservoir can be better mastered.

Description

Technical Field

[0001] The present invention relates to the field of oil and gas exploration in unconventional reservoirs such as shale and tight sandstone, and more specifically to a high-resolution dipole bending wave velocity inversion method. Background Art

[0002] With the continuous development and deepening of oil and gas field exploration and development, the requirements for logging technology and interpretation are getting higher and higher. High-resolution acoustic logging plays an important role in the division of thin layers and subdivision of thick layers. It also plays a key role in the evaluation of future logging data and the re-evaluation of logging data of old wells. When evaluating the formation, it is often necessary to estimate the acoustic parameters and rock physical properties of horizontal thin layers in order to better understand the characteristics and reserves of oil reservoirs. The conventional processing of array acoustic logging data, limited by the structure of the instrument, usually gives an average slowness curve within the span of the receiving transducer array. The highest resolution is the span between the first receiving transducer and the last receiving transducer. This often ignores the characteristics of formations with a span smaller than the array span, which is not conducive to the logging interpretation of thin interlayers.

[0003] In 1987, Hsu and Chang proposed to use the multi-source waveform similarity superposition method to process repeated subarray data to improve the resolution of the slowness curve. In 2000, Tang XM et al. proposed a signal processing method for evaluating thin formations using high-resolution acoustic slowness logging, which increased the resolution from 3.5ft to 0.5ft; then proposed a dual waveform matching method using subarrays to combine different formation spans to achieve 0.5ft resolution P-wave velocity extraction. In 2005, Li Pengju et al. proposed a processing method combining correlation spectrum with multiple shot points, that is, using the correlation spectrum method to extract the shortest co-transmitted and co-received subarrays of multipole array acoustic logging. The P-wave velocities of all the shortest co-transmitted or co-received subarrays across the same depth formation are averaged to obtain the co-transmitted P-wave velocity or the co-received P-wave velocity, respectively. In 2019, T Lei et al. proposed a new interpretation algorithm. The algorithm is based on a powerful downscaling technique. It extracts all convolution relationships between acoustic logging curves of different resolutions to obtain an overdetermined matrix, and uses the Moore-Penrose pseudo-inverse method to estimate the velocity to improve the vertical resolution of the logging curve. In 2020, Zhou Haoyi proposed to combine the bisection method, simulated annealing method and slowness-time method to extract component wave velocities. This method improves the traditional slowness-time correlation method in terms of processing efficiency and accuracy.

[0004] The above research content mainly focuses on how to use array acoustic wave data to extract acoustic wave velocity, and does not consider how to use the depth-offset multi-dimensional information and high-frequency components in the waveform data to carry out high-resolution acoustic wave velocity extraction, especially how to carry out high-resolution dipole flexural wave velocity extraction, which is a mode wave close to the formation shear wave velocity, and has practical application value for the analysis of reservoirs with soft elastic properties such as shale; no related technology has been retrieved in the prior art. In this regard, the present invention proposes a high-resolution dipole flexural wave velocity inversion method that first uses forward simulation to determine the frequency range of high-order flexural waves, and then uses high-order flexural waves to carry out high-resolution flexural wave velocity inversion on this basis. Summary of the invention

[0005] The purpose of the present invention is to provide a high-resolution dipole bending wave velocity inversion method. The method makes the logging interpretation of thin interlayers more accurate, so as to better understand the characteristics and reserves of oil reservoirs. It provides more effective data support for measuring reservoirs with soft elastic properties and has practical application value.

[0006] A high-resolution dipole bending wave velocity inversion method provided by the present invention comprises the following steps: S1. Collect conventional logging data and array acoustic logging data of the well; S2, performing curve preprocessing on conventional logging data and array acoustic logging data to obtain monopole array waveform data and dipole array waveform data; S3, using a velocity extraction method to process the monopole array waveform data to obtain low-resolution longitudinal wave velocity and longitudinal wave arrival time, and using a velocity extraction method to process the dipole array waveform data to obtain low-resolution bending wave velocity; S4. Based on the well diameter, density, longitudinal wave velocity and bending wave velocity, a numerical model of the well is established and a forward simulation is performed to obtain a simulated dipole bending waveform; S5. Perform frequency-wavenumber domain analysis on the simulated dipole bending waveform to obtain the frequency distribution range of the high-order bending wave; S6, selecting a high-order bending wave frequency distribution range to perform filtering processing on the measured dipole array waveform data to obtain high-order bending wave array waveform data; S7. Perform correlation analysis of the high-order bending wave array waveform data with bending wave arrival time constraints to obtain high-resolution dipole bending wave velocity.

[0007] In S1, the conventional logging data of the well mainly include natural gamma, well diameter, and density curve; the array acoustic logging data of the well mainly include the original waveform, gain curve and delay curve of the monopole array acoustic logging and the original waveform, gain curve and delay curve of the dipole array acoustic logging.

[0008] In S2, the curve preprocessing of array acoustic logging data mainly includes gain recovery and delay recovery. Gain recovery is performed first, and then delay recovery. During the waveform acquisition process, in order to reduce the storage volume of recorded data, the waveform data of a period of time before the arrival of the first wave is often not collected. The purpose of delay recovery is to fill zeros before the zero moment of the monopole array acoustic wave waveform, so as to obtain waveform data with accurate time.

[0009] In S3, the velocity extraction method is a correlation method. For the pre-processed monopole array acoustic wave waveform data, the correlation method is selected to extract the P-wave velocity curve therefrom; by calculating the correlation function of the array waveform in two dimensions of time and velocity, the velocity at which the maximum function value is located is the low-resolution P-wave velocity; based on the low-resolution P-wave velocity, the P-wave arrival time curve can be directly calculated on the basis of considering the well diameter and the length of the logging tool; The correlation method is used to extract the shear wave velocity curve from the preprocessed dipole array acoustic waveform data. The correlation function of the array waveform is calculated in the two dimensions of time and velocity. The velocity with the maximum function value is the low-resolution dipole bending wave velocity.

[0010] In S4, a comprehensive analysis is performed on the wellbore diameter, density, P-wave velocity and dipole bending wave velocity curves of the well to find out the layer section with no wellbore diameter collapse and no sudden change in density value, P-wave velocity and dipole bending wave velocity; The density, wellbore diameter, compressional wave velocity and dipole flexural wave velocity of the interval are taken as input parameters, and the forward simulation is carried out using the numerical simulation method to calculate the dipole flexural wave array waveform.

[0011] The numerical simulation method adopts real axis integration method or finite difference method.

[0012] In S5, the frequency-wavenumber domain analysis mainly involves performing a two-dimensional Fourier transform on the simulated dipole bending waveform data to obtain a frequency-wavenumber two-dimensional spectrum corresponding to the simulated dipole bending wave array waveform; In the frequency-wavenumber two-dimensional spectrum, the horizontal axis is set to wavenumber and the vertical axis is set to frequency; from it, a first-order bending wave mode with a lower frequency and a second-order bending wave mode with a higher frequency can be observed; the frequency band distribution range corresponding to the second-order bending wave mode is the frequency distribution range of the higher-order bending wave.

[0013] In S6, the filtering process refers to setting a digital bandpass filter and using the digital bandpass filter to process the dipole bending wave array waveform data measured in S4, thereby filtering out low-order bending wave information, and the retained waveform data is high-order bending wave array waveform data.

[0014] The frequency band passing range of the digital bandpass filter is set to the frequency band distribution range of the high-order dipole bending wave obtained in S5.

[0015] In S7, the correlation analysis of the bending wave arrival time constraint is improved on the basis of the existing correlation method. Specifically, the window opening start time is no longer a fixed value, but the dipole bending wave arrival time curve is used as the window opening start time. The purpose of this implementation is to reduce the impact of noise.

[0016] Compared with the prior art, the advantages of the present invention are: 1. The present invention proposes a new method for determining the frequency range of high-order flexural waves by forward modeling, and then using high-order flexural waves to carry out high-resolution flexural wave velocity inversion. Compared with the existing sound velocity extraction method, this method can increase the vertical resolution of flexural wave velocity to 0.3 m, making the logging interpretation of thin interlayers more accurate, so as to better understand the characteristics and reserves of oil reserves, and provide more effective data support for measuring reservoirs with soft elastic properties, which has practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The original well logging data of Well Y and the result diagram after preprocessing of array acoustic logging data provided by the embodiment of the present invention; Figure 2 The simulated dipole bending wave waveform corresponding to the forward model constructed based on the Y well logging curve provided in the embodiment of the present invention; Figure 3 A two-dimensional spectrum in the frequency-wavenumber domain corresponding to the simulated dipole bending wave waveform provided by an embodiment of the present invention; Figure 4 A comparison diagram of the high-resolution dipole bending wave velocity and the low-resolution dipole bending wave velocity provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the technical solution of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and through embodiments.

[0019] A high-resolution dipole bending wave velocity inversion method comprises the following steps: S1. Collect conventional logging data and array acoustic logging data of the well, mainly including natural gamma, wellbore diameter, and density curves; array acoustic logging data of the well, mainly including the original waveform, gain curve and delay curve of monopole array acoustic logging and the original waveform, gain curve and delay curve of dipole array acoustic logging.

[0020] Figure 1The first track is the natural gamma of Well Y; the second track is the wellbore diameter; the third track is the density; the fourth track is the depth track; the fifth track is the original monopole waveform variable density map WVPO, from which the longitudinal wave information can be observed; the eighth track is the original dipole bending waveform variable density map WVFLO, from which the shear wave information can be observed.

[0021] S2. Perform curve preprocessing on conventional logging data and array acoustic logging data to obtain wellbore curves, density curves, monopole array waveform data and dipole array waveform data; monopole array acoustic data preprocessing mainly includes gain recovery and delay recovery, firstly, gain recovery is performed, and then delay recovery is performed. In the waveform acquisition process, in order to reduce the storage volume of recorded data, the waveform data for a period of time before the arrival of the first wave is often not collected. The purpose of delay recovery is to fill zero before the zero moment of the monopole array acoustic waveform, so as to obtain waveform data at the accurate time.

[0022] Figure 1 The sixth track is the monopolar waveform density map WVP after preprocessing, from which it can be observed that the longitudinal wave signal becomes clearer after waveform preprocessing; the ninth track is the dipole bending wave density map WVFL after preprocessing, from which it can be observed that the dipole bending wave signal becomes clearer after waveform preprocessing.

[0023] S3. The velocity extraction method is used to process the monopole array waveform data to obtain low-resolution longitudinal wave velocity and longitudinal wave arrival time, and the velocity extraction method is used to process the dipole array waveform data to obtain low-resolution bending wave velocity.

[0024] For the pre-processed monopole array acoustic waveform data, the correlation method is selected to extract the P-wave velocity curve from it. This method calculates the correlation function of the array waveform in two dimensions, time and velocity, and the velocity at which the maximum function value is located is the low-resolution P-wave velocity. Based on the low-resolution P-wave velocity, the P-wave arrival time curve can be directly calculated based on the well diameter and the length of the logging instrument. Similarly, for the pre-processed dipole array acoustic waveform data, the correlation method is selected to extract the S-wave velocity curve from it. This method calculates the correlation function of the array waveform in two dimensions, time and velocity, and the velocity at which the maximum function value is located is the low-resolution dipole bending wave velocity.

[0025] Figure 1 The seventh track is a low-resolution longitudinal wave velocity curve, which is the result of processing the monopole array waveform data using the velocity extraction method. The tenth track is a low-resolution dipole bending wave velocity curve, which is the result of processing the dipole bending wave array waveform data using the velocity extraction method. It can be observed that the longitudinal resolution of the dipole bending wave velocity is low.

[0026] S4. Based on the wellbore diameter, density, longitudinal wave velocity and flexural wave velocity, a numerical model of the well is established and a forward simulation is performed to obtain a simulated dipole flexural waveform.

[0027] Taking into account the conditions such as no expansion of the wellbore diameter, no sudden changes in density, P-wave velocity and dipole flexural wave velocity, the wellbore diameter, density, P-wave velocity and dipole flexural wave velocity at a depth of 2322 m were selected to establish a numerical model of the wellbore formation. This embodiment uses the real axis integration method to implement forward simulation. Usually, a real array acoustic logging instrument will measure eight dipole flexural wave waveforms at each depth position, and they have different source distances. In the numerical simulation process, these source distances should also be used as input parameters. The simulation results are shown in Figure 2. Figure 2 As shown in the figure, the horizontal axis is time, and the vertical axis is the source distance corresponding to the 8 dipole bending wave array waveforms, the minimum source distance is 10.25 ft, and the maximum source distance is 13.75 ft. From the figure, we can observe the pure dipole bending waveform, and its arrival time is delayed as the source distance increases.

[0028] S5. Perform frequency-wavenumber domain analysis on the simulated dipole bending waveform to obtain the frequency distribution range of the high-order bending wave.

[0029] against Figure 2 The simulated dipole bending wave array waveform shown in the figure is subjected to double Fourier transform to obtain a two-dimensional spectrum in the frequency-wavenumber domain, as shown in Figure 3 As shown in the figure. The horizontal axis is the wave number and the vertical axis is the frequency. The oblique line VFL in the figure represents the projection of the shear wave velocity in the two-dimensional spectrum. The box in the figure is the first-order dipole bending wave mode, from which it can be observed that its intersection with the shear wave velocity curve is at about 5kHz. Therefore, if the second-order dipole bending wave (such as Figure 3 The position indicated by the arrow in the middle) or higher-order dipole bending waves, the waveform signals below 5kHz need to be filtered out.

[0030] S6. Select a high-order bending wave frequency distribution range to perform filtering processing on the measured dipole array waveform data to obtain high-order bending wave array waveform data.

[0031] Figure 4 The first channel is the depth channel; the second channel is the pre-processed dipole bending wave density map; the third channel is the corresponding frequency of the second waveform, from which it can be observed that the sound wave frequency band is relatively wide, with response characteristics in the range of 3-8kHz. Figure 3 The analysis results show that if high-order dipole bending waves are to be extracted, the waveform signals below 5kHz need to be filtered out. After considering the bandwidth broadening of the original waveform of the measured dipole bending waves, a digital bandpass filter with a passband frequency range of 6-8kHz is designed to obtain pure high-order dipole bending waves. Figure 4 The fourth track is the dipole bending wave after digital bandpass filtering based on the second track; the fifth track is the corresponding spectrum of the fourth track, from which it can be observed that the spectrum components below 5kHz are indeed filtered out, and only the spectrum energy above 5kHz is retained, which corresponds to the high-order dipole bending wave.

[0032] S7. Perform correlation analysis of the high-order bending wave array waveform data with bending wave arrival time constraints to obtain high-resolution dipole bending wave velocity.

[0033] The correlation analysis of the first-order bending wave arrival time constraint is performed on the high-order bending wave array waveform data. In particular, the correlation analysis method of the arrival time constraint used is improved on the basis of the existing correlation method. The specific approach is that the window start time is no longer a fixed value, but the dipole bending wave arrival time curve is used as the window start time. The purpose of this implementation is to reduce the influence of noise. After analysis, a high-resolution dipole bending wave velocity curve is obtained, such as Figure 4 As shown in the seventh track, compared with the low-resolution velocity curve corresponding to the first-order flexural wave shown in the sixth track, the high-resolution flexural wave velocity curve shows more detailed information about the formation changes, especially in Figure 4 Arrows indicate depth positions 2318.5m, 2332.5m and 2336m.

[0034] At this point, the entire high-resolution dipole bending wave velocity inversion method process is completed.

Claims

1. A high-resolution dipole bending wave velocity inversion method, characterized in that: The following steps are involved: S1. Collect conventional logging data and array acoustic logging data of the well; S2, performing curve preprocessing on conventional logging data and array acoustic logging data to obtain monopole array waveform data and dipole array waveform data; S3, using a velocity extraction method to process the monopole array waveform data to obtain low-resolution longitudinal wave velocity and longitudinal wave arrival time, and using a velocity extraction method to process the dipole array waveform data to obtain low-resolution bending wave velocity; S4. Based on the well diameter, density, longitudinal wave velocity and bending wave velocity, a numerical model of the well is established and a forward simulation is performed to obtain a simulated dipole bending waveform; S5. Perform frequency-wavenumber domain analysis on the simulated dipole bending waveform to obtain the frequency distribution range of the high-order bending wave; S6, selecting a high-order bending wave frequency distribution range to perform filtering processing on the measured dipole array waveform data to obtain high-order bending wave array waveform data; S7. Perform correlation analysis of the high-order bending wave array waveform data with bending wave arrival time constraints to obtain high-resolution dipole bending wave velocity.

2. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S1, the conventional logging data of the well mainly include natural gamma, well diameter, and density curve; the array acoustic logging data of the well mainly include the original waveform, gain curve and delay curve of the monopole array acoustic logging and the original waveform, gain curve and delay curve of the dipole array acoustic logging.

3. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S2, the curve preprocessing of array acoustic logging data mainly includes gain recovery and delay recovery. Gain recovery is performed first, and then delay recovery. During the waveform acquisition process, in order to reduce the storage volume of recorded data, the waveform data of a period of time before the arrival of the first wave is often not collected. The purpose of delay recovery is to fill zeros before the zero moment of the monopole array acoustic wave waveform, so as to obtain waveform data with accurate time.

4. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S3, the velocity extraction method is a correlation method. For the pre-processed monopole array acoustic wave waveform data, the correlation method is selected to extract the P-wave velocity curve therefrom; by calculating the correlation function of the array waveform in two dimensions of time and velocity, the velocity at which the maximum function value is located is the low-resolution P-wave velocity; based on the low-resolution P-wave velocity, the P-wave arrival time curve can be directly calculated on the basis of considering the well diameter and the length of the logging tool; The correlation method is used to extract the shear wave velocity curve from the preprocessed dipole array acoustic waveform data. The correlation function of the array waveform is calculated in the two dimensions of time and velocity. The velocity with the maximum function value is the low-resolution dipole bending wave velocity.

5. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S4, a comprehensive analysis is performed on the wellbore diameter, density, P-wave velocity and dipole bending wave velocity curves of the well to find out the layer section with no wellbore diameter collapse and no sudden change in density value, P-wave velocity and dipole bending wave velocity; The density, wellbore diameter, compressional wave velocity and dipole flexural wave velocity of the interval are taken as input parameters, and the forward simulation is carried out using the numerical simulation method to calculate the dipole flexural wave array waveform.

6. A high-resolution dipole bending wave velocity inversion method according to claim 5, characterized in that: The numerical simulation method adopts real axis integration method or finite difference method.

7. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S5, the frequency-wavenumber domain analysis mainly involves performing a two-dimensional Fourier transform on the simulated dipole bending waveform data to obtain a frequency-wavenumber two-dimensional spectrum corresponding to the simulated dipole bending wave array waveform; In the frequency-wavenumber two-dimensional spectrum, the horizontal axis is set to wavenumber and the vertical axis is set to frequency; from it, a first-order bending wave mode with a lower frequency and a second-order bending wave mode with a higher frequency can be observed; the frequency band distribution range corresponding to the second-order bending wave mode is the frequency distribution range of the higher-order bending wave.

8. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S6, the filtering process refers to setting a digital bandpass filter, the frequency band passing range of which is set to the frequency band distribution range of the high-order dipole bending wave obtained in S5, and using the digital bandpass filter to process the dipole bending wave array waveform data measured in S4, thereby filtering out low-order bending wave information, and the retained waveform data is the high-order bending wave array waveform data.

9. A high-resolution dipole bending wave velocity inversion method according to claim 8, characterized in that: The frequency band passing range of the digital bandpass filter is the frequency band distribution range of the high-order dipole bending wave obtained in S5.

10. A high-resolution dipole bending wave velocity inversion method according to claim 1, characterized in that: In S7, the correlation analysis of the bending wave arrival time constraint is performed by setting the window start time to the time of the dipole bending wave arrival time curve based on the existing correlation method to reduce the influence of noise on the result.

Images