Method for improving thin interbed oil reservoir fluid earthquake prediction

Abstract

A method for improving thin interbed oil reservoir fluid earthquake prediction precision relates to the field of oil exploration and development, and comprises the following steps: carrying out rock geophysical analysis through logging information, and screening earthquake attribute types sensitive to oil reservoir fluid; based on the sensitive attribute type, performing resolution enhancement processing on the seismic data to generate optimized seismic attribute data; stratigraphic slicing processing is carried out on the optimized seismic attribute data along a target horizon, and seismic fluid sensitive attributes are extracted; performing time calibration and regression modeling on the well logging fluid interpretation result and the seismic fluid sensitive attribute, and establishing a fluid distribution prediction equation; based on the fluid distribution prediction equation, seismic fluid sensitive attributes are converted into an oil reservoir fluid plane distribution diagram, and thin interbed fluid prediction is achieved.

Description

Technical Field

[0001] The present invention relates to the field of oil exploration and development, and particularly to a method for improving seismic prediction of fluids in thin interbedded reservoirs. Background Art

[0002] In oil exploration and development, thin interbedded reservoirs are a common and economically important type of reservoir. However, due to the fact that the thickness of thin interbeds is usually less than the seismic resolution, the formation and fluid information show a complex combined response in seismic data, resulting in relatively low prediction accuracy of traditional seismic prediction methods for fluids in thin interbedded reservoirs and making it difficult to accurately identify the oil-bearing areas, which poses a great challenge to the effective development of reservoirs. Therefore, a new method is needed to improve the accuracy of seismic prediction of fluids in thin interbedded reservoirs. Summary of the Invention

[0003] The object of the present invention is to provide a method for improving seismic prediction of fluids in thin interbedded reservoirs, which can effectively overcome the limitations of traditional methods and improve the prediction accuracy of the fluid distribution in thin interbedded reservoirs.

[0004] A method for improving the seismic prediction accuracy of fluids in thin interbedded reservoirs, comprising the following steps: Step (1): Conduct rock geophysical analysis through well logging data to screen the types of seismic attributes sensitive to reservoir fluids; Step (2): Based on the sensitive attribute types in step (1), perform resolution enhancement processing on the seismic data to generate optimized seismic attribute data; Step (3): Perform formation slicing processing on the optimized seismic attribute data along the target horizon to extract seismic fluid-sensitive attributes; Step (4): Perform time calibration and regression modeling on the well logging fluid interpretation results and the seismic fluid-sensitive attributes to establish a fluid distribution prediction equation; Step (5): Based on the fluid distribution prediction equation, convert the seismic fluid-sensitive attributes into a planar distribution map of reservoir fluids to achieve prediction of thin interbedded fluids.

[0005] Preferably, the rock geophysical analysis in step (1) includes: Through well logging curve crossplot analysis, determine the sensitive attribute types when containing different fluids, and the sensitive attribute types include at least one of amplitude type, frequency type, phase type, and wave impedance.

[0006] Preferably, the resolution enhancement processing in step (2) includes at least one of the following techniques: Spectrum decomposition technology, prestack inversion technology, frequency compensation technology, or wavelet shaping technology, which are used to improve the signal-to-noise ratio and resolution of thin interbed reflection characteristics.

[0007] Preferably, the formation slicing process in step (3) includes: Based on the principles of seismic sedimentology, isochronous formation slices or non-isochronous waveform tracking slices are used to extract the planar distribution characteristics of seismic attributes within the target interval.

[0008] Preferably, the regression modeling in step (4) includes: When the correlation between well logging fluid data and seismic attributes is insufficient, clustering analysis or support vector machines are used to classify reservoir fluid types, and a regression model is established for each type of fluid.

[0009] Preferably, the classification basis includes: At least one of reservoir physical property parameters, seismic waveform characteristics, and fluid saturation distribution.

[0010] Preferably, the fluid planar distribution map in step (5) includes: At least one of a saturation distribution map, a fluid abundance map, or an oil-water boundary prediction map, and geological structure constraints and sedimentary facies boundary corrections are introduced in its generation process.

[0011] Preferably, the method is applicable to thin interbed oil reservoirs with a burial depth greater than 1500 meters and a single layer thickness less than the tuning thickness.

[0012] Constructing a petrophysical model and fluid substitution Based on the actual rock characteristic data such as rock mineral composition, porosity, and permeability of the target thin interbed oil reservoir, an accurate petrophysical model is established. This model takes into account the rock skeleton, pore fluids in the formation, and their interactions.

[0013] Using the established petrophysical model, substitution simulations are performed on different types of fluids (such as oil, gas, water) in the pores to analyze the influence of fluid property changes on rock elastic parameters.

[0014] Seismic characteristic parameter extraction and optimization On the pre-stack seismic data volume, multiple seismic characteristic parameters related to fluids are extracted, including amplitude attributes (such as root mean square amplitude, average absolute amplitude), frequency attributes (such as dominant frequency, frequency band width), phase attributes, etc.

[0015] Using multi-parameter analysis methods such as correlation analysis and principal component analysis, the extracted seismic characteristic parameters are optimized, redundant information is removed, and characteristic parameters sensitive to fluid identification in thin interbed oil reservoirs are selected.

[0016] Establishing a seismic response template library According to the petrophysical model and fluid substitution results, combined with geostatistical methods, simulate the seismic response characteristics under different thin interbed reservoir geological models, and establish a seismic response template library. This template library covers the seismic response characteristics corresponding to thin interbed reservoirs with different thicknesses, different fluid types and saturations.

[0017] During the process of establishing the template library, factors such as absorption attenuation and scattering in the seismic wave propagation process are considered to make the template library more in line with the actual seismic data situation.

[0018] Matching identification and prediction Match the actually collected seismic data with the seismic response template library, and use the similarity calculation method to find the template that is most similar to the actual seismic response.

[0019] According to the matching results, combined with geological background information and expert experience, predict and interpret the fluid properties (oil, gas, water) and distribution in the thin interbed reservoir. Description of the drawings

[0020] Figure 1 It is a flow chart of the method for improving the seismic prediction of fluids in thin interbed reservoirs of the present invention. Detailed implementation manners

[0021] The method of the present invention will be described in detail below in conjunction with specific embodiments.

[0022] Embodiment 1 (continental sandstone-mudstone thin interbed reservoir) 1. Petrophysical analysis: For a sandstone-mudstone thin interbed with a burial depth of 1800 meters (single layer thickness 2 - 5 meters), the wave impedance attribute sensitive to the oil-water interface (amplitude change rate > 15%) is screened out through the natural gamma-resistivity cross plot.

[0023] 2. Resolution enhancement: The frequency band of the seismic data is expanded using the spectral decomposition technique (main frequency 30 - 50 Hz), and the signal-to-noise ratio of the thin layer reflection signal is increased by 40%.

[0024] 3. Stratigraphic slicing: Based on the principle of isochronous slicing, extract the planar distribution of wave impedance within the target interval, and it is found that the coincidence degree between the boundary of the channel sand body and the amplitude anomaly area reaches 85%.

[0025] 4. Regression modeling: Using the water saturation data of 16 wells, establish a linear regression model of wave impedance - saturation (R² = 0.78), and the oil-bearing probability in the low wave impedance area (< 6500 m / s•g / cm³) is predicted to exceed 70%.

[0026] 5. Map correction: Combine the fault distribution to perform spatial interpolation constraints on the prediction results, and finally the error between the fluid abundance map and the subsequent drilling verification is less than 12%.

[0027] Example 2 (marine carbonate thin interbedded reservoir) 1. Attribute screening: In the thin interbeds of reef flat bodies (single layer thickness 1 - 3 m), the frequency attenuation attribute is determined as a sensitive parameter through the density - longitudinal wave velocity cross - plot (oil layer attenuation gradient > 0.8 dB / ms).

[0028] 2. Inversion optimization: Using pre - stack elastic inversion technology (AVO attribute constraint), a Poisson's ratio data volume is inverted, and the longitudinal resolution is improved to 3 m.

[0029] 3. Non - isochronous slicing: Conduct waveform tracking slicing along the top interface of the reef body, extract the plane distribution characteristics of frequency attenuation, and identify 3 "low - frequency high - attenuation" fluid anomaly areas.

[0030] 4. Machine learning modeling: When well data is sparse, a support vector machine (kernel function RBF) is used to classify seismic attributes and core oil - bearing data, and the model verification accuracy reaches 82%.

[0031] 5. Facies - controlled mapping: Introduce sedimentary micro - facies boundaries (such as reef core / reef flank) to correct the saturation distribution in zones, and reduce the prediction error of the oil - water boundary from 20% to 8%.

[0032] Example 3 (deep tight sandstone thin interbedded reservoir) 1. Multi - attribute fusion: For tight sandstone with a burial depth of 3500 m (porosity < 8%), a fluid - sensitive factor is constructed by combining amplitude - type (root - mean - square amplitude) and frequency - type (main frequency shift) attributes.

[0033] 2. Wavelet shaping processing: Use zero - phasing processing to eliminate the interference of wavelet sidelobes, so that the peak - to - valley time difference resolution of the top and bottom reflections of thin interbeds reaches λ / 8 (λ is the wavelength).

[0034] 3. Dynamic slicing verification: Compare the differences in formation slices of two seismic data before and after development, and identify the remaining oil distribution areas (areas with an attribute change rate > 25%).

[0035] 4. Application of clustering analysis: When the fluid phase change is complex, use K - means clustering to divide the reservoir into 4 types of fluid units, and establish piece - wise regression models of attribute - saturation respectively.

[0036] 5. 3D visualization: Integrate the fluid distribution map with the structural undulation model to visually display the spatial matching relationship between high - saturation areas and structural traps.

[0037] Experimental effect verification Through the application of the above - mentioned examples: In the actual application of the oilfield, the prediction thickness accuracy of thin - interbed fluids is improved from ±5 m to ±1.5 m; The coincidence rate of fluid type discrimination has increased from 65% to 88%, and the success rate of exploration and development wells has increased by 32%. After the development plan was adjusted, the production of each well increased by 25%, which verified the industrial applicability of the technical solution.

[0038] The above embodiments are only used to illustrate the technical solutions and application effects of the present invention, and are not intended to limit the present invention. Without departing from the core technology of the present invention, those skilled in the art can modify and change the above embodiments, but these modifications and changes should fall within the protection scope of the present invention.

Claims

1. A method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs, characterized in that, It includes the following steps: Step (1): Conduct rock geophysical analysis through logging data to screen seismic attribute types sensitive to reservoir fluids; Step (2): Based on the sensitive attribute types in Step (1), perform resolution enhancement processing on seismic data to generate optimized seismic attribute data; Step (3): Conduct formation slicing processing on the optimized seismic attribute data along the target horizon to extract seismic fluid-sensitive attributes; Step (4): Perform time calibration and regression modeling on the logging fluid interpretation results and seismic fluid-sensitive attributes to establish a fluid distribution prediction equation; Step (5): Based on the fluid distribution prediction equation, convert the seismic fluid-sensitive attributes into a reservoir fluid plane distribution map to achieve thin interbed fluid prediction.

2. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 1, characterized in that, The rock geophysical analysis in Step (1) includes: determining sensitive attribute types when containing different fluids through logging curve crossplot analysis, and the sensitive attribute types include at least one of amplitude type, frequency type, phase type, and wave impedance.

3. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 2, wherein The resolution enhancement processing in Step (2) includes at least one of the following techniques: spectral decomposition technique, prestack inversion technique, frequency compensation technique, or wavelet shaping technique, which is used to improve the signal-to-noise ratio and resolution of thin interbed reflection characteristics.

4. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 3, characterized in that, The formation slicing processing in Step (3) includes: based on the principles of seismic sedimentology, using isochronous formation slicing or non-isochronous waveform tracking slicing to extract the plane distribution characteristics of seismic attributes within the target interval.

5. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 4, characterized in that, The regression modeling in Step (4) includes: when the correlation between logging fluid data and seismic attributes is insufficient, using cluster analysis or support vector machine to classify reservoir fluid types, and establishing a regression model for each type of fluid respectively.

6. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 5, characterized in that, The classification basis includes at least one of reservoir physical property parameters, seismic waveform characteristics, and fluid saturation distribution.

7. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 6, characterized in that, The fluid plane distribution map in Step (5) includes at least one of saturation distribution map, fluid abundance map, or oil-water boundary prediction map, and its generation process introduces geological structure constraints and sedimentary facies boundary correction.

8. The method for improving the seismic prediction accuracy of fluid in thin interbedded reservoirs according to claim 7, characterized in that, The method is applicable to thin interbed oil reservoirs with a burial depth greater than 1500 meters and a single layer thickness less than the tuning thickness.

Images