CN117388919A - DNN-based method for predicting transverse wave speed of tight oil reservoir - Google Patents
DNN-based method for predicting transverse wave speed of tight oil reservoir Download PDFInfo
- Publication number
- CN117388919A CN117388919A CN202311101649.8A CN202311101649A CN117388919A CN 117388919 A CN117388919 A CN 117388919A CN 202311101649 A CN202311101649 A CN 202311101649A CN 117388919 A CN117388919 A CN 117388919A
- Authority
- CN
- China
- Prior art keywords
- oil reservoir
- density
- shear
- fluid
- modulus
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 239000012530 fluid Substances 0.000 claims abstract description 77
- 238000013528 artificial neural network Methods 0.000 claims abstract description 20
- 239000011435 rock Substances 0.000 claims description 39
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 26
- 239000000203 mixture Substances 0.000 claims description 25
- 239000007787 solid Substances 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 18
- 238000011160 research Methods 0.000 claims description 17
- 229920006395 saturated elastomer Polymers 0.000 claims description 16
- 239000002245 particle Substances 0.000 claims description 12
- 239000011148 porous material Substances 0.000 claims description 12
- 239000007790 solid phase Substances 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 11
- 230000000694 effects Effects 0.000 claims description 10
- 239000007791 liquid phase Substances 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 239000012071 phase Substances 0.000 claims description 7
- 230000008878 coupling Effects 0.000 claims description 6
- 238000010168 coupling process Methods 0.000 claims description 6
- 238000005859 coupling reaction Methods 0.000 claims description 6
- 230000035699 permeability Effects 0.000 claims description 6
- 239000002023 wood Substances 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000002834 transmittance Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 abstract description 10
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000003062 neural network model Methods 0.000 abstract description 3
- 230000007547 defect Effects 0.000 abstract description 2
- 238000004519 manufacturing process Methods 0.000 abstract description 2
- 229910052500 inorganic mineral Inorganic materials 0.000 description 4
- 239000011707 mineral Substances 0.000 description 4
- 235000019738 Limestone Nutrition 0.000 description 2
- 239000006028 limestone Substances 0.000 description 2
- 239000004576 sand Substances 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000004927 clay Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000010459 dolomite Substances 0.000 description 1
- 229910000514 dolomite Inorganic materials 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- -1 sandstone Substances 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/30—Analysis
- G01V1/303—Analysis for determining velocity profiles or travel times
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/282—Application of seismic models, synthetic seismograms
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/30—Analysis
- G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/61—Analysis by combining or comparing a seismic data set with other data
- G01V2210/616—Data from specific type of measurement
- G01V2210/6169—Data from specific type of measurement using well-logging
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/62—Physical property of subsurface
- G01V2210/622—Velocity, density or impedance
- G01V2210/6222—Velocity; travel time
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/62—Physical property of subsurface
- G01V2210/624—Reservoir parameters
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/62—Physical property of subsurface
- G01V2210/624—Reservoir parameters
- G01V2210/6242—Elastic parameters, e.g. Young, Lamé or Poisson
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/60—Analysis
- G01V2210/66—Subsurface modeling
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses a method for predicting the shear wave velocity of a tight oil reservoir based on DNN, which is a method for predicting the shear wave velocity of the tight oil reservoir by considering the fact that a Biot model containing viscous fluid is utilized to characterize the elastic wave propagation of the tight oil reservoir, using a fully connected neural network to learn elastic parameters in the Biot model containing viscous fluid, further constructing a neural network model for learning the elastic wave propagation characteristics of the tight oil reservoir, solving the shear wave velocity through a plane wave analysis method, verifying by using actual logging data, and further realizing machine learning under the constraint of a wave propagation equation. The method has high prediction precision, can make up for the defects of high requirements on data quantity and data quality and lack of physical significance when pure data drive machine learns to predict the transverse wave speed, and has better popularization prospect in actual production.
Description
Technical Field
The invention belongs to the field of unconventional oil and gas reservoir transverse wave speed prediction, and particularly relates to a method for predicting a dense oil reservoir transverse wave speed based on DNN.
Background
The shear wave velocity is significant in the AVA, AVO analysis, pre-stack seismic inversion, reservoir lithology, physical properties, fluid prediction, etc. of seismic data, and is essential when seismic inversion is performed on the basis of logging data. However, in the actual production process, because of the great difficulty of interpretation and high cost, the transverse wave logging information is relatively lack, and the predicted transverse wave speed can solve the problems, so the research on the prediction of the transverse wave speed is an important direction.
Currently, there are three main methods for predicting transverse wave velocity: empirical formula methods, petrophysical model methods, and neural network methods. The empirical formula method mainly utilizes the longitudinal wave velocity information in logging data to empirically calculate the transverse wave velocity by establishing the relationship among parameters such as longitudinal wave velocity, transverse wave velocity, rock mineral composition, porosity, density, clay content and the like. Pickett et al (1963) gives a longitudinal and transverse wave empirical formula based on data analysis of a large amount of limestone; castagna et al (1985) provide empirical regression formulas for longitudinal and transverse waves between dolomite, sandstone, limestone and shale in a water-saturated state based on a large number of study log data; han et al (1986) measured the wave velocity of 75 sandstone samples at different pressures, and obtained an empirical regression equation between longitudinal and transverse wave velocities. However, the empirical formula method is only suitable for oil and gas reservoirs with single rock mineral components, is easily affected by factors such as regional limitation, has low universality and has lower prediction accuracy. The petrophysical model method improves universality and accuracy of transverse wave speed prediction by establishing relations between reservoir physical parameters, elastic parameters and the like and transverse wave speed, and Xu and White (1995,1996) are obtained by adopting micro equivalent medium theory,The combination of the model (1974) and Gassmann (1951) theory suggests a theoretical model suitable for tight sandstone reservoirs, the Xu-White model. In the calculation using the Xu-White model, the pore aspect ratio of sand and mudstone is assumed to be constant, however, a lot of researches show that the pore aspect ratio of sand and mudstone is not constant in the depth range, and is affected by temperature, depth, pressure and the like, and the variation range is large, so that the pore properties cannot be accurately described using a fixed pore aspect ratio. However, in the petrophysical modeling process, various influencing factors such as rock skeleton mineral composition and structure, rock pore structure and connectivity, rock pore fluid and the like need to be accurately determined, and an appropriate petrophysical model is selected according to different actual conditions so as to achieve the expected target. With the rapid development of artificial intelligence machine learning methods such as DNN (Deep Neural Networks, deep neural network) and the improvement of computational level, intelligent methods based on machine learning have attracted a great deal of attention in the field of oil and gas exploration. Machine learning techniques have been applied to predict shear wave velocities using log data, such as SVR (SupportVector Regression ), GRU (gated recurrent neural network, gated loop unit), LSTM (Long Short-Term Memory network), and GB (Gradient Boosting, gradient enhancement). Most machine learning methods of predicting shear wave velocity are purely data driven, often requiring training of the network using shear wave velocity, and thus are not suitable for situations where shear wave data is lacking. Furthermore, purely data-driven machine learning methods lack physical significance.
Disclosure of Invention
The technical problems to be solved by the invention are as follows: aiming at the problems in the prior art, a method for predicting the transverse wave speed of a tight oil reservoir based on DNN is provided, so as to solve the defects that the existing petrophysical model is low in prediction precision and the existing transverse wave speed model based on machine learning pure data driving lacks physical significance. According to the invention, the elastic parameters in the Biot equation containing the viscous fluid are learned by constructing the fully-connected neural network by utilizing the Biot equation containing the viscous fluid, which is proposed by Lu Minghui et al, and then the transverse wave speed is obtained by a plane wave analysis method, so that a deep neural network model under the constraint condition of the wave equation is constructed, and the prediction precision of the transverse wave speed is improved.
In order to solve the technical problems, the invention provides the following technical scheme: a method for predicting a tight oil reservoir shear wave velocity based on DNN, comprising the steps of:
s1, acquiring logging data of a target research area, dividing a training set and a testing set, and calculating fluid mixture density, fluid mixture viscosity, dissipation coefficient and solid phase, liquid phase and fluid-solid coupling phase density parameters based on the logging data of the target research area;
s2, calculating the bulk modulus and the shear modulus of the dry skeleton of the rock based on a Krief model, and calculating the bulk modulus of the mixed fluid based on a Wood model;
s3, calculating the bulk modulus and the shear modulus of the fluid saturated rock and the solid particles based on the longitudinal wave speed, the transverse wave speed and the density in the known logging data and the bulk modulus and the shear modulus of the dry skeleton of the rock obtained in the step S2;
s4, calculating elastic parameters according to the bulk modulus, the shear modulus and the bulk modulus of the mixed fluid of the dry rock skeleton obtained in the step S2 and the bulk modulus and the shear modulus of the solid particles obtained in the step S3 and the elastic wave equation of the porous medium containing viscous fluid, and solving the longitudinal wave velocity and the transverse wave velocity; taking porosity, mixed fluid bulk modulus, mixed fluid shear modulus, total density, transmittance and water saturation in logging data as input, and corresponding elastic parameters and longitudinal wave speed and transverse wave speed as output, constructing and training a neural network containing elastic parameters of viscous fluid wave propagation equation to obtain a tight oil reservoir transverse wave speed prediction model;
and S5, verifying the tight oil reservoir transverse wave speed prediction model by using a test set, predicting the tight oil reservoir longitudinal wave speed and the tight oil reservoir transverse wave speed, and calculating root mean square error and a decision coefficient in the prediction process to express the deviation between the predicted value and the actual value so as to evaluate the prediction effect of the model.
Further, the foregoing step S1 specifically includes the following substeps:
s101, acquiring logging data of a target research area, wherein the logging data comprise: the total density of the tight oil reservoir, the density of water, the density of oil, the water saturation, the viscosity of water, the viscosity of oil, the porosity, the permeability,
s102, calculating the density of the fluid mixture and the viscosity of the fluid mixture according to the following formula:
ρ f ＝S w ρ f1 +(1-S w )ρ f2 (1)
wherein: ρ f For the density of the fluid mixture ρ f1 、ρ f2 Respectively the density of water and the density of oil, S w Is the water saturation, eta is the viscosity of the fluid mixture, eta 1 、η 2 The viscosity of water and the viscosity of oil respectively;
s103, calculating to obtain solid density according to the total density of the dense oil layer of the target research area and the density of the fluid mixture in the formula (1):
wherein: ρ is the total density and φ is the porosity;
s104, calculating and obtaining dissipation coefficients and solid phase, liquid phase and fluid-solid coupling phase density parameters based on a Biot theory:
ρ 12 ＝[1-α(1+1/φ)]φρ f (5)
ρ 11 ＝(1-φ)ρ s -ρ 12 (6)
ρ 22 ＝φρ f -ρ 12 (7)
wherein: b is dissipation coefficient, kappa is permeability, ρ 11 、ρ 12 、ρ 22 Respectively solid phase density parameter, liquid phase density parameter and flow-solid phase density parameter, and alpha is average pore radius.
Further, the step S2 includes the following sub-steps:
s201, calculating the bulk modulus and the shear modulus of the rock dry skeleton based on a Krief model:
K b ＝K s (1-φ) m(φ) (8)
μ b ＝μ s (1-φ) m(φ) (9)
wherein: k (K) b 、μ b The bulk modulus, the shear modulus and the K of the dry skeleton of the rock are respectively s 、μ s Bulk modulus and shear modulus of solid particles respectively;
s202, calculating based on a Wood model to obtain the bulk modulus of the mixed fluid:
K f ＝S w K w +(1-S w )K o (11)
wherein: k (K) f 、K w 、K o Bulk modulus of the mixed fluid, bulk modulus of water, bulk modulus of oil, respectively.
Further, the step S3 includes the following sub-steps:
s301, calculating longitudinal wave velocity and transverse wave velocity in logging data based on Gassmann equation:
wherein: v (V) p-log 、V s-log Longitudinal wave velocity and transverse wave velocity, K and mu are bulk modulus and shear modulus of fluid saturated rock,
the bulk modulus and the shear modulus of the fluid saturated rock are obtained by the simultaneous formula (12) and the formula (13);
s302, obtaining the bulk modulus and the shear modulus of the fluid saturated rock based on a Gassmann equation, wherein the bulk modulus and the shear modulus of the fluid saturated rock have the following relation:
μ＝μ b (15)
the simultaneous formulas (8) - (10) and (12) - (15) give bulk modulus and shear modulus of the solid particles.
Further, in the aforementioned step S4, the elasticity parameter is calculated according to the porous medium elastic wave equation containing the viscous fluid, as follows:
N＝μ b (17)
wherein: A. n, Q, R is an elasticity parameter.
Further, in the aforementioned step S4, the longitudinal wave velocity is solved as follows:
a 0 k 4 +a 1 k 2 +a 2 ＝0 (21)
wherein: k is complex number, re (·) represents the real part of (·), V p In order to be the velocity of the longitudinal wave,
wherein: a, a 0 ＝(A+2N)(R-2ηωi)-Q 2 (23)
Solving the transverse wave velocity as follows:
b 0 k 4 +b 1 k 2 +b 2 ＝0 (26)
wherein:
b 0 ＝-Nηωi (28)
b 2 ＝a 2 (30)
further, in the step S4, the network including the elastic parameter of the viscous fluid wave propagation equation is a fully connected neural network, and includes an input layer, an output layer, and 5 hidden layers, each hidden layer has 50 units, and the activation function is a hyperbolic tangent function; optimizing the network structure by using an adaptive moment estimation optimizer, wherein the initial learning rate is 0.001; taking the mean square error as a loss function evaluation index, the training process updates the network parameters, namely the weight and the deviation, so as to minimize the loss function and the expression of the elastic parameter DNN:
wherein: θ 1 、θ 2 、θ 3 And theta 4 Is a neural network parameter, which is determined by minimizing a loss function, for the elastic parameters learned by neural networks, < +.>Parameters are input for the DNN model.
Further, in the step S4, the model for predicting the shear wave velocity of the tight oil reservoir is as follows:
wherein: θ= |θ 1 ,θ 2 ,θ 3 ,θ 4 |；X＝(φ,K,μ,ρ,κ,S w )，For the longitudinal wave speed and the transverse wave speed predicted by the network, the MSE loss function expression is:
wherein: v (V) p,i And V s,i Longitudinal and transverse velocities respectively for the ith data point in the training set,and->Longitudinal wave velocity and transverse wave velocity of the ith data point predicted by the network respectively, and n is the number of training set samples.
Further, the step S5 specifically includes: the root mean square error and the decision coefficient are calculated to express the deviation between the predicted value and the actual value to evaluate the prediction effect of the model as follows:
wherein: n is the number of samples in the test set; y is the speed;is y i Average number of (i=1, 2,., n), wherein R 2 The closer to 1 represents the modelThe better the predictive effect of (c).
Compared with the prior art, the beneficial technical effects of the invention adopting the technical scheme are as follows: the invention can realize the combination of machine learning and theoretical model by constructing the fully connected neural network model under the constraint of wave equation, thereby predicting the transverse wave speed of the reservoir in the research area, and has the following advantages compared with the prior art: (1) The method avoids the various influences of the rock skeleton mineral composition and structure, the rock pore structure and connectivity, the rock pore fluid and the like which need to be accurately determined in the process of predicting the transverse wave speed by a rock physical modeling method; (2) The existing machine learning-based method for predicting the transverse wave speed is based on pure data driving, the requirements on the data quantity size and the data quality are high, and the machine learning method based on the pure data driving has no physical significance.
Drawings
Fig. 1 is a schematic flow chart according to the present invention.
Fig. 2 is a DNN-based shear wave velocity prediction flow chart.
Fig. 3 is a graph of longitudinal and transverse wave velocity predictions for actual three-well log data using the technique of the present patent, where (a) is a well 1 prediction curve (b) is a well 2 prediction curve, and (c) is a well 3 prediction curve.
Fig. 4 is a graph of the intersection of the predicted results of the longitudinal and transverse wave velocities with the actual data for the actual three well log data using the technique of this patent, where (a) is the intersection of the predicted longitudinal wave velocities and the longitudinal wave velocities in the actual three well log data for well 1, well 2, and well 3, and (b) is the intersection of the predicted transverse wave velocities and the longitudinal wave velocities in the actual three well log data for well 1, well 2, and well 3.
FIG. 5 is a graph of relative root mean square error and determination coefficients of longitudinal and transverse wave velocity predictions and actual data for actual three-well log data using the technique of the present patent.
Detailed Description
For a better understanding of the technical content of the present invention, specific examples are set forth below, along with the accompanying drawings.
Aspects of the invention are described herein with reference to the drawings, in which there are shown many illustrative embodiments. The embodiments of the present invention are not limited to the embodiments described in the drawings. It is to be understood that this invention is capable of being carried out by any of the various concepts and embodiments described above and as such described in detail below, since the disclosed concepts and embodiments are not limited to any implementation. Additionally, some aspects of the disclosure may be used alone or in any suitable combination with other aspects of the disclosure.
As shown in FIG. 1, the invention provides a method for predicting the transverse wave speed of a tight oil reservoir based on DNN, which comprises the following steps S1 to S5, and the prediction of the transverse wave speed of a target tight oil reservoir is completed.
S1, acquiring logging data of a target research area, dividing a training set and a testing set, and calculating fluid mixture density, fluid mixture viscosity and solid density according to the logging data; then calculating dissipation coefficients and density parameters of solid phase, liquid phase and fluid-solid coupling phases;
19 well logging in the research area are selected for research, and the effective data points of the wells are 23630 in total and used for training and testing the neural network after pretreatment. Because the data volume is different for each well, three of the wells with richer data points are selected as the test set, and the remaining 14830 data points comprise the training set. Step S1 comprises the following sub-steps:
s101, acquiring logging data of a target research area, wherein the logging data comprise: the total density of the tight oil reservoir, the density of water, the density of oil, the water saturation, the viscosity of water, the viscosity of oil, the porosity, the permeability,
s102, when the pore medium saturates the two-phase fluid, calculating the density and viscosity of the fluid mixture according to a calculated fluid mixture density and viscosity formula constructed by Carcione et al, and calculating the density and viscosity of the fluid mixture by using the density of water, the density of oil and the water saturation:
ρ f ＝S w ρ f1 +(1-S w )ρ f2 (1)
wherein: ρ f For the density of the fluid mixture ρ f1 、ρ f2 Respectively the density of water and the density of oil, S w Is the water saturation, eta is the viscosity of the fluid mixture, eta 1 、η 2 The viscosity of water and the viscosity of oil respectively;
s103, calculating to obtain solid density according to the total density of the dense oil layer of the target research area and the density of the fluid mixture in the formula (1):
wherein: ρ is the total density and φ is the porosity;
s104, calculating and obtaining dissipation coefficients and solid phase, liquid phase and fluid-solid coupling phase density parameters based on a Biot theory:
ρ 12 ＝[1-α(1+1/φ)]φρ f (5)
ρ 11 ＝(1-φ)ρ s -ρ 12 (6)
ρ 22 ＝φρ f -ρ 12 (7)
wherein: b is dissipation coefficient, kappa is permeability, ρ 11 、ρ 12 、ρ 22 Respectively solid phase density parameter, liquid phase density parameter and flow-solid phase density parameter, and alpha is average pore radius.
S2, calculating the bulk modulus and the shear modulus of the dry rock framework based on a Krief model, and calculating the bulk modulus of the mixed fluid based on a Wood model, wherein the method comprises the following substeps:
s201, calculating the bulk modulus and the shear modulus of the rock dry skeleton based on a Krief model:
K b ＝K s (1-φ) m(φ) (8)
μ b ＝μ s (1-φ) m(φ) (9)
wherein: k (K) b 、μ b The bulk modulus, the shear modulus and the K of the dry skeleton of the rock are respectively s 、μ s Bulk modulus and shear modulus of solid particles respectively;
s202, calculating based on a Wood model to obtain the bulk modulus of the mixed fluid:
K f ＝S w K w +(1-S w )K o (11)
wherein: k (K) f 、K w 、K o Bulk modulus of the mixed fluid, bulk modulus of water, bulk modulus of oil, respectively.
S3, calculating the bulk modulus and the shear modulus of the fluid saturated rock and the solid particles based on the longitudinal wave speed, the transverse wave speed and the density in the known logging data and the bulk modulus and the shear modulus of the dry skeleton of the rock obtained in the step S2; step S3 comprises the following sub-steps:
s301, calculating longitudinal wave velocity and transverse wave velocity in logging data based on Gassmann equation:
wherein: v (V) p-log 、V s-log Longitudinal wave velocity and transverse wave velocity, K and mu are bulk modulus and shear modulus of fluid saturated rock,
the bulk modulus and the shear modulus of the fluid saturated rock are obtained by the simultaneous formula (12) and the formula (13);
s302, obtaining the bulk modulus and the shear modulus of the fluid saturated rock based on a Gassmann equation, wherein the bulk modulus and the shear modulus of the fluid saturated rock have the following relation:
μ＝μ b (15)
the simultaneous formulas (8) - (10) and (12) - (15) give bulk modulus and shear modulus of the solid particles.
S4, calculating elastic parameters according to the bulk modulus, the shear modulus and the bulk modulus of the mixed fluid of the dry rock skeleton obtained in the step S2 and the bulk modulus and the shear modulus of the solid particles obtained in the step S3 and the elastic wave equation of the porous medium containing viscous fluid, and solving the longitudinal wave velocity and the transverse wave velocity; taking porosity, mixed fluid bulk modulus, mixed fluid shear modulus, total density, transmittance and water saturation in logging data as input, and corresponding elastic parameters and longitudinal wave speed and transverse wave speed as output, constructing and training a neural network containing elastic parameters of viscous fluid wave propagation equation to obtain a tight oil reservoir transverse wave speed prediction model;
in step S4, the calculated elastic parameters are specifically: the elastic parameter expression of the porous medium elastic wave equation containing viscous fluid deduced on the basis of Lu Minghui is obtained as follows:
N＝μ b (17)
wherein: A. n, Q, R is an elasticity parameter.
Based on Lu Minghui derived porous medium elastic wave equation containing viscous fluid, the solved longitudinal wave velocity expression is obtained as follows:
a 0 k 4 +a 1 k 2 +a 2 ＝0 (21)
wherein: k is complex number, re (·) represents the real part of (·), V p Is the longitudinal wave velocity. Wherein:
a 0 ＝(A+2N)(R-2ηωi)-Q 2 (23)
based on Lu Minghui derived porous medium elastic wave equation containing viscous fluid, the solution transverse wave velocity expression is obtained as follows:
b 0 k 4 +b 1 k 2 +b 2 ＝0 (26)
wherein:
b 0 ＝-Nηωi (28)
b 2 ＝a 2 (30)
in step S4, the network containing elastic parameters of the viscous fluid wave propagation equation is a fully connected neural network, referring to fig. 2, the network includes an input layer, an output layer, and 5 hidden layers, each hidden layer has 50 units, and the activation function is a hyperbolic tangent function; optimizing the network structure by using an adaptive moment estimation optimizer, wherein the initial learning rate is 0.001; taking the mean square error as a loss function evaluation index, the training process updates the network parameters, namely the weight and the deviation, so as to minimize the loss function and the expression of the elastic parameter DNN:
wherein: θ 1 、θ 2 、θ 3 And theta 4 Is a neural network parameter, which is determined by minimizing a loss function, for elastic parameters learned by neural networks，/>Parameters are input for the DNN model.
And constructing a tight oil reservoir shear wave velocity prediction model by using the expression of the elastic parameter DNN learned by the neural network, wherein the expression is as follows:
wherein: θ= |θ 1 ,θ 2 ,θ 3 ,θ 4 |；X＝(φ,K,μ,ρ,κ,S w )，For the longitudinal wave speed and the transverse wave speed predicted by the network, the MSE loss function expression is:
wherein: v (V) p,i And V s,i Longitudinal and transverse velocities respectively for the ith data point in the training set,and->Longitudinal wave velocity and transverse wave velocity of the ith data point predicted by the network respectively, and n is the number of training set samples.
S5, verifying a tight oil reservoir shear wave speed prediction model by using a test set, predicting the tight oil reservoir longitudinal wave speed and the tight oil reservoir shear wave speed, and calculating root mean square error and a decision coefficient in the prediction process to express the deviation between a predicted value and an actual value so as to evaluate the prediction effect of the model, wherein the prediction effect is as follows:
wherein: n is the number of samples in the test set; y is the speed;is y i Average number of (i=1, 2,., n), wherein R 2 A closer to 1 indicates a better prediction effect of the model.
The invention aims at selecting 3 wells (1-3 wells) in 7 sections of dense oil layers with the length of the Erdos basin to conduct transverse wave speed prediction research.
Fig. 3 shows the longitudinal and transverse wave velocity prediction curves for actual three well log data using the technique of this patent, (a) is a well 1 prediction curve (b) is a well 2 prediction curve, and (c) is a well 3 prediction curve. The comparison result of the longitudinal wave velocity and the transverse wave velocity (dotted lines) obtained by the prediction of the invention and the longitudinal wave velocity and the transverse wave velocity (black) in the actual logging data can be seen, and the coincidence degree of the longitudinal wave velocity prediction result and the transverse wave velocity prediction result obtained by the prediction of the invention and the actual data is higher.
FIG. 4 shows the intersection of the longitudinal and transverse wave velocity predictions with actual data for actual three well log data using the technique of the present patent. In the figure, (a) is an intersection graph of the longitudinal wave velocity and the predicted longitudinal wave velocity in the actual logging data of the three wells 1,2 and 3, and (b) is an intersection graph of the transverse wave velocity and the predicted transverse wave velocity in the actual logging data of the three wells 1,2 and 3, as can be seen from fig. 4, the predicted result has a strong correlation with the actual data, and the scattered points basically fall near the standard line.
FIG. 5 shows the Relative root mean square error (Relative RMSE) and the determination coefficient (R) 2 ) And (5) calculating. As can be seen from FIG. 5, the present inventionThe relative root mean square error range of the longitudinal wave speed and the transverse wave speed of the predicted three wells is 1% -2.12%, the determination coefficient is between 0.94-0.988, and the overall error is small.
While the invention has been described in terms of preferred embodiments, it is not intended to be limiting. Those skilled in the art will appreciate that various modifications and adaptations can be made without departing from the spirit and scope of the present invention. Accordingly, the scope of the invention is defined by the appended claims.
Claims (9)
1. A method for predicting a tight oil reservoir shear wave velocity based on DNN, comprising the steps of:
s1, acquiring logging data of a target research area, dividing a training set and a testing set, and calculating fluid mixture density, fluid mixture viscosity, dissipation coefficient and solid phase, liquid phase and fluid-solid coupling phase density parameters based on the logging data of the target research area;
s2, calculating the bulk modulus and the shear modulus of the dry skeleton of the rock based on a Krief model, and calculating the bulk modulus of the mixed fluid based on a Wood model;
s3, calculating the bulk modulus and the shear modulus of the fluid saturated rock and the solid particles based on the longitudinal wave speed, the transverse wave speed and the density in the known logging data and the bulk modulus and the shear modulus of the dry skeleton of the rock obtained in the step S2;
s4, calculating elastic parameters according to the bulk modulus, the shear modulus and the bulk modulus of the mixed fluid of the dry rock skeleton obtained in the step S2 and the bulk modulus and the shear modulus of the solid particles obtained in the step S3 and the elastic wave equation of the porous medium containing viscous fluid, and solving the longitudinal wave velocity and the transverse wave velocity; taking porosity, mixed fluid bulk modulus, mixed fluid shear modulus, total density, transmittance and water saturation in logging data as input, and corresponding elastic parameters and longitudinal wave speed and transverse wave speed as output, constructing and training a neural network containing elastic parameters of viscous fluid wave propagation equation to obtain a tight oil reservoir transverse wave speed prediction model;
and S5, verifying the tight oil reservoir transverse wave speed prediction model by using a test set, predicting the tight oil reservoir longitudinal wave speed and the tight oil reservoir transverse wave speed, and calculating root mean square error and a decision coefficient in the prediction process to express the deviation between the predicted value and the actual value so as to evaluate the prediction effect of the model.
2. The method for predicting the shear wave velocity of a tight oil reservoir based on DNN according to claim 1, wherein step S1 comprises the following sub-steps:
s101, acquiring logging data of a target research area, wherein the logging data comprise: the total density of the tight oil reservoir, the density of water, the density of oil, the water saturation, the viscosity of water, the viscosity of oil, the porosity, the permeability,
s102, calculating the density of the fluid mixture and the viscosity of the fluid mixture according to the following formula:
ρ f ＝S w ρ f1 +(1-S w )ρ f2 (1)
wherein: ρ f For the density of the fluid mixture ρ f1 、ρ f2 Respectively the density of water and the density of oil, S w Is the water saturation, eta is the viscosity of the fluid mixture, eta 1 、η 2 The viscosity of water and the viscosity of oil respectively;
s103, calculating to obtain solid density according to the total density of the dense oil layer of the target research area and the density of the fluid mixture in the formula (1):
wherein: ρ is the total density and φ is the porosity;
s104, calculating and obtaining dissipation coefficients and solid phase, liquid phase and fluid-solid coupling phase density parameters based on a Biot theory:
ρ 12 ＝[1-α(1+1/φ)]φρ f (5)
ρ 11 ＝(1-φ)ρ s -ρ 12 (6)
ρ 22 ＝φρ f -ρ 12 (7)
wherein: b is dissipation coefficient, kappa is permeability, ρ 11 、ρ 12 、ρ 22 Respectively solid phase density parameter, liquid phase density parameter and flow-solid phase density parameter, and alpha is average pore radius.
3. A method of predicting a tight oil reservoir shear wave velocity based on DNN according to claim 1, wherein step S2 comprises the sub-steps of:
s201, calculating the bulk modulus and the shear modulus of the rock dry skeleton based on a Krief model:
K b ＝K s (1-φ) m(φ) (8)
μ b ＝μ s (1-φ) m(φ) (9)
wherein: k (K) b 、μ b The bulk modulus, the shear modulus and the K of the dry skeleton of the rock are respectively s 、μ s Bulk modulus and shear modulus of solid particles respectively;
s202, calculating based on a Wood model to obtain the bulk modulus of the mixed fluid:
K f ＝S w K w +(1-S w )K o (11)
wherein: k (K) f 、K w 、K o Bulk modulus of the mixed fluid, bulk modulus of water, bulk modulus of oil, respectively.
4. A method of predicting a tight oil reservoir shear wave velocity based on DNN according to claim 3, wherein step S3 comprises the sub-steps of:
s301, calculating longitudinal wave velocity and transverse wave velocity in logging data based on Gassmann equation:
wherein: v (V) p-log 、V s-log Longitudinal wave velocity and transverse wave velocity, K and mu are bulk modulus and shear modulus of fluid saturated rock,
the bulk modulus and the shear modulus of the fluid saturated rock are obtained by the simultaneous formula (12) and the formula (13);
s302, obtaining the bulk modulus and the shear modulus of the fluid saturated rock based on a Gassmann equation, wherein the bulk modulus and the shear modulus of the fluid saturated rock have the following relation:
μ＝μ b (15)
the simultaneous formulas (8) - (10) and (12) - (15) give bulk modulus and shear modulus of the solid particles.
5. The method for predicting the shear wave velocity of a tight oil reservoir based on DNN according to claim 4, wherein in step S4, the elasticity parameter is calculated according to the porous medium elastic wave equation containing viscous fluid, as follows:
N＝μ b (17)
wherein: A. n, Q, R is an elasticity parameter.
6. The method for predicting the shear wave velocity of a tight oil reservoir based on DNN as recited in claim 5, wherein in step S4, the longitudinal wave velocity is solved by:
a 0 k 4 +a 1 k 2 +a 2 ＝0 (21)
wherein: k is complex number, re (·) represents the real part of (·), V p In order to be the velocity of the longitudinal wave,
wherein: a, a 0 ＝(A+2N)(R-2ηωi)-Q 2 (23)
Solving the transverse wave velocity as follows:
b 0 k 4 +b 1 k 2 +b 2 ＝0 (26)
wherein:
b 0 ＝-Nηωi (28)
b 2 ＝a 2 (30)。
7. the method of predicting a tight oil reservoir shear wave velocity based on DNN of claim 6,
in the step S4, the network containing the elastic parameters of the viscous fluid wave propagation equation is a fully-connected neural network, and comprises an input layer, an output layer and 5 hidden layers, wherein each hidden layer is provided with 50 units, and an activation function is a hyperbolic tangent function; optimizing the network structure by using an adaptive moment estimation optimizer, wherein the initial learning rate is 0.001; taking the mean square error as a loss function evaluation index, the training process updates the network parameters, namely the weight and the deviation, so as to minimize the loss function and the expression of the elastic parameter DNN:
wherein: θ 1 、θ 2 、θ 3 And theta 4 Is a neural network parameter, which is determined by minimizing a loss function,for the elastic parameters learned by neural networks, < +.>Parameters are input for the DNN model.
8. The method of predicting the compressional velocity of a tight oil reservoir based on DNN of claim 7, wherein in step S4, the compressional velocity prediction model of the tight oil reservoir is of the formula:
wherein: θ= |θ 1 ,θ 2 ,θ 3 ,θ 4 |；X＝(φ,K,μ,ρ,κ,S w )，For the longitudinal wave speed and the transverse wave speed predicted by the network, the MSE loss function expression is:
wherein: v (V) p,i And V s,i Longitudinal waves respectively of the ith data point in the training setAnd the velocity of the transverse wave,and->Longitudinal wave velocity and transverse wave velocity of the ith data point predicted by the network respectively, and n is the number of training set samples.
9. The method for predicting the shear wave velocity of a tight oil reservoir based on DNN according to claim 8, wherein step S5 is specifically: the root mean square error and the decision coefficient are calculated to express the deviation between the predicted value and the actual value to evaluate the prediction effect of the model as follows:
wherein: n is the number of samples in the test set; y is the speed;is y i Average number of (i=1, 2,., n), wherein R 2 A closer to 1 indicates a better prediction effect of the model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311101649.8A CN117388919B (en) | 2023-08-29 | 2023-08-29 | DNN-based method for predicting transverse wave speed of tight oil reservoir |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311101649.8A CN117388919B (en) | 2023-08-29 | 2023-08-29 | DNN-based method for predicting transverse wave speed of tight oil reservoir |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117388919A true CN117388919A (en) | 2024-01-12 |
| CN117388919B CN117388919B (en) | 2024-05-28 |
Family
ID=89463802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311101649.8A Active CN117388919B (en) | 2023-08-29 | 2023-08-29 | DNN-based method for predicting transverse wave speed of tight oil reservoir |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117388919B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5265016A (en) * | 1989-07-19 | 1993-11-23 | Amoco Corporation | Method of shear wave velocity estimation |
| CN106680878A (en) * | 2016-12-29 | 2017-05-17 | 中国石油大学(华东) | Transverse wave speed estimation method based on improved Biot coefficient analysis |
| WO2018010628A1 (en) * | 2016-07-15 | 2018-01-18 | 河海大学 | Seismic rock physics inversion method based on a large area tight reservoir |
| CA3067013A1 (en) * | 2017-07-21 | 2019-01-24 | Landmark Graphics Corporation | Deep learning based reservoir modeling |
| WO2020146797A1 (en) * | 2019-01-10 | 2020-07-16 | Baker Hughes Oilfiled Operations Llc | Estimation of fracture properties based on borehole fluid data, acoustic shear wave imaging and well bore imaging |
| WO2021147529A1 (en) * | 2020-01-21 | 2021-07-29 | 长江大学 | Multipoint geostatistical pre-stack inversion method based on updated theory of permanence of probability ratio |
| CN115685335A (en) * | 2022-11-14 | 2023-02-03 | 中海石油(中国)有限公司深圳分公司 | Longitudinal and transverse wave velocity prediction method and device, computer equipment and storage medium |
| CN115793048A (en) * | 2022-11-16 | 2023-03-14 | 河海大学 | Shale oil bed system compact reservoir transverse wave velocity prediction method |
| CN115932965A (en) * | 2022-12-19 | 2023-04-07 | 西北大学 | Compact sandstone shear wave velocity prediction method based on Xu-White model |
-
2023
- 2023-08-29 CN CN202311101649.8A patent/CN117388919B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5265016A (en) * | 1989-07-19 | 1993-11-23 | Amoco Corporation | Method of shear wave velocity estimation |
| WO2018010628A1 (en) * | 2016-07-15 | 2018-01-18 | 河海大学 | Seismic rock physics inversion method based on a large area tight reservoir |
| CN106680878A (en) * | 2016-12-29 | 2017-05-17 | 中国石油大学(华东) | Transverse wave speed estimation method based on improved Biot coefficient analysis |
| CA3067013A1 (en) * | 2017-07-21 | 2019-01-24 | Landmark Graphics Corporation | Deep learning based reservoir modeling |
| WO2020146797A1 (en) * | 2019-01-10 | 2020-07-16 | Baker Hughes Oilfiled Operations Llc | Estimation of fracture properties based on borehole fluid data, acoustic shear wave imaging and well bore imaging |
| WO2021147529A1 (en) * | 2020-01-21 | 2021-07-29 | 长江大学 | Multipoint geostatistical pre-stack inversion method based on updated theory of permanence of probability ratio |
| CN115685335A (en) * | 2022-11-14 | 2023-02-03 | 中海石油(中国)有限公司深圳分公司 | Longitudinal and transverse wave velocity prediction method and device, computer equipment and storage medium |
| CN115793048A (en) * | 2022-11-16 | 2023-03-14 | 河海大学 | Shale oil bed system compact reservoir transverse wave velocity prediction method |
| CN115932965A (en) * | 2022-12-19 | 2023-04-07 | 西北大学 | Compact sandstone shear wave velocity prediction method based on Xu-White model |
Non-Patent Citations (3)
| Title |
|---|
| JUN WANG ET AL.: "Shear wave velocity prediction based on adaptive particle swarm optimization optimized recurrent neural network", 《JOURNAL OF PETROLEUM SCIENCE & ENGINEERING》, 2 June 2020 (2020-06-02) * |
| 刘雅杰;李生杰;王永刚;夏勇辉;: "横波预测技术在苏里格气田储层预测中的应用", 石油地球物理勘探, no. 01, 15 February 2016 (2016-02-15) * |
| 王树华 等: "基于深度前馈神经网络方法的横波速度预测", 《油气地质与采收率》, vol. 29, no. 1, 31 January 2022 (2022-01-31) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117388919B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109447532B (en) | Oil reservoir inter-well connectivity determination method based on data driving | |
| CN109471168B (en) | Method for predicting velocity and attenuation of longitudinal wave in pore fracture medium | |
| CN103235338B (en) | Rock fracture parameter inversion method | |
| CN115793048B (en) | Shale oil reservoir system tight reservoir shear wave speed prediction method | |
| CN102041995A (en) | System for monitoring complicated oil deposit flooding conditions | |
| RU2475646C1 (en) | Method of construction of geologic and hydrodynamic model of oil and gas deposits | |
| CN106150477A (en) | A kind of method determining single well controlled reserves | |
| CN103412336A (en) | Method for predicting velocity of longitudinal wave of rock system in heterogeneous reservoir | |
| CN104181585A (en) | Shear wave estimation method and system in geophysical exploration | |
| CN117388919B (en) | DNN-based method for predicting transverse wave speed of tight oil reservoir | |
| Islam et al. | Experimental modeling of gravity underflow in a sinuous submerged channel | |
| CN115685335A (en) | Longitudinal and transverse wave velocity prediction method and device, computer equipment and storage medium | |
| CN111101935A (en) | Oil shale prediction method under few-well condition | |
| CN109490988A (en) | Establish the method for being suitable for the rock physics new model of hard rock | |
| CN115932965A (en) | Compact sandstone shear wave velocity prediction method based on Xu-White model | |
| CN114675328A (en) | Organic-matter-rich shale shear wave prediction method based on changed pore aspect ratio Xu-white model | |
| CN110716235B (en) | Sand shale well logging pore structure inversion method | |
| Huang et al. | Quantitative analysis of the main controlling factors of oil saturation variation | |
| CN116520394A (en) | Method for directly predicting porosity based on seismic logging double-drive fusion multi-scale information | |
| CN110851982B (en) | Method for analyzing feasibility of oxygen reduction air flooding of medium-low permeability oil reservoir | |
| CN111622750A (en) | Calculation method for evaluating validity of glutenite reservoir through conventional logging information | |
| CN114428372B (en) | Self-adaptive rock physical modeling method | |
| CN111708985A (en) | Unconsolidated sandstone stratum logging shear wave velocity prediction method | |
| CN116819607B (en) | Transverse wave prediction method and system based on tunnel curvature | |
| CN113866824B (en) | Method for determining elastic parameters of formation between salts, storage medium and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |