A method for predicting the mechanical parameters of oil sandstone using digital core samples

Abstract

A method for predicting the rock mechanical parameters of oil sands using digital core analysis comprises the following steps: Step 1: Obtaining grayscale images of oil sand cores using an environmental scanning electron microscope. Step 2: Denoising the grayscale images of the oil sand cores. Step 3: Binarizing the denoised grayscale images to obtain a binarized image of the oil sands. Step 4: Vectorizing the binarized image of the oil sands to obtain a vector image. Step 5: Calculating the porosity of the oil sands using Matlab software. Step 6: Conducting a digital core deformation simulation experiment using COMSOL. This invention can quickly and accurately predict the rock mechanical parameters of oil sands, while saving significant manpower, resources, and time.

Description

Technical Field

[0001] This invention belongs to the field of rock mechanics, specifically relating to a method for predicting the mechanical parameters of oil sands rocks using digital core samples. Background Technology

[0002] Oil sands refer to sedimentary sands rich in natural bitumen, essentially a mixture of bitumen, water, sand, and mineral-rich clay. Bitumen content is 10-12%, water 3-5%, and minerals such as sand and clay make up 80-85%. The mineral composition of oil sands varies from region to region. For example, in the oil sands deposits of western Canada, natural bitumen may dominate in some lithologies such as siltstone and carbonates.

[0003] However, current research on the influence of oil sand porosity on the mechanical parameters of oil sand rocks is scarce. To date, there is no well-established theoretical relationship between the mechanical parameters and porosity of oil sand rocks. Furthermore, most studies focus on methods for measuring the mechanical parameters of oil sand rocks. Establishing a link between the mechanical parameters of oil sand rocks measured by digital core analysis and their porosity remains a challenge. Therefore, there is an urgent need to develop a method for predicting the mechanical parameters of oil sand rocks using digital core analysis. Summary of the Invention

[0004] The purpose of this invention is to provide a method for predicting the mechanical parameters of oil sands rocks using digital core samples.

[0005] The technical solution adopted in this invention is:

[0006] A method for predicting the mechanical parameters of oil sandstone using digital core samples, characterized by comprising the following steps:

[0007] Step 1: Obtain the core grayscale image of the oil sand using an environmental scanning electron microscope.

[0008] Step 2: Denoise the core grayscale image of the oil sands.

[0009] Step 3: Binarize the denoised core grayscale image to obtain the binarized image of the oil sand.

[0010] Step 4: Vectorize the binarized image of the oil sand to obtain a vector image of the oil sand.

[0011] Step 5: Calculate the porosity of the oil sand using Matlab software.

[0012] Step Six: Conduct Digital Core Deformation Simulation Experiments using COMSOL

[0013] Step one is mainly accomplished using a field emission scanning electron microscope (FEIQuanta 200F).

[0014] In step one, the sample is cut from a 25mm×50mm standard core using an ultrathin blade. It is a small block with a length and width of 5mm×5mm and a thickness of 2-3mm. The lower end is ground flat, and the upper end is a fresh cut surface that is pried open by tweezers, facing the direction of the electron beam. The sample is then fixed on an aluminum sample holder and placed in a low-vacuum environment wellbore.

[0015] In step one, the samples are observed in a Quanta 200F field emission scanning electron microscope at magnifications of 60x, 75x, 90x, and 120x, thereby obtaining the corresponding microstructure electron images.

[0016] In step two, a median filtering algorithm is selected to denoise the digital core grayscale image.

[0017] In step two, the basic principle of median filtering is as follows: select the pixel values ​​(A1, A2, ..., A1) of the pixel in the digital image and its surrounding pixels. n The process involves taking a window length of m (where m is an odd number), sorting the pixel values, and then selecting the middle pixel value to replace the current pixel value as the filter output. This process is repeated for all pixels in the digital image, making the surrounding pixel values ​​closer to the true values, thereby removing noise. The expression is as follows:

[0018]

[0019] In step two, the median filtering is performed using the built-in function medfilt2 in Matlab software to denoise the image.

[0020] In step three, the Otsu method is used to binarize the denoised rock grayscale image.

[0021] In step three, the Otsu method (Otsu's method) is used to select an optimal gray level (segmentation threshold). The grayscale image with 256 gray levels is divided into two parts based on the selected threshold: a background part with gray values ​​less than the threshold and a foreground part with gray values ​​greater than the threshold. Background pixels are set to 0 (black), and foreground pixels are set to 255 (white), thus achieving image binarization.

[0022] In step three, assuming the image threshold is R, the proportion of foreground elements to the entire image is W0 with an average gray value of μ0, and the proportion of background elements to the entire image is W1 with an average gray value of μ1, then...

[0023] W0 + W1 = 1 (1)

[0024] The average gray value μ of the image is:

[0025] μ=W0μ0+W1μ1 (2)

[0026] The within-class variance g is defined as:

[0027] g = W0(μ0 - μ) 2 +W1(μ1-μ) 2 (3)

[0028] Definition of between-class variance:

[0029] g = W0 W1(μ0-μ1) 2 =[μ . ω(R)-μ(R)] 2 / ω(R)[1-ω(R)] (4)

[0030] Change R between 0 and 255. When g in equation (4) reaches its maximum value, R is the threshold value to be obtained.

[0031] In step three, the image is binarized using the imbinarize function in Matlab.

[0032] In step four, the binarized image is vectorized using CorelDRAW software. This involves importing the binarized image, tracing its outline using spline curves, and generating a vector graphic.

[0033] In step five, the area of ​​black pixels (i.e., pore area) in the binarized image is calculated using Matlab. The ratio of the calculated value to the total number of pixels in the image is the porosity of the oil sand.

[0034] In step six, the microscopic deformation simulation of the oil sands is first performed using a solid mechanics physics field interface with a steady-state solver. Then, the oil sand vector diagram generated by CorelDRAW is imported into COMSOL.

[0035] In step six, it is assumed that the oil sands consist only of a framework phase and a porous phase, with the framework phase mainly composed of quartz and the porous phase filled with bitumen. The framework and porous phase regions of the oil sands are selected and assigned values ​​using appropriate materials.

[0036] In step six, the digital core deformation simulation experiment uses a plane strain model to apply fixed constraints to the bottom boundary of the digital core, applies a strain of about 10% downward in the Y direction to the top of the digital core, and constrains the top to not displace in the X direction.

[0037] In step six, the stress field, strain field, and displacement field distribution diagrams of the oil sand under different strains are obtained through simulation calculations.

[0038] In step six, the stress-strain curve of the oil sand can be obtained by calculating the reaction force at the top of the digital core, and then the elastic modulus of the oil sand can be determined. The Poisson's ratio of the oil sand can be obtained from the strain values ​​at both ends of the digital core.

[0039] In step six, the influence of porosity on the elastic modulus and Poisson's ratio of oil sand is analyzed by using the curves showing the variation of oil sand elastic modulus and Poisson's ratio with porosity.

[0040] The beneficial effects of this invention are: Digital core experiments for predicting the mechanical parameters of oil sands, conducted according to the technical solution of this invention, can quickly and accurately predict the rock mechanical parameters of oil sands, while saving significant manpower, material resources, and time. For a long time, rock mechanical parameters have been determined through triaxial compression mechanics experiments using rock cores. Although this method is highly accurate, it is complex and costly, a problem particularly prominent in petroleum engineering. This method uses digital core testing to evaluate the mechanical properties of oil sands, thereby establishing the relationship between oil sand porosity and rock mechanical parameters. The variation patterns of rock mechanical parameters in oil sands can be quickly predicted simply by conducting digital core experiments. Attached Figure Description

[0041] Figure 1 A flowchart of a method for predicting the mechanical parameters of oil sandstone using digital core samples;

[0042] Figure 2 This is the image after binarization of the oil sands;

[0043] Figure 3 This is a vectorized image of the oil sand.

[0044] Figure 4 This is a stress distribution diagram at the end of the digital core loading stage;

[0045] Figure 5 This is a strain distribution diagram at the end of the digital core loading stage;

[0046] Figure 6 This is a displacement field distribution diagram at the end of the digital core loading stage;

[0047] Figure 7 This is a stress-strain curve diagram for oil sand.

[0048] Figure 8 This is a graph showing the relationship between the elastic modulus and porosity of oil sand.

[0049] Figure 9 This is a graph showing the relationship between Poisson's ratio and porosity in oil sand.

[0050] Figure 10 For code diagrams;

[0051] Figure 11 For code diagrams;

[0052] Figure 12 This is a code diagram. Detailed Implementation

[0053] The present invention will be further described below with reference to the embodiments and accompanying drawings.

[0054] Example 1:

[0055] like Figure 1 As shown, a method for predicting the mechanical parameters of oil sandstone using digital core samples includes the following steps:

[0056] Step 1: Obtain the core grayscale image of the oil sand using an environmental scanning electron microscope.

[0057] Step 2: Denoise the core grayscale image of the oil sands.

[0058] Step 3: Binarize the denoised core grayscale image to obtain the binarized image of the oil sand.

[0059] Step 4: Vectorize the binarized image of the oil sand to obtain a vector image of the oil sand.

[0060] Step 5: Calculate the porosity of the oil sand using Matlab software.

[0061] Step Six: Conduct Digital Core Deformation Simulation Experiments using COMSOL

[0062] Step one is mainly accomplished using a field emission scanning electron microscope (FEIQuanta 200F).

[0063] In step one, the sample is cut from a 25mm×50mm standard core using an ultrathin blade. It is a small block with a length and width of 5mm×5mm and a thickness of 2-3mm. The lower end is ground flat, and the upper end is a fresh cut surface that is pried open by tweezers, facing the direction of the electron beam. The sample is then fixed on an aluminum sample holder and placed in a low-vacuum environment wellbore.

[0064] In step one, the samples are observed in a Quanta 200F field emission scanning electron microscope at magnifications of 60x, 75x, 90x, and 120x, thereby obtaining the corresponding microstructure electron images.

[0065] In step two, the basic principle of median filtering is as follows: select the pixel values ​​(A1, A2, ..., A1) of the pixel in the digital image and its surrounding pixels. n The process involves taking a window length of m (where m is an odd number), sorting the pixel values, and then selecting the middle pixel value to replace the current pixel value as the filter output. This process is repeated for all pixels in the digital image, making the surrounding pixel values ​​closer to the true values, thereby removing noise. The expression is as follows:

[0066]

[0067] In step two, the median filtering function medfilt2, a built-in function of Matlab software, is used to denoise the image. The code is as follows: Figure 10 As shown:

[0068] In step three, the Otsu method is used to binarize the denoised rock grayscale image.

[0069] In step three, the Otsu method (Otsu's method) is used to select an optimal gray level (segmentation threshold). The grayscale image with 256 gray levels is divided into two parts based on the selected threshold: a background part with gray values ​​less than the threshold and a foreground part with gray values ​​greater than the threshold. Background pixels are set to 0 (black), and foreground pixels are set to 255 (white), thus achieving image binarization.

[0070] In step three, assuming the image threshold is R, the proportion of foreground elements to the entire image is W0 with an average gray value of μ0, and the proportion of background elements to the entire image is W1 with an average gray value of μ1, then...

[0071] W0 + W1 = 1 (1)

[0072] The average gray value μ of the image is:

[0073] μ=W0μ0+W1μ1 (2)

[0074] The within-class variance g is defined as:

[0075] g = W0(μ0 - μ) 2 +W1(μ1-μ) 2 (3)

[0076] Definition of between-class variance:

[0077] g = W0 W1(μ0-μ1) 2

[0078] g = W0 W1(μ0-μ1) 2 =[μ . ω(R)-μ(R)] 2 / ω(R)[1-ω(R)] (4)

[0079] Change R between 0 and 255. When g in equation (4) reaches its maximum value, R is the threshold value to be obtained.

[0080] In step three, the image is binarized using the `imbinarize` function in Matlab, as shown in the code below. Figure 11 As shown. The generated binary image is as follows. Figure 2 As shown.

[0081] In step four, the binarized image is vectorized using CorelDRAW software. This involves importing the binarized image, tracing its contours using spline curves, and generating a vector image. The vector image is shown below. Figure 3 As shown.

[0082] In step five, the area of ​​black pixels (i.e., pore area) in the binarized image is calculated using Matlab. The ratio of this calculated value to the total number of pixels in the image is the porosity of the oil sand. The Matlab code for calculating porosity is as follows: Figure 12 As shown:

[0083] In step six, the microscopic deformation simulation of the oil sands is first performed using a solid mechanics physics field interface with a steady-state solver. Then, the oil sand vector diagram generated by CorelDRAW is imported into COMSOL.

[0084] In step six, it is assumed that the oil sands consist only of a framework phase and a porous phase, with the framework phase mainly composed of quartz and the porous phase filled with bitumen. The framework and porous phase regions of the oil sands are selected and assigned values ​​using appropriate materials.

[0085] In step six, the digital core deformation simulation experiment uses a plane strain model to apply fixed constraints to the bottom boundary of the digital core, applies a strain of about 10% downward in the Y direction to the top of the digital core, and constrains the top to not displace in the X direction.

[0086] In step six, the stress field of the oil sand under different strains is obtained through simulation calculation. Figure 4 ), strain field ( Figure 5 ) and displacement field distribution map ( Figure 6 ).

[0087] In step six, the stress-strain curve of the oil sand can be obtained by calculating the reaction force at the top of the digital core. Figure 7 The elastic modulus of the oil sand is then determined. The Poisson's ratio of the oil sand can be obtained from the strain values ​​at both ends of the digital core.

[0088] In step six, the elastic modulus of the oil sand ( Figure 8 ) and Poisson's ratio ( Figure 9 The curves showing the change in porosity were used to analyze the influence of porosity on the elastic modulus and Poisson's ratio of oil sand.

[0089] If there is a sample from the same block, by following the above operating method, testing the porosity of the sample, and then following the above operating steps, the rock mechanical parameters of the oil sands can theoretically be evaluated.

Claims

1. A method for predicting the mechanical parameters of oil sandstone using digital core samples, characterized in that, Includes the following steps: Step 1: Obtain core grayscale images of oil sands using environmental scanning electron microscopy: The sample was cut from a 25mm×50mm standard core using an ultrathin blade, resulting in a small block measuring 5mm×5mm in length and width and 2-3mm in thickness. The lower end was ground flat, and the upper end, facing the direction of the electron beam, was a freshly cut surface pried open with tweezers. The sample was then fixed on an aluminum sample holder and placed in a low-vacuum environment wellbore. The sample was observed in a Quanta 200F field emission scanning electron microscope at magnifications of 60x, 75x, 90x, and 120x, thus obtaining the corresponding electron microstructure images. Step 2: Select the median filtering algorithm to denoise the digital core grayscale image: Select the pixel values ​​A1, A2, ... A1 of the pixel in the digital image and its surrounding pixels. n Let the window length be m, where m is an odd number. Then, sort the pixel values ​​and select the pixel value at the middle position to replace the current pixel value as the filter output. This process is repeated for all pixels in the digital image to make the surrounding pixel values ​​closer to the true values, thereby removing noise. The expression is as follows: Step 3: Use the Otsu method to binarize the denoised rock grayscale image to obtain the binarized image of the oil sands; Using Otsu's method, also known as the maximum inter-class variance method, an optimal gray level is selected to divide a grayscale image with 256 gray levels into two parts according to the selected threshold: the background part with gray values ​​less than the threshold and the foreground part with gray values ​​greater than the threshold. The background pixels are set to 0 and the foreground pixels are set to 255, thus achieving image binarization. Step 4: Use CorelDRAW software to vectorize the binarized image of the oil sand. Import the binarized image, use spline curves to trace the outline of the image, and generate a vector image of the oil sand. Step 5: Use Matlab to calculate the area of ​​black pixels in the binarized image, i.e., the pore area. The ratio of the calculated value to the total number of pixels in the image is the porosity of the oil sand. Step Six: First, the microscopic deformation simulation of the oil sands was performed using the solid mechanics physics field interface with a steady-state solver. Then, the oil sand vector diagram generated by CorelDRAW was imported into COMSOL, and a digital core deformation simulation experiment was conducted using COMSOL. It was assumed that the oil sands consisted only of a framework phase and a porous phase, with the framework phase mainly composed of quartz and the porous phase filled with bitumen. The framework and porous phase regions of the oil sands were selected and corresponding materials were added and assigned values. Through simulation calculations, the stress field, strain field, and displacement field distribution maps of the oil sands under different strains were obtained. The stress-strain curve of the oil sands was obtained by calculating the reaction force at the top of the digital core, and then the elastic modulus of the oil sands was obtained. The Poisson's ratio of the oil sands was obtained by the strain values ​​at both ends of the digital core. The influence of porosity on the elastic modulus and Poisson's ratio of the oil sands was analyzed by the curves of the change of the elastic modulus and Poisson's ratio of the oil sands with porosity.

2. The method for predicting the mechanical parameters of oil sandstone using digital cores according to claim 1, characterized in that, In step two, the median filtering is performed using the built-in function medfilt2 in Matlab software to denoise the image.

3. The method for predicting the mechanical parameters of oil sandstone using digital cores according to claim 1, characterized in that, In step three, assuming the image threshold is R, the proportion of foreground elements to the entire image is W0 with an average grayscale value of μ0, and the proportion of background elements to the entire image is W1 with an average grayscale value of μ1, then... W0 + W1 = 1 (1) The average gray value μ of the image is: μ=W0μ0+W1μ1 (2) The within-class variance g is defined as: g=W0(μ0-μ) 2 +W1(μ1-μ) 2 (3) Definition of between-class variance: g=W0 W1(μ0-μ1) 2 =[μ . [ω(R)-μ(R)] 2 / ω(R)[1-ω(R)] (4) Change R between 0 and 255. When g in equation (4) reaches its maximum value, R is the threshold value to be obtained.

4. The method for predicting the mechanical parameters of oil sandstone using digital cores according to claim 1, characterized in that, In step three, the image is binarized using the imbinarize function in Matlab.

5. The method for predicting the mechanical parameters of oil sandstone using digital cores according to claim 1, characterized in that, In step six, the digital core deformation simulation experiment uses a plane strain model to apply fixed constraints to the bottom boundary of the digital core, applies a strain of about 10% downward in the Y direction to the top of the digital core, and constrains the top to not displace in the X direction.

Images