A quantitative analysis method for droplet velocity in porous media emulsion flow experiments

Abstract

The present invention relates to a method for quantitatively analyzing droplet velocity in a porous medium emulsion flow experiment, which is mainly used to analyze the flow condition of the emulsion in the porous medium from the experimental video collected by a laser scanning confocal microscope. According to the imaging principle, the blurred area generated around the high-speed moving object in the video frame reflects its movement. Therefore, the present invention first determines the segmentation threshold of the static droplet and the moving droplet area by image segmentation and determines the blur range of the moving droplet area based on the two thresholds; then identifies the boundary of the moving droplet image and determines the main axis, thereby obtaining the direction of droplet movement; finally, extracts the motion blur scale according to the direction of movement, and calculates the droplet velocity in combination with the experimental parameters. The present invention can be used to analyze and process high-speed flowing emulsion images, thereby quantitatively describing the state of emulsion flow under different conditions, providing key parameters for subsequent emulsion flow law research, and has important scientific research value.

Description

Technical Field

[0001] The present invention relates to a method for quantitatively analyzing droplet velocity in a porous medium emulsion flow experiment, which is mainly used to quantitatively analyze the emulsion flow experiment video collected in real time by a laser scanning confocal microscope in the experiment, determine the movement direction and movement speed of the emulsion droplets, and thus lay a foundation for studying the flow law of the emulsion in the porous medium. The present invention starts with an image processing method and uses the fuzzy area generated by the droplet movement in the image to quantitatively describe its velocity. It is a fast, effective, accurate, convenient, and low-cost experimental data analysis method. Background Art

[0002] The development of unconventional reservoirs such as tight oil and heavy oil has always been the focus of research at home and abroad. The use of emulsion flooding for such reservoirs has been proven to be one of the most economical development options that can effectively improve the sweep coefficient. Baldygin et al. (2014) experimentally proved that after conventional water flooding, the use of the minimum emulsion injection volume, water and emulsion alternating flooding can increase the recovery rate by 20%. Farias et al. (2016) also observed the microscopic and macroscopic improvement of the sweep coefficient of heavy oil after emulsion flooding. When the emulsion flows in the pores, the emulsion will become unstable, and a large number of droplets will gather together to block the pore channels. By blocking the high permeability area or dominant channel in the reservoir, the injected water can be transferred to the undisplaced area, thereby improving the productivity of the reservoir. In addition, most heavy oil reservoirs are produced and transported in the form of emulsions. It is easier to transport heavy oil in the form of emulsions, and it is also a more economical and effective solution. In order to effectively recover the crude oil in them, these emulsions need to be decomposed or separated. Therefore, understanding the emulsion formation and destabilization mechanism is of great significance for the development of heavy oil.

[0003] Emulsion flow experiments in porous media are an important method for studying the formation and instability mechanisms of emulsions in media. The laser scanning confocal microscope is an important instrument for completing such experiments. The instrument can observe and record the flow of emulsions in porous channels in real time, while collecting experimental video data, so as to gain a deep understanding of the effects of different external conditions, such as pressure and channel width, on the flow state of the emulsion. The key parameter used to describe the flow of emulsions is the flow velocity of the emulsion droplets in the channel, so accurately extracting the droplet velocity from the collected video is the key to studying the flow mechanism of emulsions in porous media.

[0004] At present, there are two technical solutions for the speed measurement of moving objects from the perspective of hardware and software. In terms of hardware, radio positioning (RADAR) technology and laser radar (LIDAR) technology can be used. This technical solution is usually expensive and cannot be applied to microscopic experimental research. In terms of software, the collected video frames can be processed and analyzed through image processing technology to obtain the motion information of the object in the image and thus obtain the speed parameters. This technical solution is usually low-cost and more suitable for processing experimental data. At present, the commonly used processing methods for analyzing the speed of objects in images include distance map method (DistanceMap), optical flow method (Optical Flow) and motion blur method (Motion Blur). The distance map method and optical flow method have good effects when processing slow-moving objects. However, in the porous medium emulsion flow experiment, due to the complex pore structure in the core and the large number of experimental control variables, the droplet flow rate is usually fast and varies greatly. Therefore, the distance map method and optical flow method are less applicable. In comparison, the motion blur method has greater application potential. The key to this method is to accurately determine the blur scale of the moving object. For high-speed flowing objects, the blur scale is proportional to its own speed. Lin et al. (2005) and Xu et al. (2010) both proposed a method to calculate the motion speed by using the blurred area around the object in a single frame based on the imaging principle combined with parameters such as exposure time.

[0005] For the video of the emulsion flow in the porous medium collected by the laser scanning confocal microscope, the droplet morphology and flow conditions are complex and changeable, which makes it difficult to calculate the blur scale of the droplets. This is also the biggest difficulty in using the motion blur method to calculate the velocity of the emulsion droplets in the porous medium flow experiment. The present invention introduces image processing technology into the motion blur method, uses the static droplet scale method to calculate the blur scale of the moving droplets, and then obtains the emulsion droplet flow velocity, achieving a quantitative description of the flow of droplets in the porous medium emulsion flow experiment. Summary of the invention

[0006] The purpose of the present invention is to provide a quantitative analysis method for experimental data of emulsion flow in porous media, which is mainly used to analyze the emulsion flow video data collected by a laser scanning confocal microscope. The velocity of the droplet is characterized by determining the blurred area in the image caused by the movement of the droplet. Combined with the experimental instrument parameters, the emulsion flow at different positions in the porous medium channel can be quantitatively described, providing key parameters for the study of the emulsion flow law in the porous medium.

[0007] In order to achieve the above technical objectives, a method for quantitative analysis of droplet velocity in a porous medium emulsion flow experiment is provided, characterized in that the analysis method comprises the following steps:

[0008] Step 1: Extract key images. Extract the key frames to be studied from the video data collected by the laser scanning confocal microscope, and capture different image segments based on the distribution of typical moving droplets in the image. These segments should contain complete droplets to be processed.

[0009] Step 2: Calibration image selection. Select video frames containing stationary droplets from the experimental video data. These image segments containing stationary droplets will be used as calibration images for subsequent processing. Due to the instability of the emulsion, droplets of a specific flow rate will aggregate in a specific area and get stuck in the pore channel, which is very common in flow experiments.

[0010] Step 3: Image filtering. Filter and reduce noise on the selected image segments to eliminate noise in the image, improve image quality, and facilitate subsequent processing.

[0011] Step 4: Image segmentation. Perform image segmentation on the extracted image containing the moving droplets to be processed and the calibration image containing the stationary droplets, calculate the optimal segmentation threshold of the two images, and extract the droplets in the two images from the background. It should be pointed out that the present invention does not limit the image segmentation method used to calculate the optimal threshold. Any reasonable image segmentation method can be used to extract droplets in the image. Here, only the scheme of image segmentation using the Otsu method is described.

[0012] The Otsu method uses the inter-class variance between the foreground and background obtained by image segmentation as the standard for determining the segmentation threshold, that is, the threshold at which the inter-class variance reaches the maximum value is the optimal segmentation threshold. The calculation method of the inter-class variance is:

[0013]

[0014] In the formula, σ B —Inter-class variance between foreground and background;

[0015] ω0—variance of grayscale values ​​of foreground pixels in the image;

[0016] ω1—variance of grayscale values ​​of background pixels in the image;

[0017] μ0—the mean gray value of the foreground pixels in the image;

[0018] μ1—The mean gray value of the image background pixels.

[0019] Step 5: Determine the blurred area of ​​the moving droplet. Determine the blurred area caused by the motion of the droplet based on the segmentation threshold T2 of the static droplet image and the segmentation threshold T1 of the moving droplet image. Since the grayscale value of the pixels in the motion blurred area will be smaller than the grayscale value of the pixels in the actual droplet area and larger than the grayscale value of the pixels in the background area, the grayscale value range of the pixels in the blurred area should be between the two thresholds, which can be expressed as:

[0020] I m ∈[T1,T2] (2)

[0021] In the formula, I m —Grayscale value of pixels in the blurred area;

[0022] T1—image segmentation threshold containing moving droplets;

[0023] T2—The image segmentation threshold containing stationary droplets.

[0024] Step 6: Identify the boundary of the moving droplet area. If all the adjacent pixels of a certain pixel belonging to the droplet area in the image contain both background pixels (core pores) and foreground (droplet) pixels, then the pixel can be judged as a droplet boundary pixel. The specific judgment criteria are as follows:

[0025]

[0026] Where P Edge —Boundary pixels of the droplet image;

[0027] P N —Pixels adjacent to the pixels in the droplet area;

[0028] P front —Image foreground pixels;

[0029] P back —Image background pixels.

[0030] Step 7: Calculate the equation of the main axis of the moving droplet. Select the extracted boundary pixels and calculate the distance between them. According to the motion blur theory, the longest distance between boundary pixels is the main axis of the moving droplet, which also represents the moving direction of the droplet, which can be expressed as equation (4).

[0031]

[0032] Where, D is the length of the main axis of the moving droplet;

[0033] Distance(x,y)—a function that calculates the straight-line distance between pixel x and pixel y.

[0034] —The end point of the main axis of the moving droplet;

[0035] n—Number of pixels on the droplet boundary.

[0036] The analytical expression of the linear equation of the main axis can be solved based on the extracted endpoints of the main axis of the moving droplet.

[0037] Step 8: Determine the blur scale of the moving droplet. According to the direction of the droplet movement and the extracted motion blur range, the blur scale of the moving droplet is calculated by combining the main axis equation. The main axis length within the blur range on the opposite side of the movement direction is the blur scale. The blur scale is proportional to the speed of the droplet movement.

[0038] Step 9: Calculate the velocity of the droplet. Combining imaging theory with actual equipment parameters, the velocity of the moving droplet can be calculated based on the fuzzy scale. In the porous medium emulsion flow experiment, the camera of the laser scanning confocal microscope is parallel to the pore channel and adjusted to an appropriate depth. The camera exposure time can be obtained based on the frame rate of the video, and finally the velocity of the droplet can be calculated according to formula (5).

[0039]

[0040] Where, d is the distance the object moves during the exposure time;

[0041] T—exposure time;

[0042] z—the distance between the object and the lens;

[0043] K—object blur scale;

[0044] s x —CCD pixel size in the horizontal direction;

[0045] f—focal length of the camera;

[0046] β—The angle between the blur scale and the direction of motion.

[0047] The present invention proposes a method for quantitative analysis of droplet velocity in porous medium emulsion flow experiments, and the method flow chart is as follows: Figure 1 As shown. The droplet movement speed is quantitatively calculated using the blurred area caused by the movement of the droplet in the video frame. Since there is no blurred area around the stationary droplet, the image segmentation principle is used to determine the segmentation threshold of the stationary droplet image and the moving droplet image, and thus determine the blurred range around the moving droplet. The droplet movement speed is then quantitatively calculated based on the imaging principle combined with the experimental parameters, and the direction of the droplet movement is determined by extracting the main axis of the droplet, which provides essential parameters for the subsequent study of the emulsion flow mechanism.

[0048] Beneficial Effects

[0049] Compared with the prior art, the present invention has the following significant advantages: (1) Advanced method and high precision. The threshold value obtained by static droplet image segmentation is used to calibrate the moving droplet image, and the blur range around the moving droplet is accurately obtained by combining the droplet image segmentation threshold in a moving state. (2) Low cost and strong operability. The use of image segmentation technology to process the video frames collected in the emulsion flow experiment does not require supporting equipment such as high-speed cameras, which greatly saves experimental costs; in addition, it is very normal to observe static droplets in the emulsion flow experiment, which also makes the invention highly operational. (3) Quantitative accuracy. The present invention uses the imaging principle to quantitatively calculate the droplet velocity, and at the same time, the main axis is extracted from the droplet area in the image to accurately determine the droplet flow direction. The parameters obtained according to the present invention can accurately and quantitatively describe the flow state of the droplet, providing the necessary parameters for subsequent analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Flow chart of a method for quantitative analysis of droplet velocity in a porous medium emulsion flow experiment proposed by the present invention

[0051] Figure 2 Key frames of the emulsion flow experiment video collected by laser scanning confocal microscope equipment.

[0052] Figure 3 The image segments captured from the key image contain typical moving droplets and the image segments contain stationary droplets.

[0053] Figure 4 The image segmentation result of the image fragment containing the stationary droplet.

[0054] Figure 5 The image segmentation result of the image fragment containing moving droplets.

[0055] Figure 6 Extraction process for the spindle of a moving droplet.

[0056] Figure 7 Calculated diagram of droplet velocity distribution for key frames of the emulsion flow experiment video. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0058] Figure 1 The present invention is a flowchart of a method for quantitatively analyzing droplet velocity in a porous medium emulsion flow experiment summarized according to the present invention.

[0059] Figure 2 The video key frames are collected by laser scanning confocal microscope in the emulsion flow experiment. The image contains a timestamp to record the current time and scale information for subsequent experimental data analysis. In the grayscale image, the emulsion droplets are grayish white, and the grayscale value is lower than the background. It can be seen that although the droplet outline is relatively clear in the experiment, due to the fast flow speed, motion blur occurs around the droplet, and the specific position is difficult to determine. It is impossible to calculate the speed through the droplet position relationship between adjacent frames.

[0060] Figure 3 This is an image clip containing typical moving droplets and stationary droplets captured from an image. In this image, the emulsion droplets mainly flow in the upper part, while the emulsion is unstable on the left side of the image. A large number of droplets gather and cause the flow channel to be blocked, and the emulsion droplets are stationary there. Droplet instability is a phenomenon that is focused on in emulsion flow experiments. The stationary droplets extracted from the unstable area can be used as a calibration image to calibrate the fuzzy scale of the moving droplets. In addition, since the observation conditions remain unchanged in the experiment, for images without unstable areas, unstable areas in other frames can also be used for processing.

[0061] Figure 4 and Figure 5 The image segmentation results of the image segment containing stationary droplets and the image segment containing moving droplets. This application example uses the Otsu method to determine the image segmentation threshold and convert the original grayscale image into a binary image. Due to the difference in the grayscale of the droplets (foreground) in the two image segments, the grayscale value of the blurred area is low, which makes the segmentation thresholds of the two images different. The blur scale can be calibrated according to the two thresholds.

[0062] Figure 6 This is the process of extracting the main axis of a moving droplet and calculating the fuzzy scale. The key parameters for describing the motion of an emulsion droplet are the flow velocity and flow direction of the droplet, and the flow direction of the droplet can be determined based on the direction of the main axis of the droplet. For droplets moving at high speeds, the droplets will appear elongated along the direction of motion in the image. This is due to motion blur, so the longest axis within the droplet area in the image represents the direction of motion, which is the main axis. In the figure, image segmentation, boundary recognition, and main axis extraction operations were performed on the image fragment containing a typical droplet. Finally, the size of the fuzzy scale is quantitatively calculated in combination with the scaled fuzzy area, which is used for subsequent quantitative calculation of the droplet velocity.

[0063] Figure 7 The velocity distribution streamline diagram of the droplets in the image is drawn using the method of the present invention based on typical droplets extracted from the image.

[0064] Application examples:

[0065] The present invention is applied to process the image data of the emulsion flow experiment collected by the laser scanning confocal microscope. In the experiment, a simulated formation was constructed by using a sand-filled model, and an injection experiment was carried out on the simulated formation using a 2% by mass desulfurized cellulose nanocrystal (HCNC) emulsion. In the experiment, the flow velocity of the emulsion in the pores of the simulated formation was changed by continuously adjusting the inlet pressure, and the emulsion flow under different conditions was studied. The droplet velocity in the sand-filled model is proportional to the inlet pressure. The greater the inlet pressure, the faster the droplet flow velocity in the pores.

[0066] The experimental device uses a pump to drive the fluid, and the water in the water tank is introduced into the stainless steel container for conversion, where it forms an emulsion with the nanocrystals. After that, the emulsion flows through the pressure sensor into the simulated formation, which records the inlet pressure of the emulsion. This experiment mainly observes the flow changes of the emulsion in the pressure range of 0 to 1000 psi. A laser scanning confocal microscope is used to observe the formation in real time and record the flow of the emulsion. Combined with the droplet flow video collected by the equipment, events at the pore scale can be visualized and characterized, and through pressure data, interactions at the microscale can be associated with responses at the macroscale.

[0067] In this example, the simulated stratum is a hydrophilic rock. Figure 2 The experimental time marked by the timestamp in the figure. From the data recorded by the pressure sensor, we can see that the inlet pressure at this time is 9psi. The inlet pressure is small and the droplet flow rate is slow. When the droplet flow rate is slow, the emulsion droplets have a longer contact time with the core sand particles, which makes it easier for the droplets to aggregate in the pores. This collective crowding phenomenon in the pore throat makes the emulsion droplets have a higher coalescence efficiency at a lower flow rate, which is also Figure 2 The reason why the droplets on the left side gather and remain motionless; correspondingly, at higher flow rates, even if the droplets are in close contact with the sand particles, the residence time of the emulsion droplets is not enough to affect the stability of the droplets.

[0068] In order to quantitatively describe the cause of droplet aggregation and describe the movement state of the emulsion, it is necessary to determine the flow velocity of the droplets. In this example, the method described in the present invention is used to obtain the original image ( Figure 2 ) to capture image segments containing typical moving droplets and image segments containing stationary droplets, such as Figure 3 As shown. Afterwards, the two image segments are thresholded using image segmentation processing to obtain binary images of the two image segments, as shown Figure 4 , Figure 5As shown in the figure. In this example, the segmentation thresholds of the two image segments are: 67 grayscale values ​​for the moving droplet image segment and 86 grayscale values ​​for the stationary droplet image segment. Obviously, due to motion blur, the overall grayscale value of the moving image segment will be lower than that of the stationary image segment. According to the grayscale values ​​of the two segments, it can be known that the grayscale range of the motion blur scale should be between 67 and 86 grayscale values.

[0069] After the motion blur grayscale range is determined, typical droplets in the binary image of the moving droplet image segment are selected, such as Figure 6 (1), and identify its droplet boundary according to formula (3), as follows: Figure 6 (2) As shown. Based on the boundary of the droplet area, the maximum distance between the boundary pixels is obtained by comparing the distances between the two boundary pixels, which is the main axis, as shown in Figure 6 (3) is shown. The main axis also represents the direction of motion of the droplet. In this example, the angle between the direction of motion of the droplet and the X-axis is 41.2°. After determining the direction of motion of the droplet, the blur scale of the droplet can be quantitatively calculated based on the range of the blur region calculated in the previous step, that is, the length of the main axis in the blur region on the side opposite to the motion direction, as shown in Figure 6 As shown in (4), the blur scale of the typical moving droplet selected in this example is 5.9 pixels.

[0070] According to the calculated blur scale and the laser scanning confocal microscope instrument probe parameters, the flow velocity of the droplet can be calculated according to formula (5). In this example, the size of the collected image is 232.5μm×232.5μm, and the pixel size is 512px×512px. The distance between the instrument camera and the pore droplet is 10mm, the focal length of the camera is 1.95mm, and the frame rate of the collected experimental video is 14.71 / s. According to the above parameters, the flow velocity of the droplet can be calculated as:

[0071] v=(10×5.9×0.2325 / 512) / (1 / 14.71×1.95)=0.20(mm / s) (6)

[0072] Using the same processing method, the typical moving droplets in the entire image can be processed and the blur scale and main axis direction can be calculated. According to the position of the selected droplet, the flow trend of the entire image can be drawn, such as Figure 7 As shown, the solid line is the main axis length, the dotted line represents the flow trend of the droplet, and the value in the figure is the droplet movement speed calculated based on the fuzzy scale of the typical moving droplet, in units of mm / s.

Claims

1. A method for quantitative analysis of droplet velocity in a porous medium emulsion flow experiment, characterized in that: The analytical method comprises the following steps: Step 1: Extract the key frames to be studied from the video data collected by the laser scanning confocal microscope, and capture different image segments according to the distribution of typical moving droplets in the image. These segments should contain complete droplets to be processed; Step 2: Select video frames containing stationary droplets from the experimental video data. These image segments containing stationary droplets will be used as calibration images for subsequent processing; Step 3: Filter and reduce noise on the selected video frames to eliminate noise in the image, improve image quality, and facilitate subsequent processing; Step 4: Perform image segmentation on the extracted image containing the moving droplets to be processed and the calibration image containing the stationary droplets, calculate the optimal segmentation threshold of the two images, and extract the droplets in the two images from the background; Step 5: Determine the blurred area caused by the motion of the droplet according to the segmentation threshold T2 of the static droplet image and the segmentation threshold T1 of the moving droplet image. Since the grayscale value of the pixels in the motion blurred area will be smaller than the grayscale value of the pixels in the actual droplet area and larger than the grayscale value of the pixels in the background area, the grayscale value range of the pixels in the blurred area should be between the two thresholds, which can be expressed as: I m ∈[T1,T2](1) In the formula, I m is the gray value of the pixel in the blurred area, T1 is the image segmentation threshold for images containing moving droplets, and T2 is the image segmentation threshold for images containing stationary droplets; Step 6: Identify the boundary of the moving droplet area. If all the adjacent pixels of a certain pixel belonging to the droplet area in the image contain both background pixels and foreground pixels, then the pixel can be judged as a droplet boundary pixel. The specific judgment criteria are as follows: Where P Edge is the boundary pixel of the droplet image, P N is the pixel point adjacent to the pixel point in the droplet area, P front is the image foreground pixel, P back is the image background pixel; Step 7: Select the extracted boundary pixels and calculate the distance between them. According to the motion blur theory, the longest distance between pixels is the main axis of the moving droplet, which also represents the moving direction of the droplet, which can be expressed as: Where D is the length of the main axis of the moving droplet, Distance(x,y) is the function for calculating the straight-line distance between pixel point x and pixel point y. is the endpoint of the main axis of the moving droplet, and n is the number of pixel points on the droplet boundary; the linear equation of the main axis can be obtained based on the extracted endpoint of the main axis of the moving droplet; Step 8: According to the direction of the droplet movement and the extracted motion blur range, the blur scale of the moving droplet is calculated in combination with the principal axis equation. The length of the principal axis within the blur range on the opposite side of the movement direction is the blur scale, which is proportional to the movement speed of the droplet. Step 9: Combining imaging theory with actual equipment parameters, the velocity of the moving droplets can be calculated based on the fuzzy scale. In the porous medium emulsion flow experiment, the camera of the laser scanning confocal microscope will be parallel to the pore channel and adjusted to a suitable depth. The camera exposure time can be obtained based on the frame rate of the collected video. The final droplet movement velocity can be expressed as: Where d is the distance the object moves during the exposure time, T is the exposure time, z is the distance between the object and the lens, K is the blur scale of the object, and s x is the CCD pixel size in the horizontal direction, f is the focal length of the camera, and β is the angle between the blur scale and the motion direction.

Images