Oil reservoir water injection suspended matter particle diameter median measurement method and system based on deep learning
Through deep learning-based methods, high-precision real-time monitoring of the particle diameter of suspended matter during water injection in oil fields is achieved, which solves the problem of low detection accuracy in the existing technology and improves the effectiveness evaluation ability of water injection development.
Patent Information
- Application Number
- CN202411082398.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The median online monitoring of suspended particles during water injection in the oil field is not yet mature, and there is a lack of high-precision real-time detection methods, which affects the effect of water injection development and the evaluation of formation porosity changes.
Using the median measurement method of water-injected suspended particles in reservoirs based on deep learning, the water-injected water flow image is collected through the measurement device, and image analysis and edge detection are performed using the AI edge calculation module and improved detection algorithm to calculate the diameter of the suspended particles, and through Deep learning models realize real-time monitoring and data transmission.
It improves the accuracy and real-time detection of suspended particles, reduces manual interference, enhances the representativeness and reliability of detection results, and enables optical imaging and data acquisition at the micron level.
Smart Images

Figure CN119000486A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of effective monitoring of unconventional pollutants in oil and gas production, and in particular to a method and system for measuring the median diameter of suspended matter particles in oil reservoir injection based on deep learning. Background Art
[0002] With the advancement of science and technology, water injection development has developed into a major technical means to maintain the efficient development of oil fields and an important method to improve the ultimate recovery of oil fields. The water quality index of injected water is one of the important factors affecting the effect of water injection development.
[0003] Generally speaking, the index of oilfield water injection is determined by the reservoir permeability. When suspended particles are injected into the water, they will be carried into the porous medium of the formation, and then move in the pore bends of the porous medium along with the injected water flow. After that, these suspended particles will be retained and precipitated in the pore skeleton, and finally captured, which will cause changes in the pore skeleton structure and reduce the porosity and permeability of the formation rock. This phenomenon is usually called formation damage, and the degree of damage is measured by the damage rate of permeability.
[0004] At present, the research on online monitoring of the median diameter of suspended particles is still in its infancy. The method of detecting the median diameter of suspended particles is still mainly based on manual sampling and laboratory testing. That is, the artificially obtained samples are placed in the laboratory detection instrument. When measuring by laser diffraction, the laser beam is allowed to pass through the dispersed particle sample to measure the intensity of the scattered light. The light scattering angle of large particles is small, while the light scattering angle of small particles is large. Analyze the scattered light intensity data and use the Mie light scattering theory to calculate the particle size of the particles that produce the scattered light. The particle size is expressed as the diameter of a sphere of equal volume. The more common one at present is the MASTERSIZER 3000 intelligent particle size measuring instrument, which can measure particles with a particle size of 0.01μm-3.5mm.
[0005] With the development of computer image recognition technology and related hardware, machine vision recognition systems have also been applied to various fields. The system uses a camera to shoot the target site, and the central processing unit performs regional processing and classification recognition on the captured images to determine the target object. This technology has a high degree of automation and good flexibility, and has stronger judgment capabilities compared to other types of recognition technologies. At present, machine vision recognition systems have begun to be widely used in industrial plants and pollutant identification, monitoring and other aspects. Invention patent CN112052634A discloses an analysis method and device for the influence of particle migration during water injection in low-permeability oil reservoirs, which can understand the specific process of particle movement, deposition and blockage with fluid, and the impact on the displacement process during water injection and exploitation of oil reservoirs. However, this method lacks the classification, effective collection and screening of training samples, which has a certain impact on the accuracy of the prediction results. Summary of the invention
[0006] The present invention focuses on the problems existing in the above-mentioned prior art and provides a method and system for measuring the median diameter of suspended matter particles in oil reservoir water injection based on deep learning.
[0007] In a first aspect of the present invention, a method for measuring the median diameter of suspended particles in oil reservoir water injection based on deep learning is provided, comprising the following steps:
[0008] Step S1: Power on the system of the measuring device, connect the water inlet pipe and the water outlet pipe, introduce the injection water flow into the measuring device through the conduit, and adjust the injection flow rate to a first speed;
[0009] Step S2: stratify the injection water flow passing through the window box in the measuring device, and collect reservoir water injection images corresponding to different liquid level height positions;
[0010] Step S3: judging the clarity of each layer of images collected in layers, discarding images that do not meet the clarity standards, and saving clear images that meet the set conditions in the AI edge computing module of the measuring device;
[0011] Step S4: Analyze the image stored in the AI edge computing module using an improved deep learning-based detection algorithm to identify target suspended particles;
[0012] Step S5: using the HED algorithm based on deep learning to perform edge detection on the detected target suspended particles, fitting the edges into a circle and then calculating the diameter;
[0013] Step S6: Counting the distribution of target suspended matter particles within different diameter ranges;
[0014] Step S7: adjusting the water injection flow rate to a second speed, and repeating steps S2-S6;
[0015] Step S8: Transmit the statistical target suspended matter particle diameter distribution results to the shore side through the network.
[0016] Preferably, the step S2 specifically includes:
[0017] Step S21: dividing the liquid surface into n layers according to the thickness of the liquid surface in the window box, where n≥3;
[0018] Step S22: Move the high-definition industrial camera to the position corresponding to the i-th layer of liquid surface through the focusing module, 0<i≤n;
[0019] Step S23: adjusting the parameters of the high-definition industrial camera and the parameters of the highlight light source to meet the imaging requirements;
[0020] Step S24: continuously collect m images, where m≥1000.
[0021] Preferably, the step S3 specifically includes:
[0022] Step S31: Traverse the n layers of n*m images in steps S21-S24, and for the j-th image P(i,j) of the i-th layer, use the energy gradient function to calculate its clarity F value:
[0023]
[0024] Where x and y represent the xth row and yth column of the image P(i,j), f(x,y) represents the grayscale value of the xth row and yth column of the image P(i,j), 0<j≤m;
[0025] Step S32: discard images whose clarity F value is less than the set threshold, and regard the remaining images as clear images that meet the set conditions and save them in the AI edge computing module of the measuring device.
[0026] Preferably, the step S4 specifically includes:
[0027] Step S41: Crop the clear images that meet the set conditions, and crop each image into 5×5 sub-images, a total of 25 sub-images;
[0028] Step S42: annotate the cropped sub-image, annotate the target suspended matter particles, and use the annotated sub-image image as a target detection data set;
[0029] Step S43: using an improved YOLOv8 network to train the target detection data set to obtain a detection model;
[0030] Step S44: deploy the trained detection model on the server, use TensorRT to accelerate reasoning, and realize real-time reasoning.
[0031] Preferably, the step S6 specifically includes:
[0032] Step S61: Count the number of target suspended solids whose particle diameters fall within the intervals [0, 0.5], [0.5-5], [5-10], [10-15], [15-25], [25-50], [50-100], [100-200], [200-300], and [300-400] respectively;
[0033] Step S62: Calculate the D50 value of all target suspended matter particle diameters.
[0034] Preferably, the step S7 specifically includes:
[0035] Step S71: before entering the window box, disturb the injection water flow to change its flow rate and impurity distribution pattern;
[0036] Step S72: repeat steps S2-S6, and then weight all statistically obtained D50 values of the target suspended matter particle diameters.
[0037] Preferably, the step S8 specifically includes:
[0038] Step S81: transmitting the detection results such as the D50 value of the target suspended matter particle diameter to the offshore platform network control center via the serial port and / or wired / wireless local area network;
[0039] Step S82: The detection results such as the D50 value of the target suspended matter particle diameter are transmitted from the offshore platform network control center to the shore side through satellite communication methods such as Haisatong.
[0040] In a second aspect of the present invention, a deep learning-based system for measuring the median diameter of suspended particles in oil reservoir water injection is provided, the system comprising a measuring device and an offshore platform network control center, the measuring device specifically comprising: a housing, a flow rate adjustment module, a window module, an image acquisition module, a focusing module, an algorithm processing module, a communication module and a power supply module;
[0041] The flow rate regulating device comprises a magnetic pump and a driver, wherein the driver is used to provide driving force to the magnetic pump; the magnetic pump is used to regulate the water injection flow rate;
[0042] The window module is formed by two transparent glass sheets stacked opposite to each other in a vertical direction, and is used to introduce the injection water flow from the magnetic pump;
[0043] The image acquisition module includes a high-point light source, a high-power microscope lens and a high-definition industrial camera;
[0044] The focusing module includes a focusing component and a stepping motor;
[0045] The algorithm processing module includes an embedded AI edge computing module for processing the collected images through a visual algorithm;
[0046] The offshore platform network control center is used to receive the image processing results from the communication module and transmit the image processing results to the shore side via satellite communication.
[0047] Preferably, the direction from the inlet toward the outlet of the shell is defined as the first direction, and the direction from the bottom surface toward the top surface of the shell is defined as the second direction. On a first plane formed by the first direction and the second direction, a partition assembly parallel to the first plane is provided; the partition assembly divides the shell into a first chamber and a second chamber, the flow rate regulation module, algorithm processing module, communication module and power supply module are located in the first chamber, and the window module, image acquisition module and focusing module are located in the second chamber.
[0048] Preferably, the measurement system further comprises a disturbance module, which is located between the magnetic pump and the window module and is used for disturbing the water flow flowing out of the magnetic pump and then introducing the water flow into the window module.
[0049] According to a third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.
[0050] According to a fourth aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the steps of any of the aforementioned methods when executed by the processor.
[0051] Compared with the prior art, the present invention has the following beneficial effects:
[0052] 1) The two-chamber online monitoring device of the present invention can use the window module to provide a good field of view for the camera module, and realize micron-level real-time optical imaging and collection of suspended particles in oil reservoir water injection under the premise of ensuring stable and efficient operation of each module. With the improved deep learning algorithm, the accuracy of classification and regression can be improved, a better detection model can be obtained, and the median particle size of suspended particles can be calculated more quickly and accurately.
[0053] 2) By stratifying the liquid surface thickness, continuously sampling different layers at different flow rates, and using the energy gradient function to screen the sampling results, the influence of interferences such as bubbles and oil droplets can be eliminated, making the collected target suspended matter samples more representative and improving the accuracy of the results.
[0054] 3) The design of the spoiler module can flexibly adjust the water injection flow morphology and form multiple comparative reference examples in a targeted manner, which helps to further eliminate the influence of interferences such as bubbles and oil droplets and improve the reliability of the collected images. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] Figure 1 Flow chart of the method for measuring the median diameter of suspended particles in water injection in oil reservoirs based on deep learning of the present invention;
[0056] Figure 2 A side view of a device for measuring the median diameter of suspended particles in water injection in an oil reservoir according to the present invention from a first viewing angle;
[0057] Figure 3 A side view from a second viewing angle of the device for measuring the median diameter of suspended particles in water injection in an oil reservoir according to the present invention;
[0058] Figure 4 Structural diagram of the spoiler module of the present invention.
[0059] Explanation of the accompanying drawings: shell 1, magnetic pump 2, driver 3, algorithm processing module 4, communication module 5, power module 6, spoiler module 7, spoiler 71, window module 8, highlight light source 9, high-definition industrial camera 10, focusing module 11, partition assembly 12. DETAILED DESCRIPTION
[0060] The technology described below can be transformed in many ways and can have many embodiments, which are described in detail with reference to the accompanying drawings in specific embodiments. However, this does not mean that the technology described below is limited to specific embodiments. It should be understood that the present invention includes all similar modifications, equivalents and substitutions without departing from the spirit and technical scope of the technology described below.
[0061] Example 1
[0062] like Figure 1 As shown, the present invention provides a method for measuring the median diameter of suspended particles in oil reservoir water injection based on deep learning, comprising the following steps:
[0063] Step S1: Power on the system of the measuring device, connect the water inlet pipe and the water outlet pipe, introduce the injection water flow into the measuring device through the conduit, and adjust the injection flow rate to a first speed;
[0064] Step S2: stratify the injection water flow passing through the window box in the measuring device, and collect reservoir water injection images corresponding to different liquid level height positions;
[0065] Step S3: judging the clarity of each layer of images collected in layers, discarding images that do not meet the clarity standards, and saving clear images that meet the set conditions in the AI edge computing module of the measuring device;
[0066] Step S4: Analyze the image stored in the AI edge computing module using an improved deep learning-based detection algorithm to identify target suspended particles;
[0067] Step S5: using the HED algorithm based on deep learning to perform edge detection on the detected target suspended particles, fitting the edges into a circle and then calculating the diameter;
[0068] Step S6: Counting the distribution of target suspended matter particles within different diameter ranges;
[0069] Step S7: adjusting the water injection flow rate to a second speed, and repeating steps S2-S6;
[0070] Step S8: Transmit the statistical target suspended matter particle diameter distribution results to the shore side through the network.
[0071] Wherein, the step S2 specifically includes:
[0072] Step S21: dividing the liquid surface into n layers according to the thickness of the liquid surface in the window box, where n≥3;
[0073] Step S22: Move the high-definition industrial camera to the position corresponding to the i-th layer of liquid surface through the focusing module, 0<i≤n;
[0074] Step S23: adjusting the parameters of the high-definition industrial camera and the parameters of the highlight light source to meet the imaging requirements;
[0075] Step S24: continuously collect m images, where m≥1000.
[0076] Since the window box has a certain thickness (about 1mm) and the depth of field of a high-power microscope lens is generally tens of microns, the position of the suspended matter target in the window box is random. If stratified sampling is not performed, the collected suspended matter samples will lack representativeness.
[0077] Wherein, the step S3 specifically includes:
[0078] Step S31: Traverse the n layers of n*m images in steps S21-S24, and for the j-th image P(i,j) of the i-th layer, use the energy gradient function to calculate its clarity F value:
[0079]
[0080] Where x and y represent the xth row and yth column of the image P(i,j), f(x,y) represents the grayscale value of the xth row and yth column of the image P(i,j), 0<j≤m;
[0081] Step S32: discard images whose clarity F value is less than the set threshold, and regard the remaining images as clear images that meet the set conditions and save them in the AI edge computing module of the measuring device.
[0082] The threshold is usually set to 5000 based on experience, and can also be flexibly adjusted based on on-site conditions.
[0083] Wherein, the step S4 specifically includes:
[0084] Step S41: Crop the clear images that meet the set conditions, and crop each image into 5×5 sub-images, a total of 25 sub-images;
[0085] Step S42: annotate the cropped sub-image, annotate the target suspended matter particles, and use the annotated sub-image image as a target detection data set;
[0086] Step S43: using an improved YOLOv8 network to train the target detection data set to obtain a detection model;
[0087] Step S44: deploy the trained detection model on the server, use TensorRT to accelerate reasoning, and realize real-time reasoning.
[0088] The YOLOv8 network structure mainly consists of three parts: Backbone, Neck, and Detect Head. Backbone mainly refers to the CSPDarkNet structure, but uses the C2f (CSPLayer_2Conv) module instead of the C3 module. The C2f module is optimized on the basis of CSPDarkNet, reducing the amount of calculation while improving the convergence speed and effect. The input image first passes through a series of convolutional layers and pooling layers for feature extraction to form feature maps of different scales. The Neck part adopts a two-stream FPN (Feature Pyramid Network) structure similar to PANet (Path Aggregation Network). This structure integrates feature information of different scales and improves the model's ability to detect targets of different scales. The detection head adopts a Decoupled Head design similar to YOLOX, which decouples the classification and regression tasks and processes them by different branches respectively.
[0089] In order to improve the detection performance of the network, the present invention inserts a CoT module in the Neck part. The CoT module first context-encodes the input features through 3×3 convolution to obtain a static context representation. Then, the encoded features are connected to the query (Query) and input into the self-attention mechanism of the Transformer. Through the self-attention mechanism, the model can learn the dynamic dependencies between features. Finally, the features output by the self-attention mechanism are fused with the original features to obtain an enhanced feature representation. The detection head uses the features enhanced by the Transformer attention mechanism as input to improve the accuracy of classification and regression.
[0090] For step S5, HED is a classic edge detection algorithm based on deep learning. It uses a fully convolutional network to extract features at different levels and fuse features of different scales through skip connections to obtain more accurate edge detection results. This multi-scale feature fusion method enables HED to capture edge information of different scales in the image. At the same time, it can be trained end-to-end and learn edge features directly from the image without manually designing feature extractors. This enables HED to automatically learn edge features suitable for specific tasks, improving the accuracy and robustness of detection. Then the ODEC method (overlapping two-dimensional circle method) is used to fit the circle. The core idea of this method is to overlap a series of two-dimensional circles inside the particle shape contour, starting from a node on the contour, draw a circle along the inner normal direction of the node, and continuously expand the circle until its contour touches the particle contour. Repeat this process for other nodes to ensure that all circles overlap inside the particle and cover the entire particle shape as much as possible. Finally, according to the outer contour of the overlapping two-dimensional circles, an optimal circle is fitted by the algorithm.
[0091] Wherein, the step S6 specifically includes:
[0092] Step S61: Count the number of target suspended solids whose particle diameters fall within the intervals [0, 0.5], [0.5-5], [5-10], [10-15], [15-25], [25-50], [50-100], [100-200], [200-300], and [300-400] respectively;
[0093] Step S62: Calculate the D50 value of all target suspended matter particle diameters.
[0094] D50 refers to the particle size value corresponding to the cumulative distribution percentage reaching 50%. D50 is also called the median diameter or median particle size. Assuming that the D50 of a sample is 5um, it means that among the particles that make up this sample, particles with a diameter greater than 5um account for 50%, and particles with a diameter less than 5um also account for 50%. D50 is a typical indicator for evaluating the particle size of suspended matter, and is usually used to represent the average particle size of suspended matter.
[0095] Wherein, the step S7 specifically includes:
[0096] Step S71: before entering the window box, disturb the injection water flow to change its flow rate and impurity distribution pattern;
[0097] Step S72: repeat steps S2-S6, and then weight all statistically obtained D50 values of the target suspended matter particle diameters.
[0098] Under different flow rates and water flow conditions, the impurity forms and distribution in the injection water flow will be different; in addition, the camera will also have a certain impact on the imaging at different image acquisition positions. Therefore, based on the layered operation, by collecting multiple images at different positions and flow rates, processing and analyzing them separately, and performing fusion comparison, the recognition accuracy of the target suspended matter can be maximized.
[0099] When the process of step S32 is repeated for the second time, the threshold value to be compared with the sharpness F value may be different from the threshold value to be set when step S32 is repeated for the first time.
[0100] When step S41 is repeated for the second time, each image may be cropped into 36 sub-images of 6×6, that is, the cropping size of the image is different from that of the first time step S41 is performed.
[0101] When weighting is performed on the D50 value obtained by the first statistics and the D50 value obtained by the second statistics, the respective weights may be different.
[0102] Wherein, the step S8 specifically includes:
[0103] Step S81: transmitting the detection results such as the D50 value of the target suspended matter particle diameter to the offshore platform network control center via the serial port and / or wired / wireless local area network;
[0104] Step S82: Transmit the detection results such as the D50 value of the target suspended matter particle diameter from the offshore platform network control center to the shore side through satellite communication methods such as Haisatong.
[0105] Example 2
[0106] like Figure 2-3 As shown, the present invention also provides a deep learning-based median measurement system for reservoir water injection suspended matter particle diameters, the system comprising a measuring device and an offshore platform network control center, the measuring device specifically comprising: a shell 1, a flow rate adjustment module, a window module 8, an image acquisition module, a focusing module 11, an algorithm processing module 4, a communication module 5 and a power supply module 6.
[0107] The direction from the inlet of the shell 1 toward the outlet is defined as the first direction, and the direction from the bottom surface toward the top surface of the shell 1 is defined as the second direction. On a first plane formed by the first direction and the second direction, a partition assembly 12 parallel to the first plane is provided; the partition assembly 12 divides the shell 1 into a first chamber and a second chamber, the flow rate regulation module, the algorithm processing module 4, the communication module 5 and the power supply module 6 are located in the first chamber, and the window module 8, the image acquisition module and the focusing module 11 are located in the second chamber.
[0108] The inlet of the housing 1 is connected to the water inlet pipe, and the outlet is connected to the water outlet pipe.
[0109] The flow rate regulating device includes a magnetic pump 2 and a driver 3, wherein the driver 3 is used to provide driving force to the magnetic pump 2; the magnetic pump 2 is used to adjust the water injection flow rate in response to user input, or to adjust the water injection flow rate based on a pre-set program, or to automatically adjust the water injection flow rate according to current water injection flow rate, water pressure, water quality and other parameters.
[0110] The window module 8 is a window box formed by stacking two pieces of transparent glass opposite to each other in a direction perpendicular to the first plane and the first direction, and is used to introduce the water injection flow from the magnetic pump; the interlayer spacing between the two pieces of transparent glass of the window module 8 is h, and h<1mm.
[0111] The image acquisition module includes a highlight light source 9, a high-power microscope lens and a high-definition industrial camera 10; the highlight light source 9 is located on the side of the window module 8 facing the exit, and the high-power microscope lens and the high-definition industrial camera 10 are located on the side of the window module 8 facing the entrance and are arranged in sequence.
[0112] The focusing module 11 includes a focusing component and a stepping motor. The focusing component includes a slider and a slide rail. The slider is provided with a connector for fixing the high-definition industrial camera. The stepping motor is used to control the movement of the slider.
[0113] The algorithm processing module 4 includes an embedded AI edge computing module, which is used to process the collected images through a visual algorithm; the processing of the collected images through the visual algorithm includes image cropping, target recognition, edge detection, shape fitting, and image fusion.
[0114] The offshore platform network control center is used to receive the image processing results from the communication module and transmit the image processing results to the shore side via satellite communication.
[0115] The device further includes a disturbance module 7 , which is located between the magnetic pump 2 and the window module 8 and is used to disturb the water flow flowing out of the magnetic pump 2 and then introduce it into the window module 8 .
[0116] like Figure 4As shown, one end of the spoiler module 7 is connected to the outlet of the magnetic pump 2, and the other end is connected to the inlet of the window box 8, so that the whole constitutes a closed fluid passage close to a rectangle. On the upper and lower sides or the left and right sides of the passage, a spoiler 71 with one end fixed and the other end rotatable is arranged at staggered intervals, and the rotatable end of the spoiler 71 is connected to the side of the outlet of the magnetic pump 2 toward the spoiler module 7, and the rotation angle α is adjustable, and the range is specifically 0°≤α≤45°. By adjusting the different rotation angles of different spoilers 71, the water injection flow can be effectively disturbed, and the state of bubbles, oil droplets and suspended matter in the water injection flow can be changed, so as to facilitate the effective identification and extraction of target suspended matter through subsequent image acquisition and analysis. When the spoiler module 7 does not work, the spoiler 71 is located at the initial position, α=0°.
[0117] When performing step S71, some spoilers 71 are rotated to a certain angle to achieve disturbance of the injected water flow. Which spoiler 71 is rotated and the rotation angle value are all flexibly adjusted according to actual needs.
[0118] The positions of the spoiler 71, the high-definition industrial camera 10 and the flow rate of the water injection flow can be combined and changed in various ways according to actual needs, and are not limited to the above-mentioned situations.
[0119] Example 3
[0120] The present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the aforementioned methods.
[0121] Example 4
[0122] The present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the steps of any of the aforementioned methods when executed by the processor.
[0123] Although the present invention has been described in detail above with general description and specific embodiments, some modifications or improvements can be made on the basis of the present invention. The above description is only a preferred embodiment of the present invention and is not limited to the scope of the present invention. Other changes and modifications made by those skilled in the art without departing from the spirit and scope of protection of the present invention are still included in the scope of protection of the present invention.
Claims
1. A method for measuring the median diameter of suspended particles in oil reservoir water injection based on deep learning, characterized in that: The following steps are involved: Step S1: Power on the system of the measuring device, connect the water inlet pipe and the water outlet pipe, introduce the injection water flow into the measuring device through the conduit, and adjust the injection flow rate to a first speed; Step S2: stratify the injection water flow passing through the window box in the measuring device, and collect reservoir water injection images corresponding to different liquid level height positions; Step S3: judging the clarity of each layer of images collected in layers, discarding images that do not meet the clarity standards, and saving clear images that meet the set conditions in the AI edge computing module of the measuring device; Step S4: Analyze the image stored in the AI edge computing module using an improved deep learning-based detection algorithm to identify target suspended particles; Step S5: using the HED algorithm based on deep learning to perform edge detection on the detected target suspended particles, fitting the edges into a circle and then calculating the diameter; Step S6: Counting the distribution of target suspended matter particles within different diameter ranges; Step S7: adjusting the water injection flow rate to a second speed, and repeating steps S2-S6; Step S8: Transmit the statistical target suspended matter particle diameter distribution results to the shore side through the network.
2. The measuring method according to claim 1, characterized in that: The step S2 specifically includes: Step S21: dividing the liquid surface into n layers according to the thickness of the liquid surface in the window box, where n≥3; Step S22: Move the high-definition industrial camera to the position corresponding to the i-th layer of liquid surface through the focusing module, 0<i≤n; Step S23: adjusting the parameters of the high-definition industrial camera and the parameters of the highlight light source to meet the imaging requirements; Step S24: continuously collect m images, where m≥1000.
3. The measuring method according to claim 2, characterized in that: The step S3 specifically includes: Step S31: Traverse the n layers of n*m images in steps S21-S24, and for the j-th image P(i,j) of the i-th layer, use the energy gradient function to calculate its clarity F value: Where x and y represent the xth row and yth column of the image P(i,j), f(x,y) represents the grayscale value of the xth row and yth column of the image P(i,j), 0<j≤m; Step S32: discard images whose clarity F value is less than the set threshold, and regard the remaining images as clear images that meet the set conditions and save them in the AI edge computing module of the measuring device.
4. The measuring method according to claim 3, characterized in that: The step S4 specifically includes: Step S41: Crop the clear images that meet the set conditions, and crop each image into 5×5 sub-images, a total of 25 sub-images; Step S42: annotate the cropped sub-image, annotate the target suspended matter particles, and use the annotated sub-image image as a target detection data set; Step S43: using an improved YOLOv8 network to train the target detection data set to obtain a detection model; Step S44: deploy the trained detection model on the server, use TensorRT to accelerate reasoning, and realize real-time reasoning.
5. The measuring method according to claim 4, characterized in that: The step S6 specifically includes: Step S61: Count the number of target suspended solids whose particle diameters fall within the intervals [0, 0.5], [0.5-5], [5-10], [10-15], [15-25], [25-50], [50-100], [100-200], [200-300], and [300-400] respectively; Step S62: Calculate the D50 value of all target suspended matter particle diameters.
6. The measuring method according to claim 5, characterized in that: The step S7 specifically includes: Step S71: before entering the window box, disturb the injection water flow to change its flow rate and impurity distribution pattern; Step S72: repeat steps S2-S6, and then weight all statistically obtained D50 values of the target suspended matter particle diameters.
7. The measuring method according to claim 6, characterized in that: The step S8 specifically includes: Step S81: transmitting the detection results such as the D50 value of the target suspended matter particle diameter to the offshore platform network control center via the serial port and / or wired / wireless local area network; Step S82: Transmit the detection results such as the D50 value of the target suspended matter particle diameter from the offshore platform network control center to the shore side through satellite communication methods such as Haisatong.
8. A measurement system for implementing the method for measuring the median diameter of suspended particles in water injection in oil reservoirs based on deep learning as claimed in any one of claims 1 to 7, characterized in that: The system includes a measuring device and an offshore platform network control center, wherein the measuring device specifically includes: a housing, a flow rate adjustment module, a window module, an image acquisition module, a focusing module, an algorithm processing module, a communication module and a power supply module; The flow rate regulating device comprises a magnetic pump and a driver, wherein the driver is used to provide driving force to the magnetic pump, and the magnetic pump is used to regulate the water injection flow rate; The window module is formed by two transparent glass sheets stacked opposite to each other in a vertical direction, and is used to introduce the injection water flow from the magnetic pump; The image acquisition module includes a high-light point light source, a high-power microscope lens and a high-definition industrial camera; The focusing module includes a focusing component and a stepping motor; The algorithm processing module includes an embedded AI edge computing module for processing the collected images through a visual algorithm; The offshore platform network control center is used to receive the image processing results from the communication module and transmit the image processing results to the shore side via satellite communication.
9. The measuring system according to claim 8, characterized in that The direction from the inlet toward the outlet of the shell is defined as a first direction, and the direction from the bottom surface toward the top surface of the shell is defined as a second direction. A partition assembly parallel to the first plane is provided on a first plane formed by the first direction and the second direction; the partition assembly divides the shell into a first chamber and a second chamber, the flow rate regulation module, the algorithm processing module, the communication module and the power supply module are located in the first chamber, and the window module, the image acquisition module and the focusing module are located in the second chamber.
10. The measuring system according to claim 9, characterized in that It also includes a disturbance module, which is located between the magnetic pump and the window module and is used to disturb the water flow flowing out of the magnetic pump and then introduce it into the window module.
Citation Information
Patent Citations
A nanoparticle size measurement method based on an improved Mask R-CNN
CN109948712A
Monitoring of particulate matter in water supply
US20010015096A1
Filtrate monitoring device, and filtrate monitoring system
US20100097605A1
Methods and systems for particle classification
US20230186610A1
Water quality image analysis method and system based on deep learning, and device and medium
WO2022217589A1
Cited By
Online monitoring device and method for diameter median value of oil reservoir water injection suspended matter particles
CN118777177A
