A well condition early warning method based on cuttings identification

Abstract

The application discloses a well condition early warning method based on rock debris identification, and comprises the following steps: according to the rock breaking condition of a drill bit in a drilling process, the theoretical distribution characteristics of the rock debris flow rate at different well depths are calculated; the rock debris image of a vibrating screen is collected and the flow back rock debris is identified, and based on this, the actual distribution characteristics of the rock debris flow rate at different well depths are calculated; and according to the theoretical distribution characteristics of the rock debris flow rate and the actual distribution characteristics of the rock debris flow rate at different well depths, the well condition early warning at different well depths is carried out. The application can predict the wellhead types at different well depths and timely find the complex conditions of the drilled positions at different depths in a well.

Description

Technical Field

[0001] This invention relates to the field of cuttings detection technology, and in particular to a well condition early warning method based on cuttings identification. Background Technology

[0002] Rock cuttings are rock particles carried to the surface by the drill bit during drilling. As the most direct reflection of the downhole condition, the morphology of the cuttings can be used to determine whether the drilling is proceeding normally at the drilling site. Currently, drill cuttings monitoring is mainly used in gas drilling. Furthermore, in realizing the technical solution of this invention, the inventors discovered at least the following problems in the existing technology:

[0003] (1) Existing well condition early warning can only evaluate the cleanliness of the wellbore at the location reached by the drill bit. When the drill bit drills deeper into the well, it cannot continuously monitor the cleanliness of the wellbore at the already drilled depth. In other words, it cannot achieve long-term and quantitative monitoring of the cleanliness of the wellbore in the already drilled section.

[0004] (2) Most existing drill cuttings monitoring schemes can only be applied to the gas drilling industry, thus forming a monitoring method for gas drilling cuttings return monitoring system. They cannot be used in conventional liquid drilling fluids due to different field conditions. Moreover, most existing schemes can only detect the presence or absence of drill cuttings and the quantitative statistics of drill cuttings weight, but cannot accurately identify the quantity or shape of drill cuttings, which limits the judgment of the cleanliness of the downhole well. In addition, shape analysis is more valuable for oil drilling.

[0005] (3) At present, the shape of the drill cuttings on the vibrating screen is mainly identified by manual labor, which cannot achieve real-time online judgment and is greatly affected by human subjective factors, resulting in high labor intensity.

[0006] (4) Existing wellbore stability analysis systems based on drill cuttings analysis require thorough cleaning of drill cuttings in order to achieve analysis and statistics of falling blocks. Summary of the Invention

[0007] The purpose of this invention is to provide a well condition early warning scheme based on drill cuttings morphology recognition, which can realize real-time dynamic cumulative prediction of drilled sections during the drilling process.

[0008] To address the aforementioned technical problems, this invention provides a well condition early warning method based on cuttings identification, comprising: calculating the theoretical distribution characteristics of cuttings return volume at different well depths based on the drill bit's rock breaking situation during drilling; acquiring images of cuttings from a vibrating screen and identifying the returned cuttings, and based on this, calculating the actual distribution characteristics of cuttings return volume at different well depths; and conducting well condition early warning at different well depths based on the theoretical distribution characteristics and the actual distribution characteristics of cuttings return volume at different well depths.

[0009] Preferably, the step of calculating the theoretical distribution characteristics of cuttings return volume at different well depths based on the rock breaking situation of the drill bit during drilling includes: collecting drill bit size, real-time logging data, and waterhole information; calculating the bottom hole rock breaking volume of the drill bit per unit time based on the collected data; calculating the theoretical value of the wellhead cuttings return volume based on the bottom hole rock breaking volume and the theoretical rock carrying efficiency per unit time; and calculating the real-time cumulative cuttings volume at different well depths based on the theoretical wellhead cuttings return volume.

[0010] Preferably, the theoretical rock-carrying efficiency is determined based on the drill bit type, drilling fluid inlet flow rate, and drilling fluid properties; the bottom hole rock breaking volume is calculated based on the drill bit diameter, wellbore enlargement rate, and well depth variation.

[0011] Preferably, based on the drilling fluid inlet flow rate, well depth, drill string size, and wellbore structure, the arrival time of the drilling fluid from the bottom of the well to different well depths is estimated; based on the arrival time at the different well depths, combined with the theoretical cuttings return volume at the wellhead, the theoretical value of the real-time cumulative cuttings volume at different well depths is calculated, and the theoretical distribution characteristics of the cuttings return volume at different well depths are plotted.

[0012] Preferably, during the cuttings return process, a first type of image containing information on the lateral transport status of unwashed cuttings and a second type of image containing information on the longitudinal morphological features of unwashed cuttings are acquired; the morphological features of each cuttings are extracted based on the first and second types of images, and the shape of each cuttings is identified using a preset cuttings morphology recognition model; based on the particle size and shape of each cuttings, the actual amount of cuttings returned to the wellhead per unit time is calculated; based on the late arrival time at different well depths and combined with the actual amount of cuttings returned to the wellhead, the actual value of the real-time cumulative amount of cuttings at different well depths is calculated, and the actual distribution characteristics of the cuttings return amount at different well depths are plotted.

[0013] Preferably, the rock debris morphology recognition model is constructed according to the following steps: collecting a large number of top-view morphology images and vertical cross-sectional laser information of unwashed returned rock debris, wherein the vertical cross-sectional laser information is the contour feature of the vertical cross-section corresponding to each position point along the lateral movement direction of the rock debris as the axis; extracting the corresponding lateral cross-sectional contour based on the top-view morphology image of each rock debris sample, and extracting the laser information of key vertical cross-sections based on the vertical cross-sectional laser information of each rock debris sample; classifying and labeling the three-dimensional shape of each rock debris sample; constructing a first neural network model, taking the lateral cross-sectional contour and key vertical cross-sectional laser information of each rock debris sample as input, and the corresponding three-dimensional shape as output, training, verifying and testing the first neural network model to construct the rock debris morphology recognition model.

[0014] Preferably, the step of calculating the actual amount of wellhead cuttings returned per unit time based on the particle size morphology and shape of each cuttings includes: calculating the volume of each cuttings piece based on its three-dimensional shape and particle size morphology, thereby calculating the actual amount of wellhead cuttings returned per unit time, wherein the three-dimensional shape includes circular, plate-like, and three-dimensional triangular shapes.

[0015] Preferably, the step of conducting well condition early warning for different well depths based on the theoretical distribution characteristics and actual distribution characteristics of cuttings flowback at different well depths includes: integrating the theoretical distribution characteristics and actual distribution characteristics of cuttings flowback at different well depths belonging to the same time period according to the well depth sequence; comparing the theoretical and actual values ​​of cuttings flowback at the same well depth location in the same time period, and determining the well condition type at different well depth locations in the current time period based on the comparison results, wherein, based on the difference between the theoretical and actual values, the well cleanliness level and corresponding well condition at the corresponding well depth location are determined using the difference threshold corresponding to different well cleanliness levels.

[0016] Preferably, based on the theoretical distribution characteristics of cuttings flowback at different well depths and the actual distribution characteristics of cuttings flowback at different well depths, the downhole early warning model is used to predict the well condition type at different well depths.

[0017] Preferably, the downhole early warning model is constructed using the following steps: During the cuttings runoff process, the theoretical distribution characteristics and actual distribution characteristics of cuttings runoff at different well depths are continuously collected, and the well condition type, wellbore cleanliness estimate, and cuttings carrying efficiency estimate at different well depths within the corresponding time period are labeled; a second neural network model is constructed, taking the theoretical distribution characteristics and actual distribution characteristics of cuttings runoff at different well depths as inputs, and the corresponding well condition type, wellbore cleanliness estimate, and cuttings carrying efficiency estimate as outputs, and training, validating, and testing the second neural network model to construct the downhole early warning model.

[0018] Compared with the prior art, one or more embodiments of the above solutions may have the following advantages or beneficial effects:

[0019] This invention proposes a well condition early warning method based on cuttings identification. This method theoretically analyzes the distribution of cuttings at different well depths during the flowback process using real-time logging data. It also calculates the actual distribution of flowback cuttings at different depths per unit time by analyzing the shape of the cuttings using a neural network model algorithm. By comparing theoretical and actual data, the method predicts the well condition type at different depths. Therefore, this invention not only judges wellbore cleanliness and drilling fluid cuttings carrying capacity by comparing the total volume of flowback cuttings per unit time with the theoretical value, but also predicts wellhead types at different depths, promptly identifying complex situations at drilled locations at different depths, thus enabling timely intervention and ensuring the smooth operation of drilling.

[0020] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims, and drawings. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0022] Figure 1 This is a flowchart illustrating the steps of a well condition early warning method based on cuttings identification, according to an embodiment of this application.

[0023] Figure 2 This is a schematic diagram illustrating the specific process of the well condition early warning method based on cuttings identification according to an embodiment of this application. Detailed Implementation

[0024] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples, so that the process of how the present invention uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly. It should be noted that, as long as there is no conflict, the various embodiments and features in the various embodiments of the present invention can be combined with each other, and the resulting technical solutions are all within the protection scope of the present invention.

[0025] Furthermore, the steps illustrated in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in a different order than that shown here.

[0026] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments. Unless the context clearly indicates otherwise, the singular forms “a” and “an” as used herein are also intended to include the plural. It should also be understood that the terms “comprising” and / or “including” as used herein specify the presence of the stated features, integers, steps, operations, units, and / or components, without excluding the presence or addition of one or more other features, integers, steps, operations, units, components, and / or combinations thereof.

[0027] Rock cuttings are rock particles carried to the surface by the drill bit during drilling. As the most direct reflection of the downhole condition, the morphology of the cuttings can be used to determine whether the drilling is proceeding normally at the drilling site. Currently, drill cuttings monitoring is mainly used in gas drilling. Furthermore, in realizing the technical solution of this invention, the inventors discovered at least the following problems in the existing technology:

[0028] (1) Existing well condition early warning can only evaluate the cleanliness of the wellbore at the location reached by the drill bit. When the drill bit drills deeper into the well, it cannot continuously monitor the cleanliness of the wellbore at the already drilled depth. In other words, it cannot achieve long-term and quantitative monitoring of the cleanliness of the wellbore in the already drilled section.

[0029] (2) Most existing drill cuttings monitoring schemes can only be applied to the gas drilling industry, thus forming a monitoring method for gas drilling cuttings return monitoring system. They cannot be used in conventional liquid drilling fluids due to different field conditions. Moreover, most existing schemes can only detect the presence or absence of drill cuttings and the quantitative statistics of drill cuttings weight, but cannot accurately identify the quantity or shape of drill cuttings, which limits the judgment of the cleanliness of the downhole well. In addition, shape analysis is more valuable for oil drilling.

[0030] (3) At present, the shape of the drill cuttings on the vibrating screen is mainly identified by manual labor, which cannot achieve real-time online judgment and is greatly affected by human subjective factors, resulting in high labor intensity.

[0031] (4) Existing wellbore stability analysis systems based on drill cuttings analysis require thorough cleaning of drill cuttings in order to achieve analysis and statistics of falling blocks.

[0032] To address the aforementioned technical problems, this application proposes a well condition early warning method based on cuttings identification. This method calculates the expected cuttings distribution baseline in the wellbore under normal conditions and measures the actual cuttings distribution in real time during the cuttings runoff process. Based on the comparison between the expected and actual cuttings distribution, the method determines the wellbore cleanliness and provides well condition early warning.

[0033] Figure 1This is a flowchart illustrating the steps of a well condition early warning method based on cuttings identification, according to an embodiment of this application. Figure 1 As shown, the well condition early warning method described in this embodiment of the invention includes the following steps: Step S110 calculates the theoretical distribution characteristics of cuttings return volume at different well depths based on the rock breaking situation of the drill bit during drilling; Step S120 acquires images of cuttings from a vibrating screen and identifies the returned cuttings, and calculates the actual distribution characteristics of cuttings return volume at different well depths based on this; Finally, Step S130 conducts well condition early warning for different well depths based on the theoretical distribution characteristics and the actual distribution characteristics of cuttings return volume at different well depths.

[0034] Therefore, this invention utilizes steps S110 and S120 to obtain the theoretical and actual distribution characteristics of cuttings return volume at different well depths within each time period. Based on these theoretical and actual distribution characteristics, the well condition type corresponding to different well depths is determined, and early warnings are issued for abnormal well condition types. Well condition types include, but are not limited to: normal, poor cuttings carrying capacity, stuck drill bit, cuttings falling off, unstable wellbore, and wellbore collapse.

[0035] Figure 2 This is a schematic diagram illustrating the specific process of the well condition early warning method based on cuttings identification, as described in an embodiment of this application. The following is a detailed flowchart... Figure 1 and Figure 2 The specific process of the well condition early warning method described in the embodiments of the present invention will be explained.

[0036] In step S110, step S1101 involves acquiring drill bit size, real-time logging data, and waterhole information. Specifically, in step S1101, not only drill bit size data and waterhole-related information are obtained, but real-time logging data is also acquired during the drilling process. This logging data includes, but is not limited to, real-time logging time, real-time logging depth, drilling fluid inlet flow rate data, standpipe pressure, and suspended weight.

[0037] Step S1102 calculates the bottom hole rock breaking volume per unit time based on the data collected in step S1101. Thus, this embodiment of the invention can calculate the bottom hole rock breaking volume per unit time in real time in step S1102. In step S1102, the bottom hole rock breaking volume is calculated based on the drill bit diameter, wellbore enlargement rate, and well depth changes. Specifically, this embodiment of the invention calculates the bottom hole rock breaking volume in real time according to the following expression:

[0038] V = π(D / 2) 2 *a*ΔH (1)

[0039] Where V represents the rock breaking volume at the bottom of the well, D represents the drill bit diameter, a represents the wellbore enlargement rate, and ΔH represents the change in well depth per unit time compared to the well depth per unit time.

[0040] Step S1103 calculates the theoretical value of the wellhead cuttings return volume based on the bottom hole rock breaking volume in step S1102 and the theoretical rock carrying efficiency per unit time. Specifically, in step S1103, the rock carrying volume per unit time is calculated in real time based on the bottom hole rock breaking volume in step S1102 and the theoretical rock carrying efficiency per unit time, which is the theoretical value of the wellhead cuttings return volume per unit time.

[0041] Furthermore, based on the drill bit type, drilling fluid inlet flow rate, and drilling fluid properties, the theoretical cuttings carrying efficiency k corresponding to the current drilling state is determined. Then, the theoretical cuttings carrying efficiency k is multiplied by the real-time cuttings breaking volume V at the bottom of the well to obtain the volume of cuttings carried back from the bottom of the well by the drilling fluid per unit time, which is the theoretical value of the real-time cuttings back flow at the wellhead.

[0042] Finally, step S1104 calculates the real-time cumulative amount of cuttings at different well depths based on the theoretical backflow of cuttings at the wellhead per unit time.

[0043] Specifically, in step S1104, firstly, based on the drilling fluid inlet flow rate, well depth, drill string size, and wellbore structure, the arrival time of the drilling fluid from the bottom of the well to different well depths is estimated. Within a certain drilling time period, the cutter fluid reaches the wellhead vibrating screen after the arrival time at the wellhead. Since the arrival time from the bottom of the well is less than the arrival time at the wellhead, the theoretical return volume at different well depths during the return process from the bottom of the well to the wellhead can be calculated based on the theoretical return volume calculated above. Therefore, this embodiment of the invention also calculates the theoretical value of the real-time cumulative (volume) cutter volume at different well depths based on the arrival time at different well depths and the theoretical return volume at the wellhead, thereby plotting the theoretical distribution characteristic curve of the cutter return volume at different well depths within the current unit time.

[0044] Thus, according to step S110, the present invention can not only obtain the drill bit rock breaking volume curve under different unit logging time, but also obtain the theoretical distribution characteristic curve of rock cuttings return volume formed by the theoretical values ​​of rock cuttings return volume at different well depths for each time period.

[0045] During the drilling process, this invention not only uses step S110 to calculate the theoretical distribution characteristics of cuttings flowback at different well depths in real time, but also uses step S120 to calculate the actual distribution characteristics of cuttings flowback at different well depths in real time for each unit logging time period by acquiring flowback images at vibration points.

[0046] like Figure 2 As shown, in step S1201, during the rock cuttings return process, a first type of image containing information on the lateral transport status of unwashed rock cuttings and a second type of image containing information on the longitudinal morphological characteristics of unwashed rock cuttings are acquired.

[0047] In step S1201, during the rock cuttings return process, a visible light camera positioned at the vibrating screen outlet acquires a first type of image. Simultaneously, during the rock cuttings return process, step S1201 also acquires a second type of image using a laser line scan camera positioned at the vibrating screen outlet. Therefore, this embodiment of the invention directly acquires images and performs morphological detection on the returned rock cuttings without requiring a rock cuttings cleaning process, while maintaining the accuracy of the total rock cuttings volume identification.

[0048] Furthermore, in this embodiment of the invention, a visible light camera is positioned above or at the edge of the vibrating screen's outlet. The visible light camera's field of view encompasses the entire width of the vibrating screen's outlet. Thus, the visible light camera's image field of view clearly shows the entire sand return state across the vibrating screen's outlet width, enabling the camera to detect the lateral movement speed and lateral cross-sectional profile of all rock fragments on the vibrating screen.

[0049] Furthermore, in this embodiment of the invention, the laser line scanner is positioned above or at the edge of the vibrating screen's outlet. The laser line emitted by the laser line scanner covers the entire width of the vibrating screen's outlet. In this way, all rock chips conveyed at the vibrating screen's outlet must flow through the laser line, enabling the laser line scanner to obtain the contour of each rock chip on the vertical surface of the vibrating screen, facilitating real-time three-dimensional shape recognition.

[0050] Further, in step S1202, the morphological features of each cuttings are extracted based on the first and second types of images acquired in real time during the cuttings return process in step S1201. A cuttings morphology recognition model based on a neural network model is then used to identify the shape of each cutting. The morphological features of the cuttings include: the cuttings transport velocity and the lateral area contour of each cutting based on the first type of image; and the vertical contour features of the cuttings analyzed based on the second type of image. The vertical contour features are the contour features of key vertical sections.

[0051] In step S1202, the transport speed and transverse cross-sectional profile of each rock fragment on the vibrating screen are first analyzed based on the first type of image, and the vertical profile features of each rock fragment on the vibrating screen are analyzed based on the second type of image. Then, the transverse area profile and vertical profile features of each rock fragment are input into the rock fragment morphology recognition model to output the (three-dimensional) shape of each rock fragment. The three-dimensional shape is selected from one of the following: circular (including sphere or ellipsoid), sheet-like (thin sheet and / or strip-like), and three-dimensional triangle.

[0052] Specifically, step S1202 first parses the first type of image information into an area array image, identifies all rock cuttings on the vibrating screen in the parsed area array image, and marks each identified rock cutting portion. Simultaneously, it obtains the transport velocity of each rock cutting in the XY horizontal direction and the transverse cross-sectional profile of the (drill cuttings), thus obtaining the analysis results of the first type of image information. Then, it parses the second type of image information into a laser line scan image, identifies all rock cuttings on the vibrating screen in the parsed image, and marks each identified rock cutting portion. Simultaneously, it determines the vertical profile features of each rock cutting in the YZ vertical direction (wherein, the vertical profile features of the rock cuttings are preferably formed by different transverse cross-sections of the current rock cuttings along the transverse transport direction). The vertical scan profile under the position coordinates is formed into a vertical profile sequence, and key vertical cross-sectional profile information is extracted from the vertical profile sequence to obtain the analysis results of the second type of image information. Then, the analysis results of the first type of image information and the second type of image information acquired at the same time are integrated. By identifying the same rock fragments, the migration speed, transverse cross-sectional profile, and key vertical cross-sectional profile information of each rock fragment at the current image acquisition time are obtained. Finally, the transverse cross-sectional profile and key vertical cross-sectional profile information of each rock fragment obtained above are input into the rock fragment morphology recognition model, and the corresponding three-dimensional shape is output for each rock fragment.

[0053] It should be noted that, in the embodiments of the present invention, in the process of obtaining the key vertical cross-sectional profile, it is necessary to extract the vertical cross-section where the maximum vertical height is located and multiple pairs of vertical cross-sections on both sides of the vertical interface where the maximum vertical height is located from the vertical profile sequence.

[0054] Furthermore, the rock cuttings morphology recognition model described in this embodiment of the invention is a model for classifying rock cuttings morphology based on a neural network model. This model can directly classify each rock cuttings into three-dimensional shapes based on the morphological features extracted from the first and second types of images. In practical applications, due to the highly irregular shapes of the returned drill cuttings, inaccurate statistical results may occur when calculating drill cuttings volume. Moreover, if too many sampling points are used when calculating drill cuttings volume, the volume calculation process becomes overly cumbersome, making it difficult to meet the real-time requirements of the vibrating screen return process. Therefore, this embodiment of the invention statistically divides all rock cuttings samples into three basic rock cutting shapes, which not only facilitates the statistical analysis of rock cuttings volume but also serves as a diagnostic basis for drilling safety early warning.

[0055] Specifically, the rock debris morphology recognition model is constructed through the following steps: First, a large number of unwashed, returned rock debris samples are collected, including top-view morphology images and vertical cross-sectional laser information sequences. The vertical cross-sectional laser information sequences represent the contour features of the vertical cross-sections at various points along the lateral movement direction of the rock debris. Then, the corresponding lateral cross-sectional contours are extracted from the top-view morphology images of each rock debris sample, and the laser information of key vertical cross-sections is extracted from the vertical cross-sectional laser information sequences of each sample. Next, each rock debris sample is classified and labeled according to its three-dimensional shape. Finally, a neural network model is constructed, using the lateral cross-sectional contours and key vertical cross-sectional laser information of each rock debris sample as input and the corresponding three-dimensional shape as output. The neural network model is trained, validated, and tested to construct the rock debris morphology recognition model.

[0056] In addition, during field applications, the embodiments of the present invention can periodically collect images of rock debris from the vibrating screen and measure the shape, volume, and particle size distribution of each returned rock debris, so as to continuously optimize and update the above-mentioned rock debris morphology recognition model.

[0057] After obtaining the three-dimensional shape information of each rock cuttings in real time, the process proceeds to step S1203. Step S1203 calculates the total volume of rock cuttings returned per unit time based on the morphological characteristics and three-dimensional shape of each rock cuttings obtained in step S1202, which is the actual value of the wellhead rock cuttings return volume per unit time (actual wellhead rock cuttings return volume).

[0058] After forming the three-dimensional shape of each rock fragment, step S1203 will also calculate the volume of each rock fragment based on its three-dimensional shape and morphological characteristics, and further calculate the actual amount of rock fragments returned per unit time (total volume).

[0059] Specifically, after identifying the (three-dimensional) geometric shape characteristics of a single rock cuttings, step S1203 will calculate the volume of each transported rock cuttings based on the (three-dimensional) geometric shape characteristics of each rock cuttings, combined with the rock cuttings transport speed, lateral area profile, and key vertical cross-sectional profile characteristics of each rock cutting. Then, the volume of each rock cutting is statistically analyzed using a cumulative algorithm to calculate the total volume of all returned rock cuttings per unit time, and the shape category ratio data of all returned rock cuttings per unit time is also statistically analyzed. Thus, the shape characteristics and volume data of each rock cutting, the total volume of returned rock cuttings per unit time, and the shape category ratio data are used as rock cuttings statistical results information, thereby obtaining the actual amount of wellhead rock cuttings returned per unit time.

[0060] Finally, step S1204, based on the late arrival time at different well depths and combined with the actual wellhead cuttings return volume per unit time, calculates the actual value of the real-time cumulative (volume) cuttings volume at different well depths for each unit logging time, thereby plotting the actual distribution characteristic curve of cuttings return volume at different well depths for different unit time.

[0061] Thus, the present invention can generate an actual distribution characteristic curve of cuttings flowback volume based on the actual values ​​of cuttings flowback volume at different well depths for each time period, according to step S120.

[0062] Furthermore, in this embodiment of the invention, step S130 will be used to identify and issue warnings for the well condition type at different well depths based on the theoretical distribution characteristics of cuttings flowback at different well depths obtained in step S110 and the actual distribution characteristics of cuttings flowback at different well depths obtained in step S120.

[0063] In one embodiment, in step S130, firstly, (step S1301) the theoretical distribution characteristics of cuttings flowback at different well depths within the same time period and the actual distribution characteristics of cuttings flowback at different well depths are integrated and arranged according to the well depth sequence. Then, (step S1302) the theoretical and actual values ​​of cuttings flowback at the same well depth location within the same time period are compared, and the well condition type at different well depth locations within the current time period is determined based on the comparison results.

[0064] Since the methods used to determine the well condition type at the same time period and well depth are similar, this embodiment of the invention will only use the wellhead type determination method at a certain well depth as an example for explanation. Specifically, based on the difference between the theoretical value and the actual value of cuttings return at the current well depth (denoted as the real-time difference), the well condition type at the corresponding well depth is determined using the difference threshold corresponding to different wellbore cleanliness levels.

[0065] If the aforementioned real-time difference is greater than zero and the absolute value of the real-time difference reaches or exceeds the preset abnormal difference threshold, it indicates that cuttings may not be returning normally, posing a significant risk of stuck pipe (the well condition type at the current well depth is stuck pipe). If the aforementioned real-time difference is less than zero and the absolute value of the real-time difference reaches or exceeds the preset abnormal difference threshold, it means that complex downhole conditions such as cuttings falling or wellbore collapse may have occurred (the well condition type at the current well depth is cuttings falling, wellbore collapse, or wellbore instability, etc.). Furthermore, if the absolute value of the aforementioned real-time difference is less than the preset abnormal difference threshold, it indicates that the well condition type at the current well depth is normal.

[0066] In addition, in order to improve the efficiency and ensure the accuracy of the well condition early warning method described in the embodiments of the present invention, the embodiments of the present invention will also use a pre-constructed downhole early warning model to predict the well condition type at different well depths based on the theoretical distribution characteristics of cuttings flowback at different well depths and the actual distribution characteristics of cuttings flowback at different well depths.

[0067] Furthermore, in this embodiment of the invention, the downhole early warning model is constructed using the following steps: First, during the cuttings runoff process, the theoretical distribution characteristics of cuttings runoff at different well depths obtained in step S110 and the actual distribution characteristics of cuttings runoff at different well depths obtained in step S120 are continuously collected, and the well condition type, wellbore cleanliness estimate, and (actual) cuttings carrying efficiency estimate at different well depths within the corresponding time period are labeled; then, a second neural network model is constructed, taking the theoretical distribution characteristics of cuttings runoff at different well depths and the actual distribution characteristics of cuttings runoff at different well depths as inputs, and taking the corresponding well condition type, wellbore cleanliness estimate, and (actual) cuttings carrying efficiency estimate as outputs, and training, verifying, and testing the constructed second neural network model to build the downhole early warning model.

[0068] Specifically, during drilling, the downhole early warning model can be constructed using the following method: under appropriate conditions (e.g., when the amount of training data reaches a certain quantity), the downhole early warning model can be directly used to predict well condition types, wellbore cleanliness, and actual inclined hole efficiency at different well depths by using the theoretical and actual distribution characteristics of cuttings return volume obtained from real-time logging data and real-time vibrating screen cuttings return.

[0069] ① Collect image data of rock cuttings from the vibrating screen and real-time logging data at the corresponding time;

[0070] ② Mark the well condition type, wellbore cleanliness estimate, and actual rock-carrying efficiency estimate according to the corresponding time, with the wellbore cleanliness and actual rock-carrying efficiency marked as percentages;

[0071] ③ Calculate the theoretical wellhead cuttings backflow volume and the real-time cumulative cuttings volume at different well depths using real-time logging data.

[0072] ④ Using a cuttings identification algorithm, calculate the actual amount of cuttings returned to the wellhead at the corresponding time and the actual value of the real-time cumulative amount of cuttings at different well depths (the cumulative amount of cuttings at each well depth includes the cumulative volume and particle size distribution);

[0073] ⑤ Using real-time data such as time, well depth, stand pressure, and suspended weight from the logging data, as well as theoretical cuttings quantity, cumulative cuttings quantity, actual cuttings quantity collected by cuttings visual identification, cumulative quantity difference, and particle size distribution as inputs, and the labeled wellbore cleanliness status, cuttings carrying efficiency, and wellbore cleanliness level as outputs, a second neural network model is constructed, and the second neural network model is trained using the BP algorithm to obtain the downhole early warning model.

[0074] ⑥ In the application of the downhole early warning model, the theoretical distribution characteristics and actual distribution characteristics of cuttings flowback are obtained by collecting real-time logging data and cuttings images from vibrating screens. The downhole early warning model is then used to obtain data on well condition type, cuttings carrying efficiency estimate, and wellbore cleanliness within the corresponding time period, thereby providing early warning for complex downhole conditions (abnormal well condition types).

[0075] This invention discloses a well condition early warning method based on cuttings identification. This method theoretically analyzes the distribution of cuttings at different well depths during the flowback process using real-time logging data. It also calculates the actual distribution of flowback cuttings at different well depths per unit time by analyzing the shape of the cuttings using a neural network model algorithm. By comparing theoretical and actual data, the method predicts the well condition type at different well depths. Therefore, this invention not only judges wellbore cleanliness and drilling fluid cuttings carrying capacity by comparing the total volume of flowback cuttings per unit time with the theoretical value, but also predicts wellhead types at different well depths, promptly identifying complex situations at different drilled depths and implementing timely intervention measures to ensure the smooth operation of drilling.

[0076] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0077] In the description of this invention, unless otherwise stated, "a plurality of" means two or more; the terms "upper," "lower," "left," "right," "inner," "outer," "front end," "rear end," "head," "tail," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0078] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0079] It should be understood that the embodiments disclosed herein are not limited to the specific structures, processing steps, or materials disclosed herein, but should be extended to equivalent substitutions of these features as understood by those skilled in the art. It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting.

[0080] The phrase "an embodiment" or "an embodiment" used in this specification means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. Therefore, the phrase "an embodiment" or "an embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment.

[0081] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and changes in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection of this invention shall still be determined by the scope defined in the appended claims.

Claims

1. A well condition early warning method based on cuttings identification, characterized in that, include: Based on the rock breaking situation of the drill bit during drilling, the theoretical distribution characteristics of cuttings return volume at different well depths are calculated; Images of rock cuttings from a vibrating screen are acquired and the returned rock cuttings are identified. Based on this, the actual distribution characteristics of rock cuttings returned at different well depths are calculated. Based on the theoretical and actual distribution characteristics of cuttings flowback at different well depths, well condition early warning systems are implemented at different well depths. During the rock cuttings return process, first-class images containing information on the lateral transport status of unwashed rock cuttings and second-class images containing information on the longitudinal morphological characteristics of unwashed rock cuttings were acquired. Based on the first type of image and the second type of image, extract the morphological features of each rock fragment, and use the preset rock fragment morphology recognition model to identify the shape of each rock fragment; Based on the particle size, morphology, and shape of each rock cuttings, the actual amount of rock cuttings returned to the wellhead per unit time was statistically analyzed. Based on the arrival time at different well depths and the actual amount of cuttings returned from the wellhead, the actual real-time cumulative amount of cuttings at different well depths is calculated, and the actual distribution characteristics of the cuttings return at different well depths are plotted. The rock debris morphology recognition model is constructed according to the following steps: A large number of top-view morphological images and vertical cross-sectional laser information of unwashed returned rock cuttings were collected. The vertical cross-sectional laser information is the contour feature of the vertical cross-section corresponding to each position point along the lateral movement direction of the rock cuttings as the axis. The corresponding transverse section contour is extracted from the top-view morphological image of each rock cutting sample, and the laser information of key vertical sections is extracted from the vertical section sequence laser information of each rock cutting sample. Each rock fragment sample was classified and labeled according to its three-dimensional shape; A first neural network model is constructed, taking the transverse cross-sectional contour and key vertical cross-sectional laser information of each rock debris sample as input and the corresponding three-dimensional shape as output. The first neural network model is trained, verified and tested to construct the rock debris morphology recognition model.

2. The well condition early warning method according to claim 1, characterized in that, The steps for calculating the theoretical distribution characteristics of cuttings return at different well depths based on the drill bit's rock breaking behavior during drilling include: Collect drill bit dimensions, real-time logging data, and waterhole information; Based on the collected data, calculate the amount of rock broken at the bottom of the well by the drill bit per unit time; Based on the rock breaking volume at the bottom of the well, and combined with the theoretical rock carrying efficiency per unit time, the theoretical value of the wellhead cuttings return volume is calculated. Based on the theoretical flowback rate of cuttings at the wellhead, the real-time cumulative amount of cuttings at different well depths was calculated.

3. The well condition early warning method according to claim 2, characterized in that, The theoretical rock-carrying efficiency is determined based on the drill bit type, drilling fluid inlet flow rate, and drilling fluid properties. The rock breaking volume at the bottom of the well is calculated based on the drill bit diameter, wellbore enlargement rate, and well depth variations.

4. The well condition early warning method according to claim 2, characterized in that, Based on the drilling fluid inlet flow rate, well depth, drill string size, and well structure, estimate the time it takes for the drilling fluid to travel from the bottom of the well to different well depths. Based on the late arrival time at different well depths and the theoretical cuttings runoff volume at the wellhead, the theoretical value of the real-time cumulative cuttings volume at different well depths is calculated, and the theoretical distribution characteristics of the cuttings runoff volume at different well depths are plotted.

5. The well condition early warning method according to claim 1, characterized in that, The step of calculating the actual amount of cuttings returned to the wellhead per unit time based on the particle size, morphology, and shape of each cuttings includes: Based on the three-dimensional shape of each rock cuttings and the particle size characteristics of each rock cuttings, the volume of each rock cuttings is calculated, thereby statistically analyzing the actual amount of rock cuttings returned to the wellhead per unit time. The three-dimensional shape includes circular, plate-like, and three-dimensional triangular shapes.

6. The well condition early warning method according to any one of claims 1 to 5, characterized in that, The steps for conducting well condition early warning at different well depths based on the theoretical distribution characteristics of cuttings flowback at different well depths and the actual distribution characteristics of cuttings flowback at different well depths include: The theoretical distribution characteristics of cuttings flowback at different well depths within the same time period and the actual distribution characteristics of cuttings flowback at different well depths are integrated according to the well depth sequence. Compare the theoretical and actual values ​​of cuttings return at the same well depth and time period, and determine the well condition type at different well depths within the current time period based on the comparison results. Specifically, based on the difference between the theoretical and actual values, the well cleanliness level and corresponding well condition at the corresponding well depth are determined using the difference threshold corresponding to different well cleanliness levels.

7. The well condition early warning method according to any one of claims 1 to 5, characterized in that, Based on the theoretical distribution characteristics of cuttings flowback at different well depths and the actual distribution characteristics of cuttings flowback at different well depths, a downhole early warning model is used to predict well condition types at different well depths.

8. The well condition early warning method according to claim 7, characterized in that, The downhole early warning model is constructed using the following steps: During the cuttings runoff process, the theoretical distribution characteristics of cuttings runoff volume at different well depths and the actual distribution characteristics of cuttings runoff volume at different well depths are continuously collected, and the well condition type, wellbore cleanliness estimate and cuttings carrying efficiency estimate at different well depths within the corresponding time period are marked. A second neural network model is constructed, taking the theoretical distribution characteristics of cuttings flowback at different well depths and the actual distribution characteristics of cuttings flowback at different well depths as inputs, and the corresponding well condition type, wellbore cleanliness estimation, and cuttings carrying efficiency estimation as outputs. The second neural network model is then trained, verified, and tested to construct the downhole early warning model.

Images