Method of automated biostratigraphic support for well drilling

Abstract

The invention relates to methods of biostratigraphic support for well drilling. The present method comprises the steps of generating a provisional drilling model, photographing nannofossils, carrying out a biostratigraphic analysis and establishing a biostratigraphic scale of the drilling area with biostratigraphic subdivisions, marking target and non-target regions of a microscope slide and marking objects on enlarged images, determining an algorithm for the automated photographing of nannofossils, training a neural network to identify images of nannofossils, generating a plan for sampling drilling cuttings, carrying out the planned sampling at the drilling area and numbering the samples, extracting nannofossils from a sample and preparing a microscope slide, photographing the slide, identifying the target zones thereof on a computer with the trained neural network, positioning sections of the microscope slide under a microscope while disregarding the regions that are devoid of nannofossils, photographing the sections of the microscope slide that contain nannofossils, identifying by means of the trained neural network the nannofossils extracted from the sample, determining the actual biostratigraphic subdivisions of the drilling cuttings, comparing the provisional and actual subdivisions and updating the drilling model. The technical result consists in improving the efficiency of the drilling process.

Description

[0001]

[0002] METHOD OF AUTOMATED BIOSTRATIGRAPHIC SUPPORT OF WELL DRILLING

[0003] The present invention relates to the field of computer processing of specialized data to ensure the process of well drilling support, namely to methods of biostratigraphic support, and can be used to determine the biostratigraphic subdivision and position of wells during directional drilling.

[0004] The productive strata of deposits through which directional drilling is conducted may be composed of various biostratigraphic units characterized by the remains of fossil microorganisms. The boundaries of these units are determined by both the evolutionary changes of individual taxa or microfauna (microflora) complexes and their replacement, regardless of the lithological and other physical characteristics of the constituent rocks. Biostratigraphic support for well drilling is based on the analysis of fossil microorganisms contained in drilled rock. Analysis of fossil microorganisms allows us to determine the relative geological age of the drilled rock, as well as the actual biostratigraphic level being drilled.Analysis of fossil microorganisms isolated from drill cuttings brought from the borehole to the surface is performed by specialists at the drilling site using a microscope to study the fossil microorganisms.

[0005] When supporting directional drilling, the main challenge is the presence of low-amplitude tectonic faults with displacements that are not determined by seismic surveys and are not distinguished by well logging methods (LWD / MWD) due to their similar lithological composition. Due to the unknown direction and amplitude of the displacement, it is difficult to determine the position of the drilling tool relative to the target formation and adjust the subsequent drilling trajectory. This problem can be solved by identifying the current biostratigraphic unit by determining the relative geological age of the rock based on the analysis of fossil microorganisms. PCI7RU2024 / 000362

[0006] A fossil recognition device, as well as a corresponding system and method, are known from the prior art (Russian Federation Patent for Invention No. 2586348 C2, published on June 10, 2016). The method, implemented by means of a processor, includes receiving fluid image information from an image generator in which the fluid is located in the field of view, wherein said fluid image information contains fossil image information, processing the fossil image information to identify fossil types in the fluid in the form of data associating fossil types with the formation from which the fluid was obtained, determining the location of a first well in the formation based on said data and offset records associated with a second well, and publishing said data in combination with location indications.A device for recognizing fossils comprises a fluid container having an optically transparent window and intended to contain fluid and fossils, a pump for moving fluid into and out of said container, at least one image former for receiving fluid image information from the fluid, wherein said fluid image information contains fossil image information, and a processor intended for extracting fossil image information from the fluid image information, for determining fossil types in the fluid in the form of data associating fossil types with the formation from which the fluid was obtained, and for determining the location of a first well in the formation based on said data and offset records associated with the second well. The fossil recognition system comprises a wellbore tool casing and said device.The disadvantages of this solution include the absence of a maceration process (the process of separating fossil microorganisms from the rock); the fluid contains rock fragments in which the fossil microorganisms and the rock are not separated, which complicates recognition not only by computer vision, but also by humans.

[0007] The closest to the claimed solution is a method for analyzing sedimentary samples with automatic recognition of fossil microorganisms (patent application WO2015132531 A1, published on September 11, 2015), which includes the following stages: obtaining images of samples using a microscope; pre-processing the images to extract zones of interest from them; analyzing the zones of interest using artificial neural networks to conduct a first classification of objects between groups of species of fossil microorganisms; analyzing the zones of interest using at least one morphostatistical recognition method to conduct a second classification of objects between groups of species of fossil microorganisms; and collecting the results of the first and second classifications in at least one file indicating the groups of species of fossil microorganisms, respectively assigned to the zones of interest by the first and second classifications.

[0008] A system for analyzing sedimentary samples includes at least one microscope for obtaining images of samples and is configured to obtain three images of each sample in polarized light with different polarization and an image of each sample in natural light, as well as a computer and a computing server configured to implement the following steps: pre-processing of images to extract zones of interest; analysis of zones of interest using artificial neural networks to perform a first classification of objects between groups of species of fossil microorganisms; analysis of zones of interest using at least one morphostatistical recognition method to perform a second classification of objects between groups of species of fossil microorganisms;and collecting the results of the first and second classifications in at least one file indicating groups of species of fossil microorganisms, respectively assigned to the zones of interest by the first and second classifications, wherein the analysis operations are performed on the basis of images in polarized light for the first groups of species of fossil microorganisms and on the basis of images in natural light for the second groups of species of fossil microorganisms, less birefringent than the first fossil microorganisms of the species of the first groups.

[0009] The disadvantages of this solution include the insufficiently precise selection of photographic areas containing fossil microorganisms, as the entire area selected by the operator is photographed. The distribution of fossil microorganisms, which are the target of recognition, on the studied surface is random. Furthermore, in addition to fossil microorganisms, the macerate contains various particles of organic and inorganic origin that are not of interest to the researcher and cannot be identified by this solution.

[0010] The problem addressed by the claimed invention is to ensure controlled positioning of a micropreparation for obtaining digital images of fossil microorganisms, identifying these microorganisms and performing statistical calculations of their types.

[0011] The technical result of the claimed invention consists in increasing the efficiency of the drilling process by increasing the accuracy of determining the actual biostratigraphic subdivision of rock sections subject to destruction during well drilling or during drilling of a mine working.

[0012] The technical result is achieved by a method of automated biostratigraphic support of well drilling, which includes the following stages: a) a design model of well drilling is formed on a computer, b) a micropreparation containing fragments of a drill core is placed on the microscope stage, and the fossil microorganisms contained therein are photographed with a digital camera in an amount sufficient for biostratigraphic analysis and the formation of a training and test sample for a neural network, c) a biostratigraphic analysis is carried out and a biostratigraphic scale of the drilling site with biostratigraphic subdivisions of the rock is formed, d) the images of microorganisms obtained in stage (b) are transmitted to a computer, where target and non-target areas of the micropreparation are marked, as well as objects are marked on enlarged images, e) after which an algorithm for automated photographing of fossil microorganisms is determined,characterizing the biostratigraphic units of the drilling site, f) based on the data obtained in step (e), a neural network is trained to identify images of fossil microorganisms, g) a plan for collecting samples of drill cuttings is then generated on the computer, h) after which a planned collection and numbering of samples of drill cuttings is carried out at the drilling site in accordance with the data obtained in step (g), i) fossil microorganisms are isolated from the numbered sample of drill cuttings and a microscopic specimen is prepared, j) the microscopic specimen is placed under a microscope, photographed with a digital camera, the resulting images are transmitted to a computer with a trained neural network, where the target zones of the microscopic specimen are identified, after which a signal is sent via the computer to the control unit, which generates a control signal for the motors of the microscope stage, which is then transmitted to the said motors,position sections of the micropreparation in front of the microscope objective, ignoring areas of the micropreparation that do not contain fossil microorganisms, k) photograph sections of the micropreparation that contain fossil microorganisms and transmit the images to a computer, where the trained neural network identifies fossil microorganisms isolated from a numbered sample of cuttings, i) determine the actual biostratigraphic subdivision of the drill cuttings on the computer, t) compare the planned and actual biostratigraphic subdivisions and update the design and actual drilling model.

[0013] Below we will examine in detail the terms and their definitions used in the description of the technical solution.

[0014] A microscope slide is a glass slide with a macerate applied to it (rock fragments, microorganisms, fragments of microorganisms, etc.) with a cover glass.

[0015] Intelligent photography is the acquisition of digital photographic images of target areas of a microscopic specimen containing fossil microorganisms.

[0016] The claimed invention is explained by the figures. Fig. 1 shows a diagram of a system implementing the claimed method of automated biostratification of well drilling; Fig. 2 shows photographs of a drill core sample taken from a target interval as part of an example of implementing the invention; Fig. 3 shows a photograph of non-target zones of a micropreparation; and Fig. 4 shows a photograph of target zones of a micropreparation. The following are designated by numbers: 1 - microscope, 2 - digital camera, 3 - anti-vibration table, 4 - computer, 5 - control unit.

[0017] The method of automated biostratigraphic support of directional drilling is implemented with the help of a system including a positioning system consisting of a light microscope 1 containing different objectives for different types of microorganisms and configured to change the light source (transmitted, reflected, etc.), a high-speed digital camera 2 with a matrix providing high resolution, and, to ensure the accuracy of photography, an additional anti-vibration table 3 connected via a communication channel to a computer 4, as well as a control unit 5 to which the said computer 4 is connected via a communication channel. As an anti-vibration table 3, a design known from the prior art is used (see, for example, patent for utility model No. 120583, published on September 27, 2012), selected depending on the operating conditions (offshore platform, drilling site on land) for isolating the microscope 1 from vibration.The control unit 5 is a microcontroller (programmable logic integrated circuit) and is designed to control the piezoelectric motors of the microscope stage 1 and process signals from the sensors of the microscope stage 1 (primarily optical position sensors for monitoring the movement of the stage with high accuracy, as well as force and collision sensors for preventing damage to the microscope and the sample, temperature and vibration sensors for monitoring the environment and correcting the operation of the system depending on the conditions) in real time.For this purpose, control unit 5 is equipped with piezoelectric motor drivers (electronic components for converting control signals from the control unit 5 microcontroller into commands for the motors of the microscope's mechanized stage 1) and a control panel for manual control of the positioning system, if necessary, as well as displaying the current system status (position, errors, etc.). Control unit 5 is equipped with communication interfaces (USB, Ethernet, or Wi-Fi) for connection to computer 4 for data transfer and control.Computer 4 contains software for a system for biostratigraphic support of well drilling (or separate software products), configured to form a well drilling design model, divide the drilling design model into biostratigraphic units, mark objects on an image, determine a biostratigraphic unit based on the parameters of identified fossil microorganisms by a trained neural network, form a drilling sample collection plan, pre-process images from a digital camera, identify images from a digital camera by a trained neural network, determine a biostratigraphic unit of a numbered rock sample (in particular, drill cuttings), compare the planned and actual biostratigraphic unit, identify tectonic faults, and may also contain a machine learning system for training a neural network for the purpose of detecting target zones of fossil microorganisms.

[0018] The claimed method of automated biostratigraphic support of directional drilling is carried out as follows.

[0019] During the preparatory stage, which occurs before drilling, a design geological and hydrodynamic drilling model is created with the parameters of the well being constructed (information on the geological structure of the drilling site, the planned well placement, etc.) using software such as that used by oil companies. The resulting design model is converted into biostratigraphic support system software pre-installed on computer 4 for further subdivision of the said design model into biostratigraphic units.

[0020] To subdivide the drilling project model into biostratigraphic units, biostratigraphic studies are conducted in accordance with methodological recommendations on various parts of the borehole core and, optionally, drill cuttings, related to various geological heterogeneities and located closest to the drilling site or any existing reference boreholes at the field. For this purpose, a slide containing drill core fragments is placed on the microscope stage 1. A sufficient number of microorganisms contained within are photographed at low magnification using a digital camera 2, sufficient for biostratigraphic analysis. All microorganisms that can be identified are also photographed at higher magnification, sufficient for the formation of training and test samples for the neural network. Biostratigraphic analysis is then performed using a known method, and a local biostratigraphic section is created.At this stage, there is no need to photograph all the microorganisms in the core slide. It is sufficient to identify them all, perform statistical calculations, and create a biostratigraphic section. During the study, drill core samples are obtained, from which fossil microorganisms are subsequently isolated, characterizing geological heterogeneities or mineral differences. Local and regional stratigraphic scales are also used for reference. The result of the preliminary biostratigraphic study is a biostratigraphic scale for the drilling site with biostratigraphic units characterized by the quantitative and species composition of fossil microorganisms.

[0021] The resulting images are then used in computer 4 to label target and non-target areas of the slide using image labeling software. Target areas of the slide are those containing microorganisms—not empty areas, but areas containing microorganisms, their fragments, or rock fragments. The labeling software also labels enlarged images of fossil microorganisms by highlighting objects (fossil microorganisms) in the image and assigning them to a specific species.

[0022] Next, the optimal algorithm for automated photography of fossil microorganisms that characterize the biostratigraphic units of the drilling site is determined (lenses for photographing the micropreparation are selected to determine the target areas, the magnification factor, light (transmitted, or direct, or polarized) for photographing microorganisms, and photographs of all microorganisms that can be identified are taken with a higher degree of magnification).

[0023] The next step involves training the neural network, which is implemented using a well-known machine learning system installed on computer 4 (e.g., ResNet50 and YOLOv5). The neural network is trained using a known method, with the input data being labeled images of fossil microorganisms obtained from preliminary biostratigraphic analysis of the core (and optionally drill cuttings) to identify the fossil microorganism images. Neural network training can also be performed on a separate high-performance server computer. Neural network training serves two purposes: by labeling the microscope slide images, the neural network enables subsequent positioning of the slide on the microscope stage 1, and by labeling the enlarged images of fossil microorganisms, it enables their subsequent identification in drill cuttings samples.

[0024] Based on the obtained biostratigraphic scale of the drilling site, the software pre-installed on the system's computer is configured to determine the biostratigraphic unit based on data on the number and species of fossil microorganisms characterizing various biostratigraphic units. Following this configuration, the software should determine the biostratigraphic unit based on data received from the neural network on the number and species of identified fossil microorganisms in the images.

[0025] In the next step, a drilling cuttings sampling plan is generated in computer software 4 based on the geological conditions of the drilling site, the sampling depth for the first sample is established, and the sampling interval for each subsequent sample, including the sampling depth for the final sample, is determined. Additionally, the sampling depth is determined in intervals before and after known tectonic faults. The next step of the automated biostratigraphic drilling support method is the production stage and is carried out at the drilling site as follows. A planned sampling of drilling cuttings is performed, numbering the samples by depth in accordance with the data obtained from the analysis in the previous step. Fossil microorganisms are isolated from the dispersed phase and dispersed medium of the numbered drilling cuttings sample, and microscopic preparations are prepared using standard maceration techniques for photographing the fossil microorganisms.

[0026] Next, the prepared microslide is placed under microscope 1 of the claimed system and intelligently photographed using a mechanism controlled by the microcontroller of control unit 5. This mechanism positions the sections of the microslide containing fossil microorganisms in front of the microscope objective 1, ignoring areas of the microslide that do not contain fossil microorganisms. For this purpose, computer 4 sends signals to microcontroller control unit 6 in accordance with neural network data. The neural network identifies target zones on the microslide and creates an algorithm for sequentially photographing the target zones. Based on the information received, software in microcontroller control unit 5 sends control signals to the motors of microscope stage 1, which position the microslide. The result of this step is digital images of only the portion of the microslide containing microorganisms.

[0027] The obtained images from camera 2 are sent to computer software 4, where these images are pre-processed and photographs of fossil microorganisms are loaded for further identification by a trained neural network of fossil microorganisms isolated from a numbered sample of drill cuttings.

[0028] Also, by means of the specified software, based on a statistical comparison of the type and quantity of fossil microorganisms isolated from the cuttings with the biostratigraphic scale formed during the biostratigraphic analysis of the core, a statistical calculation is made and the actual biostratigraphic subdivision of the numbered sample of drill cuttings is determined.

[0029] Next, the computer software compares the planned and actual biostratigraphic units and updates the planned and actual drilling models. If the planned and actual biostratigraphic units match, the software visualizes the drilling site model. If the planned and actual biostratigraphic units do not match, the software notifies the specialist by visualizing the planned and actual biostratigraphic levels for decision-making.

[0030] When a low-amplitude tectonic fault is detected that is not reflected in the planned drilling model, unscheduled cuttings sampling is performed at the drilling site before and after the fault. This is possible because the cuttings are brought to the surface with a certain delay. Unscheduled cuttings samples are numbered, indicating the sampling depth. Fossil microorganisms are then isolated from the dispersed phase and dispersed medium of the unscheduled cuttings samples collected before and after the fault, and microscopic slides are prepared for photographing the fossil microorganisms. The prepared slide is placed under the automated microscope 1 of the system, and the fossil microorganisms are photographed in accordance with a previously defined optimal algorithm for automated photography of fossil microorganisms characterizing biostratigraphic units before and after the fault.

[0031] Next, computer software 4 preprocesses the images and uploads photographs of fossil microorganisms to identify fossil microorganisms isolated from drill cuttings samples collected unscheduled before and after the fault using a trained neural network. The software also performs statistical calculations and determines the biostratigraphic subdivision of the unscheduled drill cuttings samples based on the number and types of identified fossil microorganisms. The biostratigraphic subdivisions before and after the fault are then compared, and the direction and amplitude of the displacement are determined.Using computer software 4, a visualization of the tectonic fault is performed and the design and actual drilling model is updated with the display of the tectonic fault, and recommendations are also generated for changing the further drilling trajectory to reach the target formation.

[0032] An example illustrating the operation of the system and implementation of the method.

[0033] To simulate real-world conditions for biostratigraphic drilling support, work traditionally performed in biostratigraphic drilling support was performed. Initially, two targets were identified: a target and a non-target, representing distinct stratigraphic units. The target is an oil-saturated light-gray fine-grained sandstone with gently undulating bedding caused by deposition of dark-gray fine-grained clayey siltstone and cross-wavy bedding. Isolated clayey intraclasts and rare synsedimentary deformations are noted within the layer. The sedimentation setting is a coastal-marine complex. At the top, the target is overlain by a facies of a shallow-water-marine sedimentary complex. This facies consists of dark-gray fine-grained clayey siltstone with fine subhorizontal and lenticular bedding caused by light-gray coarse-grained siltstone. The layer shows isolated bioturbation (Planolites).

[0034] In the core photograph, the target unit (light gray sandstone at the bottom of the first row from the left) is shown in daylight and ultraviolet light (see Fig. 2). The ultraviolet glow of the target unit indicates hydrocarbon saturation. The overlying unit (dark gray fine-grained clayey siltstone at the top of the first row from the left, the middle row of the core, and the bottom of the third row) shows no signs of hydrocarbon saturation in ultraviolet light.

[0035] Based on the results of biostratigraphic core studies, a table was compiled showing the quantitative and species abundance of microorganisms in the target and overlying non-target areas. This table serves as a local stratigraphic scale for the two facies, characterizing the local distribution of microorganisms across the section, and serving as the basis for statistical calculations and the assignment of samples to the target and non-target facies, which are separate stratigraphic units.

[0036] Table. Quantitative and species content of microorganisms in the target object and in the overlapping non-target object. Next, microscopic slides containing fossilized microorganisms isolated from drill cuttings obtained during drilling of the horizontal section of the well were prepared. The slides were preliminarily analyzed by a specialist, and those containing fossilized microorganisms from the target and non-target sites were identified.

[0037] Next, to conduct biostratigraphic analysis, a sequence of microscopic specimen examinations was determined, simulating the gradual progression of the drilling tool through the target object, with a pre-planned exit to a non-target horizon and return of the drilling trajectory to the target horizon. To simulate automated biostratigraphic support for drilling, specimens containing fossil microorganisms were sequentially photographed under a microscope with objectives of varying magnification. Photographing at low magnifications identified target areas of the specimen containing fossil microorganisms. The microscope stage and focusing were controlled manually. The images obtained in this mode were used to identify areas of the specimens rich in fossil microorganisms.Photographing slides containing fossil microorganisms with high-magnification lenses was intended for identification purposes. Sequential photography of 10 slides, followed by image processing by the system's computer, allowed for the construction of an experimental drilling model that matched the planned model. The software system successfully identified fossil microorganisms, counted them statistically, and determined their facies.

[0038] As a result of using the claimed technical solution, the accuracy of determining the actual biostratigraphic unit of drill cuttings, as well as drill core, drill cuttings in real time at each sampling point is increased and, consequently, the accuracy of determining the position of the wellbore, which is either in the target biostratigraphic unit, which is a productive formation, or in a biostratigraphic unit located above or below the target productive formation, is increased.

Claims

CLAUSES OF THE INVENTION A method for automated biostratigraphic support of well drilling, which includes the following stages: a) a design model of well drilling is formed on a computer, b) a micropreparation containing fragments of a drill core is placed on the microscope stage and the fossil microorganisms contained therein are photographed with a digital camera in an amount sufficient for biostratigraphic analysis and the formation of a training and test sample for a neural network, c) a biostratigraphic analysis is carried out and a biostratigraphic scale is formed for the drilling site with biostratigraphic subdivisions of the drill core, d) the images of microorganisms obtained in step (b) are transmitted to a computer, where target and non-target areas of the micropreparation are marked, as well as objects are marked on enlarged images, e) after which an algorithm for automated photographing of fossil microorganisms is determined,characterizing the biostratigraphic units of the drilling site, f) based on the data obtained in step (e), a neural network is trained to identify images of fossil microorganisms, d) then a plan for collecting samples of drill cuttings is generated on the computer, h) after which, at the drilling site, a planned collection and numbering of samples of drill cuttings is carried out in accordance with the data obtained in step (d), 1) fossil microorganisms are isolated from a numbered sample of drill cuttings and a microscopic preparation is prepared, ^the microscopy is placed under the microscope, photographed with a digital camera, the resulting images are transmitted to a computer with a trained neural network, where the target zones of the microscopy are identified, after which a signal is sent via the computer to the control unit, which generates a control signal for the motors of the microscope stage, which is then transmitted to the said engines position sections of the micropreparation in front of the microscope objective, ignoring areas of the micropreparation that do not contain fossil microorganisms, k) photograph sections of the micropreparation that contain fossil microorganisms and transmit the images to a computer, where a trained neural network identifies fossil microorganisms isolated from a numbered sample of cuttings, l) determine the actual biostratigraphic subdivision of the drill cuttings on the computer, g) compare the planned and actual biostratigraphic subdivisions and update the planned and actual drilling models.

Images