A kind of tank ear ranging device, monitoring method and system for coal mine vertical shaft wellbore

Abstract

The application discloses a cage ear distance measuring device, a monitoring method and a system for a coal mine vertical shaft well. The device is composed of four components: distance measuring wheel, swing arm transmission, base fixation and rotation angle sensor. The distance measuring wheel is attached to the cage guide by pre-tightening the tension spring. The swing arm rotation angle is collected by the encoder, and the cage ear gap is converted. The scheme carries a surface array camera on the cage, and the device shares a lithium battery power supply circuit with pre-charging protection. The underground image is enhanced and optimized through multi-stage filtering and CLAHE. Based on the YOLOv5s model, a crack identification network is trained to extract compliant crack parameters. The system binds the gap and crack data coordinates of the same well depth to construct a joint atlas. According to the gap threshold, three types of early warning of cage guide wear, shaft wall damage and high-risk support deformation are issued. The invention combines mechanical ranging and machine vision to realize unmanned automatic inspection of the well, and solves the pain points of high risk, poor precision and data isolation in traditional manual inspection.

Description

Technical Field

[0001] This invention relates to the field of mine safety technology, and more specifically, to a lug measuring device, monitoring method, and system for use in coal mine vertical shafts. Background Technology

[0002] Coal mine shaft cages are crucial facilities for the lifting and lowering of coal and personnel. The shafts are constantly subjected to ground pressure, water seepage, and coal dust erosion, leading to frequent safety hazards such as wear on the cage lugs and guide beams, and cracks and leaks in the shaft walls. Traditional shaft inspections rely primarily on manual inspections by personnel entering the mine, which has several drawbacks: First, the working space in the shaft is confined, posing a risk of falling objects and personnel falling into the shaft; second, gaps in the cage lugs and minute cracks in the shaft walls are only detectable visually, resulting in poor measurement accuracy and a high rate of missed detections; third, manual inspections require shutting down the hoisting system, taking approximately two hours per inspection, severely impacting normal mine production, and necessitating the long-term deployment of certified inspection teams, resulting in high labor and maintenance costs.

[0003] Currently, existing monitoring methods mostly rely on manual visual inspection or single-point laser ranging to detect gaps in the wellhead. However, underground coal slurry, dust, and pipeline obstructions can easily cause laser refraction and deviation, resulting in large errors in ranging data and making it impossible to achieve dynamic and continuous measurement throughout the entire lifting process. Well wall defects are mostly captured on-site by ordinary industrial cameras. However, the underground environment is dark and the existing lighting is insufficient, resulting in dim and noisy camera images that are difficult to capture small cracks, water seepage, and other defects. Conventional supplementary lighting equipment has high power consumption and heat dissipation difficulties, and the battery life under underground battery power supply conditions is insufficient to meet the needs of full wellbore inspection. The supporting inspection equipment generally uses conventional wireless direct charging. When the lithium battery is deeply depleted, the BMS enters sleep mode and locks out, and conventional charging cannot conduct the circuit, causing the equipment to shut down and the inspection to be interrupted.

[0004] Existing detection schemes are independent of each other. The detection of mechanical wear in the shaft and the detection of structural defects in the well wall are based on two separate sets of equipment and control systems. The data cannot be linked and matched at the same well depth, and it is impossible to realize the integrated analysis of hidden dangers between the shaft electromechanical system and the surrounding rock structure. This makes it difficult to meet the development needs of fully automated and unmanned shaft inspection in smart mines. Summary of the Invention

[0005] The main objective of this invention is to provide a lug measuring device, monitoring method, and system for coal mine vertical shafts, so as to at least solve the technical problems of high safety risks and insufficient inspection accuracy in the prior art, and to realize the development needs of fully automated and unmanned shaft inspection in smart mines.

[0006] To achieve the above objectives, a can ear ranging device, monitoring method, and system for use in coal mine vertical shafts are provided.

[0007] In a first aspect, the present invention provides a lug-based ranging device for a vertical shaft in a coal mine, the ranging device comprising:

[0008] The measuring wheel is a flat, disc-shaped rotating component. The measuring wheel is used to roll in contact with the guide beam of the coal mine vertical shaft and convert the gap change between the guide lug of the coal mine vertical shaft and the guide beam into swing displacement.

[0009] The swing arm transmission assembly is a crank arm irregular rotating component. The short end of the swing arm transmission assembly is connected to the measuring wheel by bolts. The swing arm transmission assembly is used to make the measuring wheel fit tightly against the tank track beam.

[0010] The base support and fixing assembly is a multi-component integrated assembly. The long end of the base support and fixing assembly is connected to the swing arm transmission assembly by bolts. The base support and fixing assembly is used to support the ranging device.

[0011] The corner sensor acquisition component is connected to the base support fixing component and the swing arm transmission component by bolts. The corner sensor acquisition component is used to acquire corner signals.

[0012] Specifically, the swing arm drive assembly includes:

[0013] The rotating arm is a bent component. The long end of the rotating arm is connected to the angle sensing and acquisition component, and the short end of the rotating arm is connected to the ranging wheel and the base support and fixing component. The rotating arm is used to connect the ranging wheel, the base support and fixing component and the angle sensing and acquisition component.

[0014] The pin is a cylindrical component that is inserted between the short end of the rotating arm and the measuring wheel. The pin is used to connect the measuring wheel and the rotating arm.

[0015] The drive shaft is a cylindrical component that passes between the long end of the rotating arm and the base support and fixing assembly. The shaft end of the drive shaft is connected to the angle sensing and acquisition assembly. The drive shaft is used to provide a support axis for the rotating arm.

[0016] The elastic preload spring is a helical spring with its two ends attached to the rotating arm and the base support fixing assembly, respectively. The elastic preload spring is used to continuously pull the rotating arm to deflect to one side of the tank passage.

[0017] Specifically, the base support fixing components include:

[0018] The base fixing plate is a rectangular flat plate with multiple pre-drilled mounting holes. The base fixing plate is used to provide an installation reference for the ranging device.

[0019] The bearing housing is a block-shaped housing with a through shaft hole in the center and a bearing inside. The bearing housing is mounted on the mounting hole of the base plate through the bearing. The bearing housing is also connected to the drive shaft through the bearing. The bearing housing is used to constrain the radial movement of the drive shaft.

[0020] The limiting ring is a ring-shaped sleeve. It is installed on the mounting hole of the base fixing plate by bolts. The limiting ring is fitted between the drive shaft and the bearing housing. The limiting ring is used to limit the axial displacement of the drive shaft.

[0021] The limiting bolt is an externally threaded cylindrical fastener. The limiting bolt is vertically fixed in the limiting groove of the base fixing plate and the limiting ring. The limiting bolt is used to limit the outward extension distance of the measuring wheel.

[0022] Specifically, the corner sensing and acquisition component includes:

[0023] An encoder is a cylindrical component that is coaxially connected to a drive shaft. The encoder is used to acquire rotation angle signals in real time.

[0024] The encoder mounting bracket is a U-shaped component. It is connected to the base mounting plate by bolts and is used to fix the position of the encoder.

[0025] Secondly, the present invention provides a method for inspection and monitoring of coal mine vertical shafts. The monitoring method provides an image acquisition device, which is installed on the top of a cage. The method is characterized by applying a cage-ear distance measuring device as described in the first aspect, which is installed on a cage-ear support on the outside of the cage. The monitoring method includes:

[0026] The turning angle signal is collected by the can ear ranging device, and the original underground image of the coal mine shaft is collected by the image acquisition equipment. The original underground image is filtered, enhanced and abnormal image is removed to obtain the processed underground image.

[0027] Manually annotated crack images are obtained, and a deep learning model is trained based on the manually annotated crack images to obtain a trained deep learning model. The processed underground images are then input into the trained deep learning model for recognition to obtain crack disease data of the coal mine vertical shaft.

[0028] The corner signal is converted into the gap value of the can ear, and the crack disease data of the same depth in the coal mine shaft are bound to the gap value of the can ear with coordinate correlation to obtain a joint data map.

[0029] A preset threshold for the gap between the shaft and the cannula is used to classify and judge the vertical shaft of the coal mine based on the joint data map and the threshold for the gap between the shaft and the cannula, so as to obtain early warning of potential hazards in the vertical shaft of the coal mine at different levels.

[0030] Specifically, the cornering signal is acquired using a can-ear ranging device, and the original underground images of the coal mine shaft are acquired using image acquisition equipment. These original underground images are then filtered, enhanced, and anomaly removed to obtain the processed underground images, including:

[0031] The corner signal is collected by the cage ear measuring device. The area array camera is installed on the top of the cage in the vertical shaft of the coal mine. The original underground image is collected by the area array camera combined with low frequency flashing supplementary light. The cage ear measuring device and the area array camera share the same lithium battery pack and single-chip microcomputer electronic control charging circuit.

[0032] The microcontroller in the microcontroller-controlled charging circuit uses a 30mA pre-charge branch to pre-activate the sleep BMS. After 60 seconds of pre-charging, the pre-charge branch is disconnected. After a 10ms anti-shake delay, it switches to the wireless charging main circuit. The microcontroller is equipped with dual-branch conduction interlock protection logic to prevent the pre-charge branch and the wireless charging main circuit from being powered on at the same time, which would cause a short circuit.

[0033] The original downhole image was filtered using a combination of Gaussian filtering and median filtering.

[0034] Image enhancement of raw downhole images is performed using the CLAHE algorithm;

[0035] A preset image filtering threshold is set, and the image clarity of the original downhole image is compared with the image filtering threshold. The original downhole images with image clarity less than or equal to the image filtering threshold are removed.

[0036] After filtering, image enhancement, and abnormal image removal, the downhole processed image is obtained.

[0037] Specifically, manually annotated crack images are acquired, and a deep learning model is trained based on these images to obtain a trained deep learning model. The processed images from underground are then input into the trained deep learning model for identification, thereby obtaining crack damage data for the coal mine vertical shaft, including:

[0038] Obtain at least 1500 manually labeled crack images, perform image augmentation on the manually labeled crack images to obtain at least 6000 manually labeled crack images, and divide the at least 6000 manually labeled crack images into training set and test set in a 4:1 ratio;

[0039] A deep learning model is built based on the YOLOv5s network, and the training set is used to train the deep learning model to obtain the trained deep learning model.

[0040] The processed downhole images are input into the trained deep learning model to obtain recognition results. Data of 2.5cm≤crack length≤50cm and / or 0.8mm≤crack width≤15mm and / or 0°≤crack orientation≤90° are extracted from the recognition results to obtain crack damage data.

[0041] Specifically, the corner signal is converted into a can ear gap value, including:

[0042] The downward displacement of the measuring wheel is calculated based on the rotation angle signal;

[0043] When the downward displacement is ≤0mm, the can ear gap value is 0; when 0mm < downward displacement <20mm, the can ear gap value is 20mm minus the downward displacement; when the downward displacement is ≥20mm, the downward displacement is discarded.

[0044] Specifically, a preset threshold for the gap between the shaft lugs is used. Based on the joint data map and the threshold for the gap between the shaft lugs, the coal mine vertical shaft is classified and judged to obtain early warnings of potential hazards in coal mine vertical shafts at different levels, including:

[0045] A preset threshold for the can ear gap is set. The data in the joint data map is compared with the threshold for the can ear gap. When only the can ear gap value is greater than the threshold for the can ear gap, a can ear and can passage wear warning is issued. When only crack damage data appears, a well wall damage warning is issued. When the can ear gap value is greater than the can ear gap threshold for the can ear gap and crack damage data appears, a high-risk support deformation warning is issued.

[0046] Thirdly, the present invention provides an inspection and monitoring system for coal mine vertical shafts, wherein the monitoring system is applied to the monitoring method of the second aspect, and the monitoring system includes:

[0047] The data acquisition and processing module is used to acquire corner signals through the can ear ranging device, acquire original underground images of the coal mine shaft using image acquisition equipment, and perform filtering, image enhancement, and abnormal image removal on the original underground images to obtain processed underground images.

[0048] The crack identification module is connected to the data acquisition and processing module. The crack identification module is used to acquire manually labeled crack images, train a deep learning model based on the manually labeled crack images to obtain a trained deep learning model, and input the processed underground images into the trained deep learning model for identification to obtain crack disease data of coal mine vertical shaft.

[0049] The joint data map construction module is connected to the crack identification module. The joint data map construction module is used to convert the corner signal into the gap value of the can ear, and bind the crack disease data of the same depth in the coal mine shaft with the gap value of the can ear to obtain the joint data map.

[0050] The safety early warning module is connected to the joint data map construction module. The safety early warning module is used to preset the gap threshold of the shaft ear. Based on the joint data map and the gap threshold of the shaft ear, the module classifies and judges the coal mine vertical shaft to obtain early warning of potential hazards in coal mine vertical shafts of different levels.

[0051] This application provides a can-ear ranging device, monitoring method, and system for coal mine vertical shafts. The method relies on a dedicated can-ear ranging device to collect can-path turning angle signals and convert them into can-ear gap values. Simultaneously, it acquires underground shaft images through an area array camera, and completes image preprocessing through filtering, enhancement, and impurity removal. It uses the augmented crack-annotated images to train a YOLOv5s model to accurately extract shaft crack disease data. It associates and binds can-ear gap data at the same depth with crack disease data coordinates to construct a joint data map. Through preset gap threshold classification, it outputs corresponding hidden danger warnings for three types of working conditions: can-path wear, shaft wall damage, and high-risk support deformation, realizing integrated and accurate monitoring and early warning of shaft can-path operation status and shaft wall diseases. Attached Figure Description

[0052] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0053] Figure 1 A schematic diagram of a can ear distance measuring device for a coal mine vertical shaft provided in this application;

[0054] Figure 2 A flowchart illustrating a method for inspection and monitoring of coal mine vertical shafts provided in this application;

[0055] Figure 3 This application provides a connection diagram for an inspection and monitoring system for coal mine vertical shafts.

[0056] Other illustrations

[0057] 10. Measuring wheel; 20. Swing arm transmission assembly; 201. Rotating arm; 202. Pin; 203. Drive shaft; 204. Elastic preload spring; 30. Base support and fixing assembly; 301. Base fixing plate; 302. Bearing seat; 303. Limiting ring; 304. Limiting bolt; 40. Angle sensor acquisition assembly; 401. Encoder; 402. Encoder mounting base. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, every other embodiment obtained by those skilled in the art without creative effort falls within the scope of protection of this application.

[0059] The terms "first," "second," "third," "fourth," etc. (if present) in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein.

[0060] In this invention, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0061] This application provides a can ear ranging device, monitoring method, and system for coal mine vertical shafts. The method integrates mechanical ranging and machine vision intelligent recognition technology. A dedicated can ear ranging device collects can guideway angle signals in real time and calculates can ear gap parameters. Simultaneously, a camera acquires shaft images, which are preprocessed by filtering, enhancement, and removal of inferior images. Then, a trained YOLOv5s model automatically identifies and extracts shaft crack information. Can ear gap data and crack data at the same depth are spatially correlated and fused to construct a joint data map. Based on a preset gap threshold, a graded assessment is conducted, and early warnings are output for three hidden danger scenarios: simple can guideway wear, simple shaft wall cracking, and support coupling deformation. This achieves integrated, high-precision intelligent inspection and monitoring of the vertical shaft structure and can guideway operation conditions.

[0062] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0063] Figure 1 This application provides a schematic diagram of the structure of a lug-type distance measuring device for a coal mine vertical shaft, as shown in the following figure. Figure 1As shown, this embodiment provides a can ear distance measuring device for a coal mine vertical shaft. The can ear distance measuring device includes: a distance measuring wheel 10, a swing arm transmission assembly 20, a base support and fixing assembly 30, and an angle sensing and acquisition assembly 40. The ranging wheel 10 is a flat, disc-shaped rotating component. It is used to roll in contact with the guide beam of the coal mine shaft and convert the gap changes in the guide beam into swing displacement. The swing arm transmission assembly 20 is a crank-arm shaped rotating component. The short end of the swing arm transmission assembly 20 is connected to the ranging wheel 10 by bolts. The swing arm transmission assembly 20 is used to keep the ranging wheel 10 tightly against the guide beam. The base support and fixing assembly 30 is a multi-component integrated assembly. The multi-module includes two annular modules, one rectangular module, and one cylindrical module. The long end of the base support and fixing assembly 30 is connected to the swing arm transmission assembly 20 by bolts. The base support and fixing assembly 30 is used to support the ranging device. The angle sensing and acquisition assembly 40 is a cylindrical component. The angle sensing and acquisition assembly 40 is connected to the base support and fixing assembly 30 and the swing arm transmission assembly 20 by bolts. The angle sensing and acquisition assembly 40 is used to acquire angle signals.

[0064] The entire device is fixed on the cage ear support. When no external force is applied, under the traction of the elastic pre-tension spring 204, the measuring wheel 10 extends outward by 20mm relative to the generatrix of the cage ear roller. When the cage moves up and down the shaft with the hoist, the measuring wheel 10 is in close contact with the outer wall of the cage guide beam and rolls purely without relative sliding. At this time, the encoder linear velocity is synchronized with the cage travel linear velocity, and the system determines it to be a valid measurement condition. When the gap between the cage ear and the cage guide changes, the measuring wheel is squeezed inward by the cage guide, causing the rotating arm 201 to swing and the transmission shaft 203 to rotate. The encoder 4... 01. The system collects corner data and calculates the downward displacement L of the measuring wheel through geometric conversion. Judgment rules: L≤0mm, the measuring wheel is fully compressed and the gap between the cage ear and the cage guide beam is 0; 0mm<L<20mm, the real-time gap of the cage ear = 20mm−L; when L≥20mm, the cage is mechanically limited by the 304 limiting bolt, and this data segment is invalid and not included in the gap calculation. The gap data collected by the measuring device is synchronously bound to the cage well depth coordinates and matched with the crack and defect data obtained by the subsequent well wall image acquisition equipment for joint well inspection and monitoring.

[0065] Specifically, the swing arm transmission assembly 20 includes: a rotating arm 201, a pin 202, a transmission shaft 203, and an elastic preload spring 204. The rotating arm 201 is a bent component; its long end is connected to the angle sensing and acquisition assembly 40, and its short end is connected to the measuring wheel 10 and the base support and fixing assembly 30. The rotating arm 201 connects the measuring wheel 10, the base support and fixing assembly 30, and the angle sensing and acquisition assembly 40. The pin 202 is a cylindrical component, passing between the short end of the rotating arm 201 and the measuring wheel 10. The pin 202 connects the measuring wheel 10 and the rotating arm 201. Shaft 203 is a cylindrical component. The drive shaft 203 is installed between the long end of the rotating arm 201 and the base support and fixing component 30. The shaft end of the drive shaft 203 is connected to the angle sensing and acquisition component 40. The drive shaft 203 is used to provide a support axis for the rotating arm 201. The elastic preload spring 204 is a helical spring. The two ends of the elastic preload spring 204 are respectively hooked on the rotating arm 201 and the base support and fixing component 30. The elastic preload spring 204 is used to continuously pull the rotating arm 201 to deflect to one side of the tank passage.

[0066] The swing arm transmission assembly 20 includes a rotating arm 201, a pin 202, a transmission shaft 203, and an elastic preload spring 204. The rotating arm 201 is a bent, irregularly shaped steel component with an overall crank arm structure. The short end of the rotating arm 201 has a pin mounting hole, and the long end has a bushing mounting hole. The long end of the rotating arm 201 connects to the transmission shaft 203, and the short end is fitted with the measuring wheel 10 via the pin 202. The rotating arm 201 is used to connect the measuring wheel 10, the transmission shaft 203, and the base support and fixing assembly 30 in series, converting the radial compression stroke of the measuring wheel 10 into the rotational swing angle of the transmission shaft 203. The pin 202 is a high-precision cylindrical fastener, passing between the short end of the rotating arm 201 and the center mounting hole of the measuring wheel 10. The pin 202 is used to enable the measuring wheel 10 to rotate freely and be hinged. The installation ensures that the measuring wheel 10 rolls smoothly along the tank guide beam without jamming. The drive shaft 203 is a heat-treated cylindrical rotating component that passes through the bearing inside the base support and fixing component 30. One end of the drive shaft 203 is interference-fitted to the rotating arm 201, and the other end is coaxially connected to the encoder shaft of the angle sensing and acquisition component 40. The drive shaft 203 is used to accurately transmit the deflection angle of the swing arm without any transmission backlash. The elastic pre-tension spring 204 is a high-elasticity spiral ring elastic component. Each end of the elastic pre-tension spring has a hanging ring. One end is hung on the reserved hanging hole of the rotating arm 201, and the other end is hung on the hanging protrusion of the base fixing plate. The elastic pre-tension spring 204 continuously pulls the rotating arm 201 to deflect towards the tank guide side by its own constant tension, ensuring that the measuring wheel 10 is continuously pressed against the tank guide beam under normal conditions.

[0067] Specifically, the base support and fixing assembly 30 includes: a base fixing plate 301, a bearing seat 302, a limiting ring 303, and a limiting bolt 304. The base fixing plate 301 is a rectangular flat plate with multiple pre-drilled mounting holes, providing an installation reference for the ranging device. The bearing seat 302 is a block-shaped seat with a through-hole at its center housing a bearing. The bearing seat 302 is mounted on the mounting holes of the base fixing plate 301 via the bearing and is also connected to the drive shaft 203 via the bearing. The bearing seat 302 constrains the radial movement of the drive shaft 203. The limiting ring 303 is an annular sleeve. The limiting ring 303 is installed on the mounting hole of the base fixing plate 301 by bolts. The limiting ring 303 is fitted between the drive shaft 203 and the bearing seat 302. The limiting ring 303 is used to limit the axial displacement of the drive shaft 203. The limiting bolt 304 is an external thread cylindrical fastener. The limiting bolt 304 is vertically fixed in the limiting groove of the base fixing plate 301 and the limiting ring 303. The limiting bolt 304 is used to limit the outward extension distance of the measuring wheel 10.

[0068] The base support and fixing assembly 30 includes a base fixing plate 301, a bearing seat 302, a limiting ring 303, and limiting bolts 304. The base fixing plate 301 is a thickened rectangular steel plate component. Multiple sets of assembly through holes, spring mounting bosses, and limiting grooves are pre-drilled on the surface of the base fixing plate 301. The base fixing plate 301 is locked onto the upper part of the can ear bracket by fastening bolts. The base fixing plate 301 serves as a unified installation reference component for the entire device. The bearing seat 302 is a square hollow support component. A through shaft hole is opened in the center of the bearing seat 302, and a wear-resistant rolling bearing is embedded in the shaft hole. The drive shaft 203 passes through the rolling bearing to achieve rotational assembly. The bearing seat 302 is bolted to the assembly holes of the base fixing plate 301. The bearing seat 302 is used to withstand the radial force generated by the operation of the drive shaft and the downhole impact load, limiting the transmission... The drive shaft 203 moves radially; the limiting ring 303 is an annular sleeve component (corresponding to the ring module). The limiting ring 303 is fitted on the outside of the drive shaft 203 and close to the end face of the bearing seat 302. The limiting ring 303 is positioned on the base fixing plate 301 by locking bolts. The limiting ring 303 is used to constrain the axial movement of the drive shaft 203 and prevent the shaft from sliding left and right; the limiting bolt 304 is a high-strength external thread columnar limiting component. The limiting bolt 304 is vertically tightened in the limiting groove of the base fixing plate 301. The top of the limiting bolt 304 faces the swing stroke path of the rotating arm 201. Under normal conditions, the bolt height limits the distance measuring wheel 10 to extend 20mm outward from the edge of the can ear roller in its natural state. When the distance measuring wheel is fully compressed, the rotating arm 201 abuts against the limiting bolt 304 to achieve mechanical limiting protection.

[0069] Specifically, the corner sensing and acquisition component 40 includes an encoder 401 and an encoder mount 402. The encoder 401 is a cylindrical component, coaxially connected to the drive shaft 203, and is used to acquire corner signals in real time. The encoder mount 402 is a U-shaped component, connected to the base mounting plate 301 by bolts, and is used to fix the position of the encoder 401.

[0070] The angle sensing and acquisition component 40 includes an encoder 401 and an encoder mounting base 402. The encoder 401 is a cylindrical high-precision pulse sensing component, which is an incremental rotary encoder. The encoder 401 shaft is rigidly connected to the drive shaft 203 on the same axis. The encoder 401 rotates synchronously with the drive shaft 203 and collects the angle pulse data in real time, converting the mechanical swing angle into an electrical signal and outputting it to the system main control. The encoder mounting base 402 is a U-shaped bent bracket component. The encoder mounting base 402 has fixing screw holes on both sides. The encoder mounting base 402 is locked to the mounting position on the rear side of the base fixing plate 301 by bolts. The encoder mounting base 402 covers the limiting encoder housing, isolates the displacement deviation caused by the vibration of the cage, and prevents the encoder housing from rotating synchronously with the shaft.

[0071] Figure 2 A flowchart illustrating a method for inspection and monitoring of coal mine vertical shafts provided in this application is shown below. Figure 2 As shown in this embodiment, a method for inspection and monitoring of coal mine vertical shafts is provided. The method includes: acquiring corner signals using the aforementioned ear-shaped distance measuring device; acquiring original underground images of the coal mine vertical shaft using an image acquisition device; filtering, enhancing, and removing abnormal images from the original underground images to obtain processed underground images; acquiring manually annotated crack images; training a deep learning model based on the manually annotated crack images to obtain a trained deep learning model; inputting the processed underground images into the trained deep learning model for identification to obtain crack damage data of the coal mine vertical shaft; converting the corner signals into ear-shaped gap values; and binding the crack damage data of the same depth in the coal mine vertical shaft with the ear-shaped gap values ​​through coordinate association to obtain a joint data map; setting an ear-shaped gap threshold; and classifying the coal mine vertical shaft based on the joint data map and the ear-shaped gap threshold to obtain early warnings of potential hazards in coal mine vertical shafts of different levels.

[0072] This application provides a method for inspection and monitoring of coal mine vertical shafts. The method relies on the aforementioned ear-mounted ranging device to collect transmission angle signals in real time, and simultaneously uses an image acquisition device with dual-mode supplementary lighting to acquire original images of the underground shaft wall. After filtering, noise reduction, texture enhancement, and removal of inferior images, compliant processed images are obtained. An improved deep learning network is trained based on batch manually labeled crack samples. The trained model is used to analyze and process the images to generate shaft wall crack disease data. The angle parameters output by the ranging device are then converted into measured gap values ​​of the ear-mounted components. Crack disease data corresponding to the same shaft depth are bound with ear-mounted component gap data coordinates to generate an integrated joint data map. Finally, the map information is analyzed in conjunction with a pre-set gap safety threshold to classify and determine the type of shaft hazard and output corresponding safety warnings.

[0073] This method integrates the detection of mechanical clearances in the cage ear with the visual detection of structural defects in the well wall. Relying on hardware devices, it achieves dynamic and automated data acquisition throughout the entire process along with the cage, eliminating the need for well shutdown and on-site inspection by personnel. It uses deep learning to accurately capture minute cracks in the well wall, and the correlation between the two types of monitoring data and well depth can accurately distinguish between wear of individual components and high-risk structural hazards in the well wall. All data is automatically archived in the cloud, resulting in high detection accuracy and low false negative rate. It is suitable for harsh working conditions such as dust and low light in underground mines, effectively reducing labor costs and safety risks in underground operations, and meeting the needs of automated operation and maintenance construction in smart mines.

[0074] Specifically, the aforementioned cage-ear ranging device is used to acquire corner signals, and an image acquisition device is used to acquire original underground images of the coal mine shaft. The original underground images are then filtered, enhanced, and anomaly removed to obtain a processed image. This process includes: acquiring corner signals using the cage-ear ranging device; installing a surface-array camera on the top of the cage in the coal mine shaft; and acquiring original underground images using the surface-array camera combined with low-frequency flash lighting. The cage-ear ranging device and the surface-array camera share a lithium battery pack and a microcontroller-controlled charging circuit. The microcontroller in the microcontroller-controlled charging circuit uses a 30mA pre-charge branch to pre-activate the dormant BMS. After 60 seconds of charging, the pre-charging branch is disconnected. After a 10ms anti-shake delay, the system switches to the wireless charging main circuit. The microcontroller is equipped with dual-branch conduction interlock protection logic to prevent short circuits caused by simultaneous power-on of the pre-charging branch and the wireless charging main circuit. Gaussian filtering combined with median filtering is used to filter the original downhole image. The CLAHE algorithm is used to enhance the image of the original downhole image. A preset image filtering threshold is set, and the image clarity of the original downhole image is compared with the image filtering threshold. The original downhole images with image clarity less than or equal to the image filtering threshold are removed. After filtering, image enhancement, and abnormal image removal, the processed downhole image is obtained.

[0075] This application provides a method for inspection and monitoring of coal mine vertical shafts. The method collects corner signals through a cage ear ranging device, while simultaneously installing an area array camera on the top of the cage and using LED low-frequency flashing supplementary lighting to collect original underground images. The cage ear ranging device and the area array camera share a matching lithium battery and a single-chip microcomputer-controlled charging circuit. The single-chip microcomputer first controls a 30mA small current pre-charging branch to activate the dormant BMS for 60s, disconnects the pre-charging branch, and switches to the wireless charging main circuit after a 10ms anti-shake delay. At the same time, the branch interlocking logic prevents the two power supplies from conducting synchronously and short-circuiting. After the images are collected, Gaussian filtering combined with median filtering is used for noise reduction, and the CLAHE algorithm is used to enhance crack texture. Then, unqualified images are filtered out using a preset clarity threshold. Finally, the image processing is completed to obtain the processed underground image.

[0076] The base fixing plate 301 of the entire range measuring device is fastened to the upper part of the can ear bracket. The range measuring wheel 10 is pre-extended 20mm outward relative to the outer edge of the can ear roller under no load. The swing arm assembly is continuously tensioned by the elastic pre-tension spring 204 to ensure that the range measuring wheel 10 presses against the can track beam in a normal state. The encoder 401 is locked and fixed by the encoder fixing seat 402, and the encoder shaft is rigidly connected to the transmission shaft 203 on the same axis.

[0077] The area scan camera is rigidly fixed on the mounting base at the top of the cage. LED supplementary lights are arranged in a ring around the area scan camera to form a low-frequency supplementary light module. The supplementary light source and the area scan camera are connected to a common lithium battery power supply terminal after wiring, so that the rangefinder encoder, area scan camera and supplementary light LED share the same 3.7V explosion-proof lithium battery pack.

[0078] The main control uses an STC89C52 microcontroller. The microcontroller's P1.0 pin controls the 30mA low-current pre-charge branch, and the P1.1 pin controls the 5V / 2A wireless charging main circuit. The pre-charge branch is composed of an N-channel MOSFET connected in series with a 100Ω current-limiting resistor. The wireless charging main circuit is connected to a standard 5V wireless charging module. The two branches are connected in parallel and then connected to a common lithium battery.

[0079] The logic for power supply timing control of the equipment is as follows:

[0080] Upon system power-on initialization, the microcontroller's P1.0 and P1.1 pins are at a low level by default, and both the pre-charge and main charge circuits are shut down, entering standby protection mode.

[0081] When the microcontroller pulls P1.0 high, the MOSFET turns on, the pre-charge branch starts, and a constant 30mA micro-current is used to pre-charge the depleted dormant lithium battery, waking up the dormant BMS protection board; at the same time, the built-in hardware timer starts a 60-second countdown.

[0082] When the timer reaches 60 seconds, the microcontroller pulls P1.0 low to cut off the pre-charge branch and waits for a 10ms anti-jitter delay to eliminate circuit voltage fluctuations; then it pulls P1.1 high to conduct the high-power wireless charging main circuit and switches to conventional constant current and constant voltage charging.

[0083] The microcontroller software determines the I / O port level in real time and sets interlock logic. If P1.0 and P1.1 are both high, both channels are immediately forced to shut down and lock out to prevent parallel short circuits in the dual circuits.

[0084] The electrical parameters are as follows: pre-charge current = 30mA, current limiting resistor = 100Ω; main charge specification 5V / 2A; pre-charge time = 60s, anti-shake interval = 10ms; lithium battery nominal voltage 3.7V, BMS sleep cutoff voltage < 3.0V, main charge start-up access voltage ≥ 3.2V.

[0085] During the lifting and moving of the cage, the measuring wheel 10 rolls purely without slippage against the cage guide beam. Changes in the gap between the cage lugs and the guide beam cause the measuring wheel to retract, driving the rotating arm 201 and drive shaft 203 to deflect. The encoder 401 collects the drive shaft rotation angle pulse signal in real time as the raw rotation angle data. The gap conversion calculation formula is:

[0086]

[0087] in, This indicates the actual gap between the can ear and the can guide beam, in mm; This indicates the reference distance of the measuring wheel's natural outward extension, in mm; This indicates the displacement of the measuring wheel under pressure, in mm; it is determined as... ≤0, =0; 0 < Substitute <20 into the formula; Data with a value of ≥20 is invalid.

[0088] After the area array camera acquires the original images of the well at a frame rate of 10fps and a low-frequency LED strobe illumination of 50Hz, it performs three standardized data processing steps in sequence: joint filtering and noise reduction, CLAHE image enhancement, and sharpness screening and elimination.

[0089] Gaussian filtering is used to perform pixel-by-pixel weighted smoothing of the entire image, eliminating large areas of cloud-like noise caused by underground water mist and widespread coal dust dispersion, while preserving the general outline of the well wall edges. The calculation formula is as follows:

[0090]

[0091] in, This represents the image after Gaussian filtering. Represents the pixel coordinates within the filter window; The standard deviation is a Gaussian kernel, with a fixed value used on-site. ;

[0092] Median filtering uses a 3×3 fixed-size sliding pixel window. All gray values ​​within the window are sorted from smallest to largest, and the median value is used to replace the gray value of the center pixel of the window. All pixels in the entire image are traversed to remove isolated noise points such as splashed coal dust and sporadic water stains. Gaussian smoothing is performed first, followed by median noise removal. The two-stage filtering is performed in sequence to obtain the first denoised image.

[0093] The CLAHE adaptive contrast image enhancement process involves uniformly dividing the denoised image into non-overlapping sub-image blocks of consistent size, with uniform pixel specifications for each block; performing histogram equalization on each sub-block separately, and simultaneously setting a contrast amplification upper limit to limit excessive contrast enhancement in a single image block; and using bilinear interpolation to stitch together the enhanced sub-blocks to restore the entire image.

[0094] By using CLAHE adaptive contrast image enhancement processing, the problem of excessive whitening of strong light spots and blackening of dark areas in downhole is suppressed. The grayscale difference between cracks and well wall matrix is ​​stretched to highlight the texture of fine cracks and solve the problem of details being buried due to uneven downhole lighting.

[0095] The sharpness S of a single image is calculated using the Laplacian variance method, with the variance of the second-order difference between pixel gray levels used as a quantification of sharpness; a fixed filtering threshold is preset. ;when If the image is determined to be blurry or lacks detail, the original image is directly discarded. The image is acceptable and will be kept for future use.

[0096] This method uses a unified power supply architecture to ensure uninterrupted and stable operation of ranging and camera equipment. The pre-charge control scheme solves the problem of batteries failing to start when they are depleted in the well. Multi-level image processing effectively overcomes the imaging blurring caused by the dim lighting and dust in the well. The supplementary lighting is reasonably selected and the image screening rules are clear, which greatly improves the image quality of the well wall and provides a reliable image data source for subsequent accurate crack identification.

[0097] Specifically, the process involves acquiring manually labeled crack images, training a deep learning model based on these images to obtain a trained deep learning model, and inputting the processed underground images into the trained deep learning model for recognition to obtain crack damage data for coal mine vertical shafts. This includes: acquiring at least 1500 manually labeled crack images; performing image augmentation on these images to obtain at least 6000 manually labeled crack images; dividing these 6000 images into a training set and a test set at a 4:1 ratio; constructing a deep learning model based on a YOLOv5s network; training the deep learning model using the training set to obtain a trained deep learning model; inputting the processed underground images into the trained deep learning model to obtain recognition results; and extracting data from the recognition results where 2.5cm ≤ crack length ≤ 50cm and / or 0.8mm ≤ crack width ≤ 15mm and / or 0° ≤ crack orientation ≤ 90° to obtain crack damage data.

[0098] This application provides a method for inspection and monitoring of coal mine vertical shafts. The method first collects no less than 1,500 manually annotated images of shaft wall cracks, expands them to 6,000 images through image augmentation, and then splits them into training and testing sets at a ratio of 4:1. A deep learning model with multi-spatial perspective fusion and ternary coordinate attention module is built based on the YOLOv5s CSPDarknet53 backbone network. The model is iteratively trained using the divided training set. Then, the pre-processed underground images are fed into the trained model for target detection. Crack parameter information with a length of 2.5cm to 50cm, a width of 0.8mm to 15mm, and a direction of 0° to 90° is selected from the output recognition results. Finally, the data on shaft crack defects are summarized to generate data.

[0099] Original sample collection: Real crack images of the coal mine shaft wall were collected through on-site cage image acquisition equipment and manual re-shooting. The number of original labeled samples was no less than 1,500. All samples were manually labeled by professional technicians using rectangular annotation boxes to complete the crack location, type, and boundary. The annotation information included crack pixel coordinates, length, width, and direction angle, forming a standard original crack dataset.

[0100] Image augmentation: Multi-dimensional data augmentation processing was performed on 1,500 original manually labeled images. Augmentation methods included random rotation, horizontal flipping, vertical flipping, brightness adjustment, contrast adjustment, Gaussian noise superposition, and scale scaling. Through combined augmentation, the dataset size was expanded to no less than 6,000 images, effectively improving the model's generalization ability under different lighting, angles, and dust conditions, and avoiding model overfitting.

[0101] Dataset partitioning: The 6,000 augmented and labeled images are randomly split in a fixed ratio of 4:1. 80% of the images are used as the training set for iterative updates of the deep learning model weights, and 20% are used as the test set for accuracy verification and performance evaluation during model training. This ensures that the training and test sets have no overlap or duplication and are distributed in a consistent manner.

[0102] A crack detection model is built based on YOLOv5s. The backbone feature extraction network is CSPDarknet53, which includes Focus slice layer, CBS convolution module, CSP module and SPPF spatial pyramid pooling structure. The SPPF kernel size is 5, 9 and 13. The neck adopts PANet multi-scale feature fusion structure, which fuses feature maps of three scales: 80×80, 40×40 and 20×20. The model output head has three preset anchor box sizes: small target (10,13), (16,30) and (33,23), medium target (30,61), (62,45) and (59,119), and large target (116,90), (156,198) and (373,326). The overall input size of the model is 640×640 pixels, the output class is 1, the number of parameters is 7.02M, and the inference time per frame does not exceed 20ms.

[0103] The model was trained using the PyTorch 1.12.1 framework and CUDA 11.3 environment, employing the SGD optimizer with an initial learning rate of 0.01. The learning rate was decayed to a minimum of 0.0001 using cosine annealing, with a momentum factor of 0.937, a weight decay coefficient of 0.0005, a batch size of 16, a maximum training epoch of 300, and a data warm-up epoch of 3.0. Multi-scale training was enabled, with a scale range of 320 to 640, and the loss function was CIoULoss. Training was stopped when the accuracy on the test set showed no improvement for 50 consecutive epochs. The final model achieved an mAP@0.5 of 96.2%, precision of 95.8%, recall of 94.9%, and a false negative rate of no more than 1.2%.

[0104] The downhole processed image was scaled to 640×640 pixels and input into the trained model. The confidence threshold was set to 0.25 and the NMS non-maximum suppression threshold to 0.45. The system automatically extracted fracture parameters, retaining only valid data with fracture lengths of 2.5cm to 50cm, widths of 0.8mm to 15mm, and strike angles of 0° to 90°. Data outside the range was automatically filtered out. Finally, standardized fracture defect data containing wellbore depth, fracture length, width, strike, confidence level, and coordinate location were generated.

[0105] This method addresses the problem of insufficient downhole fracture samples by expanding the sample size. The improved deep learning network can enhance the ability to extract features of small fractures, effectively reduce the probability of missed or false detections of fractures in dusty and dark environments, accurately quantify key parameters such as the length, width, and orientation of fractures, and realize automated and refined statistics of wellbore fractures. This eliminates the inefficiency and error problems of manual measurement and ensures accurate and reliable wellbore defect detection data.

[0106] Specifically, the corner signal is converted into a can ear gap value, including: calculating the downward displacement of the measuring wheel 10 based on the corner signal; when the downward displacement is ≤0mm, the can ear gap value is 0; when 0mm < downward displacement <20mm, the can ear gap value is 20mm minus the downward displacement; when the downward displacement is ≥20mm, the downward displacement is discarded.

[0107] This application provides a method for inspection and monitoring of coal mine vertical shafts. The method relies on the encoder rotation angle signal output by the can ear ranging device to calculate the downward displacement of the ranging wheel, and calculates the can ear gap according to a three-segment judgment rule: when the downward displacement is less than or equal to 0mm, the gap is directly determined to be 0; when the downward displacement is in the range of 0mm to 20mm, the actual gap is obtained by subtracting the downward displacement from 20mm; when the downward displacement is greater than or equal to 20mm, the mechanical limit is triggered and the data collected in this group is discarded, and finally the accurate can ear gap value is obtained.

[0108] During the lifting and lowering operation of the cage, the encoder 401 inside the cage ear ranging device collects the rotation angle pulse signal generated by the rotation of the transmission shaft 203 in real time and transmits it to the system's main control microcontroller. The microcontroller reads the number of pulses and, combined with the mechanical geometry of the swing arm, calculates the actual downward displacement L of the ranging wheel 10, in mm. The calculation relies on the fixed mechanical transmission ratio parameters of the device, and the linear retraction distance corresponds to the swing angle, and the calculation is performed in real time throughout the entire process.

[0109] The system calculates the can ear gap Δ according to three criteria: L≤0mm: The measuring wheel is fully compacted with no shrinkage margin, and the can ear gap Δ=0mm is directly assigned; 0mm<L<20mm: The calculation formula Δ=20−L is used, where 20mm is the preset outward reference dimension of the measuring wheel under no-load conditions, and the real-time can ear gap value is calculated; L≥20mm: The measurement exceeds the effective range, and the current displacement data is directly discarded, and no corresponding gap value is generated.

[0110] As the cage rises and falls, the hoisting system's position sensors synchronously collect real-time data on the shaft depth. Each set of calculated cage ear gap data is then bound to the currently collected shaft depth coordinates, forming a single set of correlated data between shaft depth and cage ear gap.

[0111] The fracture data obtained through deep learning model identification and screening in the early stage comes with the wellbore depth coordinates corresponding to the image acquisition time. The data of a single fracture includes the corresponding well depth, fracture length, fracture width, and fracture direction parameters.

[0112] The system uses wellbore depth as a unified index keyword to retrieve data on the gap between the ear canals and data on fracture defects with the same depth coordinates, and binds the two types of monitoring data from different sources to each other. If there are no fractures at the same well depth, only the data on the ear canals is retained; if there is no valid gap data, the information on fracture defects is saved separately.

[0113] All coordinate-bound data for the entire wellbore are aggregated and arranged from top to bottom according to the wellbore depth. An integrated joint data map is constructed with well depth as the vertical scale and gaps and cracks in the wellbore as the horizontal content. The map is stored locally and simultaneously uploaded to the cloud database to provide data support for subsequent graded hazard early warning.

[0114] This method has a simple and efficient conversion logic, low computational load, and is easy for microcontrollers to perform real-time calculations. It achieves distance measurement without additional sensors by relying on a 20mm preset extension amount, avoids invalid data beyond the range by relying on mechanical limits, and is suitable for harsh working conditions such as downhole vibration and dust. It can dynamically and continuously output quantitative gap data throughout the entire stroke.

[0115] Specifically, a preset threshold for the gap between the shaft ear and the shaft is used to classify and judge the vertical shaft of the coal mine based on the joint data map and the threshold for the gap between the shaft ear, so as to obtain early warning of hidden dangers of different levels of vertical shaft of coal mine. This includes: setting a preset threshold for the gap between the shaft ear and comparing the data in the joint data map with the threshold for the gap between the shaft ear and the shaft ear; issuing a warning of shaft ear wear when only the gap between the shaft ear and the shaft ear is greater than the threshold for the gap between the shaft ear and the shaft ear; issuing a warning of shaft wall damage when only the data of crack defects is present; and issuing a warning of high-risk support deformation when the gap between the shaft ear and the shaft ear is greater than the threshold for the gap between the shaft ear and the data of crack defects are present.

[0116] This application provides a method for inspection and monitoring of coal mine vertical shafts. The method pre-sets a safety threshold for the gap between the can ear and the shaft, and conducts graded analysis based on joint data maps. If the gap between the can ear exceeds the threshold, a wear warning for the can ear and the shaft passage is pushed. If only the data of cracks in the shaft wall are detected, a shaft wall damage warning is issued. When both the gap exceeds the standard and cracks exist at the same shaft depth, the shaft support is determined to be damaged and a high-risk support deformation warning is issued.

[0117] Based on the safety operation and maintenance specifications for coal mine vertical shaft guideways and the allowable range of normal wear of guideways, this method uniformly presets a safety threshold of 5mm for guideway clearance. This threshold is the industrial field-calibrated safety critical value: under normal operating conditions, the dynamic clearance between the guideway and the guideway beam is stable within the range of 0-5mm, which is within the normal floating wear range of the equipment; when the measured guideway clearance value is continuously greater than 5mm, it is determined that the clearance between the guideway and the guideway exceeds the standard, and there is a risk of abnormal wear. This is used as the core numerical benchmark for the judgment of hidden dangers throughout the entire process.

[0118] The system automatically retrieves the joint data map of well depth, ear gap, and crack defects that was previously bound to the depth coordinates. It then traverses and reads two types of core data at each well depth point: first, the real-time measured value of the ear gap at the corresponding well depth location; and second, the well wall crack defect data at the corresponding well depth location (including effective defect parameters such as crack length, width, and orientation). This provides complete data support for the point-by-point classification of potential hazards.

[0119] The system uses a 5mm can ear gap threshold as the judgment benchmark, and combines two types of monitoring data at a single well depth to classify three levels of hidden danger warnings. The judgment logic fully covers all working conditions.

[0120] Level 1 Warning (Ordinary Equipment Wear Warning): When the gap value of the can ear is >5mm at the same well depth and there is no valid well wall crack data, it is judged as an abnormal wear fault of the can ear and can passage, triggering the can ear and can passage wear warning. The hidden danger is only the mechanical wear of the hoisting equipment, and there is no structural risk to the well.

[0121] Level 2 Warning (Well Wall Structure Damage Warning): When the gap value of the can ear is ≤5mm (normal gap) at the same well depth, but effective well wall crack data is detected, it is determined that there is local damage to the well wall concrete and support layer, triggering the well wall damage warning. The hidden danger is local structural damage to the well, which does not affect the safe operation of the hoisting equipment at present.

[0122] Level 3 warning (high-risk support deformation warning): When the gap value of the can ear is >5mm at the same depth of the well shaft, and there is also valid data on well wall cracks, it is determined that the well shaft at this location is subjected to ground pressure and support deformation, which causes can ear displacement, excessive can ear gap, and well wall cracking. This is a major structural hazard of the well shaft and triggers a high-risk support deformation warning.

[0123] The system matches corresponding automated response mechanisms and manual operation and maintenance solutions to different levels of early warning, achieving closed-loop management of early warning and response:

[0124] Level 1 wear warning response for can ear and can passage: The system automatically saves the gap change curve at the well depth, and a light-colored pop-up window in the cloud reminds you of the wear point and wear degree. During the daily operation and maintenance phase, the point is marked as a key inspection location. No downtime is required. During the monthly routine maintenance, wear verification, gap adjustment or parts replacement of can ear and can passage are carried out. The gap change trend is continuously tracked to prevent wear from aggravating.

[0125] Secondary wellbore damage early warning response: The system highlights the location of the damaged well depth, automatically takes screenshots to save the corresponding crack images and quantitative parameters, and pushes early warning information to the mine operation and maintenance terminal; operation and maintenance personnel arrange special wellbore repair operations according to the damage level, grouting and plastering to reinforce shallow cracks, and regularly re-measure the crack expansion status to prevent local damage from continuing to expand.

[0126] Level 3 High-Risk Support Deformation Early Warning Response: The system triggers an audible and visual high-risk alarm, and a red pop-up window forcibly pushes the early warning information, simultaneously recording all data at the location (gap data, crack data, well depth coordinates); the control platform is immediately notified to pay attention to the wellbore area, and depending on the severity of the hidden danger, measures such as time-limited speed reduction and temporary shutdown for inspection can be taken. Professional teams are urgently organized to carry out wellbore support inspection, reinforcement and repair, investigate potential ground pressure deformation hazards, and completely eliminate structural safety risks to the wellbore.

[0127] All early warning records, corresponding well depth data, disease parameters, and treatment records are uploaded to the cloud database in real time for permanent archiving. The system can automatically generate reports and trend curves of wellbore hidden dangers, enabling historical data tracing and tracking of hidden danger recurrence, and providing accurate data support for the periodic operation and maintenance, structural transformation, and equipment replacement of mine shafts.

[0128] This method relies on the linkage and hierarchical early warning of two types of monitoring data to distinguish between single component failures and major structural hazards in the wellbore. The early warning levels are clearly defined, which can accurately locate the fault type and risk level, effectively reduce false alarms, and facilitate targeted maintenance by operation and maintenance personnel.

[0129] Figure 3 A connection diagram for an inspection and monitoring system for coal mine vertical shafts is provided in this application, as shown in the figure. Figure 3As shown in this embodiment, a monitoring system for coal mine vertical shafts is provided. The system includes: a data acquisition and processing module, which uses the aforementioned ear-type ranging device to acquire angle signals and an image acquisition device to acquire original underground images of the coal mine vertical shaft. The original underground images are then filtered, enhanced, and anomaly removed to obtain a processed underground image. A crack identification module, connected to the data acquisition and processing module, is used to acquire manually annotated crack images, train a deep learning model based on the manually annotated crack images to obtain a trained deep learning model, and output the processed underground image. The data is fed into a deep learning model after training to identify cracks in the coal mine vertical shaft. A joint data map construction module, connected to the crack identification module, is used to convert the corner signal into the ear gap value and bind the crack data of the same depth in the coal mine vertical shaft with the ear gap value to obtain the joint data map. A safety early warning module, connected to the joint data map construction module, is used to preset the ear gap threshold. Based on the joint data map and the ear gap threshold, the coal mine vertical shaft is classified and judged to obtain early warning of hidden dangers in coal mine vertical shafts of different levels.

[0130] This application provides an inspection and monitoring system for coal mine vertical shafts. The system consists of a data acquisition and processing module, a crack identification module, a joint data map construction module, and a safety early warning module connected in sequence. The data acquisition and processing module collects corner signals using a shaft ear ranging device and collects raw underground images using a matching image acquisition device. It then performs filtering, image enhancement, and removal of inferior images to obtain processed images. The crack identification module trains a deep learning model based on labeled samples and detects and outputs crack defect data from the processed images. The joint data map construction module converts the corner signals to obtain shaft ear gap values ​​and binds the two types of data according to the shaft depth coordinates to generate a joint data map. The safety early warning module performs hierarchical analysis based on a preset gap threshold and the map, and outputs various types of shaft hazard warnings.

[0131] The system adopts a modular serial architecture to achieve synchronous collection and integrated correlation analysis of mechanical wear data of the tank ear and structural defects of cracks in the well wall. It breaks the limitation that the two types of detection data are independent and cannot be analyzed collaboratively. It achieves automatic quantitative identification of cracks by relying on deep learning, and accurately distinguishes three types of hidden dangers: equipment wear, well wall damage, and support deformation by relying on threshold grading mechanism. It has a high degree of automation and requires less manual intervention, which makes it convenient for mines to carry out maintenance according to the warning level. It effectively improves the efficiency of safety inspection of vertical shafts and the accuracy of hidden danger investigation.

Claims

1. A can-ear distance measuring device for coal mine vertical shafts, characterized in that, The ranging device includes: The measuring wheel (10) is a flat disc-shaped rotating component. The measuring wheel (10) is used to roll in contact with the guide beam of the coal mine shaft and convert the gap change between the guide ear of the coal mine shaft and the guide beam into swing displacement. The swing arm transmission assembly (20) is a crank arm irregular rotating component. The short end of the swing arm transmission assembly (20) is connected to the measuring wheel (10) by bolts. The swing arm transmission assembly (20) is used to make the measuring wheel (10) fit tightly against the tank track beam. The base support and fixing assembly (30) is a multi-component integrated assembly. The base support and fixing assembly (30) is connected to the long end of the swing arm transmission assembly (20) by bolts. The base support and fixing assembly (30) is used to support the ranging device. Angle sensing acquisition component (40) is connected to the base support fixing component (30) and the swing arm transmission component (20) respectively by bolts. The angle sensing acquisition component (40) is used to acquire angle signals.

2. The lug-type distance measuring device for a coal mine vertical shaft according to claim 1, characterized in that, The swing arm transmission assembly (20) includes: Rotating arm (201), the rotating arm (201) is a bent component, the long end of the rotating arm (201) is connected to the corner sensing and acquisition component (40), and the short end of the rotating arm (201) is connected to the measuring wheel (10) and the base support and fixing component (30). The rotating arm (201) is used to connect the measuring wheel (10), the base support and fixing component (30) and the corner sensing and acquisition component (40). Pin (202), the pin (202) is a cylindrical component, the pin (202) is inserted between the short end of the rotating arm (201) and the measuring wheel (10), the pin (202) is used to connect the measuring wheel (10) and the rotating arm (201). A drive shaft (203) is a cylindrical component. The drive shaft (203) is installed between the long end of the rotating arm (201) and the base support and fixing assembly (30). The shaft end of the drive shaft (203) is connected to the angle sensing and acquisition assembly (40). The drive shaft (203) is used to provide a support axis for the rotating arm (201). The elastic preload spring (204) is a helical spring. The two ends of the elastic preload spring (204) are respectively attached to the rotating arm (201) and the base support fixing assembly (30). The elastic preload spring (204) is used to continuously pull the rotating arm (201) to deflect to one side of the tank passage.

3. A lug-type distance measuring device for a coal mine vertical shaft according to claim 2, characterized in that, The base support fixing assembly (30) includes: The base fixing plate (301) is a rectangular flat plate with multiple pre-drilled mounting holes. The base fixing plate (301) is used to provide an installation reference for the ranging device. The bearing housing (302) is a block-shaped housing. The bearing housing (302) has a through shaft hole in the center and a bearing is built in it. The bearing housing (302) is mounted on the mounting hole of the base fixing plate (301) through the bearing. The bearing housing (302) is also connected to the transmission shaft (203) through the bearing. The bearing housing (302) is used to constrain the radial movement of the transmission shaft (203). A limiting ring (303) is an annular sleeve-shaped ring. The limiting ring (303) is installed on the mounting hole of the base fixing plate (301) by bolts. The limiting ring (303) is fitted between the transmission shaft (203) and the bearing seat (302). The limiting ring (303) is used to limit the axial displacement of the transmission shaft (203). The limiting bolt (304) is an external thread cylindrical fastener. The limiting bolt (304) is vertically fixed in the limiting groove of the base fixing plate (301) and the limiting ring (303). The limiting bolt (304) is used to limit the outward extension distance of the measuring wheel (10).

4. A lug-type distance measuring device for a coal mine vertical shaft according to claim 3, characterized in that, The corner sensing acquisition component (40) includes: The encoder (401) is a cylindrical component and is coaxially connected to the transmission shaft (203). The encoder (401) is used to collect the rotation angle signal in real time. The encoder mounting base (402) is a U-shaped component. The encoder mounting base (402) is connected to the base mounting plate (301) by bolts. The encoder mounting base (402) is used to fix the position of the encoder (401).

5. A method for inspection and monitoring of coal mine vertical shafts, the method providing an image acquisition device and installing the image acquisition device on the top of a cage, characterized in that, The monitoring method employs the ear-measuring device according to any one of claims 1-4, wherein the ear-measuring device is installed on the ear-support on the outside of the cage, and the monitoring method includes: The turning angle signal is collected by the can ear ranging device, and the original underground image of the coal mine shaft is collected by the image acquisition device. The original underground image is filtered, enhanced and abnormal image is removed to obtain the processed underground image. Manually annotated crack images are obtained, and a deep learning model is trained based on the manually annotated crack images to obtain a trained deep learning model. The processed underground images are then input into the trained deep learning model for recognition to obtain crack disease data of the coal mine vertical shaft. The angle signal is converted into the gap value of the can ear, and the crack disease data of the same depth in the coal mine shaft are bound to the gap value of the can ear with coordinate correlation to obtain a joint data map. A preset gap threshold for the shaft is used. Based on the joint data map and the gap threshold, the coal mine vertical shaft is classified and judged to obtain early warning of potential hazards of the coal mine vertical shaft at different levels.

6. The method for inspection and monitoring of coal mine vertical shafts according to claim 5, characterized in that, The process involves acquiring angle signals via the can-ear ranging device, acquiring original underground images of the coal mine shaft using image acquisition equipment, and performing filtering, image enhancement, and anomaly removal on the original underground images to obtain processed underground images. This includes: The corner signal is collected using the cage ear ranging device. An area array camera is installed on the top of the cage in the vertical shaft of the coal mine. The original underground image is collected using the area array camera combined with low-frequency flashing supplementary lighting. The cage ear ranging device and the area array camera share a lithium battery pack and a single-chip microcomputer electronic control charging circuit. The microcontroller in the microcontroller-controlled charging circuit uses a 30mA pre-charge branch to pre-activate the sleep BMS. After 60 seconds of pre-charging, the pre-charge branch is disconnected. After a 10ms anti-shake delay, it switches to the wireless charging main circuit. The microcontroller is equipped with dual-branch conduction interlock protection logic to prevent the pre-charge branch and the wireless charging main circuit from being powered on at the same time, which would cause a short circuit. The original downhole image is filtered using a combination of Gaussian filtering and median filtering. The CLAHE algorithm is used to enhance the original downhole image. A preset image filtering threshold is set, and the image clarity of the original downhole image is compared with the image filtering threshold. The original downhole images whose image clarity is less than or equal to the image filtering threshold are removed. After the filtering, image enhancement, and abnormal image removal processes, the downhole processed image is obtained.

7. The method for inspection and monitoring of coal mine vertical shafts according to claim 5, characterized in that, The process of acquiring manually annotated crack images, training a deep learning model based on these images to obtain a trained deep learning model, and inputting the processed underground images into the trained deep learning model for identification to obtain crack damage data of the coal mine vertical shaft includes: At least 1,500 manually labeled crack images are obtained, and image augmentation is performed on the manually labeled crack images to obtain at least 6,000 manually labeled crack images. The at least 6,000 manually labeled crack images are divided into a training set and a test set in a 4:1 ratio. The deep learning model is constructed based on the YOLOv5s network, and the deep learning model is trained using the training set to obtain the trained deep learning model. The downhole processed image is input into the trained deep learning model to obtain the recognition result. Data of 2.5cm≤crack length≤50cm and / or 0.8mm≤crack width≤15mm and / or 0°≤crack orientation≤90° are extracted from the recognition result to obtain the crack disease data.

8. A method for inspection and monitoring of coal mine vertical shafts according to claim 5, characterized in that, The step of converting the angle signal into a can ear gap value includes: The downward displacement of the measuring wheel (10) is calculated based on the rotation angle signal; When the downward displacement is ≤0mm, the can ear gap value is 0; when 0mm < the downward displacement < 20mm, the can ear gap value is 20mm minus the downward displacement; when the downward displacement is ≥20mm, the downward displacement is discarded.

9. A method for inspection and monitoring of coal mine vertical shafts according to claim 5, characterized in that, The preset ear gap threshold, based on the joint data map and the ear gap threshold, is used to classify and judge the coal mine vertical shaft to obtain early warning of potential hazards in the coal mine vertical shaft at different levels, including: A preset ear gap threshold is established. The data in the joint data map is compared with the ear gap threshold. When only the ear gap value is greater than the ear gap threshold, an ear wear warning is issued. When only the crack defect data appears, a well wall damage warning is issued. When the ear gap value is greater than the ear gap threshold and the crack defect data appears, a high-risk support deformation warning is issued.

10. A monitoring and inspection system for coal mine vertical shafts, characterized in that, The monitoring system employs the monitoring method according to claim 5, and the monitoring system comprises: The data acquisition and processing module is used to acquire corner signals through the can ear ranging device, acquire original underground images of the coal mine shaft using image acquisition equipment, and perform filtering, image enhancement and abnormal image removal on the original underground images to obtain processed underground images. A crack identification module is connected to the data acquisition and processing module. The crack identification module is used to acquire manually annotated crack images, train a deep learning model based on the manually annotated crack images to obtain a trained deep learning model, and input the processed underground image into the trained deep learning model for identification to obtain crack disease data of the coal mine vertical shaft. A joint data map construction module is connected to the crack identification module. The joint data map construction module is used to convert the corner signal into the ear gap value and bind the crack disease data of the same depth in the coal mine shaft with the ear gap value to obtain the joint data map. A safety early warning module is connected to the joint data map construction module. The safety early warning module is used to preset the gap threshold of the shaft ear, and based on the joint data map and the gap threshold of the shaft ear, to classify and judge the coal mine vertical shaft to obtain early warning of hidden dangers of the coal mine vertical shaft at different levels.

Images