A detection method for preventing secondary loading of the main shaft skip based on deep learning of laser point clouds

By installing lidar outside the unloading outlet of the main shaft skip in a coal mine and using point cloud deep learning technology to detect residual material, the problem of secondary loading of the main shaft skip in a coal mine has been solved, improving safety and production efficiency.

CN118111327BActive Publication Date: 2026-11-10ZHONGMEI KEGONG INTELLIGENT STORAGE TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410214006.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-02-27
Publication Date
2026-11-10
Estimated Expiration
2044-02-27

AI Technical Summary

Technical Problem

Existing technologies are insufficient for effectively detecting, in real time, whether there is secondary loading in the skip of the main shaft of a coal mine, leading to safety hazards and low production efficiency.

Method used

A laser point cloud-based deep learning method is adopted. The bottom of the skip is scanned by a laser radar installed outside the skip unloading outlet to obtain point cloud data. The location and volume of residual material are obtained through machine learning and calculation, and a warning signal is issued in real time to prevent secondary loading.

Benefits of technology

This enables precise detection of residual material in the skip, eliminates the need for secondary loading, and improves the operational safety and production efficiency of the main shaft hoist.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118111327B_ABST
    Figure CN118111327B_ABST
Patent Text Reader

Abstract

This invention relates to a detection method for preventing secondary loading of a skip in a mine shaft based on laser point cloud deep learning. The method includes: establishing a baseline point cloud model; waiting for the skip to reach the unloading position signal; acquiring skip position information; filtering point cloud data; acquiring point cloud data of a vertical profile; calculating the cross-sectional area of ​​the residual material; calculating the volume of the residual material; and making a residual material assessment. This invention utilizes a laser radar installed outside the skip to scan the location at the skip outlet where residual material is most likely to be generated, acquiring point cloud data of the residual material. The data is then analyzed, and the location and volume of the residual material are calculated to determine its impact on skip operation. Real-time and accurate residual material quantity and early warning signals are issued to prevent secondary loading, improve the safety of skip operation, effectively avoid overload accidents caused by secondary loading, and ensure the safe operation of mine hoists.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to a detection method for preventing secondary loading of skips in main shafts based on laser point cloud deep learning. It is a detection method in the loading process of transportation equipment and a loading detection method for coal mine vertical shaft hoisting machinery. Background Technology

[0002] As a crucial large-scale piece of equipment for transporting coal in vertical shafts of coal mines, the main shaft hoist's ability to meet the requirements of safe, reliable, efficient, and long-term operation directly impacts the normal safe production and economic benefits of the coal mine. Among the many factors affecting the safe and reliable operation of the main shaft hoist, skip overloading is one of the main factors, and its consequences are quite serious. Due to limitations in technical detection methods and the non-fixed operating position of the equipment, the reliable detection and eradication of skip overloading in the main shaft has not been completely resolved. Therefore, in-depth research and development of overload prevention technologies are of significant practical importance. To ensure the safety of main shaft skip hoisting, various mines have adopted a variety of overload prevention technologies, including fixed-weight loading technology, secondary loading prevention technology, and real-time wire rope tension monitoring technology, which have to a certain extent avoided overloading and safety accidents caused by overloading. However, due to the heavy lifting workload and harsh physical environment of the main shaft, and because the density and viscosity of raw coal produced in coal mines vary greatly due to the influence of moisture and gangue content, it is inevitable that coal will remain unloaded from the skip, i.e., the phenomenon of "incomplete coal unloading." This can lead to overloading or even "secondary loading." The "secondary loading" refers to the situation where, when the skip reaches the shaft opening for unloading, the coal in the skip has not been completely unloaded, and a certain amount of coal remains in the skip before the unloading gate is closed. Then, the skip is reloaded when it descends to the bottom of the shaft.

[0003] Currently, the main method for detecting "incomplete coal unloading" is "post-event detection." This involves the skip being lifted after receiving material and moving away from the unloading position, then, during the uniform speed operation phase, comparing the changes in motor current or wire rope tension caused by the difference in skip weight to determine if "incomplete coal unloading" has occurred, issuing a warning before the next loading. This method has two main problems: first, false alarms can occur due to changes in the load; second, as it is a "post-event" detection method, if incomplete unloading is detected, the incomplete skip (heavy skip) is already in the lowering operation state, posing safety hazards and impacting production efficiency. Another method for real-time overload detection is to add a tension sensor at the wire rope connection to the skip. This method can obtain the skip's weight in real time to determine whether unloading is incomplete or overloaded. However, this method is costly, and because the tension sensor moves up and down with the skip, power supply and data transmission are difficult, resulting in a high failure rate and maintenance workload, hindering its effective widespread application. Therefore, finding an economical, effective, and reliable method for detecting "incomplete coal unloading" is a problem that needs to be solved. Summary of the Invention

[0004] To overcome the problems of existing technologies, this invention proposes a detection method for preventing secondary loading of the skip in the main shaft based on laser point cloud deep learning. After the skip is unloaded, the method uses lidar to scan and detect the coal storage at the bottom of the skip, issuing real-time and accurate coal quantity and early warning signals to prevent secondary loading.

[0005] The objective of this invention is achieved as follows: a detection method for preventing secondary loading of the skip in the main shaft based on laser point cloud deep learning. The system used in the method includes: at least one lidar installed outside the skip's unloading outlet when the skip is stopped at the unloading position; the lidar is electrically connected in sequence to a point cloud data computing server and a skip hoisting controller; the steps of the method are as follows:

[0006] Step 1, Establishing the Reference Point Cloud Model: Point cloud data of the unloading port of the empty skip is acquired and reference data learning is performed to establish a reference point cloud model of the empty unloading port. The process is as follows:

[0007] 1) Use lidar to scan the skip discharge port that is in the unloading position and has completely unloaded, as well as the lower part of the skip below the discharge port, to obtain lidar point cloud data;

[0008] 2) Establish a three-dimensional coordinate system, with the lidar as the spatial zero point;

[0009] 3) Locate the spatial coordinates of the lower edge of the skip and several key points within the skip unloading port from the point cloud data;

[0010] 4) Establish a baseline data model for the discharge port based on key points;

[0011] Step 2, wait for the skip to reach the unloading position signal: After the skip lifting controller detects that the skip has reached the position and unloads the material, it sends the skip arrival signal to the point cloud data computing server;

[0012] Step 3, Obtain the position information of the winnowing basket: After receiving the winnowing basket arrival signal, the point cloud data computing server starts point cloud data filtering and obtains the position of the lower edge of the winnowing basket by the winnowing basket lower edge recognition method;

[0013] Step 4, Filter point cloud data: Filter the point cloud data inside the current unloading port using machine learning;

[0014] Step 5, obtain point cloud data of vertical profile: subdivide the interior of the discharge port according to the vertical profile of the laser beam, and filter out the laser point cloud of multiple profiles.

[0015] Step 6, Calculate the area of ​​the residual material profile: Perform polygon calculations on the point cloud of each profile to find the area;

[0016] Step 7, Calculate the residual volume: Summate the areas of each cross section using calculus to calculate the residual volume;

[0017] Step 8, make a residual material assessment: assess the residual material volume and send the conclusion to the skip hoisting controller.

[0018] The advantages and beneficial effects of this invention are as follows: This invention utilizes a lidar installed outside the skip to scan the location at the skip outlet where residual material is most likely to be generated, obtains point cloud data of the residual material in the skip, analyzes the data, calculates the location and volume of the residual material, thereby determining the impact of the residual material on the skip operation, and issues real-time and accurate residual material quantity and early warning signals to prevent secondary loading, improve the safety of skip operation, effectively avoid overload accidents caused by secondary loading, and ensure the safe operation of the mine hoist. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0020] Figure 1 This is a schematic diagram of the system and its skip structure used in the method described in the embodiments of the present invention, and a cross-sectional view of the skip elevation with a lidar as the 0 point.

[0021] Figure 2 This is a flowchart of the method described in the embodiments of the present invention;

[0022] Figure 3 The laser radar scanning coordinate system of the skip outlet in the method described in this embodiment of the invention is the orthographic coordinate system. Figure 1 View in the direction of arrow B;

[0023] Figure 4 The lidar scanning miscellaneous side-view coordinate system of the method described in the embodiments of the present invention is... Figure 1 Simplified elevation view of the central bucket;

[0024] Figure 5 This is a point cloud distribution diagram of the xy coordinate system at the outlet of the skip discharge port when the residual material curve is approximately L2 and z = 0 under a certain unloading cycle of the method described in this embodiment of the invention.

[0025] Figure 6 This is a coordinate schematic diagram of calculating the cross-sectional area of ​​the residual material at the skip discharge port in the xy coordinate system under a certain unloading cycle of the method described in this embodiment of the invention;

[0026] Figure 7 This is a schematic diagram of the differential area of ​​the residual material cross-section under a certain unloading cycle of the method described in the embodiment of the present invention. Detailed Implementation

[0027] Example:

[0028] This embodiment is a detection method for preventing secondary loading of the skip in the main shaft based on laser point cloud deep learning. The system used in the method includes: at least one lidar 2 installed outside the skip's unloading outlet 101 when the skip 1 is stopped at the unloading position. The lidar is installed at a position slightly below the unloading outlet, preferably so that it can scan the interior of the unloading outlet with the maximum area. The lidar is electrically connected in sequence to the point cloud data computing server and the skip hoisting controller, as follows: Figure 1 As shown; the connection method can be wired or wireless network, such as Ethernet.

[0029] The skip described in this embodiment is a container for carrying coal or other ores, used to lift mined coal or ore from the bottom of the mine to the surface. The skip in this embodiment is a tall cylinder with a rectangular horizontal cross-section and a discharge port on one side of the bottom. An inclined plate 102 is provided at the bottom. Material enters the skip from the upper part of the tall cylinder and is discharged from the discharge port on one side of the bottom of the skip under the action of the inclined plate. Figure 1 Arrow A indicates the flow process of coal. Of course, other types of skips can also be used, as long as the lidar installed outside the skip's discharge port can illuminate the residual material inside the skip, the method described in this embodiment can be used for detection.

[0030] One or two multi-line lidar sensors are placed at a certain position directly opposite the bottom discharge port of the skip. The lidar sensors scan the discharge port below the skip and acquire three-dimensional point cloud data of the discharge port and its interior. The three-dimensional point cloud data of the lidar sensor at the discharge port after discharge is compared and matched with the point cloud data of the discharge port after normal discharge to determine the material adhesion status at the discharge port below the skip, including: blockage, partial coal adhesion, and complete discharge.

[0031] The point cloud data computing server and the skip lifting controller can be a general industrial computer or a PLC or other electronic device with digital computing and storage capabilities.

[0032] A challenge in measuring skips is the variable unloading position (adjusted daily during maintenance based on production conditions) and the expansion and contraction of the wire rope over working hours, resulting in a range of variation (±0.4m) in the skip's unloading position. This makes point cloud matching difficult (locating the unloading port). The solution is to use lidar to locate the lower edge of the skip, measure its deviation from the standard unloading position (data from the skip's running direction), then shift the baseline point cloud data (the point cloud after coal unloading) according to this deviation, and finally compare and match the data.

[0033] The principle of using lidar to scan the residual material in the skip and calculate the amount of residual material is as follows: Figure 1As shown in the side cross-section of the skip, the area that the lidar can scan through the discharge port is indicated by the shaded area. If coal is detected in this area, the amount of residual coal can be estimated based on the distribution data of the coal measurement points. The storage curves L1, L2, and L3 represent the intersections of the three possible storage surfaces with the skip cross-section. The portion coinciding with the shaded area is the location where the laser point lands on the residual coal. Figure 1 The small and medium-sized circles indicate the area of ​​the residual material scanned by the laser scanning line. Curve fitting easily yields the cross-sectional area of ​​the material.

[0034] The specific steps of the method are as follows, and the process is as follows: Figure 2 As shown:

[0035] Step 1, Establishing the Reference Point Cloud Model: Point cloud data of the unloading port of the empty skip is acquired and reference data learning is performed to establish a reference point cloud model of the empty unloading port. The process is as follows:

[0036] 1) Use lidar to scan the skip discharge port that is in the unloading position and the lower part of the skip below the discharge port to obtain lidar point cloud data.

[0037] 2) Establish a three-dimensional rectangular coordinate system, with the lidar as the origin, see... Figure 3 , Figure 4 .

[0038] 3) Locate the spatial coordinates of the lower edge of the skip and several key points within the skip's unloading opening from the point cloud data; such as... Figure 3 , Figure 4 Key reference data points for the six discharge ports: P0, P2, P3, P5, P7, and P8.

[0039] The volume of a material is calculated by slicing it according to the laser lines falling on it, utilizing the multi-line scanning mechanism of lidar. To measure spatial volume, a three-dimensional coordinate system must first be established.

[0040] A three-dimensional coordinate system is established using the spatial location of the lidar as the origin 0. Figure 3 To establish the ZY coordinate system (ZY plane on paper) for viewing the Winnowing Basket from the perspective of a lidar, Figure 4 Using a lidar-based side-view of the skip in an xy coordinate system (where the paper plane is the xy plane), a longitudinal section of the skip's discharge port is selected, with z = 0. Different laser lines can determine different z values; the more laser beams scanned to the discharge port, the more slices can be divided. In the zy coordinate system, laser points P7 and P8 are pre-acquired for the stationary empty skip. These two points are the intersections of the extended lines of the two vertical sides of the discharge port and the lower edge of the skip. The values ​​Z7 and Z8 are obtained. Since the left and right edges of the skip do not shift when discharging at the wellhead, the range of z values ​​for the laser points falling on the coal inside the discharge port is Z7. <z i<Z8, which is one of the laser point cloud filtering conditions.

[0041] 4) Establish a discharge opening reference data model from key points; this model is a point cloud filtering condition inside the discharge opening for matching and comparison, as shown in formula (1), the analysis is as follows:

[0042] In the xy coordinate system at z=0, the laser points P0, P1, P2, P3, P4 of the empty skip in a static state are acquired in advance, where P0 is the point on the lower edge of the skip, P1 is the lower edge of the skip coal sliding plate, P2 is the upper edge of the skip coal sliding plate, P3 is the upper edge of the skip coal discharge opening, and P4 is a certain point above the skip discharge opening. From the coordinates of P1 and P3, the spatial coordinate curve of the coal sliding plate can be obtained, and then P5 can be obtained from point P3, that is, the point of P3 on the inclined coal sliding plate along the y-axis direction. From this, it is not difficult to conclude that in the xy coordinate system, the x value range of the laser points falling on the coal material in the discharge opening is X3<x i <X2, the y value range is This range is the second laser point cloud filtering condition.

[0043] Step 2, wait for the signal that the skip reaches the discharge position (and completes discharging): After the skip hoisting controller detects that the skip is in place and completes discharging, it sends the skip in-place signal to the point cloud data calculation server.

[0044] Step 3, acquire skip position information: After receiving the skip in-place signal, the point cloud data calculation server starts point cloud data filtering, and acquires the lower edge position p0 of the skip through the skip lower edge identification method.

[0045] The reference point cloud data of the discharge opening includes P0, P3, P5, P2 and the inclined line segment P5P2, and the filtering conditions for the discharge opening point cloud (reference point cloud) are the following formulas:

[0046]

[0047] After obtaining the reference point cloud of the discharge opening, offset matching is performed on the real-time skip discharge opening, because the discharge position of the skip is not always fixed.

[0048] Step 4, filter point cloud data: Screen the current internal spatial point cloud data of the discharge opening through machine learning; the internal spatial point cloud data of the front discharge opening is the residual material point cloud, and the filtering method is shown in formula (2).

[0049] Acquire point cloud data for the skip discharge opening in a certain discharge cycle (taking the xy coordinate system of a certain laser line when z=0), as Figure 5As shown. First, the algorithm finds the lower edge point p0(x0, y0) of the skip, and its offset from the aforementioned discharge port reference point P0 is (y0-Y0). With this offset, it is not difficult to deduce the reference point cloud data P0', P3', P5', and P2' of the discharge port under this discharge cycle, as well as the diagonal line segment P5'P2'. Therefore, the selection criteria for the internal point cloud of the discharge port under this cycle can be derived:

[0050]

[0051] The selection criteria in Formula 2 can be used to find the laser points p1(x1,y1), p2(x2,y2), ..., pN(x) inside the discharge port from the laser point cloud of the laser beam. N ,y N Based on these laser points, calculate the area of ​​the coal material under that cross-section, such as... Figure 5 The ghosting part. For ease of solution, it can be arranged as follows: Figure 6 The calculation of the ghosting part involves projecting point pN onto the back plate of the bucket along the oblique line P2'P5', resulting in P6'. The y-coordinate of point P6' is easily calculated as (Y2+y0-Y0+y0). N -y1). For polygons p1(x1,y1), p2(x2,y2)...pN(xN,yN), P2', P6', calculus is used to find the area S1. Then, by numerically integrating multiple sections of the discharge port point cloud along the z-axis, the volume of the residual material can be obtained, i.e., V=ΣS i ×Δz.

[0052] Step 5, obtain point cloud data of vertical profile: subdivide the inside of the discharge port according to the laser beam with vertical profile in z-axis coordinate, and filter out laser point clouds of multiple profiles.

[0053] Step 6, calculate the area of ​​the residual material profile: For each profile point cloud, construct polygons p1(x1,y1), p2(x2,y2)...pN(xN,yN), P2', P6' and calculate the area Si; use calculus to calculate the area. Refer to [link to algorithm details]. Figure 6 , Figure 7 .

[0054] Step 7, calculate the residual volume: perform calculus on the area of ​​each section and sum the results to calculate the residual volume.

[0055] To find the area S1 using calculus, we can project p1(x1,y1), p2(x2,y2)...pN(xN,yN) onto the back plate of the bucket along the oblique lines P2'P5', and then calculate the area S of each of the multiple polygons derived from the calculus. i Finally, we get S1 = Σs i .like Figure 7As shown, based on the method for calculating the area of ​​a polygon in a two-dimensional coordinate system (finding the area by decomposing multiple trapezoids), it is not difficult to derive the area S of a certain subdivided polygon with a cross-sectional area. i :

[0056]

[0057] The algorithm for finding the lower edge point of the skip is as follows: The minimum value Y of the lower edge position after the skip has unloaded is determined manually. min Based on the fact that the vertical fluctuation range of the skip unloading position is ±0.4 meters, the range Y of the lower edge of the skip unloading position can be determined. min <y i <(Y min +0.8). For example, the lowest possible y-value of the lower edge of the skip after unloading is -0.2 meters. Given that the upper and lower fluctuation range of the skip's unloading position is ±0.4 meters, the y-value range of the lower edge of the skip can be determined to be -0.2 to -0.6 meters. By filtering the laser point cloud according to Equation 4, laser point p0 can be found.

[0058]

[0059] It should be noted that the unloading position of the skip in each cycle of unloading only changes in the y-axis direction, and there is no offset in the x-axis and z-axis.

[0060] Step 8, make a residual material assessment: assess the residual material volume and send the conclusion to the skip hoisting controller.

[0061] The size and location of the residual material are assessed based on the calculation results to determine the degree of harm caused by the residual material and whether it needs to be cleaned up. The assessment results are then sent to the skip controller so that the skip controller can take actions such as continuing operation, issuing an alarm, or stopping operation.

[0062] Finally, it should be noted that the above is only used to illustrate the technical solution of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred arrangement, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solution of the present invention (such as the form of the winnowing basket, the application of various formulas, the order of steps, etc.) without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A detection method for preventing secondary loading of the main shaft skip based on laser point cloud deep learning, wherein the system used in the method includes: At least one lidar is installed outside the skip's unloading outlet when the skip is stopped at the unloading position, and the lidar is electrically connected in sequence to a point cloud data computing server and a skip lifting controller; the method is characterized by the following steps: Step 1, Establishing the Reference Point Cloud Model: Point cloud data of the unloading port of the empty skip is acquired and reference data learning is performed to establish a reference point cloud model of the empty unloading port. The process is as follows: 1) Use lidar to scan the skip discharge port that is in the unloading position and has completely unloaded, as well as the lower part of the skip below the discharge port, to obtain lidar point cloud data; 2) Establish a three-dimensional coordinate system, with the lidar as the spatial zero point; 3) Locate the spatial coordinates of the lower edge of the skip and several key points within the skip unloading port from the point cloud data; 4) Establish a baseline data model for the discharge port based on key points; Step 2, wait for the skip to reach the unloading position signal: After the skip lifting controller detects that the skip has reached the position and unloads the material, it sends the skip arrival signal to the point cloud data computing server; Step 3, Obtain the position information of the winnowing basket: After receiving the winnowing basket arrival signal, the point cloud data computing server starts point cloud data filtering and obtains the position of the lower edge of the winnowing basket by the winnowing basket lower edge recognition method; Step 4, Filter point cloud data: Filter the point cloud data inside the current unloading port using machine learning; Step 5, obtain point cloud data of vertical profile: subdivide the interior of the discharge port according to the vertical profile of the laser beam, and filter out the laser point cloud of multiple profiles. Step 6, Calculate the area of ​​the residual material profile: Perform polygon calculations on the point cloud of each profile to find the area; Step 7, Calculate the residual volume: Summate the areas of each cross section using calculus to calculate the residual volume; Step 8, make a residual material assessment: assess the residual material volume and send the conclusion to the skip hoisting controller.

Citation Information

Patent Citations

  • Assembly line measurement device and measurement method for measuring amount of wood chips

    CN107664477A

  • Bulk material loading uniform material distribution self-adaptive adjusting method

    CN114067076A