Curved chute blanking trajectory control method and device and storage medium

Abstract

The application discloses a curved chute material falling trajectory control method and device and a storage medium, and relates to the field of conveying automation. Three-dimensional point cloud data of a material surface on a lower conveying belt is obtained through laser radar scanning; filtering and interpolation fitting are performed to obtain a continuous and smooth material cross-section profile curve; each frame of material cross-section profile curve is combined with a belt reference height curve obtained through calibration in the empty state, three-dimensional cross-section characteristics of a real-time material flow are solved, and a horizontal gravity center offset of the material relative to the gravity center of the center of the lower conveying belt is calculated; when the gravity center offset exceeds a preset allowable threshold, a turning plate control instruction is generated, an independently driven electric turning plate is controlled to act in cooperation until the gravity center offset corresponding to the material flow state returns to within the allowable threshold. The scheme can realize real-time sensing of the curved chute material falling state, accurate calculation of the horizontal gravity center offset of the material, automatic driving of an actuator for dynamic closed-loop correction, and effective prevention of material scattering and equipment failure caused by improper material falling.

Description

Technical Field

[0001] This application relates to the field of conveying automation, and in particular to a method, device and storage medium for controlling the trajectory of a curved chute material discharge. Background Technology

[0002] In bulk material conveying systems such as mines, ports, and power plants, curved chutes are critical transfer nodes connecting upstream and downstream conveyors. Due to the complex internal flow field of curved chutes, they are highly susceptible to misalignment (i.e., material deviating from the centerline of the chute or the center of the downstream conveyor belt) caused by factors such as material particle size, humidity, flow rate fluctuations, and chute wear. Misalignment can lead to serious consequences such as downstream conveyor belt misalignment, material spillage, accelerated localized wear of the liner, and even belt tearing.

[0003] While current belt conveyor technology utilizes 3D vision or LiDAR to collect belt load point clouds and drive correction motors to adjust idlers to correct belt misalignment, this approach only provides passive mechanical correction to already misaligned conveyor belts. It is a reactive measure and cannot fundamentally address the source-level material spillage and uneven impact issues caused by internal material flow trajectory deviations at upstream curved chute transfer nodes. Furthermore, the interior of curved chutes is characterized by high-concentration dust and extremely variable instantaneous flow rates. Conventional belt point cloud screening and edge-finding algorithms are prone to distortion in reconstructing the irregular, undulating material flow from the chute, failing to provide millisecond-level dynamic calculations of the centroid of the material drop section.

[0004] Currently, the adjustment of material flow through chutes largely relies on manual experience to periodically adjust the guide vanes, or on simple mechanical baffles for passive limiting. This lack of quantitative perception of the three-dimensional material flow pattern prevents the achievement of millisecond-level dynamic closed-loop correction. Therefore, there is an urgent need for an intelligent control solution capable of real-time sensing of the material flow status of curved chutes and automatically reversing the flow direction to adjust the actuator. Summary of the Invention

[0005] This application provides a method, device, and storage medium for controlling the trajectory of material falling through a curved chute; it can sense the material falling status of the curved chute in real time, accurately calculate the lateral center of gravity offset of the material, and automatically drive the actuator to perform dynamic closed-loop correction, effectively preventing material spillage and equipment failure caused by improper material falling.

[0006] On one hand, this application provides a method for controlling the trajectory of a curved chute discharge based on lidar feedback. The curved chute is equipped with multiple independently drivable electric flaps. The control method includes:

[0007] The laser radar installed above the lower conveyor belt is controlled to scan in real time to obtain three-dimensional point cloud data of the material surface on the lower conveyor belt; The three-dimensional point cloud data is filtered to remove dust interference, and a piecewise polynomial interpolation algorithm is used to fit the filtered discrete height point cloud into a continuous and smooth material cross-sectional profile curve. The cross-sectional profile curves of the material obtained in each frame are combined with the belt reference height curve obtained by no-load calibration to eliminate the interference of the deformation reference of the lower conveyor belt itself, calculate the three-dimensional cross-sectional characteristics of the real-time material flow, and calculate the offset of the material's lateral center of gravity relative to the center of gravity of the lower conveyor belt. When the center of gravity offset exceeds the preset allowable threshold, a corresponding flap control command is generated based on the pre-built mapping model, and multiple independently driven electric flaps distributed on the inner side of the curved chute head cover and both sides of the material flow surface of the drop pipe are controlled to work together until the center of gravity offset corresponding to the material flow state returns to within the allowable threshold.

[0008] Specifically, acquiring the three-dimensional point cloud data of the material surface on the lower conveyor belt includes: The lidar is controlled to emit infrared laser pulses along the width of the lower conveyor belt at a preset transverse scanning frequency, and to receive echo pulses diffusely reflected from the material surface. Based on the time difference between laser pulse emission and reception at each measurement point and the speed of light The distance values ​​of each measurement point are calculated using the following formula. ;

[0009] The distance value sequence obtained from a single transverse scan is used as a discrete height point cloud of a material cross section. Multiple frames of the discrete height point cloud are then stitched together along the movement direction of the lower conveyor belt to reconstruct the three-dimensional point cloud data.

[0010] Specifically, the filtering process for the 3D point cloud data includes: The lidar is controlled to activate the multi-echo acquisition mode. Threshold filtering is performed on multiple echo signals received in the same measurement direction. Only the distance value corresponding to the main echo signal with the highest intensity and shortest delay is retained. Secondary reflection noise caused by dust is eliminated to obtain a pure discrete height point cloud of the material cross section.

[0011] Specifically, before performing multi-echo threshold filtering on the 3D point cloud data, the process further includes: Straight-through filtering: Set the effective ranging interval Only retain distance values satisfy The point cloud is filtered to remove invalid point clouds that are far away from the material area, such as those on the frame or rollers. Statistical outlier filtering: Iterate through the average neighborhood distance of each point cloud and remove isolated dust noise points whose average neighborhood distance deviates from the overall mean by more than 3 standard deviations.

[0012] Specifically, the piecewise polynomial interpolation algorithm uses the Akima piecewise cubic polynomial interpolation algorithm, including: Let the sequence of discrete height point cloud sampling nodes after filtering be... ,in The horizontal axis is monotonically increasing. And the ordinate Mapped to real-time height sampling value of material ; In any subinterval Construct a piecewise cubic polynomial:

[0013] In the formula, The fitted continuous and smooth cross-sectional profile curve of the material. Calculate the slope of the line segment between every two adjacent sampling points. :

[0014] For internal nodes derivative value for:

[0015] Define the interval step size and interval elevation difference The coefficients of the cubic polynomial for the current interval are obtained by solving the endpoint constraints:

[0016]

[0017]

[0018]

[0019] in, interpolation nodes The smooth first derivative at point .

[0020] Specifically, calculating the center of gravity offset includes: Based on the material cross-sectional profile curve And the reference height curve of the unloaded belt In the effective material flow lateral range The cross-sectional area of ​​a single frame of material is calculated using numerical integration. Cross-sectional area moment and the lateral center of gravity coordinates of the material The formulas are expressed as follows:

[0021]

[0022]

[0023] The lateral centroid coordinate of the material With respect to the calibrated belt center reference coordinates By comparison, the center of gravity offset is obtained. .

[0024] Specifically, generating corresponding flip-board control commands based on a pre-built mapping model includes: During the system debugging phase, orthogonal experiments were conducted to record the center of gravity offset response data under different combinations of electric flap angles, and a correspondence table reflecting the flap action and material flow response was constructed and stored. During operation, based on the real-time calculated center of gravity offset The corresponding table is queried to match and determine the target flap angle combination to generate the flap control command; The model predictive control logic is used to correct the corresponding table online, taking the current center of gravity offset and the historical angle state of each electric flap as input.

[0025] Specifically, the plurality of electric flaps includes a middle flap disposed at the head of the curved chute, and a left flap and a right flap disposed on both sides of the discharge pipe; the independently driven electric flaps coordinate to move, including: When it is determined that the center of gravity offset is biased towards the left side of the lower conveyor belt, a flap control command is generated to control the left flap to close and the right flap to open. When it is determined that the center of gravity offset is biased towards the right side of the lower conveyor belt, a flap control command is generated to control the left flap to open and the right flap to close. When it is determined that the center of gravity offset is in the center, control each of the electric flaps to maintain the current opening degree in a steady state.

[0026] On the other hand, this application provides a curved chute material discharge trajectory control device based on lidar feedback, wherein the curved chute is equipped with multiple independently drivable electric flaps, and the device includes: The lidar scanning module is used to control the lidar installed above the lower conveyor belt to scan in real time and acquire three-dimensional point cloud data of the material surface on the lower conveyor belt. The continuous contour reconstruction module is used to filter the three-dimensional point cloud data to remove dust interference. It uses a piecewise polynomial interpolation algorithm to fit the filtered discrete height point cloud into a continuous and smooth material cross-sectional contour curve. The center of gravity offset calculation module is used to combine the material cross-sectional profile curves obtained in each frame with the belt reference height curve obtained by no-load calibration to eliminate the interference of the deformation reference of the lower conveyor belt itself, solve the three-dimensional cross-sectional characteristics of the real-time material flow, and calculate the center of gravity offset of the material's lateral center of gravity relative to the center of the lower conveyor belt. The dynamic closed-loop correction module is used to generate corresponding flap control commands based on a pre-built mapping model when the center of gravity offset exceeds a preset allowable threshold. It then controls multiple independently driven electric flaps distributed on the inner side of the curved chute head cover and both sides of the material flow surface of the drop pipe to work together until the center of gravity offset corresponding to the material flow state returns to within the allowable threshold.

[0027] In another aspect, this application also provides a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the above-described method for controlling the trajectory of a curved chute based on lidar feedback.

[0028] The beneficial effects of the technical solution provided in this application include at least the following: by introducing high-frequency sensing from lidar and supplementing it with targeted filtering and piecewise polynomial interpolation reconstruction, the interference of high-concentration dust secondary multipath reflection on the ranging signal is effectively overcome. It can reconstruct the three-dimensional surface morphology of the rapidly falling and irregularly accumulated abrupt material flow at high frequency and accurately. Combined with the unloaded dynamic reference of the downstream conveyor belt, the real-time cross-sectional area and center of gravity offset of the high dynamic material flow in physical space are accurately calculated through numerical integration, providing high-precision quantitative control input for upstream active flow control.

[0029] A dynamic closed-loop correction mechanism is constructed, which can proactively issue a flap control command based on the calculated offset at the instant (or before) the falling material contacts the downstream conveyor belt. By controlling the coordinated action of multiple independently driven electric flaps within the curved chute, the direction of material flow is directly changed inside the chute and at the material source. This "pre-emptive, proactive, and source-based" flow control and modulation capability fundamentally eliminates the generation of asymmetric impact loads and protects the downstream conveying system.

[0030] Based on a mapping model, millisecond-level dynamic coordinated adjustment of multiple valve plate angle changes is achieved, enabling the material drop trajectory adjustment to dynamically converge with changes in operating conditions, eliminating material drop deviation caused by moisture adhesion or sudden changes in flow rate. Through continuous closed-loop feedback control, not only can abnormal operating conditions such as material blockage and local overflow be effectively prevented and dealt with, but also the risks of abnormal wear and tear on the conveyor belt caused by deviation and spillage are avoided from the source, significantly improving the automation level, operational stability, and overall economic benefits of the bulk material transfer system. Attached Figure Description

[0031] Figure 1 This is a flowchart of the curved chute feeding trajectory control method based on lidar feedback provided in the embodiments of this application; Figure 2 This is a schematic diagram of a possible form of a curved chute transfer control system. Figure 3 This is a schematic diagram illustrating the acquisition of three-dimensional cross-sectional features of material flow; Figure 4 This is a flowchart of the three-dimensional reconstruction of the material flow cross section and the calculation of the material flow centroid and height; Figure 5 A schematic diagram of the left and right flip panels is shown; Figure 6 A schematic diagram of the middle flap is shown; Figure 7 A logic block diagram showing the coordinated control of multiple flip panels is provided. Figure 8 The timing diagram of the closed-loop correction control is shown; Figure 9 A schematic diagram of a curved chute discharge trajectory control device based on lidar feedback is shown. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0033] In this article, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0034] In bulk material conveying systems, curved chutes serve as crucial transfer nodes connecting upstream and downstream conveyors. Their internal flow field is highly susceptible to variations in material particle size distribution, humidity, and instantaneous flow fluctuations, easily causing the trajectory of the falling material flow to deviate from the design centerline. This deviation leads to asymmetrical impacts on the downstream conveyor belt, resulting in severe spillage, misalignment, and localized, intense wear on the chute liner. Current technologies primarily rely on manual periodic adjustment of internal guide plates or passive blocking with fixed baffles. These methods are not only slow to adjust, but also suffer from significant challenges due to the extremely narrow space at the curved chute's discharge port, the dynamic overlapping state of the internal material flow during high-speed descent, and the presence of extremely high concentrations of airborne dust. Traditional non-contact measurement methods (such as conventional edge-finding or filtering algorithms for long-distance, stable conveyor belts) are prone to contour reconstruction distortion under such extremely confined spaces and multiphase, multi-flow disturbances due to abrupt depth changes and secondary dust reflections. This makes it impossible to obtain high-precision instantaneous centroid parameters of the material flow cross-section, thus hindering millisecond-level closed-loop active modulation of the falling material flow.

[0035] For example, in port coal transfer operations, when handling lump coal with high moisture content, the material may experience momentary irregular accumulation and sudden slippage inside the curved chute due to adhesion. At this time, the concentration of dust flying below the discharge port surges instantaneously, and the surface of the material flow exhibits violent and irregular undulations due to high-speed descent. Conventional point cloud reconstruction algorithms designed for flat conveyor belt loads suffer severe distortion when dealing with such abrupt point clouds with severe multipath reflections and outliers. They cannot accurately and continuously solve for the true cross-sectional area integral and three-dimensional morphological changes of the slipping material flow in space within a short time, thus failing to provide accurate control input to the upstream flow control actuator. This results in the downstream conveying system being subjected to uncontrollable asymmetric off-center load impacts for extended periods, causing not only conveyor belt spillage but also frequent unplanned shutdowns of the transfer system due to momentary blockages or localized overloads.

[0036] If the above problems are not solved, passive correction methods such as setting correction rollers on the lower belt conveyor will not only fail to eliminate the abnormal wear of the liner and localized material spillage caused by the internal flow direction deviation of the chute, but also the frequent asymmetric material impacts will accelerate the structural fatigue of the curved chute itself and reduce the overall service life of the core transfer equipment.

[0037] Therefore, this application provides a method for controlling the trajectory of material falling through a curved chute based on lidar feedback. This method is applied to a control system including a lidar sensing unit, a data processing and control unit, and multiple independently driveable electric flaps arranged along the curved chute. The control method includes... Figure 1 The loop execution steps shown are as follows: Step 110, LiDAR scanning stage: Control the LiDAR sensing unit installed above the lower conveyor belt to scan in real time and obtain three-dimensional point cloud data of the material surface on the lower conveyor belt; Figure 2 This is a schematic diagram of a possible form of the curved chute transfer control system provided in this application embodiment. The end section of the curved chute serves as the discharge port, aligned with the conveyor belt. A lidar sensing unit is installed behind the guide chute of the lower conveyor belt to ensure complete coverage of the material flow area. It is a sensor that uses a laser beam to detect information such as target distance and speed. In this embodiment, the unit is configured to scan the material surface in real time to obtain its three-dimensional spatial data, providing raw data for subsequent material flow analysis.

[0038] Electric flaps are actuators installed inside curved chutes; these flaps are driven by independent motors and can change their angle or position, thereby physically intervening in and adjusting the flow direction of materials inside the curved chute, and achieving precise control of the material drop trajectory.

[0039] During the lidar scanning stage, the lidar performs multiple scans along the width of the conveyor belt, and these two-dimensional scan lines are stitched together along the direction of the conveyor belt's movement to construct three-dimensional point cloud data of the material surface.

[0040] Step 120, Continuous contour reconstruction stage: The data processing and control unit filters the three-dimensional point cloud data to remove dust interference, and uses a piecewise polynomial interpolation algorithm to fit the filtered discrete height point cloud into a continuous and smooth material cross-sectional contour curve. This step aims to extract accurate material morphology information from the raw data. To reduce dust and vibration interference, filtering and smoothing processes are required to obtain accurate material cross-sectional profile curves.

[0041] Specifically, in terms of filtering, a simple distance threshold filtering method can be used. This involves setting an effective measurement range and directly removing point cloud data that exceeds this range to eliminate some background noise from distant locations or interference from the sensor itself at close range. For continuous contour reconstruction, linear interpolation or a simple cubic spline interpolation algorithm can be used to connect the filtered discrete height point clouds to form a continuous curve, approximating the cross-sectional contour of the material.

[0042] Step 130, Center of gravity offset calculation stage: Combine the reconstructed material cross-sectional profile curves of each frame with the belt reference height curve obtained from the no-load calibration to eliminate the interference of the deformation reference of the lower conveyor belt itself, solve the three-dimensional cross-sectional characteristics of the real-time material flow, and calculate the center of gravity offset of the material's lateral center of gravity relative to the center of the lower conveyor belt. The belt reference height curve obtained from the no-load calibration needs to be measured before feeding. It represents the belt surface height (thickness) curve measured under its own weight, which is usually parabolic. Figure 3 This is a schematic diagram illustrating the acquisition of three-dimensional cross-sectional features of material flow. The difference between the reconstructed material cross-sectional contour curves of each frame and the belt reference height curve obtained from no-load calibration yields the real-time three-dimensional cross-sectional features of the material flow. After acquiring the three-dimensional cross-sectional features, the offset of the material's lateral center of gravity relative to the center of the next-level conveyor belt can be calculated based on the contour shape. For calculating the offset, the material's cross-sectional contour can be approximated as several simple geometric shapes (such as parabolic shapes). Then, its center of gravity coordinates are determined through geometric calculation methods and compared with the preset conveyor belt center coordinates to obtain the offset.

[0043] Step 140, Dynamic Closed-Loop Correction Stage: When the center of gravity offset exceeds the preset allowable threshold, the corresponding flap control command is generated based on the pre-built mapping model, and multiple independently driven electric flaps distributed on the inner side of the curved chute head cover and both sides of the material flow surface of the drop pipe are controlled to work together to change the material flow direction in the curved chute; the system continuously monitors the adjusted material flow state to form a closed-loop feedback until the center of gravity offset returns to within the allowable threshold.

[0044] This step is the core of achieving automatic deviation correction. One implementation method is to construct a simple lookup table as a mapping model, which records the correspondence between different center of gravity offset ranges and preset flap opening combinations. When a center of gravity offset is detected, the system queries this lookup table to obtain the corresponding flap opening command.

[0045] Regarding the coordinated action of the flappers, a simple logic can be set. For example, when the center of gravity shifts to the left, only the right flapper is controlled to close slightly to adjust the material flow to the right; when the center of gravity shifts to the right, the left flapper is controlled to perform a similar operation. After the system performs the adjustment, a fixed delay can be set to allow the material flow to stabilize before performing the next scan and calculation to determine the adjustment effect. Finally, through multi-round closed-loop feedback control, the center of gravity offset is brought back to within the allowable threshold.

[0046] For example, in a large bulk material transfer station, iron ore needs to be transferred from the upper conveyor to the lower conveyor belt via a curved chute. Due to frequent fluctuations in the particle size, moisture content, and conveying flow rate of the iron ore, the landing position of the iron ore on the lower conveyor belt often deviates from the center, resulting in problems such as spillage and increased local wear on the conveyor belt.

[0047] This solution uses lidar sensing to acquire 3D point cloud data of the iron ore surface on the conveyor belt. Then, points with abnormal distances from the material surface are identified and removed to filter out spurious echoes caused by dust. After filtering, the discrete point cloud data is input into a piecewise polynomial interpolation algorithm, which fits these discrete height points into a continuous and smooth iron ore cross-sectional profile curve. This curve accurately depicts the current accumulation shape of the iron ore on the conveyor belt.

[0048] Furthermore, the reconstructed iron ore cross-sectional profile curve is subtracted from the belt reference height curve obtained through pre-calibration under no-load conditions to eliminate the interference of conveyor belt deformation on the measurement results, thereby accurately calculating the three-dimensional cross-sectional characteristics of the real-time iron ore flow, including its cross-sectional area and lateral centroid coordinates. This lateral centroid coordinate is then compared with the pre-calibrated conveyor belt center reference coordinates to calculate the centroid offset of the iron ore flow relative to the conveyor belt center. For example, if the calculated centroid offset is 5 cm to the left, it indicates that the iron ore flow is biased towards the left side of the conveyor belt.

[0049] When the detected center of gravity offset exceeds a preset allowable threshold (e.g., ±3 cm), the system activates a dynamic closed-loop correction mechanism. Based on a pre-built mapping model, the currently calculated center of gravity offset (e.g., 5 cm to the left) is converted into corresponding flap control commands. These commands are then sent to multiple independently driven electric flaps distributed within the curved chute. For example, if the iron ore flow deviates to the left, the system might instruct the electric flap on the right to slightly close, while the electric flap on the left remains open or slightly opens. By altering the guiding effect of the chute's inner wall on the material, this adjusts the iron ore flow to the right, bringing it back to the center. The entire process is monitored in real-time, forming a real-time closed-loop feedback. This process continues until the center of gravity offset of the iron ore flow returns to within the allowable threshold range, ensuring that the iron ore falls stably and accurately into the central area of ​​the next conveyor belt.

[0050] In summary, conventional non-contact measurement or edge-finding algorithms are mostly designed for long-distance, stable conveyor belts and cannot adapt to the extreme and harsh conditions of narrow space at the discharge port of curved chutes, high-speed falling and dynamic overlap of materials, and high concentrations of airborne dust. This embodiment, by introducing high-frequency sensing from lidar and supplementing it with targeted filtering and piecewise polynomial interpolation reconstruction, effectively overcomes the interference of secondary multipath reflection from high-concentration dust on the ranging signal. It can reconstruct the three-dimensional surface morphology of the rapidly falling and irregularly accumulated abruptly changing material flow with high frequency and accuracy. Combined with the unloaded dynamic reference of the downstream conveyor belt, the real-time cross-sectional area and center of gravity offset of the highly dynamic material flow in physical space are accurately calculated through numerical integration, providing high-precision quantitative control input for upstream active flow control.

[0051] Traditional belt idler correction technology is a "reactive correction after the fact," meaning that the belt can only be forcibly corrected mechanically after the downstream conveyor belt has already experienced asymmetrical load and belt misalignment. This method not only has significant lag in adjustment but also fails to eliminate the physical impact and localized abnormal wear caused by unilateral material load on the belt and idlers. In contrast, the dynamic closed-loop correction mechanism constructed in this embodiment can proactively issue a flap control command based on the calculated offset at the moment (or before) the falling material contacts the downstream conveyor belt. By controlling the coordinated action of multiple independently driven electric flaps within the curved chute, the flow direction of the falling material is directly changed inside the chute and at the material source. This "pre-emptive, proactive, and source-based" flow control and modulation capability fundamentally eliminates the generation of asymmetrical impact loads and protects the downstream conveying system.

[0052] Due to the highly time-varying nature of particle size, moisture content, and instantaneous flow rate of bulk materials, fixed guide plates or baffles cannot adaptively adjust. This embodiment, based on a pre-built and online-correctable mapping model, achieves millisecond-level dynamic coordinated adjustment of multi-valve angle changes, enabling the material drop trajectory adjustment to dynamically converge with changes in operating conditions, eliminating material drop deviation caused by moisture adhesion or sudden flow changes. Through continuous closed-loop feedback control, it not only effectively prevents and responds to abnormal operating conditions such as material blockage and local overflow, but also avoids the risk of abnormal wear and tear on the conveyor belt caused by deviation and spillage, significantly improving the automation level, operational stability, and overall economic benefits of the bulk material transfer system.

[0053] In some embodiments, acquiring three-dimensional point cloud data of the material surface on the lower conveyor belt specifically includes the following steps: 1. Control the lidar to emit infrared laser pulses along the width of the lower conveyor belt at a preset transverse scanning frequency, and receive the echo pulses diffusely reflected from the material surface; 2. Based on the time difference between laser pulse emission and reception at each measurement point and the speed of light The distance values ​​of each measurement point are calculated using the following formula. ;

[0054] 3. The distance value sequence obtained from a single transverse scan is used as a discrete height point cloud of a material cross section, and multiple frames of discrete height point clouds are stitched together along the movement direction of the lower conveyor belt to reconstruct three-dimensional point cloud data.

[0055] This process involves spatially aligning and combining discrete height point clouds of multiple material cross-sections obtained from continuous laser scanning at different times, based on the real-time speed and direction of the conveyor belt. Its purpose is to utilize the conveyor belt's motion as a third dimension, expanding the two-dimensional cross-sectional scanning data into three-dimensional spatial data. This step provides a comprehensive three-dimensional spatial data foundation for subsequent material cross-sectional contour curve fitting, feature decomposition, and center of gravity offset calculation.

[0056] Traditional methods for filtering 3D point cloud data to remove dust interference often encounter problems in actual bulk material transport environments such as mines and ports, where high concentrations of dust often exist inside curved chutes. This dust generates a large number of secondary reflection noise points, causing distortion of the discrete height point cloud data of the material cross-section acquired by the lidar. Consequently, it affects the accurate reconstruction of the material contour and the precise calculation of the center of gravity offset.

[0057] This application performs filtering processing on 3D point cloud data to remove dust interference. The specific steps are as follows: The lidar is controlled to activate the multi-echo acquisition mode. Threshold filtering is performed on multiple echo signals received in the same measurement direction. Only the distance value corresponding to the main echo signal with the highest intensity and shortest delay is retained. Secondary reflection noise caused by dust is eliminated to obtain a pure discrete height point cloud of the material cross section.

[0058] Material surfaces typically have a larger reflective area and more stable reflection characteristics than individual dust particles, resulting in higher echo signal strength. Meanwhile, the actual material surface is usually farther from the lidar than the dust particles in front of it, but its latency is the shortest relative to the lidar's "primary" reflector. However, in dusty environments, dust particles may reflect before the material surface, leading to even shorter latency.

[0059] This scheme cleverly combines the criteria of "highest intensity" and "shortest latency." Highest intensity ensures that a target with a large reflective cross-section is identified, rather than a weak dust reflection; shortest latency ensures that among multiple high-intensity reflections, the closest real material surface to the sensor is selected. This dual-screening mechanism effectively distinguishes between echoes from the real material surface and secondary reflections caused by dust, thus eliminating these noise points. Through this processing, this scheme can obtain a clean discrete height point cloud of the material cross-section.

[0060] Before performing multi-echo threshold filtering, if the original point cloud is processed directly, the complex material feeding environment of the chute, the interference point cloud in non-material areas such as the frame and rollers, and the isolated noise points of dust randomly distributed in space will significantly increase the computational burden of data processing, and may lead to a decrease in the accuracy of subsequent main echo signal recognition, thereby affecting the reconstruction accuracy of the material cross-sectional contour curve.

[0061] Therefore, before performing multi-echo threshold filtering on the 3D point cloud data, this application also includes the following filtering operation: Straight-through filtering: Set the effective ranging interval Only retain distance values satisfy The point cloud is filtered to remove invalid point clouds that are far away from the material area, such as those on the frame or rollers. Pass filtering serves as the first line of defense, utilizing a preset effective ranging interval. The raw point cloud undergoes coarse-grained spatial filtering. This step quickly and efficiently removes long-distance invalid point clouds generated by fixed structures such as conveyor frames and rollers from the data collected by the LiDAR sensing unit, ensuring that the dataset for subsequent processing is mainly concentrated in areas where materials may exist. Based on this, the point cloud data after pass-through filtering then enters the statistical outlier filtering stage.

[0062] Statistical outlier filtering: Traverse the average neighborhood distance of each point cloud and remove isolated dust noise points whose average neighborhood distance deviates from the overall mean by more than 3 standard deviations.

[0063] Statistical outlier filtering focuses on processing local noise. It identifies and removes isolated noise points that are significantly inconsistent with the distribution of the surrounding point cloud, caused by factors such as dust, by traversing the average neighborhood distance of each point cloud and based on statistical principles (such as 3 times the standard deviation).

[0064] This two-stage filtering strategy first macroscopically defines the effective scope of data processing, and then microscopically purifies the noise within the data. This orderly and complementary preprocessing provides highly pure and focused data input for subsequent threshold filtering in the multi-echo acquisition mode of the LiDAR sensing unit, greatly reducing the computational burden of multi-echo threshold filtering and significantly improving the accuracy and robustness of main echo signal identification. This lays a solid foundation for accurately reconstructing the cross-sectional profile curve of the material.

[0065] In this embodiment, the piecewise polynomial interpolation algorithm employs the Akima piecewise cubic polynomial interpolation algorithm. It uses discrete sampling height points from the lidar as interpolation nodes to piecewise fit smooth, continuous contour curves, accurately reproducing the undulating surfaces of coal piles and gravel, thus improving the accuracy of material flow cross-sectional area integration, 3D morphology reconstruction, and cross-sectional centroid calculation. Specifically, it includes: S1: Let the discrete sampling node sequence of the lidar be: The horizontal axis is strictly monotonically increasing. In any subinterval Construct a piecewise cubic polynomial:

[0066] In the formula: The coefficients of the cubic polynomial in the current interval; This is the height function for the reconstructed, continuous, smooth material profile.

[0067] S2: Calculate the slope of the line segment between every two adjacent sampling points:

[0068] S3: Nodal first derivative The core feature of the algorithm is that the nodal derivative is fitted by a weighted average of the slopes of four adjacent segments, ensuring a smooth curve and suppressing oscillations. For internal nodes... :

[0069] Boundary node-specific correction formula, beginning:

[0070]

[0071] Tail end:

[0072]

[0073] in: interpolation nodes The smooth first derivative at point .

[0074] S4: Formula for solving the coefficients of a piecewise cubic polynomial Define the interval step size: Elevation difference between intervals: ; Solve for the coefficients of the interval polynomial using the endpoint function values ​​and the endpoint first derivative constraints:

[0075] S5: Final interpolation expression make This yields a formula for calculating smooth contours that can be directly used in program iteration:

[0076] S6: Extracting key feature parameters from continuous contour curves based on interpolation reconstruction: Replacement relationships in material flow profile reconstruction: Raw radar sampling data: ; : Horizontal coordinate of the belt; Real-time height sampling value of the material; Algorithm substitution: It outputs a continuous and smooth material cross-sectional height curve, which is used for numerical integration calculation of material cross-sectional area, frame-by-frame stacking reconstruction of material flow three-dimensional morphology, and centroid offset calculation of cross-sectional area moment.

[0077] Compared to directly calculating the center of gravity of discrete points, the above scheme significantly improves the accuracy and stability of the center of gravity offset calculation. This allows the dynamic closed-loop correction to receive more reliable material flow status information, thereby generating more precise flap control commands and achieving fine-grained adjustment of the material flow direction within the curved chute. This is of great significance for solving the problem of uneven material spillage and impact at the source caused by the internal material flow trajectory deviation at the upstream curved chute transfer node, fundamentally improving the operating efficiency and equipment life of the bulk material conveying system.

[0078] In one possible implementation, the center of gravity offset can be calculated using the following method: A. Based on the material cross-sectional profile curve And the reference height curve of the unloaded belt In the effective material flow lateral range The cross-sectional area of ​​a single frame of material is calculated using numerical integration. Cross-sectional area moment and the lateral center of gravity coordinates of the material Their formulas are expressed as follows:

[0079]

[0080]

[0081] B. Adjust the horizontal center of gravity coordinates of the material. With respect to the calibrated belt center reference coordinates By comparing the values, the center of gravity offset is obtained. .

[0082] Numerical integration calculation of the cross-sectional area of ​​a single frame of material Cross-sectional area moment and the lateral center of gravity coordinates of the material This is a method for approximating the integral of a continuous function using discrete points. In this case, it's used to transform a continuous material cross-sectional profile curve into physically meaningful cross-sectional parameters. (Single-frame material cross-sectional area) Reflects the instantaneous flow rate of the material; cross-sectional area moment This reflects the degree of eccentricity in the material distribution; the lateral center of gravity coordinates of the material. This directly provides the actual centroid position of the material along the width of the conveyor belt. Calculating these parameters is a crucial step in quantifying the material's off-center loading condition.

[0083] Commonly used numerical integration methods include the trapezoidal rule and Simpson's rule. The calibrated belt center reference coordinates... It represents the ideal center position of the conveyor belt in the width direction and serves as a reference for determining the center of gravity offset. This is typically determined through precise measurement and calibration during system installation or commissioning; for example, it can be set to half the geometric width of the conveyor belt. Center of gravity offset. It directly quantifies the degree and direction of material off-center loading relative to the center of the conveyor belt.

[0084] Figure 4 This is a flowchart of the 3D reconstruction of the material flow cross-section and the calculation of the material flow's center of gravity and height. The calculation of the center of gravity coordinates... and center of gravity offset Subsequently, a risk assessment will be conducted. This application has a dual assessment mechanism for material stacking deviation and overload, which is divided into the lateral center of gravity coordinate of the material and the peak value of the longitudinal height of the profile. When the lateral center of gravity deviation exceeds the set threshold, an off-center load warning will be issued; similarly, when the peak value of the longitudinal height exceeds the set stacking height, the risk of overflow will be determined.

[0085] This mapping from geometric shape to physical centroid provides high-precision real-time data support for subsequent dynamic closed-loop correction, significantly improving the response speed and correction accuracy of the control system. This effectively avoids problems such as belt misalignment, material spillage, and increased local wear of liners in the lower-level conveyor belts, thereby improving the operational stability and economic benefits of the bulk material conveying system.

[0086] In some of the embodiments described above in this application, a control method for correcting deviation by sensing with lidar and driving an electric flapper is proposed. However, in its implementation, due to the complexity and nonlinearity of material flow, it is difficult to cope with rapidly changing working conditions by relying solely on a preset static mapping model, resulting in insufficient response accuracy and real-time adaptability of flapper adjustment, and failing to achieve precise closed-loop control of complex material flow trajectories.

[0087] Therefore, this application can set a mapping model to generate corresponding flip-board control instructions, including: A. During the system debugging phase, record the center of gravity offset response data corresponding to different combinations of electric flap angles through orthogonal experiments, and construct and store a correspondence table reflecting the flap action-material flow response; B. During the operation phase, based on the real-time calculated center of gravity offset The corresponding table is queried to match and determine the target flap angle combination to generate flap control instructions; C. A model predictive control strategy is adopted, using the current center of gravity offset and the historical angle state of each electric flap as input, to correct the corresponding table online.

[0088] The pre-built mapping model is established through a series of controlled experiments before the system's formal operation, serving as the initial decision-making basis for guiding the relationship between the flapper action and the material flow response. This model can be a lookup table, a set of empirical formulas, or a simplified mathematical model. Model predictive control strategies can utilize the system's dynamic model to predict system behavior over a future period and optimize the current control input accordingly to achieve the preset control objective. This algorithm uses the current center of gravity offset... Using the historical angle status of each electric flap as input, it can capture the dynamic characteristics of the material flow and the hysteresis effect of the flap movement.

[0089] Online correction of the correspondence table refers to the process during system operation where the model predictive control algorithm dynamically adjusts or updates the mapping relationship in the correspondence table based on the difference between the actual observed material flow response and the model prediction results, thereby enabling the control strategy to adapt to constantly changing operating conditions.

[0090] Traditional chute material flow regulation relies heavily on manual experience or simple mechanical baffles, lacking the ability to quantitatively perceive the three-dimensional shape of the material flow and thus failing to achieve millisecond-level dynamic closed-loop correction. In some embodiments described above, this application proposes using the coordinated action of multiple independently driven electric flaps to change the material flow direction within the curved chute to achieve correction. These electric flaps include a central flap at the head of the curved chute, and left and right flaps located on either side of the material flow tube. Figure 5 A schematic diagram of the left and right flip panels is shown. Figure 6 A schematic diagram of the middle flap is shown. Figure 7 The diagram shows the logic block diagram for the coordinated control of multiple flaps, controlling the independently driven electric flaps to coordinate their actions. The specific logic is as follows: When it is determined that the center of gravity offset is biased to the left side of the lower-level conveyor belt, a flap control command is generated to control the left flap to close and the right flap to open. When it is determined that the center of gravity offset is biased to the right side of the lower-level conveyor belt, a flap control command is generated to control the left flap to open and the right flap to close. When the center of gravity offset is determined to be in the middle, control each electric flap to maintain the current opening steady state.

[0091] Electric flappers can be pneumatically or hydraulically driven, employing either a translational or swinging baffle. The position of the baffle is changed by controlling air or hydraulic pressure. The intermediate flapper is typically located at the inlet or middle of a curved chute, primarily used for initial guidance or diversion of the material flow entering the chute. The left and right flappers are symmetrically positioned on either side of the discharge pipe; they are key actuators for lateral material flow correction, guiding the material flow direction by altering the resistance on both sides.

[0092] When the center of gravity shifts to the left side of the downstream conveyor belt, the system generates specific flap control commands. For example, this command might instruct the left flap to close, reducing its opening or closing it completely, thus creating greater resistance on the left side of the chute; simultaneously, it instructs the right flap to open, increasing its opening or opening completely to reduce resistance on the right. This asymmetrical resistance distribution forces the material flow to shift to the right, pulling the center of gravity back to the center of the conveyor belt. Conversely, when the center of gravity shifts to the right side of the downstream conveyor belt, the system generates the opposite flap control commands, opening the left flap and closing the right flap. This creates greater resistance on the right side of the chute, guiding the material flow to shift to the left to correct the center of gravity shift on the right. When the center of gravity shift is centered, i.e., the material flow is in the ideal center position or within the allowable fluctuation range, it maintains its current steady-state opening. This means the flaps will not perform additional actions to avoid unnecessary adjustments and energy consumption, while maintaining system stability.

[0093] For extreme working conditions (such as signs of material blockage or foreign object jamming), this system also includes a safety interlock mechanism for abnormal working conditions: When the calculated cross-sectional area of ​​a single frame of material is obtained A sudden change occurs that exceeds 200% of the normal operating average, or a center of gravity shift is detected for three consecutive control cycles. When all exceed the allowable threshold, it is judged as an abnormal working condition; an audible and visual alarm signal is immediately issued, the issuance of automatic adjustment commands is interrupted, all electric flaps are controlled to reset to the preset safe angle position, and a linkage shutdown request is sent to the lower-level conveyor system.

[0094] Through the above technical solution, this application effectively solves the problems of delayed adjustment response and uncoordinated actions caused by the lack of clear control logic in traditional solutions. This solution achieves precise intervention in material flow by functionally dividing the electric tilting plate and establishing refined collaborative control logic based on the direction of center of gravity offset. This not only improves the targeting and efficiency of the correction action and avoids unnecessary adjustments, but also significantly enhances the stability of the entire control system and the service life of the actuators. Ultimately, this solution can more effectively prevent problems such as downstream conveyor belt deviation, material spillage, and accelerated local wear of the liner, ensuring the safe and stable operation of the bulk material conveying system.

[0095] Using a curved chute transfer point between the crushing station and the main conveyor belt in an iron ore beneficiation plant as a practical application scenario, the real-time control process of the system of this invention is demonstrated. Combined with... Figure 3 , Figure 5 , Figure 6 and Figure 8 The system deployment is as follows: LiDAR sensing unit: Installed behind the material guide chute of the lower conveyor belt, covering the material flow area of ​​the entire bandwidth, used to collect three-dimensional point cloud data of the material transported by the conveyor belt in real time; Edge controller: Communicates with the lidar sensing unit and is responsible for core functions of data processing and control, including point cloud analysis, command generation and coordination; Multi-flipper drive unit: It consists of 3 sets of electric flippers with wear-resistant lining plates inside, which are respectively set inside the head cover of the curved chute and on both sides of the discharge pipe. Each flipper is equipped with an independent servo motor and absolute encoder, which can achieve stepless adjustment from 0 to 60°. Flip-board angle feedback unit: It communicates bidirectionally with the multi-flip-board drive execution unit and the edge controller to provide real-time feedback on the actual angle information of the flip-board for closed-loop control calibration.

[0096] The following details the collaborative flow and control logic of each unit in the system, based on the timing nodes of the timing diagram (T0 - T100ms): Time T0 (initial state): The system enters the starting point of a complete control loop. At this time, the lidar sensing unit activates the infrared laser pulse emission mode and begins the initial scan of the material on the downstream conveyor belt; the edge controller is in standby mode, waiting for the lidar sensing unit to transmit point cloud data.

[0097] T20ms time (data acquisition and preliminary processing): The lidar sensing unit completes one full lateral scan cycle (full-range scan within the bandwidth). After the scan is completed, the raw point cloud data (containing the spatial coordinates of all points on the material surface) is transmitted to the edge controller.

[0098] After receiving the raw point cloud, the edge controller immediately initiates a three-stage filtering preprocessing: Straight-through filtering: Set the effective distance measurement range (e.g., 0-2000mm) and filter out invalid points at long distances (e.g., frame, idler rollers, etc.); Statistical outlier filtering: Traverse each point cloud neighborhood (5×5 point matrix), calculate the average distance and standard deviation, and remove isolated dust noise points that deviate from the mean by 3 times the standard deviation; Multi-echo threshold filtering: Enable multi-echo acquisition mode to retain only the primary reflection signal from the material surface and filter out secondary reflection noise from dust.

[0099] After three-stage filtering, a clean point cloud of material cross-sectional height is obtained.

[0100] T30ms timeframe (point cloud reconstruction and centroid calculation): The edge controller uses Akima piecewise cubic polynomial interpolation to transform the filtered discrete point cloud into a smooth and continuous material profile curve. This algorithm uses a weighted fitting of the slopes of four adjacent segments to ensure the reconstructed curve is both smooth and suppresses oscillations, accurately reproducing the undulating surfaces of real materials such as coal piles and sand.

[0101] Next, the edge controller calculates the centroid offset of the material cross section: by integrating the area of ​​the smooth profile curve, the geometric center of the cross section is obtained, and compared with the belt center baseline, the centroid offset is obtained (such as 80mm right offset in the example).

[0102] T80ms time (instruction generation and cooperative control): When the center of gravity offset exceeds a preset threshold (e.g., 80mm), the edge controller initiates model predictive control (MPC) and generates flap adjustment commands. These commands include: increase the opening of the left flap by 12°, decrease the opening of the right flap by 8°, and keep the middle flap in its original position. Simultaneously, the edge controller sends coordinated control commands to the multi-flap drive execution unit, specifying the target adjustment angle for each flap.

[0103] T100ms timeframe (execution and closed-loop feedback): The multi-flip-board drive execution unit completes the flip-board angle adjustment action (from the initial angle to the target angle) within 180ms. After the adjustment is completed, the flip-board angle feedback unit immediately feeds back the actual angle information of each flip-board to the edge controller.

[0104] The edge controller adjusts subsequent control strategies (such as optimizing the flap adjustment range) based on the difference between the actual angle and the target angle, forming a closed-loop feedback until the center of gravity offset returns to the allowable range (such as within 15mm).

[0105] Based on the above embodiments, this application can achieve the following beneficial effects: Precise perception: Using lidar as the sensing unit, the limitations of dust and changing lighting environments are overcome, and millimeter-level three-dimensional reconstruction of the material flow cross section is achieved.

[0106] Dynamic correction: The traditional "open-loop manual adjustment" is upgraded to "closed-loop automatic control". Multiple electric flaps work together to cope with complex and ever-changing working conditions and fundamentally eliminate material misalignment.

[0107] Preventive maintenance: The historical data of material flow deviation and the frequency of flap operation recorded by the system can be used as a basis for predicting the wear of chute liners and managing equipment health, thereby extending the equipment's lifespan.

[0108] High adaptability: The control algorithm supports online learning and parameter self-tuning, and can adapt to different material characteristics (such as sticky mineral powder, large ore, etc.) and different chute structures without frequent recalibration.

[0109] Figure 9 This is a schematic diagram of the structure of the curved chute material discharge trajectory control device based on lidar feedback provided in this application embodiment. The curved chute is equipped with multiple independently drivable electric flaps. The device includes: The lidar scanning module 910 is used to control the lidar installed above the lower conveyor belt to scan in real time and obtain three-dimensional point cloud data of the material surface on the lower conveyor belt. The continuous contour reconstruction module 920 is used to filter the three-dimensional point cloud data to remove dust interference, and uses a piecewise polynomial interpolation algorithm to fit the filtered discrete height point cloud into a continuous and smooth material cross-sectional contour curve. The center of gravity offset calculation module 930 is used to combine the material cross-sectional profile curves obtained in each frame with the belt reference height curve obtained by no-load calibration to eliminate the interference of the deformation reference of the lower conveyor belt itself, calculate the three-dimensional cross-sectional characteristics of the real-time material flow, and calculate the center of gravity offset of the material's lateral center of gravity relative to the center of the lower conveyor belt. The dynamic closed-loop correction module 940 is used to generate corresponding flap control commands based on a pre-built mapping model when the center of gravity offset exceeds a preset allowable threshold, and to control multiple independently driven electric flaps distributed on the inner side of the curved chute head cover and both sides of the material flow surface of the drop pipe to work together until the center of gravity offset corresponding to the material flow state returns to within the allowable threshold.

[0110] This application also discloses a computer-readable storage medium. Specifically, the computer-readable storage medium is used to store a computer program, which, when executed by a processor, implements the methods described in the above-described method embodiments. Those skilled in the art will understand that implementing all or part of the processes in the methods described in the above-described embodiments of this application can be accomplished by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0111] This specific embodiment is merely an explanation of the present invention and is not intended to limit the invention. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they fall within the scope of the claims of the present invention.

Claims

1. A method for controlling the trajectory of a curved chute for material feeding based on lidar feedback, characterized in that, The curved chute is equipped with multiple independently driveable electrically operated flaps, and the control method includes: The laser radar installed above the lower conveyor belt is controlled to scan in real time to obtain three-dimensional point cloud data of the material surface on the lower conveyor belt; The three-dimensional point cloud data is filtered to remove dust interference, and a piecewise polynomial interpolation algorithm is used to fit the filtered discrete height point cloud into a continuous and smooth material cross-sectional profile curve. The cross-sectional profile curves of the material obtained in each frame are combined with the belt reference height curve obtained by no-load calibration to eliminate the interference of the deformation reference of the lower conveyor belt itself, calculate the three-dimensional cross-sectional characteristics of the real-time material flow, and calculate the offset of the material's lateral center of gravity relative to the center of gravity of the lower conveyor belt. When the center of gravity offset exceeds the preset allowable threshold, a corresponding flap control command is generated based on the pre-built mapping model, and multiple independently driven electric flaps distributed on the inner side of the curved chute head cover and both sides of the material flow surface of the drop pipe are controlled to work together until the center of gravity offset corresponding to the material flow state returns to within the allowable threshold.

2. The control method according to claim 1, characterized in that, The acquisition of three-dimensional point cloud data of the material surface on the lower conveyor belt includes: The lidar is controlled to emit infrared laser pulses along the width of the lower conveyor belt at a preset transverse scanning frequency, and to receive echo pulses diffusely reflected from the material surface. Based on the time difference between laser pulse emission and reception at each measurement point and the speed of light The distance values ​​of each measurement point are calculated using the following formula. ; The distance value sequence obtained from a single transverse scan is used as a discrete height point cloud of a material cross section. Multiple frames of the discrete height point cloud are then stitched together along the movement direction of the lower conveyor belt to reconstruct the three-dimensional point cloud data.

3. The control method according to claim 1, characterized in that, The filtering process for the 3D point cloud data includes: The lidar is controlled to activate the multi-echo acquisition mode. Threshold filtering is performed on multiple echo signals received in the same measurement direction. Only the distance value corresponding to the main echo signal with the highest intensity and shortest delay is retained. Secondary reflection noise caused by dust is eliminated to obtain a pure discrete height point cloud of the material cross section.

4. The control method according to claim 3, characterized in that, Before performing multi-echo threshold filtering on the 3D point cloud data, the method further includes: Straight-through filtering: Set the effective ranging interval Only retain distance values satisfy The point cloud is filtered to remove invalid point clouds that are far away from the material area, such as those on the frame or rollers. Statistical outlier filtering: Iterate through the average neighborhood distance of each point cloud and remove isolated dust noise points whose average neighborhood distance deviates from the overall mean by more than 3 standard deviations.

5. The control method according to any one of claims 2 to 4, characterized in that, The piecewise polynomial interpolation algorithm uses the Akima piecewise cubic polynomial interpolation algorithm, including: Let the sequence of discrete height point cloud sampling nodes after filtering be... ,in The horizontal axis is monotonically increasing. And the ordinate Mapped to real-time height sampling value of material ; In any subinterval Construct a piecewise cubic polynomial: In the formula, The fitted continuous and smooth cross-sectional profile curve of the material. Calculate the slope of the line segment between every two adjacent sampling points. : For internal nodes derivative value for: Define the interval step size and interval elevation difference The coefficients of the cubic polynomial for the current interval are obtained by solving the endpoint constraints: in, interpolation nodes The smooth first derivative at point .

6. The control method according to claim 5, characterized in that, Calculating the center of gravity offset includes: Based on the material cross-sectional profile curve And the reference height curve of the unloaded belt In the effective material flow lateral range The cross-sectional area of ​​a single frame of material is calculated using numerical integration. Cross-sectional area moment and the lateral center of gravity coordinates of the material The formulas are expressed as follows: The lateral centroid coordinate of the material With respect to the calibrated belt center reference coordinates By comparison, the center of gravity offset is obtained. .

7. The control method according to claim 6, characterized in that, The generation of corresponding flip-board control commands based on the pre-built mapping model includes: During the system debugging phase, orthogonal experiments were conducted to record the center of gravity offset response data under different combinations of electric flap angles, and a correspondence table reflecting the flap action and material flow response was constructed and stored. During operation, based on the real-time calculated center of gravity offset The corresponding table is queried to match and determine the target flap angle combination to generate the flap control command; The model predictive control logic is used to correct the corresponding table online, taking the current center of gravity offset and the historical angle state of each electric flap as input.

8. The control method according to claim 1, characterized in that, The plurality of electric flaps includes a middle flap located at the head of the curved chute, and a left flap and a right flap located on both sides of the discharge pipe; the independently driven electric flaps coordinate to move, including: When it is determined that the center of gravity offset is biased towards the left side of the lower conveyor belt, a flap control command is generated to control the left flap to close and the right flap to open. When it is determined that the center of gravity offset is biased towards the right side of the lower conveyor belt, a flap control command is generated to control the left flap to open and the right flap to close. When it is determined that the center of gravity offset is in the center, control each of the electric flaps to maintain the current opening degree in a steady state.

9. A curved chute material discharge trajectory control device based on lidar feedback, characterized in that, The curved chute is equipped with multiple independently driveable electrically operated flaps, and the device includes: The lidar scanning module is used to control the lidar installed above the lower conveyor belt to scan in real time and acquire three-dimensional point cloud data of the material surface on the lower conveyor belt. The continuous contour reconstruction module is used to filter the three-dimensional point cloud data to remove dust interference. It uses a piecewise polynomial interpolation algorithm to fit the filtered discrete height point cloud into a continuous and smooth material cross-sectional contour curve. The center of gravity offset calculation module is used to combine the material cross-sectional profile curves obtained in each frame with the belt reference height curve obtained by no-load calibration to eliminate the interference of the deformation reference of the lower conveyor belt itself, solve the three-dimensional cross-sectional characteristics of the real-time material flow, and calculate the center of gravity offset of the material's lateral center of gravity relative to the center of the lower conveyor belt. The dynamic closed-loop correction module is used to generate corresponding flap control commands based on a pre-built mapping model when the center of gravity offset exceeds a preset allowable threshold. It then controls multiple independently driven electric flaps distributed on the inner side of the curved chute head cover and both sides of the material flow surface of the drop pipe to work together until the center of gravity offset corresponding to the material flow state returns to within the allowable threshold.

10. A computer-readable storage medium storing at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded by a processor and executed by the method for controlling the trajectory of a curved chute based on lidar feedback as described in any one of claims 1-9.

Images