Self-adaptive material taking method and system of stacker-reclaimer based on dynamic boundary model

Abstract

The invention relates to a self-adaptive reclaiming method and system for a stacker-reclaimer based on a dynamic boundary model. The method comprises the following steps: S1, dynamic boundary modeling; s2, safe trajectory planning; s3, load self-adaptive adjustment is carried out; wherein the step S2 and the step S3 are executed in parallel and are mutually coupled to form a double-constraint control loop. The system comprises a sensing module, a control module and an execution module. According to the invention, the crossing from static presetting to dynamic sensing and from passive response to active safety is realized, the problems of non-uniform material taking, high collision risk and low efficiency are effectively solved, and the operation safety, efficiency and intelligent level are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for bulk material handling equipment, and in particular to an adaptive material handling method and system for a stacker-reclaimer based on a dynamic boundary model. Background Technology

[0002] Stacker-reclaimers are key pieces of equipment in large bulk material yards for stacking and retrieving materials. Their level of automation and intelligence directly affects the efficiency and safety of the entire bulk material handling system.

[0003] Existing flow control methods for stacker-reclaimers, such as the Chinese invention patent "A Flow Control Method, Device, and Stacker-Reclaimer (CN110980318B)," adjust the rotational speed of the bucket wheel or conveyor arm by comparing the real-time flow rate with a fixed preset value. While this method achieves stable flow control to a certain extent, it still has significant drawbacks:

[0004] First, its control is based on a fixed preset flow rate value, which cannot adapt to sudden changes in material handling resistance caused by material pile collapse, landslides, and changes in material particle size and moisture during the material handling process, resulting in a rigid control strategy.

[0005] Secondly, this is a passive response control, which only adjusts after the flow deviation occurs, resulting in response lag and is prone to system oscillation or overshoot;

[0006] Finally, its perception dimension is singular, relying only on parameters such as flow rate or motor current, lacking real-time perception and understanding of the three-dimensional shape of the work object—the material pile. This causes the control system to "not see" the real environmental changes, and cannot fundamentally solve safety and efficiency problems such as "gnawing at the material pile" (collapse caused by impacting the root of the material pile) or "empty digging" (the material head does not contact the material).

[0007] In addition, another type of automated material handling scheme based on a preset fixed path (such as the trapezoidal method or the layered method) relies entirely on the initial, static material pile model. Once the actual material pile shape deviates from the model due to natural collapse or multiple material handling operations, the automation process will fail and manual intervention is still required. The degree of automation and reliability are not ideal.

[0008] Therefore, existing technologies urgently need an intelligent material handling solution that can achieve closed-loop linkage of "perception-modeling-decision-execution". Its core lies in the ability to perceive and predict the dynamic changes in the shape of the material pile in real time, and adjust the material handling strategy autonomously, safely and efficiently accordingly. Summary of the Invention

[0009] This invention aims to overcome the aforementioned deficiencies of existing technologies and provides an adaptive material handling method and system for stacker-reclaimers based on a dynamic boundary model. This invention constructs a dynamic boundary model in real time and, based on this model, implements dual constraint control of safe trajectory planning and adaptive load adjustment, ultimately achieving safe, efficient, and adaptive material handling.

[0010] To achieve the above objectives, the present invention adopts the following technical solution:

[0011] The adaptive material handling method for stacker-reclaimers based on dynamic boundary models includes the following steps:

[0012] S1. Dynamic boundary modeling:

[0013] The three-dimensional point cloud data of the material reclaiming working face is acquired in real time by a three-dimensional scanning device installed at the front end of the stacker-reclaimer cantilever.

[0014] The three-dimensional point cloud data is registered with the pre-stored benchmark model and three-dimensional difference calculation is performed to identify the difference area between the actual outline of the material pile and the model outline, and based on this, a dynamic and effective material picking boundary model that reflects the current actual working conditions is updated and generated.

[0015] S2, Safe Trajectory Planning:

[0016] Based on the effective material reclaiming boundary model and combined with the pre-stored material repose angle parameters, the dynamic safety distance between the material reclaiming head and the material pile boundary is calculated, and the movement trajectory of the material reclaiming head is planned or adjusted in real time to ensure that the material reclaiming operation is always within the safety boundary.

[0017] S3, Load Adaptive Adjustment:

[0018] Real-time monitoring of the load current of the material handling drive motor;

[0019] Using load current as feedback, the pitch angle of the feed head and / or the travel speed of the whole machine are dynamically adjusted by an adaptive PID controller to keep the feed power stable within the preset target range.

[0020] Steps S2 and S3 are executed in parallel and coupled to each other, forming a dual-constraint control loop. The control logic of this loop is as follows:

[0021] First priority: When step S2 determines that there is a risk of collision or collapse, its output trajectory correction command will override the load adjustment command of step S3.

[0022] Second priority: In the absence of the aforementioned risks, the expected trajectory output by step S2 and the optimized load parameters output by step S3 are coupled and calculated by a central processing unit to generate the final control command to drive the actuator to move.

[0023] Specifically, this also includes: S4, iterative learning and prediction:

[0024] Record the dynamic boundary model sequence, load current data, and corresponding operating parameters over multiple consecutive operating cycles;

[0025] Based on the recorded data, a boundary evolution prediction model is trained using a recurrent neural network to predict the evolution trend of the stockpile boundary in the next operating cycle.

[0026] The predicted boundary trends are used as prior information and applied to the initial path planning of the next work cycle to achieve forward control.

[0027] Specifically, in S4, a boundary evolution prediction model is trained by a recurrent neural network. Specifically, a long short-term memory network is used to perform time-series modeling of the dynamic boundary model sequence to predict the stockpile profile at a specific time step in the future.

[0028] Specifically, in S1, the 3D scanning equipment consists of a LiDAR and a 3D stereo camera.

[0029] Specifically, the baseline model is either a three-dimensional model of the stockpile established based on the initial scan, or a stockpile morphology model predicted based on historical operation data.

[0030] Specifically, in S2, the dynamic safety distance is calculated as follows: based on the slope angle of the effective material reclaiming boundary model and the angle of repose of the material, the minimum safe operating distance to prevent the material pile from collapsing is calculated, and this distance is used as a rigid constraint condition for trajectory planning.

[0031] In particular, in S3, the operating parameters of the material handling head are dynamically adjusted, which specifically includes decoupling control of the tilting mechanism, slewing mechanism and the overall traveling mechanism of the material handling arm.

[0032] The load adaptive adjustment loop is configured to independently and collaboratively adjust the pitch angle, rotation speed and travel speed according to changes in load current, so that the material handling system maintains constant power or optimal power while maintaining the stability of the movement of each mechanism.

[0033] An adaptive material handling system for a stacker-reclaimer based on a dynamic boundary model is used to execute an adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model, including:

[0034] The perception module includes:

[0035] The 3D scanning unit includes a lidar and a 3D camera installed on the end of the stacker reclaimer cantilever near the reclaiming head, used to scan the reclaiming working surface from the front to obtain complete point cloud data of the material pile outline;

[0036] The load detection unit includes a current sensor and a power transmitter installed in the power circuit of the material handling drive motor, which are used to directly measure the real-time load of the material handling drive motor.

[0037] The pose sensing unit includes:

[0038] The tilt sensor is installed on the cantilever of the stacker-reclaimer and is used to directly measure the pitch angle of the cantilever.

[0039] A rotary encoder, installed in the slewing drive mechanism of the cantilever, is used to detect the slewing angle of the cantilever.

[0040] A high-precision positioning module is used to detect the position of the entire machine, including the Gray busbar laid along the entire length of the stacker-reclaimer track, and the address detector installed on the stacker-reclaimer traveling mechanism;

[0041] The control module, connected to each unit of the sensing module via an industrial bus, is mounted on an industrial computer and includes:

[0042] The core processing unit serves as the core of the system's computation;

[0043] The dynamic boundary modeling unit runs on the core processing unit and is configured to perform point cloud registration and 3D difference calculation to complete the dynamic boundary modeling steps.

[0044] The safety trajectory planning unit runs on the core processing unit and is configured to perform dynamic safety distance calculation and trajectory planning to complete the safety trajectory planning steps.

[0045] The load adaptive adjustment unit runs on the core processing unit and is configured to run an adaptive PID control algorithm to complete the load adaptive adjustment steps.

[0046] The decision fusion unit, running on the core processing unit, is configured to receive the outputs of the safety trajectory planning unit and the load adaptive adjustment unit, and perform priority judgment and instruction coupling calculation based on the logic of the dual constraint control loop to generate the final control instruction.

[0047] The digital twin unit, running in the industrial computer or a host computer communicating with it, is used to build a mirror model in virtual space that is synchronized with the physical material yard and equipment status based on real-time data, and to simulate, verify and optimize control commands.

[0048] The execution module, connected to the control module via signals, is used to receive and execute control commands, and includes:

[0049] Pitch drive mechanism, used to control the pitch angle of the cantilever;

[0050] A slewing drive mechanism is used to control the slewing angle of the cantilever.

[0051] The walking drive mechanism is used to control the movement of the entire machine along the track;

[0052] The material handling drive mechanism is used to drive the material handling head to rotate.

[0053] The beneficial effects of this invention are:

[0054] Significantly improved safety: By constructing a dynamic boundary model in real time and calculating the dynamic safety distance in conjunction with the material's angle of repose, it can proactively prevent collisions between the material take-off head and the material pile, as well as the collapse of the material pile, fundamentally solving safety hazards such as "gnawing on the material pile".

[0055] High material handling efficiency and stability: Through load adaptive adjustment and multi-mechanism decoupling control, constant power or optimal power material handling is achieved within the equipment's load-bearing limit, ensuring stable flow rate while avoiding equipment overload and improving overall material handling efficiency.

[0056] High degree of automation and intelligence: The system possesses a fully closed-loop autonomous operation capability encompassing "perception-modeling-decision-execution". A unique "dual-constraint control loop" ensures the synergistic optimization of safety and efficiency, while priority logic enables the system to make correct decisions under complex operating conditions, greatly reducing the need for manual intervention.

[0057] Strong adaptability and foresight: Through iterative learning and LSTM prediction models, the system can learn the evolution law of the material pile and predict future boundary changes, thereby achieving forward-looking control, reducing response delay, and making the control process smoother and more precise.

[0058] The system boasts high reliability: It employs a fusion perception system combining LiDAR and 3D cameras, along with simulation verification using digital twin technology, forming a multi-layered reliability guarantee that ensures stable operation of the system under various working conditions. Attached Figure Description

[0059] Figure 1 This is a flowchart of the method of the present invention;

[0060] Figure 2 This is a block diagram of the sensing module structure of the system of the present invention;

[0061] Figure 3 This is a block diagram of the control module structure of the system of the present invention;

[0062] Figure 4 This is a block diagram of the execution module structure of the system of the present invention;

[0063] Figure 5 This is a block diagram of the overall structure of the system of the present invention;

[0064] The following will describe in detail, with reference to the accompanying drawings, embodiments of the present invention. Detailed Implementation

[0065] The present invention will be further described below with reference to embodiments:

[0066] like Figures 1-5 As shown, the adaptive material handling method for stacker-reclaimers based on a dynamic boundary model includes the following steps:

[0067] S1. Dynamic boundary modeling:

[0068] The three-dimensional point cloud data of the material reclaiming working face is acquired in real time by a three-dimensional scanning device installed at the front end of the stacker-reclaimer cantilever.

[0069] The three-dimensional point cloud data is registered with the pre-stored benchmark model and three-dimensional difference calculation is performed to identify the difference area between the actual outline of the material pile and the model outline, and based on this, a dynamic and effective material picking boundary model that reflects the current actual working conditions is updated and generated.

[0070] The 3D scanning equipment includes LiDAR and a 3D camera. The baseline model is either a 3D model of the stockpile established based on the initial scan, or a stockpile shape model predicted based on historical operation data.

[0071] Specifically, after system startup, the initial 3D model of the material pile for this operation is loaded from memory as the reference model. If no initial model exists, the 3D scanning unit can be controlled to perform a global scan to establish one. During the material handling operation, the 3D scanning unit scans the material handling working surface in real time at a frequency of 10Hz. The dynamic boundary modeling unit in the control module receives the point cloud data and first performs filtering and noise reduction preprocessing. Then, the real-time point cloud is registered with the reference model to find the optimal alignment position between the two in space. Next, 3D difference calculation is performed, comparing point by point to identify areas where the difference exceeds a set threshold (e.g., 0.1 meters). These areas are where the shape of the material pile has changed. Finally, the reference model is updated based on the difference areas to generate the effective material handling boundary model for the current moment. This model is a digital model containing 3D information of the safe material handling contour.

[0072] This step offers the following advantages: It achieves a fundamental shift from "static preset" to "dynamic perception," overcoming the shortcomings of existing technologies that rely on fixed models or preset paths, enabling the control system to "see" and understand the real and continuous changes in the operating environment. It provides precise environmental context for intelligent decision-making: This dynamic boundary model serves as the data foundation and prerequisite for all subsequent safety planning and efficiency optimization, ensuring the accuracy and reliability of decisions.

[0073] S2, Safe Trajectory Planning:

[0074] Based on the effective material reclaiming boundary model and combined with the pre-stored material repose angle parameters, the dynamic safety distance between the material reclaiming head and the material pile boundary is calculated, and the movement trajectory of the material reclaiming head is planned or adjusted in real time to ensure that the material reclaiming operation is always within the safety boundary.

[0075] The calculation method for dynamic safety distance is as follows: based on the slope angle of the effective material reclaiming boundary model and the angle of repose of the material, the minimum safe operating distance to prevent the material pile from collapsing is calculated, and this distance is used as a rigid constraint condition for trajectory planning.

[0076] Specifically, the safety trajectory planning unit reads the effective material handling boundary model. Based on the current material type being handled (e.g., coal, ore), it retrieves the angle of repose from the database (e.g., the angle of repose for coal is 38°). Combining this with the slope angle of the boundary model, it calculates the dynamic safety distance required to prevent collapse (e.g., the safety distance increases when the slope is steeper). Based on this safety distance and the material handling task, it plans the desired movement trajectory of the material handling head (including pitch, rotation, and travel settings). It continuously checks whether the current equipment posture and the planned trajectory meet the safety constraints; if not, it marks it as "a safety risk exists."

[0077] This step offers the following advantages: It achieves proactive safety protection, upgrading safety control from a passive mode of "alarming after overload" to a proactive mode of "calculating risks and avoiding them in advance." By introducing the key physical attribute of the angle of repose, the calculation of safe distances becomes more scientific and personalized (for different materials), fundamentally preventing "pile-eating" and pile collapse accidents. It ensures absolute safety boundaries during operations: By integrating safe distances as rigid constraints into trajectory planning, it guarantees that the material reclaiming head always operates within an absolutely safe area, greatly improving the inherent safety level of the system.

[0078] S3, Load Adaptive Adjustment:

[0079] Real-time monitoring of the load current of the material handling drive motor;

[0080] Using load current as feedback, the pitch angle of the feed head and / or the travel speed of the whole machine are dynamically adjusted by an adaptive PID controller to keep the feed power stable within the preset target range.

[0081] Among them, dynamically adjusting the operating parameters of the material receiving head specifically includes decoupling control of the pitch mechanism, slewing mechanism and overall machine travel mechanism of the material receiving arm;

[0082] The load adaptive adjustment loop is configured to independently and collaboratively adjust the pitch angle, rotation speed and travel speed according to changes in load current, so that the material handling system maintains constant power or optimal power while maintaining the stability of the movement of each mechanism.

[0083] Specifically, the load adaptive adjustment unit reads the load current of the material handling drive motor in real time. It compares the current value with a preset "target power range." Using an adaptive PID controller (whose P, I, and D parameters can be dynamically adjusted according to the error magnitude), it calculates the adjustment amount required for stable power in the pitch angle, slewing speed, or travel speed.

[0084] This step offers the following advantages: It enables intelligent and precise control of equipment load: Through an adaptive PID algorithm, the system can adapt to different material resistance characteristics, achieving smooth, overshoot-free power regulation, avoiding equipment overload and mechanical shock, and protecting equipment lifespan. It improves efficiency and stability: Through multi-mechanism decoupled control, the system can flexibly optimize between various objectives such as maintaining constant power or pursuing maximum efficiency, ensuring stable and high-efficiency material intake flow and overcoming the limitations of single adjustment methods.

[0085] Steps S2 and S3 are executed in parallel and coupled to each other, forming a dual-constraint control loop. The control logic of this loop is as follows:

[0086] First priority: When step S2 determines that there is a risk of collision or collapse, its output trajectory correction command will override the load adjustment command of step S3.

[0087] Second priority: In the absence of the aforementioned risks, the expected trajectory output by step S2 and the optimized load parameters output by step S3 are coupled and calculated by a central processing unit to generate the final control command to drive the actuator to move.

[0088] Specifically, the decision fusion unit receives the "safety status" and "desired trajectory" from S2, and the "load adjustment parameters" from S3.

[0089] Applying the first priority logic: If S2 reports "there is a safety risk", the decision fusion unit will ignore the load adjustment parameters of S3, directly use the trajectory correction command calculated by S2, and immediately control the equipment to move away from the danger zone.

[0090] Applying the second priority logic: If there is no safety risk, the decision fusion unit couples the desired trajectory of S2 with the load adjustment parameters of S3 for calculation. For example, based on the trajectory planned by S2, the travel speed is fine-tuned according to the output of S3 to achieve the highest efficiency under the premise of safety. Finally, the fused control command is generated.

[0091] Before an instruction is sent to the physical actuators, it can first be sent to the digital twin unit. The digital twin unit simulates the execution of the instruction in a virtual environment, predicts the equipment status and material yard changes after execution, and performs collision and boundary checks. If the verification passes, the instruction is then issued; if a problem is found, it is fed back to the decision-making unit for recalculation. The execution module receives the final instruction, drives the actions of each mechanism, and completes one control cycle.

[0092] The advantages of this step are: It resolves the conflict between safety and efficiency control: It clearly stipulates that safety is always the top priority, while maximizing operational efficiency within safe limits. It enhances the system's decision-making intelligence and reliability: This loop enables the system to act like an experienced operator, weighing pros and cons and making the most rational decisions under complex operating conditions, significantly enhancing the reliability and intelligence of the system's behavior.

[0093] S4. Iterative Learning and Prediction:

[0094] Record the dynamic boundary model sequence, load current data, and corresponding operating parameters over multiple consecutive operating cycles;

[0095] Based on the recorded data, a boundary evolution prediction model is trained using a recurrent neural network to predict the evolution trend of the stockpile boundary in the next operating cycle.

[0096] The predicted boundary trends are used as prior information and applied to the initial path planning of the next work cycle to achieve forward control.

[0097] Specifically, the boundary evolution prediction model is trained by a recurrent neural network. This involves using a long short-term memory network to perform time-series modeling of the dynamic boundary model sequence in order to predict the stockpile profile at a specific future time step.

[0098] Specifically, once a reclaiming shift or a complete reclaiming layer is completed, the system initiates a learning cycle. It records all valid boundary model sequences and load data within this cycle. Using this data, a time-series prediction model is trained using a Long Short-Term Memory (LSTM) network. This model learns "how the stockpile boundary evolves step by step under a specific reclaiming pattern." The trained model is then used in the next work cycle. At the start of the cycle, it can predict the boundary evolution trend for the entire cycle based on the initial state and provide this trend as "prior knowledge" to the safety trajectory planning unit, thereby achieving more forward-looking path planning and further optimizing operational efficiency and stability.

[0099] The advantages of this step include: It upgrades from "real-time response" to "proactive prediction": the system not only responds to the current state but also predicts changes several steps ahead, allowing for better path planning in advance, reducing control delays and abrupt stops and starts, resulting in a smoother and more efficient workflow. It also possesses continuous self-optimization capabilities: by constantly learning from its own operational history, the system becomes increasingly "intelligent," making more accurate predictions, optimizing paths, and continuously improving its level of automation over time.

[0100] An adaptive material handling system for a stacker-reclaimer based on a dynamic boundary model is used to execute an adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model, including:

[0101] The perception module includes:

[0102] The 3D scanning unit includes a LiDAR and a 3D camera mounted on the end of the stacker-reclaimer cantilever near the reclaiming head. It is used to scan the reclaiming working surface from the front to obtain complete point cloud data of the stockpile outline. Specifically, the 3D scanning unit consists of a LiDAR and a 3D camera, rigidly mounted together within a protective cover approximately 5 meters from the reclaiming head at the front end of the cantilever. The LiDAR is responsible for acquiring high-precision geometric point clouds at a greater distance and under complex lighting conditions, while the 3D camera provides high-resolution texture and dense point cloud details at close range. Both are jointly calibrated, ensuring a unified coordinate system.

[0103] The load detection unit includes a current sensor and a power transmitter installed in the power circuit of the material handling drive motor, which are used to directly measure the real-time load of the material handling drive motor. Specifically, the current sensor and the power transmitter are directly connected in series in the main circuit of the material handling drive motor to collect the motor's operating current and power in real time.

[0104] The pose sensing unit includes:

[0105] The tilt sensor is installed on the cantilever of the stacker-reclaimer and is used to directly measure the pitch angle of the cantilever.

[0106] A rotary encoder, installed in the slewing drive mechanism of the cantilever, is used to detect the slewing angle of the cantilever.

[0107] A high-precision positioning module is used to detect the position of the entire machine, including the Gray busbar laid along the entire length of the stacker-reclaimer track, and the address detector installed on the stacker-reclaimer traveling mechanism;

[0108] Specifically, the tilt sensor is directly mounted on the main steel beam of the cantilever to measure the pitch angle. The rotary encoder is integrated into the motor tail of the cantilever slewing drive mechanism. The high-precision positioning module adopts the Gray busbar positioning system, with the Gray busbar laid along the entire length of the track. The address detector is installed on the traveling mechanism of the stacker-reclaimer gantry, providing absolute position information for the entire machine with a positioning accuracy of ±2mm.

[0109] The control module, connected to each unit of the sensing module via an industrial bus, is mounted on an industrial computer and includes:

[0110] The core processing unit serves as the core of the system's computation;

[0111] The dynamic boundary modeling unit runs on the core processing unit and is configured to perform point cloud registration and 3D difference calculation to complete the dynamic boundary modeling steps. Specifically, it uses the PCL point cloud library and custom algorithms to implement point cloud registration (such as the ICP algorithm) and difference calculation.

[0112] The safety trajectory planning unit runs on the core processing unit and is configured to perform dynamic safety distance calculation and trajectory planning to complete the safety trajectory planning steps; specifically, it has a built-in path planning algorithm and a material repose angle database.

[0113] The load adaptive adjustment unit runs on the core processing unit and is configured to run an adaptive PID control algorithm to complete the load adaptive adjustment steps.

[0114] The decision fusion unit, running on the core processing unit, is configured to receive the outputs of the safety trajectory planning unit and the load adaptive adjustment unit, and perform priority judgment and instruction coupling calculation based on the logic of the dual constraint control loop to generate the final control instruction.

[0115] The digital twin unit, running in the industrial computer or a host computer communicating with it, is used to build a mirror model in virtual space that is synchronized with the physical material yard and equipment status based on real-time data, and to simulate, verify and optimize control commands.

[0116] The execution module, connected to the control module via signals, is used to receive and execute control commands, and includes:

[0117] Pitch drive mechanism (servo motor + reducer) is used to control the pitch angle of the cantilever;

[0118] The rotary drive mechanism (servo motor + slewing bearing) is used to control the rotation angle of the cantilever.

[0119] The walking drive mechanism (variable frequency motor + walking wheel set) is used to control the movement of the whole machine along the track;

[0120] The material handling drive mechanism (high-power variable frequency motor + transmission mechanism) is used to drive the material handling head to rotate.

[0121] By combining the above methods and systems, the present invention achieves:

[0122] From "blind control" to "visual control": Through dynamic boundary modeling, the system can "see" the real shape changes of the material pile, and control decisions are based on precise environmental perception.

[0123] From "passive" to "active": Through safe trajectory planning and priority control, risks can be proactively avoided, rather than responding only after overload or collision occurs.

[0124] From "single objective" to "multi-objective coordination": The dual-constraint control loop perfectly coordinates the sometimes conflicting objectives of "space safety" and "load efficiency", achieving optimal overall system performance.

[0125] From "experience-dependent" to "intelligent learning": Iterative learning and prediction capabilities enable the system to continuously optimize itself, gradually reducing its reliance on preset parameters and human experience, and continuously improving its level of intelligence.

[0126] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0127] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0128] The present invention has been described above by way of example. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any improvements made by adopting the inventive concept and technical solution of the present invention, or direct application to other occasions without modification, are all within the protection scope of the present invention.

Claims

1. An adaptive material handling method for stacker-reclaimers based on a dynamic boundary model, characterized in that, Includes the following steps: S1. Dynamic boundary modeling: The three-dimensional point cloud data of the material reclaiming working face is acquired in real time by a three-dimensional scanning device installed at the front end of the stacker-reclaimer cantilever. The three-dimensional point cloud data is registered with the pre-stored benchmark model and three-dimensional difference calculation is performed to identify the difference area between the actual outline of the material pile and the model outline, and based on this, a dynamic and effective material picking boundary model that reflects the current actual working conditions is updated and generated. S2, Safe Trajectory Planning: Based on the effective material reclaiming boundary model and combined with the pre-stored material repose angle parameters, the dynamic safety distance between the material reclaiming head and the material pile boundary is calculated, and the movement trajectory of the material reclaiming head is planned or adjusted in real time to ensure that the material reclaiming operation is always within the safety boundary. S3, Load Adaptive Adjustment: Real-time monitoring of the load current of the material handling drive motor; Using load current as feedback, the pitch angle of the feed head and / or the travel speed of the whole machine are dynamically adjusted by an adaptive PID controller to keep the feed power stable within the preset target range. Steps S2 and S3 are executed in parallel and coupled to each other, forming a dual-constraint control loop. The control logic of this loop is as follows: First priority: When step S2 determines that there is a risk of collision or collapse, its output trajectory correction command will override the load adjustment command of step S3. Second priority: In the absence of the aforementioned risks, the expected trajectory output by step S2 and the optimized load parameters output by step S3 are coupled and calculated by a central processing unit to generate the final control command to drive the actuator to move.

2. The adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model according to claim 1, characterized in that, Also includes: S4. Iterative Learning and Prediction: Record the dynamic boundary model sequence, load current data, and corresponding operating parameters over multiple consecutive operating cycles; Based on the recorded data, a boundary evolution prediction model is trained using a recurrent neural network to predict the evolution trend of the stockpile boundary in the next operating cycle. The predicted boundary trends are used as prior information and applied to the initial path planning of the next work cycle to achieve forward control.

3. The adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model according to claim 2, characterized in that, In S4, a boundary evolution prediction model is trained by a recurrent neural network. Specifically, a long short-term memory network is used to perform time-series modeling of the dynamic boundary model sequence to predict the stockpile profile at a specific time step in the future.

4. The adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model according to claim 1, characterized in that, In S1, the 3D scanning equipment consists of a lidar and a 3D stereo camera.

5. The adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model according to claim 1, characterized in that, The baseline model is either a three-dimensional model of the stockpile established based on the initial scan, or a stockpile morphology model predicted based on historical operation data.

6. The adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model according to claim 1, characterized in that, In S2, the dynamic safety distance is calculated as follows: based on the slope angle of the effective material reclaiming boundary model and the angle of repose of the material, the minimum safe operating distance to prevent the material pile from collapsing is calculated, and this distance is used as a rigid constraint condition for trajectory planning.

7. The adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model according to claim 1, characterized in that, In S3, the operating parameters of the material handling head are dynamically adjusted, specifically including the decoupling control of the tilting mechanism, slewing mechanism and the overall walking mechanism of the material handling arm; The load adaptive adjustment loop is configured to independently and collaboratively adjust the pitch angle, rotation speed and travel speed according to changes in load current, so that the material handling system maintains constant power or optimal power while maintaining the stability of the movement of each mechanism.

8. An adaptive material handling system for a stacker-reclaimer based on a dynamic boundary model, used to execute the adaptive material handling method for a stacker-reclaimer based on a dynamic boundary model as described in any one of claims 1-7, characterized in that, include: The perception module includes: The 3D scanning unit includes a lidar and a 3D camera installed on the end of the stacker reclaimer cantilever near the reclaiming head, used to scan the reclaiming working surface from the front to obtain complete point cloud data of the material pile outline; The load detection unit includes a current sensor and a power transmitter installed in the power circuit of the material handling drive motor, which are used to directly measure the real-time load of the material handling drive motor. The pose sensing unit includes: Tilt sensor, installed on the cantilever of stacker-reclaimer, is used to directly measure the pitch angle of the cantilever; A rotary encoder, installed in the slewing drive mechanism of the cantilever, is used to detect the slewing angle of the cantilever. A high-precision positioning module is used to detect the position of the entire machine, including the Gray busbar laid along the entire length of the stacker-reclaimer track, and the address detector installed on the stacker-reclaimer traveling mechanism; The control module, connected to each unit of the sensing module via an industrial bus, is mounted on an industrial computer and includes: The core processing unit serves as the core of the system's computation; The dynamic boundary modeling unit runs on the core processing unit and is configured to perform point cloud registration and 3D difference calculation to complete the dynamic boundary modeling steps. The safety trajectory planning unit runs on the core processing unit and is configured to perform dynamic safety distance calculation and trajectory planning to complete the safety trajectory planning steps. The load adaptive adjustment unit runs on the core processing unit and is configured to run an adaptive PID control algorithm to complete the load adaptive adjustment steps. The decision fusion unit, running on the core processing unit, is configured to receive the outputs of the safety trajectory planning unit and the load adaptive adjustment unit, and perform priority judgment and instruction coupling calculation based on the logic of the dual constraint control loop to generate the final control instruction. The digital twin unit, running in the industrial computer or a host computer communicating with it, is used to build a mirror model in virtual space that is synchronized with the physical material yard and equipment status based on real-time data, and to simulate, verify and optimize control commands. The execution module, connected to the control module via signals, is used to receive and execute control commands, and includes: Pitch drive mechanism, used to control the pitch angle of the cantilever; A slewing drive mechanism is used to control the slewing angle of the cantilever. The walking drive mechanism is used to control the movement of the entire machine along the track; The material handling drive mechanism is used to drive the material handling head to rotate.

Images