Material stacking and taking machine based on millimeter wave radar control and material taking method

Abstract

A material piling and taking machine based on millimeter-wave radar control comprises a rack, a walking mechanism, a cantilever beam and a material taking mechanism, and relates to the technical field of bulk material processing equipment, millimeter-wave radar arrays are deployed on a front-end support, a middle pitching device and a tail vehicle of the cantilever beam to form a three-dimensional sensing network covering a material pile, the radar arrays support the MIMO technology, and the radar arrays support the material taking mechanism. Single equipment integrates more than or equal to 16-transmitting and 8-receiving antennas, point cloud data acquisition of a material pile within 200 meters can be realized, the resolution ratio is less than or equal to 0.05 meter, three-point deployment of the front end, the middle part and the tripper car of a millimeter wave radar array cantilever beam is adopted, a detection range of more than or equal to 180-degree horizontal coverage and 0-30-degree vertical pitching is formed, three-dimensional point cloud data of 0.05-meter resolution ratio of the material pile within 200 meters can be acquired, and the detection precision of the material pile within 200 meters is improved. Compared with the single-point detection precision of + / -20cm of a traditional laser radar, multi-parameter coupling sensing of a height field, a density field and a flow velocity field of a material pile is realized; an MIMO radar technology single-device 16-transmitting 8-receiving antenna and an STAP space-time adaptive processing algorithm are utilized.

Description

Technical Field

[0001] The present invention relates to the technical field of bulk material processing equipment, and in particular to a material stacker and reclaimer controlled by millimeter-wave radar and a material reclaiming method. Background Art

[0002] Traditional stacker-reclaimers primarily rely on detection methods such as LiDAR and ultrasonic sensors to achieve their functions. However, this reliance on these methods has exposed numerous technical bottlenecks in practical applications. The first is environmental adaptability. LiDAR, a common detection technology, suffers from severe degradation of its detection range in dusty environments, with the degradation rate exceeding 50%. This significantly reduces the effective detection range of LiDAR in dusty environments, hindering the accurate acquisition of material-related information. Ultrasonic detection also presents challenges. It is extremely sensitive to ambient temperature and humidity. Changes in temperature and humidity can cause material level errors exceeding ±20cm, significantly compromising the accuracy of material level data obtained through ultrasonic detection and making it difficult to meet the precision requirements of actual operations. Furthermore, control lag is significant. The PID control method used in traditional stacker-reclaimers has a long control cycle, typically ≥200ms. This long control cycle prevents the equipment from responding to dynamic changes in the material pile shape in real time. In actual operation, the shape of the pile constantly changes with material stacking and reclaiming operations. Due to control lag, the bucket wheel may become overloaded, meaning it grabs more material than it can handle. This not only affects the normal operation of the equipment but can also cause damage. Furthermore, reclaiming blind spots are prone to occur, meaning some materials cannot be effectively retrieved, reducing the efficiency and integrity of reclaiming. Finally, multi-sensor coordination is lacking. Existing stacker-reclaimer equipment mostly uses a single-point detection method, capturing only relevant data at a single point and lacking the ability to model the three-dimensional contour of the pile. Without the ability to construct a three-dimensional contour model of the pile, reclaiming path planning cannot be based on comprehensive and accurate pile information and must rely on manual experience. However, manual experience has limitations, with varying levels of experience among operators. Furthermore, manual planning struggles to accurately and in real time adapt to the dynamic changes of the pile. This results in unscientific and irrational reclaiming path planning, ultimately leading to low reclaiming efficiency and failure to meet the requirements of efficient production. To address these technical issues, a new technical solution is proposed. Summary of the Invention

[0003] In response to the shortcomings of the existing technology, the present invention realizes the full-process automated control of the stacker and reclaimer through the fusion perception of the millimeter-wave radar array and the collaboration of intelligent algorithms, breaking through the bottleneck of detection accuracy and response speed of traditional equipment under complex working conditions.

[0004] The present invention provides a stacking and reclaiming machine based on millimeter-wave radar control, which includes a frame, a traveling mechanism, a cantilever beam and a reclaiming mechanism. A millimeter-wave radar array is deployed on the front-end support, middle pitching device and tail car of the cantilever beam to form a three-dimensional perception network covering the stockpile. The radar array supports MIMO technology, and a single device integrates no less than 16 transmitting and 8 receiving antennas, enabling the acquisition of stockpile point cloud data within 200 meters with a resolution of no more than 0.05 meters.

[0005] As a preferred technical solution, the reclaiming mechanism includes a bucket wheel and a conveyor. The rotational speed of the bucket wheel is 0 - 120 rpm, the excavation depth is 0 - 3 meters, and a dynamic mapping relationship is established with the material density ρ(x,y) detected by the radar, specifically as follows:

[0006] When ρ ≤ 1.2 t / m 3 the rotational speed of the bucket wheel automatically increases to 100 - 120 rpm, and the maximum excavation depth is 3 meters;

[0007] When 1.2 t / m 3 <ρ ≤ 1.8 t / m 3 the rotational speed of the bucket wheel remains at 60 - 100 rpm, and the excavation depth is limited to 1.5 - 2.5 meters;

[0008] When ρ > 1.8 t / m 3 the rotational speed of the bucket wheel decreases to ≤ 60 rpm, and the excavation depth is ≤ 1.5 meters.

[0009] As a preferred technical solution, the traveling mechanism adopts a dual-drive motor and encoder feedback closed-loop control, with a positioning accuracy of ±5 mm, supports linear interpolation and circular arc trajectory planning, a maximum traveling speed of 1.5 m / s, and a minimum adjustment step size of 0.1 m.

[0010] As a preferred technical solution, the cantilever beam realizes pitching adjustment from -15° to +18° through a pitching device, with an adjustment speed of 0.5° / s. The pitching angle is associated with the height difference Δh of the stockpile detected by the radar. When Δh > 5 meters, the layered reclaiming mode is automatically started, and the thickness of each layer of reclaiming is ≤ 2 meters.

[0011] As a preferred technical solution, the control system integrates an FPGA + ARM heterogeneous computing platform, with a data processing delay of ≤ 20 ms. It has a built-in stockpile volume calculation module, which uses the triangular prism volume integration method based on point cloud data, with a calculation accuracy of ±3%, and generates a stockpile volume change curve V(t) in real time.

[0012] A method for reclaiming materials of a stacker-reclaimer controlled by millimeter-wave radar, which synchronously collects the echo signals of the material stack by using a millimeter-wave radar array, suppresses dust clutter through the STAP space-time adaptive processing algorithm, combines the triangulation method to reconstruct the three-dimensional model of the material stack to form a hopper blockage warning model, and generates a digital twin containing the height field H(x,y) and the density distribution ρ(x,y).

[0013] As a preferred technical solution, the STAP space-time adaptive processing algorithm is an improved A path planning algorithm, and the heuristic function of the algorithm is: Where α is the slope weight factor of 0.5 - 1.0, β is the density weight factor of 0.3 - 0.8, and the constraint conditions include that the bucket wheel excavation depth ≤ the tangent value of the natural angle of repose of the material (tanθ ≤ 1.5), and the equipment load current ≤ 85% of the rated current.

[0014] As a preferred technical solution, during the reclaiming process, the material flow velocity u is measured by the radar Doppler effect, and combined with the continuity equation Predict the change of the material stack shape, adjust the traveling speed v, the bucket wheel height h and the rotation frequency n 500 ms in advance, and realize the feedforward control of the reclaiming parameters.<>

[0015] As a preferred technical solution, for the hopper blockage warning model, through the sudden change detection of the conveyor current and the analysis of the radar material flow velocity field, when the blockage probability > 70%, the traveling mechanism automatically retreats 0.5 - 1.0 meters, and the bucket wheel reverses at -60 rpm for 5 - 10 seconds to clear the material.

[0016] As a preferred technical solution, a reinforcement learning algorithm is introduced, and the reclaiming strategy is trained based on historical operation data. The optimization goal is to maximize the reclaiming efficiency and minimize the energy consumption, reduce the comprehensive energy consumption by 15% - 20%, and automatically generate an optimal reclaiming sequence table.

[0017] Compared with the prior art, the beneficial effects of the present invention are:

[0018] 1. The millimeter-wave radar array is deployed at three positions, namely the front end, the middle part and the tail car of the cantilever beam, forming a detection range with a horizontal coverage of ≥180° and a vertical pitch of 0 - 30°. It can obtain three-dimensional point cloud data with a resolution of 0.05 m of the material stack within 200 m. Compared with the single-point detection accuracy of ±20 cm of the traditional lidar, it realizes the multi-parameter coupling perception of the "height field, density field, and flow velocity field" of the material stack; using the MIMO radar technology with 16 transmit and 8 receive antennas of a single device and the STAP space-time adaptive processing algorithm, when the dust concentration > 1000 mg / m 3The detected signal attenuation in the environment is less than 5%, while that of traditional lidar exceeds 50%, solving the problem of detection failure in high-dust scenarios; by measuring the material flow velocity u through the radar Doppler effect and combining with the continuity equation, the change of the stockpile shape can be predicted 500 ms in advance. Compared with the traditional open-loop control with a lag response control period of ≥200 ms, the feedforward adjustment of the reclaiming parameters is realized, avoiding the bucket wheel from hitting the steep slope or falling into the blind area of the stockpile.

[0019] 2. Based on the three-dimensional point cloud model, the improved A* algorithm introduces the slope factor α and the density weight β into the heuristic function, which can automatically avoid generating a "Z"-shaped detour path when the height difference of the stockpile steep bank is >5 meters, and dynamically adjust the reclaiming order according to the material density. Compared with the traditional fixed-path mode, the reclaiming efficiency is increased by more than 30%, and the blind area rate is <0.5%; through the dynamic mapping of the bucket wheel speed, excavation depth and material density, when ρ > 1.8 t / m 3 the speed is ≤60 rpm and the depth is ≤1.5 meters, suppressing the equipment load fluctuation by ±10%, avoiding the motor burnout or mechanical wear caused by overload in the traditional scheme, and reducing the equipment loss by 25%; real-time monitoring of the bucket wheel cutting depth error of ±2 cm and the conveyor current, when the probability of hopper blockage is detected to be >70%, automatically trigger the "retreat and clear material" mechanism, the traveling mechanism retreats 0.5 - 1.0 meters, and the bucket wheel rotates in reverse for 5 - 10 seconds. Compared with the traditional manual intervention, the fault handling time is shortened by 80%, reducing the downtime loss.

[0020] 3. Introduce the reinforcement learning algorithm, train the reclaiming strategy based on 2000 working condition samples, and comprehensively reduce the energy consumption by 15% - 20% by optimizing the parameter combinations such as the traveling speed and the bucket wheel speed. For example, in the coal reclaiming scenario, the system automatically identifies the high-sulfur coal area and preferentially reclaims the coal to balance the coal blending demand, while reducing the idling time; the cooperation between the boom pitching device adjusted from -15° to +18° and the layered reclaiming mode with each layer thickness ≤2 meters can adapt to the vertical drop of the stockpile up to 40 meters. Compared with the traditional equipment that requires manual adjustment of the pitching angle, the automation degree is increased by 100%, and the manual safety risk is reduced; the data processing delay of the control system integrated with the FPGA and ARM heterogeneous computing platform is ≤20 ms, supporting the real-time calculation accuracy of the stockpile volume of ±3% and the dynamic management of the inventory, meeting the digital warehousing requirements of scenarios such as mines and ports.

[0021] 4. In the application of a coal yard with an annual output of 5 million tons, the reclaiming efficiency reaches 1500 t / h, which is 40% higher than that of traditional equipment, saving 800,000 yuan in labor costs annually, and the comprehensive efficiency is increased by 28%; reducing the dust disturbance through accurate reclaiming, and the dust emission is reduced by more than 30%; the full-process automatic control avoids manual intervention, and the incidence of safety accidents is reduced by 90%, meeting the requirements of green industry and intrinsic safety. Description of the Drawings

[0022] Figure 1Structural schematic diagram of the invention;

[0023] Figure 2 Structural block diagram of the invention;

[0024] Figure 3 Top view of the walking direction of the invention; Detailed implementation manners

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0026] A stacker-reclaimer controlled by millimeter-wave radar includes a frame 1, a traveling mechanism 7, a cantilever beam 3 and a reclaiming mechanism 11. A millimeter-wave radar array is deployed on the front-end bracket 9, the middle pitching device 2 and the tail car 10 of the cantilever beam 3 to form a three-dimensional perception network covering the stockpile. The radar array supports MIMO technology, and a single device integrates no less than 16 transmitting and 8 receiving antennas, enabling the acquisition of stockpile point cloud data within 200 meters with a resolution of no more than 0.05 meters. The radar array consists of several sensors 4. The MIMO radar array consists of configured antennas and brackets 9, pitching devices 2 and tail cars 10. The configured antenna is a single radar device integrating 16 transmitting antennas (Tx) and 8 receiving antennas (Rx), adopting a uniform linear array (ULA) layout. The spacing between transmitting antennas is ≤λ / 2 (λ is the radar operating wavelength, e.g., for a 24 GHz radar, λ = 12.5 mm), and the spacing between receiving antennas is =λ / 2. The 3 radars deployed on the front-end bracket 9, the middle pitching device 2 and the tail car 10 form a distributed MIMO array. Each device operates at the same frequency, and signal synchronous transmission and reception are achieved through a time synchronization module with an accuracy of ≤10 ns. It adopts a linear frequency modulation continuous wave system, with a sweep bandwidth of ≥4 GHz, achieving a high range resolution of ΔR = c / (2B) = 3.75 cm, where c is the speed of light and B is the bandwidth. The transmitting antennas adopt orthogonal waveforms to ensure that the echo signals of each channel can be separated and avoid mutual interference. The front-end radar covers the 0-180° area in front of the cantilever beam, the middle radar covers the ±90° area, and the tail car radar covers the rear 0-180° area, forming an overlapping coverage network with a 360° full circumferential detection blind area of ≤10°. By adjusting the angle of the cantilever beam from -15° to +18° through the pitching device 2 and combining with the vertical beam width of the radar itself, a 0-30° vertical pitching detection range is achieved, covering the full height from the top to the bottom of the stockpile. The transmitting antennas simultaneously send N cycles of linear frequency modulation signals, and the frequency changes with time as f(t) = f0 + Kt, t ∈ [0, T], where f0 is the starting frequency, K = B / T is the frequency modulation slope, and T is the sweep period. The 16 transmitting antennas are respectively loaded with orthogonal phase encoding, so that the echoes of different antennas can be separated by matched filtering at the receiving end. The 8 receiving antennas synchronously collect the echo signals, and the sampling rate of each channel is ≥2B to ensure no signal aliasing. After mixing the received signal with the local oscillator signal, the beat frequency f b = 2KR / c is obtained, where R is the target distance. For distance measurement, a fast Fourier transform is performed on the beat signal, and the peak frequency corresponds to the target distance: For velocity measurement, through the phase difference of echoes in adjacent periods Solve the target radial velocity v; estimate the direction of arrival (DOA). Using the phase difference Δφ = 2πdsinθ / λ (d is the antenna spacing, θ is the incident angle) of the receiving antenna array, estimate the horizontal angle θ and vertical angle φ of the target through the Multiple Signal Classification (MUSIC) algorithm or Compressive Sensing algorithm. Taking the radar installation position as the origin, establish a coordinate system, and the three-dimensional coordinates of the target are: (Rcosφcosθ, Rcosφsinθ, Rsinφ). The data of multiple radars are unified to the device global coordinate system through coordinate transformation to form point cloud stitching. Adopt the Space-Time Adaptive Processing (STAP) algorithm, use space-time two-dimensional filtering to suppress fixed clutter, and retain the dynamic stockpile echo. According to the relationship between the radar cross-section and the dielectric constant of the material, establish a mapping model: ρ(x,y) = k.ln(σ 0 ) + b, where σ 0 is the radar backscattering coefficient, and k and b are coefficients calibrated through material samples; through the global clock) or the hardware synchronization line, ensure that the starting time deviation of the FM cycle of the 3 radars ≤ 1 μs to avoid signal aliasing. Use the Iterative Closest Point (ICP) algorithm to transform the point clouds of each radar to the same coordinate system; perform voting screening on the point clouds in the overlapping detection area to eliminate outliers and improve data reliability. When the device is close to the stockpile (such as the distance < 50 meters), automatically switch to the high-resolution mode (resolution 0.02 meters) to densely sample the surface details of the stockpile. When detecting at a long distance (such as 200 meters), adopt the wide-beam mode to expand the single-scan coverage range and shorten the modeling time (the full-stockpile scan ≤ 10 seconds). The radar operating frequency randomly jumps within the range of 24 GHz ± 500 MHz to avoid industrial radio frequency interference, emits vertically polarized waves, and receives horizontally / vertically dual-polarized signals to distinguish metal structures and bulk materials through polarization characteristics. The advantages that can be obtained under this principle are shown in the following table:

[0027]

[0028] As a preferred technical solution, the material taking mechanism includes a bucket wheel and a conveyor 6. The rotation speed of the bucket wheel 5 is 0 - 120 rpm, the excavation depth is 0 - 3 meters, and a dynamic mapping relationship is established with the material density ρ(x,y) detected by the radar. Specifically: when ρ ≤ 1.2 t / m 3 , the rotation speed of the bucket wheel 5 automatically increases to 100 - 120 rpm, and the maximum excavation depth is 3 meters; when 1.2 t / m 3 <ρ ≤ 1.8 t / m 3 , the rotation speed of the bucket wheel 5 remains at 60 - 100 rpm, and the excavation depth is limited to 1.5 - 2.5 meters; when ρ > 1.8 t / m 3When the rotational speed of the bucket wheel 5 is reduced to ≤60 rpm, the excavation depth is ≤1.5 m. Based on the material density field ρ(x,y) detected in real time by the millimeter-wave radar, an empirical formula for the load torque T of the bucket wheel 5 and the density is established: T = k1·ρ + k2·v·h, where k1 and k2 are equipment structure coefficients, v is the traveling speed (m / s), and h is the excavation depth (m), ensuring that the current I of the bucket wheel motor is ≤0.85I 额定 , to avoid overload damage and maximize the material intake efficiency Q = π·n·h·B·ρ, where n is the rotational speed and B is the width of the bucket wheel. For low-density materials (ρ ≤ 1.2 t / m 3 ), when the radar detects a low-density area, the control system sends a rotational speed set signal n ref = 120 rpm to the bucket wheel drive motor (frequency conversion motor), and at the same time sends an instruction to the lifting cylinder to lower the bucket wheel to h = 3 m, and monitors the motor current in real time. If I > 0.8I 额定 , automatically reduce the rotational speed to 110 rpm until the current is stable; for medium-density materials (1.2 t / m 3 < ρ ≤ 1.8 t / m 3 ), the rotational speed is maintained at 60 - 100 rpm, and the traveling speed v (0.5 - 1.2 m / s) is adjusted to balance the material intake efficiency and the load. The excavation depth is limited to 1.5 - 2.5 m to avoid excessive bearing stress on the bucket wheel shaft due to excessive depth. A fuzzy PID controller is used, with the input being the density deviation Δρ = ρ - 1.5 t / m 3 and the current deviation ΔI = I - 0.7I 额定 , and the output being the rotational speed adjustment amount Δn and the depth adjustment amount Δh; a typical adjustment case: when ρ = 1.6 t / m 3 and I = 0.75I 额定 , automatically reduce n from 80 rpm to 70 rpm and adjust h from 2.5 m to 2.0 m; for high-density materials (ρ > 1.8 t / m 3 ), the rotational speed is ≤60 rpm (low-speed high-torque mode) to reduce the impact load of the material; the excavation depth is ≤1.5 m to reduce the single material intake volume and avoid conveyor blockage, triggering the "heavy load protection mode". The bucket wheel motor switches to the vector control mode to increase the low-frequency torque output; the traveling mechanism synchronously reduces the speed to ≤0.5 m / s to ensure uniform material transportation, and the conveyor current I conv ≤ 0.9 Iconv额度 . The conveyor speed v conv is proportionally linked to the rotational speed n of the bucket wheel, and the formula is v conv = k3·n + b (k3 = 0.02 m·s -1 / rpm, b = 0.5 m / s) to ensure smooth material transportation; the load is monitored in real time by a torque sensor installed on the conveyor roller. When the torque mutation rate dT / dt > 50 Nkm / s, it is determined as a potential blockage, and the bucket wheel is triggered to reverse for material cleaning.

[0029] As a preferred technical solution, the traveling mechanism 7 adopts a dual-drive motor and encoder feedback closed-loop control, with a positioning accuracy of ±5 mm, supports linear interpolation and circular arc trajectory planning, has a maximum traveling speed of 1.5 m / s, and a minimum adjustment step size of 0.1 m. The drive system uses two AC servo motors (power ≥ 30 kW) to independently drive the wheels on both sides, and realizes speed coupling through a synchronous gearbox. The motor is equipped with an absolute encoder (resolution 24 bits, accuracy ±0.01 mm / pulse), which real-time feeds back the wheel rotation angle θ. The closed-loop control loop includes a position loop, a speed loop, and a current loop. Among them, the position loop receives the target coordinates (X ref , y ref ) of the control system, and generates displacement commands θ ref for each axis through an interpolation algorithm; the speed loop uses a PI controller to adjust the motor speed so that the actual speed v = rkdθ / dt tracks the target speed v ref ; the current loop realizes real-time adjustment of the motor torque through vector control to suppress load imbalance. Multi-sensor fusion is supplemented with a laser rangefinder (accuracy ±2 mm) for absolute position calibration, which is triggered once every 5 meters of travel to eliminate the cumulative error of the encoder. An inertial navigation unit (IMU) is used to monitor the traveling attitude. When the detected yaw angle Δψ > 0.5°, correction is performed through differential control. For linear interpolation, the motion between two points uses a trapezoidal speed curve (acceleration and deceleration time ≤ 1 s), and the formula is where (a = 0.5 m / s 2 ) is the acceleration, and (v_{max} = 1.5 m / s) is the maximum speed; for circular arc interpolation, NURBS spline fitting is used to decompose the circular arc into small straight line segments (step size 0.1 m), and the centripetal acceleration a n = v 2 / R ≤ 0.3 m / s 2 is adjusted in real-time to ensure smooth motion.

[0030] As a preferred technical solution, the cantilever beam 3 realizes pitch adjustment from -15° to +18° through a pitching device, with an adjustment speed of 0.5° / s. The pitch angle is related to the height difference Δh of the stockpile detected by the radar. When Δh > 5 meters, the layered reclaimer mode is automatically started, and the reclaimer thickness per layer ≤ 2 meters. This device can also achieve multi-mode switching. In the automatic mode, it automatically executes linear / arc motion according to the path planning result, with a minimum adjustment step size of 0.1 m, and the corresponding encoder pulse number N = 0.1 m / (πD / 2 24 ) ≈ 8388 pulses; in the manual mode, it supports joystick input, with a speed resolution of 0.01 m / s, which is used for equipment debugging and emergency obstacle avoidance. Reclaimer efficiency analysis:

[0031] Material density ρ Bucket wheel speed n Excavation depth h Travel speed v Reclaiming efficiency Q Unit energy consumption E <![CDATA[1.0 t / m 3 > 120 rpm 3.0m 1.2 m / s 1800 t / h 0.35 kWh / t <![CDATA[1.5t / m 3 > 80 rpm 2.0m 1.0 m / s 1508 t / h 0.38 kWh / t <![CDATA[2.0t / m 3 > 50 rpm 1.2m 0.5 m / s 942 t / h 0.42 kWh / t

[0032] As a preferred technical solution, the control system integrates an FPGA + ARM heterogeneous computing platform with a data processing delay ≤ 20 ms. It has a built-in stockpile volume calculation module that uses the triangular prism volume integration method based on point cloud data, with a calculation accuracy of ±3%, and generates a real-time stockpile volume change curve V(t).

[0033] A method for reclaiming a stacker-reclaimer based on millimeter-wave radar control, which uses a millimeter-wave radar array to synchronously collect stockpile echo signals, suppresses dust clutter through the STAP space-time adaptive processing algorithm, combines the triangulation method to reconstruct the three-dimensional model of the stockpile to form a hopper blockage warning model, and generates a digital twin including the height field H(x, y) and density distribution ρ(x, y).

[0034] As a preferred technical solution, the STAP space-time adaptive processing algorithm is an improved A path planning algorithm, and the heuristic function of the algorithm is: where α is the slope weight factor of 0.5 - 1.0, β is the density weight factor of 0.3 - 0.8, and the constraint conditions include that the bucket wheel excavation depth ≤ the tangent value of the natural angle of repose of the material (tanθ ≤ 1.5), and the equipment load current ≤ 85% of the rated current.

[0035] As a preferred technical solution, during the reclaiming process, the material flow velocity u is measured through the radar Doppler effect, and combined with the continuity equation to predict the change of the stockpile shape, adjust the traveling speed v, the bucket wheel height h and the rotation frequency n 500 ms in advance, and realize the feedforward control of the reclaiming parameters.

[0036] As a preferred technical solution, for the hopper 8 blockage warning model, through the sudden change detection of the conveyor current and the analysis of the radar material flow velocity field, when the blockage probability > 70%, the traveling mechanism automatically retreats 0.5 - 1.0 meters, and the bucket wheel reverses at -60 rpm for 5 - 10 seconds to clear the material.

[0037] As a preferred technical solution, a reinforcement learning algorithm is introduced to train the reclaiming strategy based on historical operation data. The optimization goal is to maximize the reclaiming efficiency and minimize the energy consumption, reduce the comprehensive energy consumption by 15% - 20%, and automatically generate an optimal reclaiming sequence table.

[0038] This device adopts a three-dimensional sensing array layout, and three groups of millimeter-wave radar sensors 4 are respectively deployed on the front bracket 9 of the cantilever beam 3, the middle pitching device 2 and the tail car 10, forming a detection range with a horizontal coverage of ≥ 180° and a vertical pitch of 0 - 30°, which can realize the acquisition of stockpile point cloud data within 200 meters with a resolution of 0.05 meters.

[0039] The sensor 4 adopts MIMO radar technology. A single device integrates a 16-transmit and 8-receive antenna array, supports the simultaneous detection of 2,000 target points, and meets the requirements for high-density point cloud reconstruction of dynamic stockpiles.

[0040] The intelligent actuator, the material taking mechanism 11, adopts the linkage design of the bucket wheel 5 and the conveyor 6. The rotational speed of the bucket wheel (0 - 120 rpm) and the excavation depth of 0 - 3 meters can be adjusted in real time according to the material density detected by the radar; the traveling mechanism 7 is equipped with a dual-drive motor and encoder feedback, with a positioning accuracy of ±5 mm, supports linear interpolation and circular trajectory planning, and the pitching device 2 realizes the pitching adjustment of the cantilever beam from -15° to +18°, with an adjustment speed of 0.5° / s, and cooperates with the radar data to achieve layered material taking; an integrated FPGA + ARM heterogeneous computing platform, with a data processing delay of ≤20 ms, supports real-time point cloud filtering, stockpile volume calculation (accuracy ±3%), and path optimization algorithms; the control software has a built-in multi-objective optimization model, aiming at maximizing the material taking efficiency and equipment load balance, and dynamically generates the optimal combinations of the traveling speed (v: 0.1 - 1.5 m / s), the bucket wheel height (h: -3 m to 0 m), and the rotation frequency (n: 0 - 120 rpm).

[0041] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A stacking and reclaiming machine controlled by a millimeter-wave radar, comprising a frame, a traveling mechanism, a cantilever beam and a reclaiming mechanism, characterized in that: Deploy a millimeter-wave radar array at the front-end support of the cantilever beam, the middle pitching device, and the tail car to form a three-dimensional perception network covering the stockpile. The radar array supports MIMO technology, with a single device integrating no less than 16 transmit and 8 receive antennas, enabling the acquisition of stockpile point cloud data within 200 meters, and the resolution is less than or equal to 0.05 meters.

2. The stacker-reclaimer based on millimeter-wave radar control according to claim 1, wherein: The reclaimer mechanism includes a bucket wheel and a conveyor. The rotation speed of the bucket wheel is 0 - 120 rpm, the excavation depth is 0 - 3 meters, and a dynamic mapping relationship is established with the material density ρ(x, y) detected by the radar. Specifically: When ρ ≤ 1.2 t / m 3 the bucket wheel speed automatically increases to 100 - 120 rpm, and the maximum excavation depth is up to 3 m; When 1.2 t / m 3 <ρ ≤ 1.8 t / m 3 , the bucket wheel speed remains at 60 - 100 rpm, and the excavation depth is limited to 1.5 - 2.5 meters; When ρ > 1.8 t / m 3 the bucket wheel speed is reduced to ≤ 60 rpm, and the excavation depth is ≤ 1.5 m.

3. The stacker-reclaimer based on millimeter-wave radar control according to claim 1, characterized in that: The traveling mechanism adopts a dual-drive motor and encoder feedback closed-loop control, with a positioning accuracy of ±5 mm, supports linear interpolation and circular arc trajectory planning, the maximum traveling speed is 1.5 m / s, and the minimum adjustment step size is 0.1 m.

4. The stacker-reclaimer based on millimeter-wave radar control according to claim 1, characterized in that: The cantilever beam realizes pitching adjustment from -15° to +18° through the pitching device, with an adjustment speed of 0.5° / s. The pitching angle is correlated with the height difference Δh of the stockpile detected by the radar. When Δh > 5 meters, the layered reclaiming mode is automatically activated, and the reclaiming thickness per layer is ≤2 meters.

5. The stacking and reclaiming machine based on millimeter-wave radar control according to claim 1, wherein: The control system integrates an FPGA + ARM heterogeneous computing platform, with a data processing delay ≤20 ms. It has a built-in stockpile volume calculation module, which uses the triangular prism volume integration method based on point cloud data, with a calculation accuracy of ±3%, and generates a real-time stockpile volume change curve V(t).

6. A method for reclaiming materials of a stacking and reclaiming machine based on millimeter-wave radar control, characterized in that: Synchronously collect the stockpile echo signals using the millimeter-wave radar array, suppress dust clutter through the STAP space-time adaptive processing algorithm, combine the triangulation method to reconstruct the three-dimensional model of the stockpile to form a hopper blockage warning model, and generate a digital twin including the height field H(x, y) and density distribution ρ(x, y).

7. The material taking method of the stacking and reclaiming machine based on millimeter wave radar control according to claim 6, characterized in that: Using the STAP space-time adaptive processing algorithm to improve the A path planning algorithm, the heuristic function of the algorithm is as follows: where α is the slope weight factor of 0.5 - 1.0, β is the density weight factor of 0.3 - 0.8, and the constraint conditions include that the bucket wheel excavation depth ≤ the tangent value of the natural angle of repose of the material (tanθ ≤ 1.5), and the equipment load current ≤ 85% of the rated current.

8. The material fetching method of the stacking and fetching machine based on millimeter-wave radar control according to claim 6, characterized in that: During the material fetching process, the material flow velocity u is measured through the radar Doppler effect, and combined with the continuity equation predict the change of the stockpile shape, adjust the walking speed v, the bucket wheel height h and the rotation frequency n 500 ms in advance, and realize the feedforward control of the fetching parameters.

9. The material taking method of the stacker-reclaimer controlled based on millimeter wave radar according to claim 6, characterized in that: For the hopper blockage warning model, through the detection of sudden changes in the conveyor current and the analysis of the radar material flow velocity field, when the blockage probability > 70%, the traveling mechanism automatically retreats 0.5 - 1.0 meters, and the bucket wheel rotates in reverse at -60 rpm for 5 - 10 seconds for material clearing.

10. The material fetching method of the stacking and fetching machine controlled based on millimeter wave radar according to claim 6, characterized in that: Introduce a reinforcement learning algorithm, train the reclaiming strategy based on historical operation data, with the optimization goals of maximizing the reclaiming efficiency and minimizing the energy consumption, achieving a 15% - 20% reduction in the comprehensive energy consumption, and automatically generating an optimal reclaiming sequence table.

Images