Material guiding device and control system

Abstract

The invention relates to the technical field of metallurgical equipment control, in particular to a material guide device and a control system, and the device comprises a material track planning module, an unpowered dust suppression module, a speed cooperative control module, a dynamic feedback regulation module and a genetic algorithm optimization module. The material trajectory planning module constructs a sliding trajectory model based on a discrete element simulation algorithm, and generates guiding curved surface geometric parameters and damper rigidity control instructions; the dynamic feedback regulation module collects a vibration spectrum through a high-frequency piezoelectric sensor array, and outputs a classification result through matching of a convolutional neural network and an abnormal vibration mode library; the genetic algorithm optimization module constructs a multi-objective fitness function to generate a global optimization parameter set, drives guide curved surface inclination angle adjustment, guide grid damping compensation and rotating speed control logic co-evolution, and forms a closed-loop parameter collaboration mechanism. According to the invention, systematic cooperation of material sliding track optimization, dust dynamic suppression and equipment vibration energy control is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of metallurgical equipment control, and particularly to a feeding device and a control system. Background Art

[0002] In the technical field of bulk material transportation, traditional feeding devices generally have problems such as large impact force, dust dispersion, and high energy consumption during the material transfer process. In the prior art, the transfer of materials from the feeding end to the receiving end mostly relies on free fall or rigid guiding structures, resulting in violent collisions between the materials and the receiving equipment, increasing the material breakage rate, exacerbating equipment wear, and accompanied by a large amount of dust pollution. Conventional dust removal solutions rely on external power systems, leading to increased equipment complexity and operating energy consumption. In addition, the insufficient speed matching degree between the receiving belt and the materials easily causes slip friction, further shortening the service life of the equipment. Therefore, there is an urgent need for a feeding device and a supporting control system that can achieve slow-speed sliding guidance of materials, autonomous dust suppression, and optimize energy utilization, so as to simultaneously improve the material transportation efficiency, environmental protection performance, and equipment operation economy. Summary of the Invention

[0003] Aiming at the deficiencies of the prior art, the present invention provides a feeding device and a control system, which solve the problems of violent impact, dust pollution, and excessive energy consumption between the materials and the receiving equipment caused by the traditional feeding device using free fall or rigid guiding structures, and at the same time, the slip friction caused by the speed mismatch between the receiving belt and the materials exacerbates equipment wear.

[0004] To solve the above technical problems, the specific technical solutions of the present invention are as follows: In a first aspect, a feeding device provided by the present invention includes: A material trajectory planning module, which constructs a material sliding trajectory model based on the discrete element simulation algorithm, and generates guiding surface geometric parameters and corresponding damper stiffness control instructions according to the simulation results of the model; A non-powered dust suppression module, which receives the guiding surface geometric parameters output by the material trajectory planning module, dynamically configures the interlayer inclination angles of the multi-layer tapered diversion grids according to the geometric parameters, and completes the inertial separation of dust particles by implementing velocity gradient attenuation on the airflow induced by the sliding materials through the inertial settlement cavity; A speed coordination control module, which receives the damper stiffness control instructions sent by the material trajectory planning module, and calculates the target speed adjustment amount of the driving motor of the receiving belt in combination with the material flow velocity distribution data captured by the lidar scanning device in real time; A dynamic feedback adjustment module, which continuously collects the contact stress spectrum through the high-frequency piezoelectric pressure sensor array deployed on the support beam of the receiving belt, extracts the time-frequency characteristics of the stress spectrum through a convolutional neural network, and outputs the equipment abnormal vibration classification result after matching the preset vibration mode database; The genetic algorithm optimization module receives the abnormal vibration classification results of the dynamic feedback adjustment module and the rotational speed adjustment amount data of the speed coordination control module, constructs a multi-dimensional fitness evaluation function by integrating real-time operating state parameters, generates an optimization parameter set including the dynamic inclination adjustment amount of the guiding surface, the damper stiffness correction coefficient, and the deflection compensation value of the diversion grid, and synchronously feeds back the optimization parameter set to the material trajectory planning module for updating the slip trajectory model, the passive dust suppression module for adjusting the diversion grid structure parameters, and the speed coordination control module for calibrating the rotational speed control logic, forming a cross-module parameter co-evolution mechanism.

[0005] Further, for the feeding device of the present invention, the passive dust suppression module includes a diversion grid inclination angle adjustment unit, which receives the dynamic distribution density data in the material slip trajectory model output by the material trajectory planning module, calculates the cross-sectional area change amount of the air flow channel between the layers of the diversion grid based on the density data, and synchronously adjusts the inclination angle and spacing of each layer of the diversion grid to make the spatial distribution of the air flow velocity gradient consistent with the inertial settling path of the dust particles.

[0006] Further, for the feeding device of the present invention, the dynamic feedback adjustment module includes a convolutional neural network sub-module, which receives the real-time stress spectrum data collected by the high-frequency piezoelectric pressure sensor array, performs time-frequency domain feature decomposition on the spectrum data to generate a multi-dimensional feature vector, and outputs a classification result including the vibration energy level and the fault type to the genetic algorithm optimization module by comparing with the abnormal vibration waveform templates in the preset vibration mode database.

[0007] Further, for the feeding device of the present invention, the genetic algorithm optimization module assigns weight coefficients of impact acceleration, energy consumption efficiency, and dust concentration in the multi-objective fitness function according to the vibration energy level and the fault type classification result, uses the elitist retention strategy to screen out a PID control parameter set that meets the preset dynamic balance condition, and injects the parameter set into the rotational speed adjustment algorithm of the speed coordination control module to generate a closed-loop feedback control command.

[0008] Further, for the feeding device of the present invention, the material trajectory planning module includes a dynamic damping compensation unit, which receives the damper stiffness correction coefficient output by the genetic algorithm optimization module, calculates the curvature radius adjustment amount at the end of the guiding surface in real time based on the coefficient, and dynamically corrects the geometric shape of the surface through a mechanical actuator, so that the difference between the material slip end speed and the target belt line speed calculated by the speed coordination control module converges to a preset allowable range.

[0009] Furthermore, in the material guiding device described in the present invention, the dynamic damping compensation unit receives the material slip speed deviation value calculated by the speed collaborative control module through the industrial bus, and when the deviation value exceeds a preset threshold, a parameter reconfiguration request signal is sent to the genetic algorithm optimization module, triggering a global optimization parameter set generation process based on the current operating state.

[0010] Furthermore, the material guiding device described in the present invention, the adaptive crossover probability generation unit of the genetic algorithm optimization module receives the belt load fluctuation rate data collected by the dynamic feedback adjustment module and the real-time dust removal efficiency parameters of the unpowered dust suppression module, dynamically calculates the crossover operator weight distribution according to the correlation analysis results between the fluctuation rate and the efficiency parameters, and generates a genetic evolution strategy that accelerates the convergence of the global optimal solution.

[0011] Furthermore, in the material guiding device described in the present invention, the optimization parameter set output by the adaptive cross probability generation unit corrects the guide surface inclination adjustment amount through the material trajectory planning module, and simultaneously sends the guide grid interlayer damping coefficient compensation instruction to the unpowered dust suppression module, so that the surface inclination adjustment and damping coefficient correction are coordinated to perform impact energy consumption suppression and dust removal efficiency improvement based on the same optimization goal.

[0012] Furthermore, the material guiding device described in the present invention, the laser radar scanning path planning sub-module of the speed collaborative control module receives the sliding trajectory model predicted landing point distribution data output by the material trajectory planning module, dynamically calculates the laser radar scanning area coverage range and sampling frequency adjustment amount according to the spatial density change of the distribution data, and generates a scanning control instruction that matches the real-time material flow state.

[0013] In a second aspect, the present invention provides a material guiding control system, which is applied to the material guiding device, comprising: The material trajectory planning module, the unpowered dust suppression module, the speed collaborative control module, the dynamic feedback adjustment module and the genetic algorithm optimization module are connected through an industrial bus to form a closed-loop control network based on real-time data exchange, wherein the dynamic feedback adjustment module synchronously uploads the vibration spectrum data collected by the high-frequency piezoelectric pressure sensor array and the speed deviation data calculated by the speed collaborative control module to the genetic algorithm optimization module; The genetic algorithm optimization module integrates the vibration spectrum data, the rotation speed deviation data and the real-time dust removal efficiency parameters to generate a global optimization parameter set including the dynamic inclination adjustment amount of the guide surface, the damper stiffness correction coefficient and the guide grid interlayer damping compensation value, and distributes the parameter set to the material trajectory planning module, the unpowered dust suppression module and the speed coordination control module through the industrial bus; The unpowered dust suppression module adjusts the structure parameters of the diversion grid according to the received damping compensation value between the layers of the diversion grid. The material trajectory planning module updates the slip trajectory model based on the dynamic inclination adjustment amount of the guiding surface and generates a new damping control instruction. The speed cooperative control module calibrates the rotational speed adjustment algorithm using the damper stiffness correction coefficient to achieve multi-objective cooperative optimization of the material slip trajectory tracking accuracy, dust suppression efficiency, and equipment vibration energy.

[0014] Advantages of the present invention; The present invention improves the technical effect through multi-module cooperative control and closed-loop optimization mechanism: The material trajectory planning module dynamically adjusts the geometric parameters of the guiding surface based on discrete element simulation, and combines with the hydraulic actuator to correct the curvature radius in real time, so that the material slip trajectory contacts the receiving equipment flexibly, reducing the collision impact energy; The unpowered dust suppression module forms a tapered air flow channel through the adaptive adjustment of the inclination angle between the layers of the diversion grid, and cooperates with the baffle array of the inertial settling cavity to achieve hierarchical dust collection, suppressing the diffusion of fine particles carried by the air flow; The speed cooperative control module integrates the lidar flow velocity data and the PID parameters optimized by the genetic algorithm, dynamically calculates the adjustment amount of the belt drive rotational speed, so that the difference between the material slip end speed and the belt line speed converges to a stable interval, reducing friction loss; The dynamic feedback adjustment module uses a high-frequency sensor array and a convolutional neural network to identify abnormal vibration modes, triggers the damper stiffness compensation and the diversion grid structure adjustment, and synchronously suppresses the equipment vibration energy; The genetic algorithm optimization module constructs a multi-objective fitness function, generates a global optimization parameter set through weight dynamic allocation and elite retention strategy, drives the cooperative evolution of the guiding surface, dust suppression structure and rotational speed control logic, and forms a continuous improvement link for impact suppression, dust control and energy consumption optimization, systematically solving the technical defects of rigid impact, dust pollution and low energy efficiency of traditional feeding devices. Description of the drawings

[0015] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the drawings.

[0016] Figure 1 It is a system architecture diagram of a control system for a feeding device provided by an embodiment of the present invention.

[0017] Figure 2 It is an external structure schematic diagram of a feeding device provided by an embodiment of the present invention. Detailed implementation manners

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention. The technical solutions provided by the embodiments of the present invention will be described in detail below with reference to the drawings. To better understand the objectives of the present invention, the present invention will be further described in detail below.

[0019] In a first aspect, please refer to Figures 1 to 2 , a material guiding device provided by the present invention includes: A material trajectory planning module that constructs a material slip trajectory model based on the discrete element simulation algorithm and generates guiding surface geometric parameters and corresponding damper stiffness control instructions according to the simulation results of the model; A passive dust suppression module that receives the guiding surface geometric parameters output by the material trajectory planning module, dynamically configures the interlayer inclination angles of the multi-layer tapered diversion grids according to the geometric parameters, and performs velocity gradient attenuation on the airflow induced by the slipping materials through the inertial sedimentation cavity to complete the inertial separation of dust particles; A velocity coordination control module that receives the damper stiffness control instructions sent by the material trajectory planning module, combines the material flow velocity distribution data captured in real time by the lidar scanning device, and calculates the target rotational speed adjustment amount of the receiving belt drive motor; A dynamic feedback adjustment module that continuously collects contact stress spectra through a high-frequency piezoelectric pressure sensor array deployed on the receiving belt support beam, extracts time-frequency features from the stress spectra through a convolutional neural network, and outputs the equipment abnormal vibration classification results after matching the preset vibration mode database; A genetic algorithm optimization module that receives the abnormal vibration classification results of the dynamic feedback adjustment module and the rotational speed adjustment amount data of the velocity coordination control module, constructs a multi-dimensional fitness evaluation function by integrating real-time operating state parameters, generates an optimization parameter set including the dynamic inclination adjustment amount of the guiding surface, the damper stiffness correction coefficient, and the diversion grid deflection compensation value, and synchronously feeds back the optimization parameter set to the material trajectory planning module for updating the slip trajectory model, the passive dust suppression module for adjusting the diversion grid structure parameters, and the velocity coordination control module for calibrating the rotational speed control logic, forming a cross-module parameter co-evolution mechanism.

[0020] The material trajectory planning module numerically simulates the sliding motion of material particles through the discrete element simulation algorithm and establishes a three-dimensional dynamic model of the material sliding trajectory. By calculating the collision frequency and energy loss of materials with different particle sizes, this model derives the geometric profile parameters of the optimal guiding surface. Based on the material distribution density and velocity field characteristics in the simulation results, it generates corresponding damper stiffness control instructions, which include damper response time and force gradient parameters, providing a motion control benchmark for subsequent modules.

[0021] After receiving the geometric parameters of the guiding surface, the passive dust suppression module analyzes the spatial curvature distribution data of the surface and dynamically adjusts the layer inclination configuration scheme of the multi-layer tapered diversion grille. The matching relationship between the inclination angles and layer spacing of each layer of the diversion grille is adaptively adjusted according to the predicted dust diffusion path in the material sliding trajectory model. An array of diversion plates is staggered in the inertial settling cavity, and by changing the cross-sectional shape of the cavity, the air flow velocity shows a stepwise decay. Dust particles impact and adsorb on the adsorption layer along the preset trajectory under inertia, achieving particle size classification separation.

[0022] The velocity coordinated control module fuses the force control parameters in the damper stiffness control instructions with the real-time material flow velocity distribution data obtained by the lidar scanning device. The sliding window algorithm is used to perform spatial interpolation calculation on the flow velocity data to generate a dynamic load distribution map for each driving section of the receiving belt. Combining the damper stiffness parameters and the load distribution characteristics, the target speed adjustment amount of the driving motor is calculated through the proportional-integral-derivative control algorithm, and the adjustment amount includes speed increment and acceleration constraint conditions.

[0023] The dynamic feedback adjustment module collects the contact stress spectrum of the support beam of the receiving belt at a millisecond-level sampling frequency through a high-frequency piezoelectric pressure sensor array. The original spectrum data is preprocessed and then input into a convolutional neural network, which extracts time-frequency domain feature vectors through multiple convolutional kernels. The feature vectors are matched with the abnormal waveform templates in the preset vibration mode database, and a classification result including vibration energy level, fault type, and location information is output. The classification result includes the energy contribution degree weights of each vibration source.

[0024] After receiving the vibration classification results and rotational speed adjustment data, the genetic algorithm optimization module constructs a multi-dimensional fitness evaluation function that includes the mean impact acceleration, unit energy consumption index, and dust concentration change rate. A dynamic weight allocation mechanism is introduced into the function to automatically adjust the priority coefficients of each index according to the equipment operation stage. The PID control parameter set that meets the multi-objective balance condition is selected through the elitist retention strategy. The parameter set includes the dynamic inclination adjustment amount of the guiding surface, the stiffness correction coefficient of the damper, and the deflection compensation value of the diversion grille. The optimized parameter set is synchronously distributed to each execution module through the industrial bus, triggering the model retraining of the material trajectory planning module, the structural parameter calibration of the passive dust suppression module, and the rotational speed closed-loop correction of the speed coordinated control module, forming a cross-module collaborative parameter evolution link.

[0025] Specifically, for the feeding device described in the present invention, the passive dust suppression module includes a diversion grille inclination angle adjustment unit that receives the dynamic distribution density data in the material slip trajectory model output by the material trajectory planning module, calculates the cross-sectional area change amount of the air flow channel between the diversion grille layers based on the density data, and synchronously adjusts the inclination angle and spacing of each layer of the diversion grille to make the spatial distribution of the air flow velocity gradient consistent with the inertial settlement path of the dust particles.

[0026] The diversion grille inclination angle adjustment unit receives the dynamic distribution density data output by the material slip trajectory model through the industrial bus. This data includes the instantaneous density distribution characteristics of the material flow at different positions on the guiding surface. The built-in spatial interpolation algorithm of the adjustment unit performs three-dimensional grid processing on the density data to generate a gas-solid two-phase flow density field distribution map of the axial distribution area of the diversion grille, and identifies the coordinate distribution characteristics of the high dust concentration area.

[0027] Based on the gradient change characteristics of the density field distribution map, the calculation module uses the flow channel cross-sectional area adaptive algorithm to calculate the cross-sectional area adjustment amount of the air flow channel between the diversion grille layers. The algorithm dynamically matches the spatial correspondence between the cross-sectional area of the air flow channel and the inertial migration path of the dust according to the dust concentration gradient difference between adjacent diversion grille layers. The calculation result is converted into the control parameters of the diversion grille actuator, including the inclination angle adjustment angle of each layer of the diversion plate and the layer spacing expansion amount.

[0028] After receiving the control parameters, the diversion grille synchronous adjustment mechanism drives the servo motor and the linear module to perform the linkage adjustment of the inclination angle and spacing. Each layer of the diversion plate forms a tapered interlayer channel structure under the hydraulic drive. The upper diversion plate adopts a small inclination angle and wide spacing layout to reduce the initial air flow velocity, and the lower diversion plate switches to a large inclination angle and narrow spacing configuration to enhance the turbulence effect. The adjusted interlayer channel of the diversion grille forms a stepped velocity decay gradient, forcing the dust particles carried by the air flow to break away from the main body of the air flow under the action of inertia.

[0029] The guide plate array built into the inertial settling chamber forms a synergistic effect with the adjusted guide grid structure. The corrugated guide surface set in the front section of the cavity induces the centrifugal movement of dust particles by changing the direction of airflow. The contraction channel in the middle section accelerates the airflow to form a negative pressure adsorption effect. The expansion section at the end achieves the inertial deposition of dust through a sudden drop in speed. The dynamic adjustment of the cross-sectional area of the airflow channel between the guide grid layers makes the sedimentation path of dust particles accurately match the structural characteristics of the cavity. Large-size particles complete collision deposition in the front section of the cavity, and fine dust is captured and filtered in the negative pressure area at the end.

[0030] The airflow velocity gradient control module monitors the dust concentration and airflow velocity distribution at the outlet of the guide grid in real time, and feeds back the detection data to the adjustment unit to form a closed-loop control. When it is detected that the airflow velocity in a specific area deviates from the preset gradient by more than the threshold, the fine-tuning compensation mechanism of the guide grid inclination is triggered. During the compensation process, the angle of the guide plate in the high dust concentration area is adjusted first, and the secondary sedimentation of dust particles is induced by local airflow acceleration, maintaining the dynamic consistency of the airflow velocity gradient and the dust sedimentation path.

[0031] Specifically, the material guiding device described in the present invention, the dynamic feedback adjustment module includes a convolutional neural network sub-module, which receives real-time stress spectrum data collected by the high-frequency piezoelectric pressure sensor array, performs time-frequency domain feature decomposition on the spectrum data to generate a multidimensional feature vector, and outputs a classification result including vibration energy level and fault type to the genetic algorithm optimization module by comparing the abnormal vibration waveform template in the preset vibration mode database.

[0032] The high-frequency piezoelectric pressure sensor array collects the contact stress spectrum data of the receiving belt support beam at a microsecond sampling frequency. The sensors in the array are distributed in an equidistant matrix to cover the key stress concentration area of the belt support beam. The collected raw spectrum data is processed by anti-aliasing filtering and transmitted to the input buffer area of the convolutional neural network submodule through the industrial bus.

[0033] The convolutional neural network submodule uses multi-scale convolution kernels to decompose the time-domain stress spectrum in the time-frequency domain. The first convolution kernel extracts the short-time energy distribution characteristics of the stress waveform, the second convolution kernel focuses on the high-frequency resonance components in the spectrum, and the third convolution kernel identifies the waveform distortion characteristics of the periodic impact signal. The output of each layer is reduced in dimension by pooling operation and then spliced to form a 256-dimensional feature vector, which contains the time-domain mutation, frequency-domain energy distribution and nonlinear harmonic component information of the stress spectrum.

[0034] The eigenvector is input into the matching engine of the preset vibration mode database, and the database stores the reference vibration waveform templates under the typical fault states of the device. The matching engine uses the dynamic time warping algorithm to calculate the similarity index between the input eigenvector and each template, and the waveform shape similarity weight and the energy distribution matching degree weight are introduced in the similarity calculation process. After the matching result is normalized, the vibration energy level score and the fault type probability distribution are generated. The energy level score reflects the potential damage degree of the vibration to the device structure, and the fault type probability distribution identifies the fault possibilities of mechanical looseness, bearing wear or belt deviation.

[0035] The classification result output module encapsulates the vibration energy level score and the fault type probability distribution into a structured data packet and transmits it to the parameter input interface of the genetic algorithm optimization module through the data bus. The time stamp and the sensor position coding information are attached to the data packet. The genetic algorithm optimization module establishes a vibration feature evolution time series model according to the time stamp and locates the spatial distribution of the abnormal vibration source in combination with the position coding information. The field definition of the structured data packet forms a mapping relationship with the input parameters of the multi-objective fitness function of the genetic algorithm optimization module, realizing the automatic conversion of the vibration feature data into the optimization parameters.

[0036] The preset vibration mode database has a built-in self-learning mechanism. When the optimization parameters fed back by the genetic algorithm optimization module cause the vibration energy level score to continuously decrease, the database template update process is triggered. In the update process, the current eigenvector is associated and stored with the optimized device state data and participates in the subsequent matching calculation as a new reference template. The self-learning mechanism retains the effective vibration feature data of the most recent 12 hours through a rolling time window and dynamically optimizes the fault mode characterization ability of the database.

[0037] Specifically, for the material guiding device described in the present invention, the genetic algorithm optimization module distributes the weight coefficients of the impact acceleration, the energy consumption efficiency and the dust concentration in the multi-objective fitness function according to the vibration energy level and the fault type classification result, and uses the elitist retention strategy to screen out the PID control parameter set that meets the preset dynamic balance condition, and injects the parameter set into the speed regulation algorithm of the speed coordination control module to generate a closed-loop feedback control instruction.

[0038] After receiving the vibration energy level score and the fault type probability distribution data, the genetic algorithm optimization module analyzes the potential damage coefficient in the energy level score and the device deterioration trend index corresponding to the fault type. The impact acceleration tolerance threshold, the energy consumption efficiency reference value and the dust concentration critical value are preset in the multi-objective fitness function. According to the correlation analysis result of the damage coefficient and the deterioration trend index, the weight coefficients of each optimization target are dynamically distributed. When the damage coefficient exceeds the safety threshold, the impact acceleration weight is increased; when the deterioration trend points to mechanical wear, the energy consumption efficiency weight is increased; when the dust concentration approaches the critical value, the dust suppression weight is preferentially distributed.

[0039] The elite retention strategy creates parameter combinations including the PID proportional band, integral time, and derivative coefficient during the initial population generation stage, and calculates the comprehensive scores of each parameter set through the fitness function. In each iteration, the top 10% of the elite individuals are directly retained and enter the next-generation population, and the remaining individuals generate new parameter sets through crossover and mutation operations. The fitness score calculation introduces a dynamic adjustment mechanism for the weight coefficient. When the scores of the optimal individuals do not improve for three consecutive generations, the weight coefficient reallocation process is triggered to break through the local optimal solution.

[0040] The selected PID control parameter sets are injected into the speed regulation algorithm of the speed coordination control module through the data bus. The parameter sets include the proportional action intensity, integral response speed, and derivative prediction depth parameters. The speed regulation algorithm synchronously updates the new parameters with the control period and loads the updated PID parameters at the beginning of the next control period to perform speed closed-loop regulation. The control instruction generation module calculates the torque compensation value and acceleration constraint conditions of the drive motor based on the PID output, and generates a closed-loop feedback control instruction including the target speed curve and the dynamic adjustment threshold.

[0041] The closed-loop feedback data acquisition module captures the equipment vibration spectrum and dust concentration change data after speed regulation in real time, and feeds back the operating state parameters to the genetic algorithm optimization module to form a data closed-loop. The feedback data includes the evaluation indicators of the actual effect of the PID parameters, which are used to optimize the initial population generation strategy of the next round of genetic algorithm. The data closed-loop mechanism maintains the multi-objective dynamic balance of shock suppression, energy consumption control, and dust treatment through the synergistic effect of periodic parameter optimization and real-time feedback regulation.

[0042] Specifically, for the material guiding device described in the present invention, the material trajectory planning module includes a dynamic damping compensation unit, which receives the damper stiffness correction coefficient output by the genetic algorithm optimization module, and based on this coefficient, calculates the curvature radius adjustment amount at the end of the guiding surface in real time, and dynamically corrects the surface geometry through a mechanical actuator, so that the difference between the velocity at the end of material slip and the target belt line velocity calculated by the velocity coordination control module converges to a preset allowable range.

[0043] The dynamic damping compensation unit receives the damper stiffness correction coefficient data packet sent by the genetic algorithm optimization module through the industrial bus, and analyzes the stiffness correction gradient value and the action time window parameter in the data packet. The curvature calculation engine built in the compensation unit calculates the instantaneous adjustment amount of the end curvature radius based on the mapping relationship between the correction coefficient and the current guiding surface geometric parameters, and the adjustment amount includes the curvature center offset coordinates and the curvature change rate constraint conditions.

[0044] The amount of curvature radius adjustment is transmitted to the control unit of the hydraulic drive mechanism. The control unit generates a servo valve opening control command according to the curvature change rate constraint. The hydraulic actuator drives the deformable alloy plate at the end of the guiding surface to generate elastic deformation. During the deformation process, the coordinates of the curvature center move along a preset trajectory, so that the tangent angle at the end of the surface matches the material slip direction in real time. The strain sensor array arranged on the surface of the alloy plate monitors the stress distribution during the deformation process, and feeds the deformation calibration data back to the curvature calculation engine to form a closed-loop correction.

[0045] The speed difference monitoring module continuously collects the difference data between the material slip end speed and the target belt line speed. The difference data calculates the standard deviation and the mean offset through the sliding time window algorithm. When it is monitored that the standard deviation exceeds the preset threshold, a dynamic compensation mechanism for the damper stiffness correction coefficient is triggered, and a parameter re-optimization request signal is sent to the genetic algorithm optimization module. The compensation mechanism preferentially adjusts the curvature segment corresponding to the area where the difference continuously increases, and induces the gradient change of the material slip speed through local curvature mutation.

[0046] The closed-loop convergence verification module periodically compares the speed difference distribution characteristics before and after adjustment. The verification data includes the difference convergence rate, the steady-state fluctuation range, and the number of abnormal mutations. The verification result is fed back to the model training unit of the material trajectory planning module through the data bus. The training unit updates the damping correlation parameters in the slip trajectory model based on the actual convergence effect. The parameter set after the model update is synchronized to the initial population generation strategy of the genetic algorithm optimization module, forming a complete control link of parameter optimization - execution correction - effect verification - model iteration.

[0047] Specifically, for the material guiding device of the present invention, the dynamic damping compensation unit receives the material slip speed deviation value calculated by the speed coordination control module through the industrial bus. When the deviation value exceeds the preset threshold, a parameter reconfiguration request signal is sent to the genetic algorithm optimization module, triggering the generation process of the global optimization parameter set based on the current operating state.

[0048] The dynamic damping compensation unit receives the data packet of the material slip speed deviation value transmitted by the speed coordination control module through the industrial bus. The data packet contains the real-time speed difference, the historical trend data, and the corresponding timestamp information. The deviation value analysis module calculates the standard deviation and the mean offset of the difference data through the sliding time window algorithm, and determines the dynamic threshold range of the current deviation value in combination with the equipment operating condition evaluation model.

[0049] When the parsing module detects that the real-time speed difference continuously exceeds the dynamic threshold by three standard deviations, the parameter reconfiguration request generation mechanism is triggered. The request signal generation unit encapsulates the spatial coordinates, duration, and fluctuation characteristics of the deviation value overrun area into a structured request message, and transmits it to the instruction receiving interface of the genetic algorithm optimization module through a priority queue. The time-sensitive network protocol is used during the message transmission process to ensure transmission timeliness and avoid transmission delays of control instructions.

[0050] After receiving the request message, the genetic algorithm optimization module starts an emergency optimization thread to parse the abnormal area feature parameters in the message. The optimization thread calls the current device status snapshot data to construct a temporary fitness function, and strengthens the weight coefficient of the corresponding guiding surface segment in the abnormal area in the function. A local optimization parameter set for the abnormal area is generated through the elite population accelerated iteration algorithm. The parameter set includes the guiding surface curvature compensation value, the damper stiffness gradient adjustment amount, and the guide vane deflection angle correction coefficient.

[0051] The global optimization parameter set distribution module fuses the local optimization parameters with the conventional optimization parameters. After eliminating parameter conflicts through a data verification mechanism, it synchronizes them to the dynamic damping compensation unit and the speed coordination control module via an industrial bus. The dynamic damping compensation unit preferentially loads the curvature compensation value and the stiffness gradient adjustment amount, and drives the hydraulic actuator to complete the geometric correction of the guiding surface within milliseconds. The corrected surface geometric parameters are fed back to the material trajectory planning module in real time, triggering an incremental update operation of the slip trajectory model.

[0052] After the parameters are loaded, the speed difference closed-loop monitoring module starts a high-frequency sampling mode to collect the material slip speed data in the correction area. The monitoring data is analyzed by trend analysis to generate an optimization effect evaluation report, which includes the difference convergence rate, the steady-state fluctuation range, and the secondary overrun probability index. The evaluation report is transmitted back to the historical database of the genetic algorithm optimization module through the data bus, and is used to optimize the threshold determination logic and parameter generation strategy of subsequent request messages.

[0053] Specifically, for the material guiding device described in the present invention, the adaptive crossover probability generation unit of the genetic algorithm optimization module receives the belt load volatility data collected by the dynamic feedback adjustment module and the real-time dust removal efficiency parameters of the passive dust suppression module, and dynamically calculates the weight distribution of the crossover operator according to the correlation analysis result of the volatility and efficiency parameters, and generates a genetic evolution strategy that accelerates the convergence of the global optimal solution.

[0054] The adaptive crossover probability generation unit synchronously receives the data stream of the belt load volatility and the real-time dust removal efficiency parameter set of the passive dust suppression module uploaded by the dynamic feedback regulation module through the industrial bus. The data preprocessing module calculates the sliding window variance of the volatility data, extracts the load fluctuation amplitude characteristics and cycle characteristics, and simultaneously analyzes the dust capture rate and pressure drop loss indicators in the dust removal efficiency parameters to form a multi-dimensional feature matrix and input it into the correlation analysis engine.

[0055] The correlation analysis engine uses the mutual information entropy algorithm to calculate the coupling relationship coefficient between the load fluctuation amplitude and the dust removal efficiency parameters. When the load fluctuation amplitude increases accompanied by a decrease in the dust removal efficiency, it is determined that there is a risk of the diversion grid failure caused by uneven material distribution, and a high-priority optimization instruction is generated. The coupling relationship coefficient is matched with the preset working condition mode library, and the correlation probability distribution of belt deviation, diversion grid blockage, or material adhesion is output as the decision basis for the crossover operator weight allocation.

[0056] The crossover operator weight dynamic calculation module adjusts the search direction of the genetic algorithm according to the correlation probability distribution. When the probability of diversion grid blockage increases, the weight coefficient of the crossover operator for the geometric parameters of the guiding surface is increased; when the probability of material adhesion dominates, the operation frequency of the crossover operator for the damper stiffness parameters is increased. The weight coefficient forms a probability distribution matrix through normalization processing to control the crossover recombination priority of different gene segments and guide the population to rapidly evolve towards the high fitness region.

[0057] The genetic evolution strategy execution unit adopts an adaptive multi-point crossover operation and dynamically selects crossover points on the basis of retaining elite individuals. The selection density of the crossover points is positively correlated with the weight coefficient distribution, and intensive crossover operations are performed on the gene segments corresponding to the high-weight parameters. After each iteration, the population diversity index is evaluated. When the index is lower than the critical value, the weight coefficient rebalancing mechanism is triggered to prevent premature convergence of the algorithm.

[0058] The optimized genetic evolution strategy is output to the parameter set generation module to generate a control parameter package containing the crossover probability gradient value, the elite retention ratio, and the mutation threshold. The control parameter package is injected into the main thread of the genetic algorithm optimization through the data bus, and the main thread updates the evolution strategy configuration table according to the parameter package. The configuration table update signal synchronously triggers the model pre-training of the material trajectory planning module, and the optimized parameter set is pre-loaded to reduce the iteration delay.

[0059] The strategy effect feedback loop real-time collects the convergence speed of the genetic algorithm and the fitness value of the global optimal solution, and transmits the evolution efficiency index back to the adaptive crossover probability generation unit. After trend analysis of the feedback data, the weight mapping rule of the correlation analysis engine is dynamically corrected. When it is detected that the fitness value has not improved for three consecutive generations, the standby weight allocation strategy is automatically switched, and the population initialization process is restarted to break through the local optimal constraint.

[0060] Specifically, in the material guiding device described in the present invention, the optimization parameter set output by the adaptive cross probability generation unit corrects the guide surface inclination adjustment amount through the material trajectory planning module, and simultaneously sends the guide grid interlayer damping coefficient compensation instruction to the unpowered dust suppression module, so that the surface inclination adjustment and damping coefficient correction are coordinated to perform impact energy consumption suppression and dust removal efficiency improvement based on the same optimization goal.

[0061] The adaptive crossover probability generation unit encapsulates the optimization parameter set into a structured data packet containing the guide surface inclination adjustment priority mark and the damping coefficient compensation mark, and distributes it to the data receiving port of the material trajectory planning module and the unpowered dust suppression module through the industrial bus. A time synchronization mark is set in the data packet to ensure that the parameter loading actions of the two modules are executed within the same control cycle.

[0062] The parameter parsing unit of the material trajectory planning module extracts the inclination adjustment amount in the data packet and performs dynamic calculation based on the current guide surface geometry model. The calculation process uses the inverse kinematics algorithm to derive the displacement control amount of the hydraulic actuator, which includes the stroke step length and speed curve parameters of each adjustment segment. The inclination adjustment instruction is sent to the distributed hydraulic drive unit to drive the guide surface segmented adjustment mechanism to change the surface inclination angle according to the preset gradient, and the stress change data at the surface joint is monitored in real time during the adjustment process.

[0063] After receiving the damping coefficient compensation instruction, the unpowered dust suppression module activates the interlayer damping adjustment mechanism of the guide grid. The instruction parsing unit identifies the target guide grid level according to the compensation mark and extracts the damping correction coefficient of the corresponding level. The guide grid actuator uses a piezoelectric ceramic brake to adjust the opening and closing gap of the interlayer damping plate. The gap adjustment amount is nonlinearly related to the damping correction coefficient. During the adjustment process, the interlayer airflow velocity distribution data is synchronously collected to verify the damping correction effect.

[0064] The collaborative control logic module monitors the execution progress data of the tilt angle adjustment and damping correction. When it is detected that both adjustment actions have entered the steady-state stage, the joint evaluation process of impact energy consumption and dust removal efficiency is started. The impact energy consumption evaluation is based on the mechanical impact spectrum energy integral value collected by the high-frequency vibration sensor, and the dust removal efficiency evaluation uses the outlet dust concentration gradient data obtained by the laser particle counter. The evaluation results are compared with the optimization target setting value for deviation, and the parameter fine-tuning instructions are generated and fed back to the adaptive crossover probability generation unit.

[0065] The closed-loop collaborative optimization mechanism establishes real-time linkage between parameter adjustment and effect verification through the data bus. When the dust removal efficiency improvement does not meet expectations, the damping correction coefficient of the upper layer of the guide grid is increased first; if the impact energy consumption suppression effect is insufficient, the inclination adjustment priority of the middle section of the guide surface is increased. The collaborative optimization instruction triggers the local re-optimization thread of the genetic algorithm, implements a directed mutation operation based on the established parameter set, and generates an enhanced optimization parameter subset.

[0066] The parameter version management unit stores the iteration mark and effect association of the optimized parameter set. Each version parameter set is associated with the corresponding impact energy consumption suppression rate and dust removal efficiency improvement rate indicators. The historical data comparison module periodically analyzes the version iteration effect. When it detects that there is no significant improvement in the key indicators of three consecutive versions, it triggers the global optimization target weight redistribution process to redefine the coordinated optimization balance point of impact energy consumption and dust removal efficiency.

[0067] Specifically, the material guiding device described in the present invention, the laser radar scanning path planning sub-module of the speed collaborative control module receives the sliding trajectory model predicted landing point distribution data output by the material trajectory planning module, dynamically calculates the laser radar scanning area coverage range and sampling frequency adjustment amount according to the spatial density changes of the distribution data, and generates a scanning control instruction that matches the real-time material flow state.

[0068] The laser radar scanning path planning submodule receives the sliding trajectory model predicted landing point distribution data packet transmitted by the material trajectory planning module through the industrial bus. The data packet contains the spatial coordinate matrix, timestamp information and density distribution feature vector of the predicted landing point. The data verification unit performs frame verification and time synchronization on the received data, removes abnormal timestamp data and stores it in the ring buffer. The buffer sets a sliding window mechanism to retain the valid data of the last 5 seconds.

[0069] The spatial density change analysis engine processes the point distribution data of the buffer zone in three-dimensional grids to construct a dynamic density field model. The model uses a density gradient algorithm to identify the boundary coordinates of high-density aggregation areas and calculate the density change rate of each dimension. The material flow classification rules are embedded in the density field model. When the longitudinal density gradient is detected to be three times greater than the lateral gradient, it is determined to be a laminar-dominated state; when the lateral density fluctuation amplitude continues to be greater than the preset threshold, it is marked as a turbulent mixing state.

[0070] The scanning area solution module uses the convex hull algorithm to calculate the geometric center coordinates and boundary range of the minimum coverage area based on the output results of the density field model. A safety margin factor is introduced in the process of solving the area coverage range, and a 10% buffer area is extended outside the predicted landing point. The sampling frequency adjustment amount is dynamically calculated based on the density gradient change rate. The high-gradient area uses a millisecond sampling frequency, and the low-gradient area switches to a hundred-millisecond intermittent sampling mode to balance data accuracy and system load.

[0071] The scan control instruction generation unit converts the calculation result into a control parameter sequence of the lidar device. The parameter sequence includes a set of scan path point coordinates, a dwell time, and an optical intensity adjustment instruction. The set of path point coordinates generates a smooth scan trajectory through the cubic spline interpolation algorithm, avoiding detection blind spots caused by sudden stops and starts of the mechanical mechanism. After being encapsulated by the protocol, the control parameter sequence is transmitted to the lidar drive unit through the real-time Ethernet. After parsing the instruction, the drive unit controls the galvanometer motor and the laser emitter to perform the scan action.

[0072] During the execution of the scan action, the lidar point cloud data acquisition module captures the material flow state characteristics in real time, and the acquired data is transmitted back to the model correction unit of the material trajectory planning module through the data pipeline. The model correction unit compares the spatial matching degree between the predicted landing point distribution and the actual scan data. When the matching degree is lower than the set threshold, it triggers the parameter recalibration process of the slip trajectory model, forming a two-way data closed-loop of scan control and model optimization.

[0073] The dynamic response compensation mechanism is started after the scan control instruction is issued, and monitors the deviation value between the actual movement trajectory of the lidar actuator and the theoretical path. When it is detected that the response delay of the galvanometer motor exceeds the allowable range, the feedforward compensation algorithm for path point coordinates is started, and the corrected coordinate instruction is issued in advance in the next control cycle. The compensation algorithm establishes a motor dynamic response model based on historical deviation data, predicts the trajectory offset under different motion states, and implements pre-compensation.

[0074] In a second aspect, the present invention provides a material guiding control system applied to the material guiding device, including: The material trajectory planning module, the passive dust suppression module, the speed coordination control module, the dynamic feedback adjustment module, and the genetic algorithm optimization module are connected through an industrial bus to form a closed-loop control network based on real-time data exchange. Among them, the dynamic feedback adjustment module synchronously uploads the vibration spectrum data collected by the high-frequency piezoelectric pressure sensor array and the rotational speed deviation data calculated by the speed coordination control module to the genetic algorithm optimization module; The genetic algorithm optimization module fuses the vibration spectrum data, the rotational speed deviation data, and the real-time dust removal efficiency parameters to generate a global optimization parameter set including the dynamic inclination adjustment amount of the guiding surface, the stiffness correction coefficient of the damper, and the interlayer damping compensation value of the diversion grid, and distributes the parameter set to the material trajectory planning module, the passive dust suppression module, and the speed coordination control module through the industrial bus; The unpowered dust suppression module adjusts the structural parameters of the diversion grid according to the received damping compensation value between the layers of the diversion grid. The material trajectory planning module updates the slip trajectory model based on the dynamic inclination adjustment amount of the guiding surface and generates a new damping control command. The speed co-control module calibrates the rotational speed adjustment algorithm using the damper stiffness correction coefficient to achieve multi-objective collaborative optimization of the material slip trajectory tracking accuracy, dust suppression efficiency, and equipment vibration energy.

[0075] The closed-loop control network of the material guiding control system establishes a real-time data exchange channel between modules through the industrial bus. The high-frequency piezoelectric pressure sensor array deployed in the dynamic feedback adjustment module captures the vibration spectrum data of the receiving belt support beam at a millisecond-level sampling frequency. After preprocessing, the data is time-stamped and aligned with the rotational speed deviation value calculated by the speed co-control module to form a synchronous data stream and upload it to the input interface of the genetic algorithm optimization module. A redundant check mechanism is used during the data synchronization process to eliminate transmission errors and ensure the spatio-temporal correlation between the vibration characteristics and the rotational speed control state.

[0076] After receiving the vibration spectrum data, rotational speed deviation data, and the real-time dust removal efficiency parameters uploaded by the unpowered dust suppression module, the data fusion engine of the genetic algorithm optimization module constructs a multi-dimensional feature matrix including the time-domain vibration energy integral, rotational speed tracking error rate, and dust concentration gradient. The fusion engine calculates the contribution weight of each parameter through the weighted entropy value analysis method, and the weight distribution result is matched with the current operating mode of the equipment: the vibration suppression weight is strengthened under high-load conditions, and the dust removal efficiency weight is preferentially allocated during steady-state operation. Based on the weight distribution, a parameter combination of the dynamic inclination adjustment amount of the guiding surface, the damper stiffness correction coefficient, and the damping compensation value between the layers of the diversion grid is generated, and the parameter set is distributed to each execution module through the multicast protocol of the industrial bus.

[0077] After receiving the dynamic inclination adjustment amount, the material trajectory planning module starts the incremental update process of the slip trajectory model. During the update process, the backpropagation algorithm is used to adjust the geometric parameters of the guiding surface, and combined with the damper stiffness correction coefficient issued by the genetic algorithm optimization module, a new control command including the curvature radius compensation value and the hydraulic drive displacement amount is generated. The new command is sent to the surface adjustment mechanism through the distributed control unit to drive the deformable guiding plate to achieve millisecond-level dynamic adjustment of the surface geometry.

[0078] The unpowered dust suppression module analyzes the damping compensation value between the layers of the diversion grid and adjusts the gap of the damping plates of each layer of the diversion grid through the piezoelectric actuator. The compensation value is mapped to a non-linear function of the gap opening. A high compensation value corresponds to a small gap to enhance the turbulence effect, and a low compensation value enlarges the gap to reduce the pressure loss. During the adjustment process, the air velocity distribution between the layers is monitored in real time, the damping correction effect is verified through the feedback data, and the actual adjustment amount is sent back to the parameter correction interface of the genetic algorithm optimization module.

[0079] The speed collaborative control module injects the damper stiffness correction coefficient into the rotational speed regulation algorithm to update the response parameters of the proportional-integral-derivative controller. The real-time material flow rate data captured by the lidar scanning device interacts with the algorithm parameters to calculate the target rotational speed curve of the drive motor. The rotational speed control command is sent to the frequency converter via the fieldbus, and the drive motor executes rotational speed tracking according to the updated acceleration constraint and torque compensation value. Meanwhile, the actual rotational speed deviation data is fed back to the closed-loop network.

[0080] Multi-objective collaborative optimization achieves dynamic balance through the data closed-loop mechanism. The tracking accuracy of the material slip trajectory is guaranteed by the collaborative effect of the curved surface geometry adjustment and rotational speed control; the dust suppression efficiency is optimized by the damper damping correction of the diversion grille and the air flow control of the inertial sedimentation cavity; the vibration energy of the equipment is jointly suppressed by the extraction of vibration frequency spectrum characteristics and the adjustment of stiffness parameters. The execution effect data of each module is periodically summarized to the genetic algorithm optimization module to trigger the iterative update of the parameter set, forming a continuous evolution link of "data acquisition - parameter optimization - execution feedback".

[0081] The specific implementation manner of the present invention relates to the material guiding control process in metallurgical equipment, and realizes the optimization of the material slip trajectory, dust suppression and equipment vibration control through multi-module collaboration. The material trajectory planning module constructs a three-dimensional slip trajectory model based on the discrete element simulation algorithm. By calculating the collision frequency and energy loss of materials with different particle sizes, the geometric contour parameters of the guiding surface and the damper stiffness control command are generated. The material distribution density and velocity field characteristics extracted from the simulation results are used as basic data to provide a motion control reference for subsequent modules. The guiding surface parameters are transmitted to the passive dust suppression module via the industrial bus to drive the dynamic adjustment of the interlayer inclination angle of the multi-layer tapered diversion grille. The spacing and inclination angle of each layer of the diversion grille form a stepped air flow channel according to the adaptive adjustment of the dust diffusion path, and the inertial separation of dust particles is realized in combination with the deflector array of the inertial sedimentation cavity.

[0082] After receiving the damper stiffness control command, the speed collaborative control module combines the real-time material flow rate distribution data captured by the lidar scanning device, and uses the sliding window algorithm to perform spatial interpolation on the flow rate to generate the dynamic load distribution map of each drive section of the receiving belt. The target rotational speed adjustment amount of the drive motor is calculated through the proportional-integral-derivative control algorithm, and the adjustment amount includes the rotational speed increment and the acceleration constraint condition. The dynamic feedback adjustment module collects the contact stress spectrum of the receiving belt support beam at a millisecond-level sampling frequency through a high-frequency piezoelectric pressure sensor array. After extracting the time-frequency domain feature vector through a convolutional neural network, it is matched with the abnormal waveform template in the preset vibration mode database, and outputs the vibration energy level and the fault type classification result.

[0083] The genetic algorithm optimization module receives the vibration classification results and rotational speed adjustment data, and constructs a multi-dimensional fitness evaluation function including the mean impact acceleration, unit energy consumption index, and dust concentration change rate. Through the dynamic weight allocation mechanism, the priorities of each index are adjusted according to the equipment operation stage, and the elite retention strategy is used to screen the PID control parameter set that meets the multi-objective balance condition. The optimized parameter set includes the dynamic inclination adjustment amount of the guiding surface, the stiffness correction coefficient of the damper, and the deflection compensation value of the diversion grille, and is distributed to each execution module synchronously through the industrial bus. The material trajectory planning module updates the slip trajectory model based on the dynamic inclination adjustment amount, and drives the hydraulic actuator to correct the geometric shape of the guiding surface; the passive dust suppression module adjusts the damper plate gap according to the interlayer damping compensation value of the diversion grille to optimize the air flow control; the speed coordination control module calibrates the rotational speed adjustment algorithm and generates a closed-loop feedback control command.

[0084] The closed-loop control network ensures the data synchronization accuracy through the redundancy check mechanism. The dynamic damping compensation unit monitors the difference between the material slip speed and the target belt line speed in real time, and triggers the parameter reconfiguration process when the deviation exceeds the threshold. The adaptive crossover probability generation unit analyzes the correlation between the belt load volatility and the dust removal efficiency parameters, dynamically adjusts the weight distribution of the crossover operator of the genetic algorithm, and accelerates the convergence of the global optimal solution. The lidar scanning path planning sub-module dynamically calculates the scanning area coverage and sampling frequency according to the predicted landing point distribution data of the slip trajectory model, generates control commands matching the real-time material flow state, and corrects the response delay of the galvanometer motor through the feedforward compensation algorithm.

[0085] During the implementation process, the guide plate array of the inertial sedimentation cavity and the diversion grille structure cooperate. The front-wave guide surface induces the centrifugal movement of dust, the middle-section contraction channel forms a negative pressure adsorption, and the last-section expansion area realizes inertial deposition. The vibration spectrum data of the high-frequency piezoelectric pressure sensor array updates the database template through the self-learning mechanism, and continuously optimizes the fault identification accuracy. The multi-objective collaborative optimization dynamically balances the slip trajectory tracking accuracy, dust suppression efficiency, and equipment vibration energy through the data closed-loop mechanism, forming a continuous evolution link of "data acquisition, parameter optimization, and execution feedback" to meet the high-precision material guiding and low-energy consumption operation requirements of metallurgical equipment.

[0086] The genetic algorithm optimization module determines the weights through the following technical solutions: Dynamic weight allocation mechanism: Receive the vibration energy level score and fault type probability distribution data of the dynamic feedback adjustment module, and analyze the potential damage coefficient and deterioration trend index of the equipment. Construct a multi-objective fitness function according to the real-time operation state parameters, and define the impact acceleration tolerance threshold, energy consumption efficiency benchmark value, and dust concentration critical value. When the vibration energy level exceeds the safety threshold, increase the weight coefficient of the impact acceleration; when a mechanical wear type fault is detected, increase the energy consumption efficiency weight; if the dust concentration approaches the critical value, give priority to allocating the dust suppression weight.

[0087] Relevance analysis engine: The mutual information entropy algorithm is used to calculate the coupling relationship coefficient between the belt load volatility and the dust removal efficiency parameter. When the load fluctuation amplitude increases accompanied by a decrease in the dust removal efficiency, the failure risk of the diversion grille is determined, and a high-priority optimization instruction is generated. By matching the preset operating condition mode library (such as scenarios like belt deviation, diversion grille blockage, etc.), the relevance probability distribution is output as the basis for weight allocation.

[0088] Elite retention strategy screening: Individuals containing PID parameter combinations are generated in the initial population, and the comprehensive score is calculated through the fitness function. In each round of iteration, the top 10% of the elite individuals are directly retained and enter the next generation, and their weight allocation scheme is used as the benchmark. The remaining individuals generate new parameter sets through crossover and mutation. When calculating the fitness score, a weight dynamic adjustment mechanism is introduced. If the optimal individual has not improved for three consecutive generations, the weights are reallocated to break through the local optimum.

[0089] Feedback closed-loop correction: After the optimized parameter set is injected into the execution module, the vibration spectrum, dust concentration change, and energy consumption data after speed regulation are collected in real time. Through the effect evaluation indicators (such as the difference convergence rate, steady-state fluctuation range), it is transmitted back to the genetic algorithm module to correct the subsequent weight mapping rules. The historical data comparison module periodically analyzes the effect of version iteration. When there is no significant improvement in the key indicators, the global weight reallocation process is triggered.

[0090] Multi-objective collaborative constraint: The weight allocation needs to meet the dynamic balance of shock suppression, energy consumption control, and dust treatment. For example, when adjusting the interlayer damping compensation value of the diversion grille, if the dust removal efficiency does not meet the expectation, the upper layer damping weight is increased; if the shock energy consumption suppression is insufficient, the priority of adjusting the inclination angle of the middle section of the guiding surface is increased, and multi-objective optimization is achieved through parameter collaboration.

[0091] Through real-time data perception, relevance analysis, and feedback closed-loop, the present invention makes the weight allocation match the equipment operating conditions, meets the requirements of the patent law for the clarity and feasibility of the technical solution, and effectively supports the technical features of the multi-objective fitness function in the claims.

[0092] Embodiment 1: Dynamic adjustment of the guiding surface and speed collaborative control; Deploy a material guiding device in the metallurgical production line. The material trajectory planning module uses the discrete element simulation algorithm to construct an iron ore powder sliding trajectory model. By calculating the collision energy distribution of materials with different particle sizes, generate the geometric parameters of the guiding surface (curvature radius 12.5 cm, inclination angle 35°) and the corresponding damper stiffness control command (stiffness coefficient 1800 N / mm). The dynamic damping compensation unit receives the stiffness correction coefficient sent by the genetic algorithm optimization module, and calculates the adjustment amount of the curvature radius at the end of the guiding surface in real time (±0.8 cm), and drives the hydraulic actuator to complete the geometric correction of the surface within 0.5 seconds. The speed coordination control module combines the real-time flow velocity data scanned by the lidar (peak flow velocity 2.8 m / s) to calculate the target rotational speed adjustment amount of the receiving belt driving motor (+15 rpm), so that the difference between the sliding end velocity of the material and the belt line velocity is stabilized within the range of ±0.2 m / s. After implementation, the detection shows that the impact acceleration of the material decreases and the belt deviation rate decreases.

[0093] Example 2: Dynamic dust suppression and air flow control of the diversion grille; For the copper concentrate transportation scenario, the passive dust suppression module receives the dynamic distribution density data (high density area > 1.8 g / cm³) output by the material trajectory planning module. The diversion grille tilt angle adjustment unit calculates the adjustment amount of the cross-sectional area of the interlayer air flow channel (the upper layer is widened by 8 mm, and the lower layer is narrowed by 5 mm) based on the density gradient difference, and synchronously adjusts the tilt angle of the diversion grille (the upper layer 22° → 18°, the lower layer 30° → 35°). The baffle plate array in the inertial sedimentation cavity makes the air flow velocity decay step by step from 4.5 m / s to 1.2 m / s through the contraction-expansion structure. Dust particles (>50 μm) complete inertial deposition in the front section of the cavity, and fine dust (<10 μm) is captured in the negative pressure area at the end section. The measured data shows that the outlet dust concentration drops from 120 mg / m³ to 28 mg / m³, and the pressure loss is controlled within 350 Pa.

[0094] Example 3: Multi-objective genetic algorithm optimization and closed-loop feedback; In the continuous bauxite conveying system, the genetic algorithm optimization module integrates vibration spectrum data (energy integral value of 0.45 J), rotational speed deviation (±1.2 rpm), and dust removal efficiency parameter (88%) to construct a multi-objective fitness function. The weight distribution is determined through the mutual information entropy algorithm: the impact acceleration weight is 0.6, the energy consumption efficiency is 0.25, and the dust concentration is 0.15. The elite retention strategy is used to screen out the PID parameter set (proportional band of 220, integral time of 1.8 s, derivative coefficient of 0.05). After injecting it into the speed collaborative control module, the tracking error rate of the driving motor speed is reduced from 5.7% to 1.3%. The adaptive crossover probability generation unit dynamically adjusts the crossover operator weight according to the belt load volatility (standard deviation of 0.35 t / h), shortening the generation time of the global optimization parameter set to 8 seconds per time. The closed-loop feedback data shows that after optimization, the vibration energy of the equipment is reduced by 58%, and the energy consumption per ton of materials is decreased by 12%, achieving continuous and stable operation for 72 hours. The present invention solves the problems of rigid impact and dust pollution of traditional feeding devices through the collaborative control of the dynamic guiding surface and air flow. The material trajectory planning module generates the geometric parameters of the guiding surface based on the discrete element simulation algorithm, and dynamically adjusts the curvature radius and inclination angle of the surface through the hydraulic actuator, so that the sliding trajectory of the material is flexibly matched with the receiving equipment, reducing the collision energy. The passive dust suppression module dynamically configures the inclination angle between the diversion grating layers according to the surface parameters to form a tapered air flow channel, and combines the diversion plate array of the inertial sedimentation cavity to achieve the attenuation of the air flow velocity gradient, inducing the inertial deposition of dust particles in a specific area and suppressing the diffusion of dust to the external environment.

[0095] Aiming at the problem of energy consumption and speed matching, the speed collaborative control module analyzes the material flow velocity distribution data scanned by the lidar in real time, and dynamically adjusts the rotational speed of the receiving belt in combination with the PID parameters optimized by the genetic algorithm, so that the difference between the velocity at the end of the material slip and the belt line velocity converges. The dynamic feedback adjustment module collects the vibration spectrum of the equipment through a high-frequency piezoelectric sensor. After identifying the abnormal vibration mode through the convolutional neural network, it triggers the stiffness correction of the damper and the adjustment of the diversion grating structure, reducing the friction loss caused by speed mismatch of the equipment and synchronously optimizing the energy consumption efficiency.

[0096] The closed-loop control network realizes systematic improvement through multi-module data fusion and adaptive optimization. The genetic algorithm optimization module integrates vibration energy, dust removal efficiency, and rotational speed deviation data, constructs a multi-objective fitness function to dynamically allocate weight coefficients, generates a global optimization parameter set and distributes it to each execution module. The material trajectory planning, dust suppression structure, and rotational speed control co-evolve based on real-time feedback data, forming a continuous iterative link of "parameter optimization, execution correction, and effect verification", achieving a dynamic balance between suppressing impact, reducing dust, and improving energy efficiency, and systematically solving the multi-dimensional defects of traditional feeding devices.

Claims

1. A material guiding device, characterized in that, Including: A material trajectory planning module that constructs a material slip trajectory model based on a discrete element simulation algorithm and generates guide surface geometric parameters and corresponding damper stiffness control instructions according to the simulation results of the model; A passive dust suppression module that receives the guide surface geometric parameters output by the material trajectory planning module, dynamically configures the interlayer inclination angles of a multi-layer tapered diversion grid according to the geometric parameters, and implements velocity gradient attenuation on the airflow induced by the slipping material through an inertial sedimentation cavity to complete the inertial separation of dust particles; A velocity coordination control module that receives the damper stiffness control instructions sent by the material trajectory planning module, combines the material flow velocity distribution data captured in real time by a lidar scanning device, and calculates the target rotational speed adjustment amount of the receiving belt drive motor; A dynamic feedback adjustment module that deploys a high-frequency piezoelectric pressure sensor array on the receiving belt support beam to continuously collect contact stress spectra, extracts time-frequency features from the stress spectra through a convolutional neural network, and outputs the device abnormal vibration classification result after matching a preset vibration mode database; A genetic algorithm optimization module that receives the abnormal vibration classification result of the dynamic feedback adjustment module and the rotational speed adjustment amount data of the velocity coordination control module, fuses real-time operating state parameters to construct a multi-dimensional fitness evaluation function, generates an optimization parameter set including the dynamic inclination angle adjustment amount of the guide surface, the damper stiffness correction coefficient, and the diversion grid deflection compensation value, and synchronously feeds back the optimization parameter set to the material trajectory planning module for updating the slip trajectory model, the passive dust suppression module for adjusting the diversion grid structure parameters, and the velocity coordination control module for calibrating the rotational speed control logic, forming a cross-module parameter collaborative evolution mechanism.

2. The material guiding device according to claim 1, wherein: The passive dust suppression module includes a diversion grid inclination angle adjustment unit that receives the dynamic distribution density data in the material slip trajectory model output by the material trajectory planning module, calculates the cross-sectional area change amount of the airflow channels between the layers of the diversion grid based on the density data, and synchronously adjusts the inclination angles and spacings of each layer of the diversion grid to make the airflow velocity gradient consistent with the spatial distribution of the inertial sedimentation paths of the dust particles.

3. The material guiding device according to claim 1, characterized in that: The dynamic feedback adjustment module includes a convolutional neural network sub-module that receives the real-time stress spectrum data collected by the high-frequency piezoelectric pressure sensor array, performs time-frequency domain feature decomposition on the spectrum data to generate multi-dimensional feature vectors, and outputs a classification result including the vibration energy level and the fault type to the genetic algorithm optimization module by comparing with the abnormal vibration waveform templates in the preset vibration mode database.

4. The material guiding device according to claim 3, characterized in that: The genetic algorithm optimization module assigns weight coefficients of impact acceleration, energy consumption efficiency, and dust concentration in the multi-objective fitness function according to the vibration energy level and the fault type classification result, uses the elitist retention strategy to screen out a PID control parameter set that meets the preset dynamic balance conditions, and injects the parameter set into the rotational speed adjustment algorithm of the velocity coordination control module to generate a closed-loop feedback control instruction.

5. The material guiding device according to claim 1, characterized in that: The material trajectory planning module includes a dynamic damping compensation unit that receives the damper stiffness correction coefficient output by the genetic algorithm optimization module, calculates the adjustment amount of the curvature radius at the end of the guiding surface in real time based on the coefficient, and dynamically corrects the geometric shape of the surface through a mechanical actuator, so that the difference between the end velocity of the material slip and the target belt line velocity calculated by the velocity coordination control module converges to a preset allowable range.

6. The material guiding device according to claim 5, characterized in that: The dynamic damping compensation unit receives the material slip velocity deviation value calculated by the velocity coordination control module through an industrial bus. When the deviation value exceeds a preset threshold, it sends a parameter reconfiguration request signal to the genetic algorithm optimization module, triggering the generation process of the global optimization parameter set based on the current operating state.

7. The feeding device according to claim 1, characterized in that: The adaptive crossover probability generation unit of the genetic algorithm optimization module receives the belt load volatility data collected by the dynamic feedback adjustment module and the real-time dust removal efficiency parameters of the passive dust suppression module, and dynamically calculates the weight distribution of the crossover operator according to the correlation analysis results of the volatility and efficiency parameters, generating a genetic evolution strategy that accelerates the convergence of the global optimal solution.

8. The material guiding device according to claim 7, characterized in that: The optimization parameter set output by the adaptive crossover probability generation unit corrects the adjustment amount of the guiding surface inclination angle through the material trajectory planning module, and simultaneously sends a guiding grid interlayer damping coefficient compensation instruction to the passive dust suppression module, enabling the inclination angle adjustment of the surface and the damping coefficient correction to synergistically execute the impact energy consumption suppression and dust removal efficiency improvement based on the same optimization goal.

9. The material guiding device according to claim 1, wherein: The lidar scanning path planning sub-module of the velocity coordination control module receives the predicted landing point distribution data of the slip trajectory model output by the material trajectory planning module, and dynamically calculates the coverage range of the lidar scanning area and the adjustment amount of the sampling frequency according to the spatial density change of the distribution data, generating a scanning control instruction that matches the real-time material flow state.

10. A material guiding control system is applied to the material guiding device described in any one of claims 1 to 9, and is characterized in that, Including: The material trajectory planning module, the passive dust suppression module, the velocity coordination control module, the dynamic feedback adjustment module, and the genetic algorithm optimization module are connected through an industrial bus to form a closed-loop control network based on real-time data exchange. Among them, the dynamic feedback adjustment module synchronously uploads the vibration spectrum data collected by the high-frequency piezoelectric pressure sensor array and the rotational speed deviation data calculated by the velocity coordination control module to the genetic algorithm optimization module; The genetic algorithm optimization module fuses the vibration spectrum data, the rotational speed deviation data, and the real-time dust removal efficiency parameters to generate a global optimization parameter set including the dynamic inclination angle adjustment amount of the guiding surface, the damper stiffness correction coefficient, and the guiding grid interlayer damping compensation value, and distributes the parameter set to the material trajectory planning module, the passive dust suppression module, and the velocity coordination control module through an industrial bus; The passive dust suppression module adjusts the guiding grid structure parameters according to the received guiding grid interlayer damping compensation value. The material trajectory planning module updates the slip trajectory model based on the dynamic inclination angle adjustment amount of the guiding surface and generates a new damping control instruction. The velocity coordination control module calibrates the rotational speed adjustment algorithm using the damper stiffness correction coefficient, realizing the multi-objective collaborative optimization of the material slip trajectory tracking accuracy, dust suppression efficiency, and equipment vibration energy.

Images