Roasting, activating and magnetic separation combined iron removal system

Abstract

The invention relates to the technical field of magnetic separation, in particular to a roasting activation magnetic separation combined iron removal system which comprises a pretreatment module, a magnetic characteristic dynamic detection module, a magnetic field parameter self-adaptive matching module, a dynamic magnetic compensation module, a separation path optimization module and a cross-module cooperative control module. The sorting process is rehearsed through a virtual twinborn model, parameters are triggered to be reset synchronously, and particle agglomeration interference is restrained in combination with dynamic partition of a sorting cavity and a flexible electrode array. The modules achieve magnetic field dynamic adaptation, magnetic attenuation compensation and sorting path optimization through closed-loop cooperation, the problem that a static magnetic field is mismatched with non-uniform magnetic particles is effectively solved, and the medium magnetic particle capture efficiency, the iron element recovery rate and the concentrate grade are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of magnetic separation, in particular to a roasting-activated magnetic separation combined iron removal system. BACKGROUND

[0002] Roasting-activated magnetic separation combined iron removal is a composite separation process based on the regulation of mineral physical and chemical properties. The core principle is to change the crystal structure or surface magnetism of target minerals through high-temperature roasting, so that the magnetic difference between the target minerals and impurity minerals (such as iron-containing minerals) is significantly increased. Then, the roasting products are subjected to magnetic separation using a gradient magnetic field to achieve efficient separation and enrichment of iron elements. In the roasting stage, weakly magnetic minerals (such as hematite) can be converted into strongly magnetic magnetite or metallic iron phase under the action of reducing atmosphere or specific additives, while gangue minerals remain weakly magnetic or non-magnetic due to the absence of phase change. In the magnetic separation stage, the iron-containing minerals with enhanced magnetism are captured in the concentrate area under the action of the magnetic field, while the non-magnetic components enter the tailings. Finally, through the synergistic effect of the two-stage process, the recovery rate and grade of iron elements are improved. This technology is suitable for the extraction of iron components from low-grade iron ore, metallurgical slag or secondary resources, and its efficiency depends on the optimized matching of roasting temperature, atmosphere control and magnetic field strength.

[0003] In the roasting-activated magnetic separation combined iron removal system, the capture efficiency of the gradient magnetic field on the target magnetic particles is limited due to the non-uniformity of the magnetic distribution of the roasting products and the weakly magnetic oxide layer remaining on the particle surface: the magnetite or metallic iron phase generated after roasting has a significant gradient difference in magnetic susceptibility due to local oxidation or incomplete phase change. The magnetic field distribution of traditional magnetic separation equipment cannot dynamically adapt to such non-uniform magnetic particle groups, causing part of the moderately magnetic particles to enter the non-target product channel during the separation process due to the deviation of the magnetic force and fluid drag force ratio from the critical threshold, resulting in a simultaneous decrease in iron recovery rate and concentrate grade. This problem is caused by the mismatch between the static design of the magnetic field parameters in the magnetic separation link and the dynamic magnetic characteristics of the roasting products. SUMMARY

[0004] To address the deficiencies in the prior art, the present application provides a roasting-activated magnetic separation combined iron removal system. The present application solves the problem that the static magnetic field distribution of existing magnetic separation equipment cannot adapt to the non-uniform magnetic characteristics of roasting products, causing moderately magnetic particles to enter the non-target product channel due to the imbalance between magnetic force and drag force, thereby reducing the iron recovery rate and concentrate grade.

[0005] To solve the above technical problems, the specific technical solutions of the present application are as follows:

[0006] The roasting-activated magnetic separation combined iron removal system provided by the present application comprises: a pretreatment module, a magnetic characteristic dynamic detection module, a magnetic field parameter self-adaptive matching module, a dynamic magnetic compensation module, a separation path optimization module and a cross-module collaborative control module.

[0007] The pre-treatment module is used for cooling and crushing the roasted mixed material and separating the residues to generate a non-uniform magnetic material flow and transmit the non-uniform magnetic material flow to the magnetic characteristic dynamic detection module;

[0008] The magnetic characteristic dynamic detection module includes a magnetic susceptibility sensor array and an X-ray diffraction probe, and is used for collecting the magnetic susceptibility distribution and the mineral phase composition data of the material in real time, generating a magnetic susceptibility distribution thermodynamic map and a mineral phase classification matrix after filtering processing, and pushing the thermodynamic map and the classification matrix to the magnetic field parameter adaptive matching module and the separation path optimization module;

[0009] The magnetic field parameter adaptive matching module constructs a magnetic field gradient requirement model of a separation cavity based on the magnetic susceptibility distribution thermodynamic map, drives a gradient-adjustable permanent-magnet-electromagnetic composite magnetic field generator to switch a magnetic field mode, and generates a low-gradient magnetic field in a strong magnetic zone and a high-gradient magnetic field in a medium-low magnetic zone corresponding to the magnetic susceptibility distribution;

[0010] The dynamic magnetic compensation module receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, predicts the magnetic decay amount by tracking the particle residence time, adjusts the current intensity of the electromagnetic compensation coil to inject a reverse magnetic field, and sends a deflector angle compensation instruction to the separation path optimization module;

[0011] The separation path optimization module integrates the magnetic susceptibility distribution thermodynamic map, the magnetic field distribution data and the particle motion image, solves the regulation and control parameters of the deflector angle and the flushing water flow rate through a trajectory state observer and a multi-objective optimization algorithm, and sends the regulation and control parameters to the deflector execution mechanism of the separation cavity;

[0012] The cross-module collaborative control module receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, the regulation and control parameters of the separation path optimization module and the compensation instruction of the dynamic magnetic compensation module, pre-rehearses the separation process through a virtual twin model, and triggers the synchronous reset of the magnetic field parameters and the separation path when the deviation between the separation index of the virtual twin model and the actual separation index exceeds a set threshold.

[0013] Further, the roasting activation magnetic separation combined iron removal system provided by the application includes:

[0014] The crushing and screening unit is used for receiving the roasted mixed material and crushing the roasted mixed material to a target particle size range through a high-pressure roller crusher;

[0015] The airflow separation unit receives the material output by the crushing and screening unit, removes the unreacted additive residues through a cyclone separator, and generates a non-uniform magnetic material flow;

[0016] The non-uniform magnetic material flow is transmitted to the magnetic characteristic dynamic detection module through a belt conveyor.

[0017] Further, the roasting and activation magnetic separation combined iron removal system of the present application, the magnetic characteristic dynamic detection module is configured to:

[0018] The magnetic susceptibility sensor array is distributed along the feed inlet of the magnetic separation module in a ring shape, and the magnetic susceptibility distribution of the material is scanned in real time and a magnetic susceptibility distribution heat map is generated;

[0019] The X-ray diffraction probe analyzes the mineral phase composition online, and generates a mineral phase classification matrix containing the proportion of magnetite and hematite;

[0020] The magnetic susceptibility distribution heat map and the mineral phase classification matrix are transmitted to the magnetic field gradient demand model construction node of the magnetic field parameter adaptive matching module through the industrial Ethernet.

[0021] Further, the roasting and activation magnetic separation combined iron removal system of the present application, the magnetic field parameter adaptive matching module in:

[0022] The gradient-adjustable permanent magnet-electromagnetic composite magnetic field generator includes a permanent magnet array and an electromagnetic compensation coil;

[0023] The permanent magnet array is activated according to the strong magnetic area parameters output by the magnetic field gradient demand model, generating a low-gradient strong magnetic field;

[0024] The electromagnetic compensation coil is enabled according to the medium-low magnetic area parameters output by the magnetic field gradient demand model, generating a high-gradient medium-strong magnetic field;

[0025] The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the magnetic decay amount predicted by the dynamic magnetic compensation module;

[0026] The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the magnetic susceptibility decay prediction value.

[0027] Further, the roasting and activation magnetic separation combined iron removal system of the present application, the sorting path optimization module is configured to:

[0028] The trajectory state observer calculates the particle group motion trajectory through the magnetic force-gravity-drag force balance equation based on the magnetic susceptibility distribution heat map and the magnetic field distribution data;

[0029] The multi-objective optimization algorithm takes the concentrate grade and tailings loss rate as constraint conditions, and iteratively optimizes the regulation parameters of the deflector angle and the flushing water flow rate;

[0030] The regulation parameters are converted into step pulse signals of the deflector actuator by a double-encoder servo motor, and are synchronously sent to the cross-module collaborative control module.

[0031] Further, the roasting and activation magnetic separation combined iron removal system of the present application, the cross-module collaborative control module is configured to:

[0032] The virtual twin model receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, the regulation parameters of the sorting path optimization module and the compensation instructions of the dynamic magnetic compensation module, and performs a discrete element simulation to preview the sorting process;

[0033] When the deviation between the sorting index output by the virtual twin model and the actual sorting index exceeds a set threshold, the magnetic field parameters and the sorting path are synchronously reset;

[0034] The timestamp alignment module corrects the timing error of the regulation instructions of the magnetic field parameter adaptive matching module and the guide plate action instructions of the sorting path optimization module.

[0035] Further, the roasting and activation magnetic separation combined iron removal system also comprises:

[0036] The sorting cavity dynamic partition unit divides the sorting cavity into three sections, i.e., front, middle and rear, and an independent photoelectric sensor is arranged in each section to monitor the material flow rate;

[0037] A flexible electrode array is embedded in the inner wall of the sorting cavity for applying a low-frequency alternating electric field to destroy the magnetic particle agglomeration;

[0038] The electric field strength data of the flexible electrode array is linked with the trajectory state observer of the sorting path optimization module to correct the particle agglomeration interference term in the magnetic-gravity-drag force balance equation.

[0039] The present application has the following beneficial effects:

[0040] The present application can realize the real-time acquisition of the magnetic susceptibility distribution and the mineral phase composition data of the material through the magnetic characteristic dynamic detection module, generate a magnetic susceptibility distribution thermodynamic diagram and a mineral phase classification matrix, provide accurate input for the magnetic field parameter adaptive matching module, drive the gradient-adjustable permanent magnet-electromagnetic composite magnetic field generator to dynamically generate a partition gradient magnetic field adapted to the magnetic characteristics of the material, effectively solve the mismatch problem between the static magnetic field and the non-uniform magnetic particles, predict the magnetic decay amount based on the particle residence time through the dynamic magnetic compensation module, adjust the current injection of the electromagnetic compensation coil to generate a reverse magnetic field to offset the dynamic decay effect of the particle magnetism and maintain the stability of the magnetic force during the sorting process, iteratively solve the optimal parameters of the guide plate angle and the flushing water flow rate through the trajectory state observer and the multi-objective optimization algorithm of the sorting path optimization module, combine the virtual twin model preview and real-time calibration of the cross-module collaborative control module, realize the dynamic balance control of the magnetic force and the fluid drag force, prevent the medium magnetic particles from deviating from the target channel, and further improve the sorting accuracy through the cooperation of the sorting cavity dynamic partition unit and the flexible electrode array to suppress the particle agglomeration and correct the trajectory model interference term. The closed-loop cooperation of the above modules significantly improves the capture efficiency of the medium magnetic particles, and finally realizes the synchronous improvement of the iron element recovery rate and the concentrate grade. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, for those skilled in the art, other drawings can also be obtained from the drawings without any creative labor.

[0042] Figure 1 The system architecture diagram of the roasting activation and magnetic separation combined iron removal system provided for the embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present application more clear, the following will combine the specific embodiments of the present application and the corresponding drawings to clearly and completely describe the technical solutions of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative labor are within the protection scope of the present application. The following will combine the drawings to specifically describe the technical solutions provided by the embodiments of the present application. In order to better understand the purpose of the present application, the following will further describe the present application in detail.

[0044] Please refer to Figure 1 The present application provides a roasting activation and magnetic separation combined iron removal system, which comprises:

[0045] The roasting activation and magnetic separation combined iron removal system provided by the present application comprises a pretreatment module, a magnetic characteristic dynamic detection module, a magnetic field parameter adaptive matching module, a dynamic magnetic compensation module, a separation path optimization module and a cross-module cooperative control module.

[0046] The pretreatment module is used for cooling and crushing the mixture after roasting and separating the residues to generate a non-uniform magnetic material stream and transmit it to the magnetic characteristic dynamic detection module.

[0047] The magnetic characteristic dynamic detection module comprises a magnetic susceptibility sensor array and an X-ray diffraction probe, which real-time collects the magnetic susceptibility distribution and mineral phase composition data of the material, generates a magnetic susceptibility distribution thermodynamic map and a mineral phase classification matrix after filtering processing, and pushes the thermodynamic map and the classification matrix to the magnetic field parameter adaptive matching module and the separation path optimization module.

[0048] The magnetic field parameter adaptive matching module constructs a magnetic field gradient demand model of the separation cavity based on the magnetic susceptibility distribution thermodynamic map, drives the gradient-adjustable permanent magnet-electromagnetic composite magnetic field generator to switch the magnetic field mode, and generates a low-gradient magnetic field in the strong magnetic zone and a high-gradient magnetic field in the medium-low magnetic zone corresponding to the magnetic susceptibility distribution.

[0049] The dynamic magnetic compensation module receives the magnetic field distribution data from the magnetic field parameter adaptive matching module, predicts the magnetic attenuation amount by tracking the particle residence time, adjusts the current intensity of the electromagnetic compensation coil to inject a reverse magnetic field, and sends a deflector angle compensation instruction to the sorting path optimization module.

[0050] The sorting path optimization module integrates the magnetic susceptibility distribution thermodynamic map, the magnetic field distribution data, and the particle motion image, solves the regulation and control parameters of the deflector angle and the flushing water flow rate through the trajectory state observer and the multi-objective optimization algorithm, and sends the regulation and control parameters to the deflector execution mechanism of the sorting cavity.

[0051] The cross-module collaborative control module receives the magnetic field distribution data from the magnetic field parameter adaptive matching module, the regulation and control parameters from the sorting path optimization module, and the compensation instruction from the dynamic magnetic compensation module, and preforms the sorting process through the virtual twin model. When the deviation between the sorting index of the virtual twin model and the actual sorting index exceeds the set threshold, the magnetic field parameters and the sorting path are triggered to reset synchronously.

[0052] After the pre-treatment module receives the roasted mixture, the material temperature is reduced to room temperature by the inert gas cooling system to avoid interference with the subsequent magnetic separation detection. The cooled material enters the high-pressure roller crusher for mechanical crushing, with a crushing pressure controlled within 50-100 MPa, so that the particle size distribution of the material is within the target range of 0.1-1 mm. The crushed product is screened by a vibrating screen to remove particles with a particle size larger than 1 mm, and then conveyed to the air flow separation unit by a screw conveyor. In the air flow separation unit, the cyclone separator separates the unreacted additive residues at a tangential wind speed of 10-15 m / s, and the clean non-uniform magnetic material stream is transmitted to the magnetic characteristic dynamic detection module by a belt conveyor.

[0053] The magnetic susceptibility sensor array of the magnetic characteristic dynamic detection module is arranged equidistantly around the feed inlet of the sorting cavity. Each sensor node scans the magnetic susceptibility distribution of the material at a sampling frequency of 100 Hz, generating a two-dimensional magnetic susceptibility distribution thermodynamic map. The X-ray diffraction probe simultaneously analyzes the mineral phase of the material, calculates the mass proportion of magnetite and hematite through the characteristic diffraction peak intensity, and forms a mineral phase classification matrix. After the magnetic susceptibility distribution thermodynamic map and the mineral phase classification matrix are filtered by the Kalman filtering algorithm to eliminate noise interference, they are transmitted to the magnetic field gradient demand model construction node of the magnetic field parameter adaptive matching module through the industrial Ethernet, and are also pushed to the trajectory prediction unit of the sorting path optimization module.

[0054] The magnetic field parameter adaptive matching module calculates the magnetic field gradient requirement of each region in the sorting cavity based on the received magnetic susceptibility distribution heat map. In the gradient-adjustable permanent-magnetic-electromagnetic composite magnetic field generator, the permanent magnet array is activated according to the strong magnetic region parameters to generate a low-gradient strong magnetic field with an axial gradient of ≤5T / m and a tangential strength of ≥1.2T; the electromagnetic compensation coil is enabled according to the medium-low magnetic region parameters to generate a high-gradient medium-strong magnetic field with an axial gradient of ≥8T / m and a tangential strength of 0.8-1.0T. The magnetic field distribution data is fed back to the decay prediction unit of the dynamic magnetic compensation module in real time.

[0055] The dynamic magnetic compensation module records the residence time of the material in the sorting cavity through the RFID marker particle tracking system, and predicts the real-time decay amount of the particle magnetism in combination with the magnetic susceptibility decay curve in the historical experiment database. The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the predicted value to inject a reverse magnetic field to offset the magnetic decay effect. The deflector angle compensation instruction is generated based on the decay amount calculation result and is sent to the regulation and execution unit of the sorting path optimization module through the CAN bus protocol.

[0056] The sorting path optimization module integrates the magnetic susceptibility distribution heat map, the magnetic field distribution data, and the particle motion image captured by the high-speed CCD camera to construct a multi-scale sorting state atlas. The trajectory state observer calculates the theoretical motion trajectory of the particle group based on the magnetic force-gravity-drag force balance equation, and iteratively solves the optimal regulation and control parameters of the deflector angle and the flushing water flow rate in combination with the adaptive weight particle swarm optimization algorithm, with the concentrate grade ≥85% and the tailings loss rate ≤5% as the constraint conditions. The regulation and control parameters are converted into step pulse signals by the double encoder servo motor to drive the deflector actuator to adjust to the target position, and the parameters are simultaneously sent to the data verification unit of the cross-module collaborative control module.

[0057] The cross-module collaborative control module receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, the regulation and control parameters of the sorting path optimization module, and the compensation instruction of the dynamic magnetic compensation module, and builds a virtual twin model through discrete element simulation to preview the sorting process. When the concentrate grade or the tailings loss rate output by the virtual twin model deviates from the actual sorting index by more than 5%, a synchronous reset instruction of the magnetic field strength and the deflector angle is triggered. The time stamp alignment module corrects the timing error of the magnetic field regulation and control instruction and the deflector action instruction, with a delay of 0.5 seconds for the front regulation and control instruction, 1 second for the middle, and 2 seconds for the rear, to avoid sorting disorder caused by instruction conflict.

[0058] The dynamic partitioning unit of the sorting chamber divides the chamber into three sections: front, middle, and back. Photoelectric sensors are deployed in each section to monitor material flow rate. A flexible electrode array embedded in the chamber wall applies a 1-5 Hz low-frequency alternating electric field to disrupt magnetic particle agglomeration. This electric field intensity data is transmitted in real time to the trajectory state observer of the sorting path optimization module, which corrects the particle agglomeration interference term in the magnetic-gravitational drag force balance equation, thereby improving trajectory prediction accuracy.

[0059] The non-uniform magnetic material flow output by the pre-processing module provides a physical carrier for the dynamic detection of magnetic characteristics. Its particle size and cleanliness directly affect the accuracy of magnetic susceptibility detection. The thermal map and classification matrix generated by the dynamic detection module of magnetic characteristics provide a data basis for the adaptive matching of magnetic field parameters. The construction of the magnetic field gradient demand model directly determines the distribution pattern of the permanent magnet and the electromagnetic field. The dynamic magnetic compensation module predicts the attenuation based on the magnetic field distribution data. Its compensation instructions work together with the control parameters of the sorting path optimization module to ensure the dynamic balance between the magnetic force and the fluid drag force during the sorting process. The cross-module collaborative control module integrates the data of the entire system through the virtual twin model to achieve closed-loop optimization of the sorting process. The dynamic partitioning of the sorting cavity and the flexible electrode array further enhance the particle dispersion and trajectory control, ultimately achieving efficient recovery of iron elements and improvement of the concentrate grade.

[0060] Specifically, the roasting activation magnetic separation combined iron removal system of the present invention, the pretreatment module includes:

[0061] A crushing and screening unit, which is used to receive the roasted mixed material and crush it into a target particle size range through a high-pressure roller crusher;

[0062] an air flow separation unit, which receives the material output from the crushing and screening unit, removes unreacted additive residues through a cyclone separator, and generates a non-uniform magnetic material flow;

[0063] The non-uniform magnetic material flows through a belt conveyor and is transported to a magnetic characteristic dynamic detection module.

[0064] In the pretreatment module, the crushing and screening unit receives the roasted mixed material, which first enters the feed port of a high-pressure roller crusher. The high-pressure roller crusher applies vertical pressure through a hydraulic system, squeezing and crushing the material between the rollers. The crushed particles are then screened by a vibrating screener with the screen aperture set to the upper limit of the target particle size range. Oversized particles on the screen are returned to the crushing chamber for secondary crushing, and qualified particles below the screen are discharged via a screw conveyor. The material output from the crushing and screening unit enters the airflow sorting unit. The cyclone separator in the airflow sorting unit generates a tangential airflow through a centrifugal fan. Unreacted additive residues are thrown toward the separator wall due to density differences under the action of centrifugal force and discharged from the system through the slag discharge port. The purified non-uniform magnetic material flow is driven by the airflow into the aggregate bin.

[0065] The non-uniform magnetic material stream falls from the bottom outlet of the collecting bin to the belt conveyor, and the transmission speed of the belt conveyor is dynamically adjusted according to the real-time processing capacity of the magnetic characteristic dynamic detection module. The surface of the belt is made of anti-static material to suppress particle adsorption, and baffles are arranged on both sides to prevent material scattering. After the material flow spreads evenly through the belt conveyor, it is continuously conveyed to the feed inlet of the magnetic characteristic dynamic detection module. The photoelectric sensor at the feed inlet detects the material flow and feeds back to the pressure regulation system of the crushing and screening unit, forming a closed-loop control of the crushing particle size.

[0066] The crushing and screening unit controls the particle size of the roasting product within the range required by the magnetic separation process through the synergistic effect of mechanical crushing and screening, providing a size-homogeneous material basis for subsequent magnetic detection. The gas-solid two-phase flow density difference is used in the pneumatic separation unit to separate the residues, avoiding the interference of unreacted additives on the detection accuracy of the magnetic characteristics. The belt conveyor maintains the spatial distribution stability of the material flow during transmission, and the anti-static design and baffle structure reduce particle agglomeration and dispersion, ensuring that the magnetic susceptibility data obtained by the detection module reflect the true material properties. The pressure regulation system of the crushing and screening unit responds to the flow feedback of the detection module, dynamically optimizes the crushing intensity, and forms a pre-processing closed loop of "particle size control-residue separation-stable transmission", providing material input conditions that meet the requirements of the magnetic separation process for the downstream modules.

[0067] Specifically, the roasting activation and magnetic separation combined iron removal system described in the present application is configured as:

[0068] The magnetic susceptibility sensor array is distributed in a ring shape along the feed inlet of the magnetic separation module, and scans the magnetic susceptibility distribution of the material in real time and generates a magnetic susceptibility distribution thermogram;

[0069] The X-ray diffraction probe analyzes the mineral phase composition online and generates a mineral phase classification matrix containing the proportion of magnetite and hematite;

[0070] The magnetic susceptibility distribution thermogram and the mineral phase classification matrix are transmitted to the magnetic field gradient demand model construction node of the magnetic field parameter self-adaptive matching module through the industrial Ethernet.

[0071] The magnetic susceptibility sensor array of the magnetic characteristic dynamic detection module is arranged in a ring shape at equal intervals along the feed inlet of the magnetic separation module, each sensor node scans the magnetic susceptibility of the material surface at a preset frequency, and generates a two-dimensional magnetic susceptibility distribution thermal map. The sensor array uses a non-contact electromagnetic induction principle, detects the magnetic field disturbance signal when the material passes through, and maps the spatial distribution characteristics of the magnetic susceptibility in real time. The high magnetic susceptibility area in the thermal map corresponds to the magnetite enriched particle group, and the low magnetic susceptibility area represents the distribution of hematite or impurity minerals. The X-ray diffraction probe is located downstream of the feed inlet of the magnetic separation module, an X-ray beam emitted by the X-ray diffraction probe penetrates the material flow, and the mineral phase composition is analyzed by receiving the diffraction spectrum, the mass proportion of magnetite and hematite is calculated, and a mineral phase classification matrix is formed. The row and column dimensions of the classification matrix correspond to different spatial partitions of the material flow and the mineral phase type proportion data, respectively.

[0072] The magnetic susceptibility distribution thermal map and the mineral phase classification matrix are spatio-temporally aligned by a data fusion middleware, after eliminating the time sequence difference of sensor collection, and are transmitted to the magnetic field parameter adaptive matching module through an industrial Ethernet. After receiving the above data, the magnetic field gradient demand model construction node maps the magnetic susceptibility distribution of the thermal map to the spatial coordinate system of the separation cavity, combines the mineral phase proportion information of the classification matrix, and generates the magnetic field gradient demand parameters of each region of the separation cavity. The high magnetic susceptibility area of the thermal map drives the model to preferentially allocate strong magnetic field resources, and the hematite proportion data of the mineral phase classification matrix triggers the medium and low magnetic field gradient regulation strategy, realizing the dynamic adaptation of the magnetic field parameters and the material characteristics.

[0073] The ring-shaped layout of the magnetic susceptibility sensor array covers the full cross section of the material flow, and its spatial resolution capability provides basic data for the generation of the thermal map, directly reflecting the non-uniform characteristics of the magnetic distribution of the material. The mineral phase analysis data of the X-ray diffraction probe and the spatial information of the thermal map are complementary, and the classification matrix quantitatively describes the proportion relationship of different mineral phases, providing the basis for the mineral composition of the magnetic field gradient demand model. The data fusion middleware solves the spatio-temporal alignment problem of multi-source heterogeneous data, ensuring that the data received by the magnetic field parameter module is consistent and timely. The high reliability transmission of the industrial Ethernet avoids data packet loss or delay, ensuring that the model construction node can obtain the latest detection results in real time. The magnetic field gradient demand model converts the magnetic characteristic data into the magnetic field regulation instructions of the separation cavity, forming a closed loop link of "detection-analysis-modeling", and providing accurate input for subsequent magnetic field dynamic matching.

[0074] Specifically, in the magnetic field parameter adaptive matching module of the roasting activation magnetic separation combined iron removal system,

[0075] The gradient-adjustable permanent magnet-electromagnetic composite magnetic field generator includes a permanent magnet array and an electromagnetic compensation coil;

[0076] The permanent magnet array is activated according to the strong magnetic area parameters output by the magnetic field gradient demand model, and generates a low-gradient strong magnetic field;

[0077] The electromagnetic compensation coil is enabled according to the medium-low magnetic region parameter output by the magnetic field gradient demand model, and a high-gradient medium-strong magnetic field is generated;

[0078] The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the predicted magnetic attenuation amount of the dynamic magnetic compensation module;

[0079] The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the predicted magnetic attenuation amount of the dynamic magnetic compensation module;

[0080] In the magnetic field parameter adaptive matching module, the permanent magnet array of the gradient-adjustable permanent-magnetic-electromagnetic composite magnetic field generator is arranged axially along the sorting cavity, and each permanent magnet unit is independently adjusted in the magnetic pole direction through a magnetic circuit controller. When the magnetic field gradient demand model outputs a strong magnetic region parameter, the magnetic circuit controller activates the permanent magnet unit in the corresponding region to form a low-gradient strong magnetic field with an axial gradient of ≤5T / m and a tangential intensity of ≥1.2T, which is used to capture magnetite particles with high magnetic susceptibility. The electromagnetic compensation coil is nested in the gap between the permanent magnet array, and after receiving the medium-low magnetic region parameter, the coil current is adjusted through a power amplifier to generate a high-gradient medium-strong magnetic field with an axial gradient of ≥8T / m and a tangential intensity of 0.8-1.0T, which is used for sorting hematite and oxidized layer particles.

[0081] The dynamic magnetic compensation module receives the particle residence time data in the sorting cavity in real time, combines the predicted magnetic susceptibility decay value, and generates a current correction instruction for the electromagnetic compensation coil. The current correction instruction is converted into a voltage signal by a PID controller to drive the power amplifier to dynamically adjust the coil current intensity, and inject a reverse magnetic field to offset the decay effect of the particle magnetism. The reverse magnetic field strength has a linear relationship with the predicted magnetic attenuation amount, so that the magnetic force acting on the particles during the sorting process remains stable. The corrected current parameter is fed back to the magnetic field gradient demand model synchronously to update the magnetic field distribution parameters in the next cycle.

[0082] The spatial layout of the permanent magnet array and the electromagnetic compensation coil realizes the partition magnetic field control of the sorting cavity. The strong magnetic region parameter drives the permanent magnet to generate a low-gradient strong magnetic field, which is suitable for the sorting requirements of particles with high magnetic susceptibility. The medium-low magnetic region parameter generates a high-gradient magnetic field through the electromagnetic compensation coil, which improves the capture efficiency of particles with weak magnetism. The prediction data of the dynamic magnetic compensation module and the current correction form a closed-loop regulation, which can offset the influence of particle magnetic decay on the sorting accuracy in real time. The magnetic field gradient demand model iteratively optimizes the parameters according to the compensation results, realizes the dynamic matching of the magnetic field distribution and the material characteristics, and finally improves the recovery rate of iron elements and the concentrate grade through the synergistic effect of permanent magnet and electromagnetic field.

[0083] Specifically, the roasting and activation magnetic separation combined iron removal system provided by the present application is characterized in that the sorting path optimization module is configured to:

[0084] The trajectory state observer calculates the particle group motion trajectory through the magnetic gravity drag force balance equation based on the magnetic susceptibility distribution thermodynamic map and magnetic field distribution data.

[0085] The multi-objective optimization algorithm takes the concentrate grade and tailings loss rate as constraint conditions, and iteratively optimizes the guide plate angle and flushing water flow rate control parameters.

[0086] The control parameters are converted into stepping pulse signals of the guide plate actuator by a double encoder servo motor and are synchronously sent to the cross-module collaborative control module.

[0087] In the sorting path optimization module, the trajectory state observer receives the magnetic susceptibility distribution thermodynamic map transmitted by the magnetic characteristic dynamic detection module and the magnetic field distribution data output by the magnetic field parameter adaptive matching module, and registers the spatial resolution of the thermodynamic map with the spatial coordinates of the magnetic field distribution. The magnetic gravity drag force balance equation calculates the motion trajectory of the particle group in the sorting cavity based on the registered data. The magnetic force term in the equation is determined by the local magnetization intensity of the magnetic susceptibility distribution thermodynamic map and the magnetic field gradient data, the gravity term is calculated according to the particle density and the cavity inclination angle parameters, and the drag force term is derived through the material flow rate and fluid viscosity parameters. The particle trajectory calculation result forms a three-dimensional motion trajectory map, which labels the theoretical motion path of particles with different magnetic intensities.

[0088] The multi-objective optimization algorithm takes the concentrate grade ≥ 85% and tailings loss rate ≤ 5% as constraint conditions, and constructs a joint optimization model of the guide plate angle and the flushing water flow rate. The algorithm uses the non-dominated sorting genetic algorithm (NSGA-II) to perform multi-generation iteration on the parameter combinations, and selects the control parameters with the highest comprehensive score in the Pareto optimal solution set. The optimization range of the guide plate angle is set to 15°-45°, and the optimization interval of the flushing water flow rate is 0.5-2.0 m / s. The input parameters of the trajectory state observer are updated after each iteration, forming a closed-loop control link of "trajectory prediction-parameter optimization-data feedback".

[0089] The control parameters are converted into stepping pulse signals of the guide plate actuator by a double encoder servo motor. The absolute position encoder calibrates the initial angle of the guide plate, and the incremental encoder drives the guide plate to rotate to the target angle according to the stepping pulse signal, with an angle positioning accuracy of within ±0.5°. The flushing water flow rate parameter is converted into a water pressure adjustment instruction by a proportional integral valve, and the valve opening degree and flow rate are in a linear relationship. The control parameters and execution state data are synchronously sent to the data verification unit of the cross-module collaborative control module through the CAN bus protocol for real-time simulation calibration of the virtual twin model.

[0090] The trajectory state observer provides a theoretical prediction basis for the particle motion trajectory by fusing the magnetic susceptibility and magnetic field data, and the output trajectory atlas directly determines the search space of the multi-objective optimization algorithm. The multi-objective optimization algorithm generates the optimal control parameters of the guide plate and flushing water under the constraint of trajectory prediction, and the parameter conversion link accurately maps the digital instructions to the actions of the physical execution mechanism. The high-precision control of the double-encoder servo system and the linear response characteristics of the proportional-integral valve ensure the strict consistency of the sorting action and the optimization result. The synchronous feedback mechanism of the control parameters enables the cross-module collaborative control module to real-time check the simulation accuracy of the virtual twin model, forming a full-process closed loop of "trajectory prediction-parameter execution-model calibration", and finally realizing the dynamic optimization of the sorting path and the improvement of the iron element recovery rate.

[0091] Specifically, the roasting activation magnetic separation combined iron removal system described in the present application, the cross-module collaborative control module is configured to:

[0092] The virtual twin model receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, the control parameters of the sorting path optimization module, and the compensation instructions of the dynamic magnetic compensation module, and preplays the sorting process through discrete element simulation;

[0093] When the deviation between the sorting indicators output by the virtual twin model and the actual sorting indicators exceeds the set threshold, the magnetic field parameters and the sorting path are reset synchronously;

[0094] The timestamp alignment module corrects the timing error of the control instructions of the magnetic field parameter adaptive matching module and the guide plate action instructions of the sorting path optimization module.

[0095] The virtual twin model of the cross-module collaborative control module integrates the magnetic field distribution data of the magnetic field parameter adaptive matching module, the guide plate control parameters of the sorting path optimization module, and the compensation instructions of the dynamic magnetic compensation module through the data bus, and constructs the digital twin of the sorting process. Discrete element simulation simulates the motion trajectory of the particle group under the action of dynamic magnetic field and fluid based on the physical properties (density, magnetic susceptibility) of the material particles and the geometric parameters of the sorting cavity, and generates virtual sorting indicators (concentrate grade, tailings loss rate). The simulation results of the virtual twin model are compared with the sensor data (such as concentrate grade detector, tailings flowmeter) of the actual sorting module in real time, and when the deviation exceeds the set threshold (for example, the deviation of concentrate grade is ±5%), the parameter reset instruction is triggered.

[0096] The parameter reset instruction is sent to the magnetic field parameter adaptive matching module and the sorting path optimization module through the control bus. The magnetic field parameter module recalculates the magnetic field gradient requirement model according to the reset instruction, adjusts the activation strategy of the permanent magnet array and the electromagnetic compensation coil; the sorting path module updates the optimization solution set of the deflector angle and the flushing water flow rate based on the reset magnetic field parameters, generates new control parameters and issues them for execution. The timestamp alignment module detects the time sequence delay of the magnetic field control instruction and the deflector action instruction by analyzing the time label of each module instruction. The instruction queue is aligned by using the clock synchronization protocol, the front magnetic field control instruction is delayed for 0.5 seconds, the middle instruction is delayed for 1 second, and the rear instruction is delayed for 2 seconds, so that the problem of asynchronous sorting action caused by communication delay or calculation time is eliminated.

[0097] The discrete element simulation of the virtual twin model provides a rehearsal verification for the sorting process, and the deviation of the virtual sorting index output from the actual detection data reflects the deficiency of the system dynamic adaptation capability, triggering parameter reset to iteratively optimize the sorting strategy. The synchronous execution of the parameter reset instruction ensures the coordinated update of the magnetic field distribution and the deflector control, avoiding the system imbalance caused by single module adjustment. The timestamp alignment module solves the time sequence conflict problem of multiple module instructions through a hierarchical delay strategy, ensuring the precise matching of magnetic field strength change and deflector angle adjustment in different sections of the sorting cavity. The cross-module collaborative control module realizes the adaptive optimization of the sorting system under complex working conditions through the closed-loop link of "simulation rehearsal-deviation detection-parameter reset-time sequence alignment", ultimately improving the stability of iron element recovery rate and concentrate grade.

[0098] Specifically, the roasting and activation magnetic separation combined iron removal system described in the present application further comprises:

[0099] The sorting cavity dynamic partition unit divides the sorting cavity into three sections: front, middle and rear, and an independent photoelectric sensor is arranged in each section to monitor the material flow rate;

[0100] A flexible electrode array is embedded in the inner wall of the sorting cavity for applying a low-frequency alternating electric field to break up magnetic particle agglomeration;

[0101] The electric field strength data of the flexible electrode array is linked to the trajectory state observer of the sorting path optimization module to correct the particle agglomeration interference term in the magnetic-gravitational-drag force balance equation.

[0102] The sorting cavity dynamic partition unit divides the sorting cavity into three sections, front, middle and back, along the material flow direction. The inner wall of each section is symmetrically installed with an array of photoelectric sensors. The photoelectric sensors emit infrared beams and receive reflected signals to monitor the flow rate and distribution uniformity of the materials in each section in real time. The flow rate data is transmitted to the central processor of the cross-module collaborative control module through the RS-485 bus. The front section sensors focus on detecting the dispersion state of the materials when they first enter the cavity. The middle section sensors track the movement trend of the particle groups under the action of the magnetic field. The rear section sensors monitor the distribution of the materials before sorting is completed. The three sections of data are combined to generate a flow rate distribution map and pushed to the sorting path optimization module.

[0103] The flexible electrode array is composed of multiple conductive sheets carried by insulating substrates and embedded along the circumference of the inner wall of the sorting cavity with a spacing of 10-20 mm between adjacent electrode sheets. The electrode array is connected to a low-frequency signal generator, which outputs a 1-5 Hz sinusoidal alternating electric field. The electric field strength is dynamically adjusted according to the real-time detection of the degree of particle agglomeration. The degree of agglomeration is obtained through image analysis of the particles captured by a high-speed camera system. When the proportion of agglomerated particles in the image exceeds a threshold value, the signal generator increases the electric field frequency or voltage amplitude to enhance the dispersion effect of the electric field force between the electrodes. The electric field strength data of the electrode array is converted to digital after analog-to-digital conversion and transmitted to the trajectory state observer of the sorting path optimization module through industrial Ethernet.

[0104] After receiving the electric field strength data, the trajectory state observer introduces a particle agglomeration disturbance term into the magnetic-gravitational-drag force balance equation. The disturbance term is quantified through an empirical relationship model between the electric field strength and the degree of particle agglomeration. The modified equation recalculates the theoretical motion trajectory of the particle groups. The modified trajectory data is fed back to the multi-objective optimization algorithm to iteratively update the control parameters of the deflector angle and the flushing water flow rate. The electric field control instructions of the flexible electrode and the sorting path optimization parameters are synchronized through a timestamp alignment module to ensure the timing consistency of the electric field application and the deflector action.

[0105] The three-section monitoring of the dynamic partition of the sorting cavity provides spatial positioning basis for the electric field regulation of the flexible electrode array. The front section flow rate data triggers the initial electric field strength setting of the electrode array, and the middle and rear section data dynamically adjust the electric field parameters to adapt to the particle motion state. The electric field strength data of the flexible electrode array directly modifies the calculation model of the trajectory state observer, eliminating the trajectory prediction deviation caused by particle agglomeration. The modified trajectory data drives the sorting path optimization module to generate more accurate control parameters, forming a collaborative control link of "partition monitoring-electric field regulation-trajectory correction-parameter optimization". The cross-module collaborative control module integrates the partition flow rate data and the electric field regulation instructions, eliminates the delay error of multi-module collaboration through timing alignment, and finally improves the control accuracy of particle motion and the recovery efficiency of iron elements in the sorting cavity.

[0106] For each model in the technical solution of the present application, in combination with the technical solution content and the technical characteristics of the field, the following is explained and described:

[0107] Magnetic field gradient requirement model: this model is the core calculation unit of the magnetic field parameter adaptive matching module, and its input is the magnetic susceptibility distribution heat map and the mineral phase classification matrix output by the magnetic characteristic dynamic detection module. The model uses finite element simulation technology to register the spatial resolution of the magnetic susceptibility distribution heat map with the geometric coordinate system of the separation cavity, and combines the mass proportion information of magnetite and hematite in the mineral phase classification matrix to quantify the demand for magnetic field gradient in different regions of the separation cavity: the high magnetic susceptibility region (magnetite enrichment area) corresponds to the demand for low gradient strong magnetic field (axial gradient ≤ 5T / m, tangential strength ≥ 1.2T), and the medium and low magnetic susceptibility region (hematite or oxidized layer particle region) corresponds to the demand for high gradient medium strength magnetic field (axial gradient ≥ 8T / m, tangential strength 0.8-1.0T). The output of the model is the control parameter of the gradient adjustable permanent magnet electromagnetic composite magnetic field generator, which drives the permanent magnet array and the electromagnetic compensation coil to generate adaptive partition magnetic field respectively, so as to realize the dynamic mapping of the magnetic field distribution and the non-uniform magnetic characteristics of the material.

[0108] Virtual twin model: this model is deployed in the cross-module collaborative control module, and uses discrete element simulation technology as the core, and the input includes the magnetic field distribution data of the magnetic field parameter adaptive matching module, the control parameters (guide plate angle, flushing water flow rate) of the separation path optimization module, and the compensation instructions (electromagnetic coil current correction value) of the dynamic magnetic compensation module. Based on the physical properties (density, magnetic susceptibility) of the material particles and the geometric parameters (cavity length, inclination angle) of the separation cavity, the model simulates the motion trajectory of the particle group under the action of dynamic magnetic field and fluid drag force, and outputs virtual separation indicators (such as concentrate grade, tailings loss rate). The model compares the virtual indicators with the actual indicators in real time through the sensor data (concentrate grade detector, tailings flowmeter) of the actual separation module, and when the deviation between the virtual indicators and the actual indicators exceeds 5%, the synchronous reset instructions of the magnetic field parameters and the separation path are triggered, realizing the rehearsal verification and dynamic calibration of the separation process.

[0109] Trajectory state observer: this model is the trajectory prediction unit of the separation path optimization module, and the input includes the magnetic susceptibility distribution heat map (reflecting the magnetic distribution of particles), the magnetic field distribution data (providing magnetic field gradient information), and the particle motion image captured by the high-speed CCD camera (recording the actual motion state). Based on the balance relationship of magnetic force, gravity and drag force, the model calculates the theoretical motion trajectory of the particle group through multi-source data fusion: the magnetic force term is determined by the local magnetization intensity of the particle and the magnetic field gradient, the gravity term is derived from the particle density and the cavity inclination angle, and the drag force term is calculated from the material flow rate and the fluid viscosity. The model output is a three-dimensional particle motion trajectory map, which labels the theoretical motion path of particles with different magnetic intensity, providing trajectory constraint conditions for subsequent multi-objective optimization algorithms.

[0110] Multi-objective optimization algorithm: This algorithm is the parameter solving unit of the separation path optimization module. It takes the concentrate grade (constraint ≥ 85%) and the tailings loss rate (constraint ≤ 5%) as double objectives, and the deflector angle (15°-45°) and the flushing water flow rate (0.5-2.0 m / s) as optimization variables. It uses the non-dominated sorting genetic algorithm (NSGA-II) for iterative optimization. The algorithm selects the highest comprehensive score of the control parameter combination from the Pareto optimal solution set. The input is the particle trajectory pattern output by the trajectory state observer, and the output is the control instruction of the deflector actuator and the flushing water valve (deflector angle positioning accuracy ± 0.5°, flushing water flow rate is linearly related to valve opening). The algorithm realizes the dynamic adjustment of the separation path through the "trajectory prediction-parameter optimization-data feedback" closed loop.

[0111] Particle agglomeration interference term correction model: This model is integrated into the trajectory state observer of the separation path optimization module. The input is the electric field intensity data of the flexible electrode array (1-5 Hz sinusoidal alternating electric field) and the particle agglomeration degree analyzed by the high-speed camera system (agglomerated particles account for a certain percentage). The model quantifies the interference of particle agglomeration on trajectory calculation through an empirical relationship model: when the agglomeration degree exceeds the threshold value, the electric field intensity is increased to enhance the dispersion effect, and the interference term value is increased accordingly. The drag term in the magnetic-gravitational-drag force balance equation and the particle motion resistance coefficient are corrected. The corrected trajectory data is fed back to the multi-objective optimization algorithm to update the control parameters of the deflector angle and the flushing water flow rate, improving the trajectory prediction accuracy.

[0112] The specific implementation process of the roasting activation magnetic separation combined iron removal system is as follows:

[0113] For the mixture of low-grade iron ore or metallurgical slag after high-temperature roasting, first, the pretreatment module is used for cooling, crushing and residue separation. The mixture is cooled to room temperature by inert gas and then enters the high-pressure roller crusher, with the crushing pressure controlled in the range of 50-100 MPa, to break the material size to the target interval of 0.1-1 mm. The crushed product is screened by a vibrating screen to remove particles with a size larger than the target size (the screen mesh size is the upper limit of the target size), and the screened particles are returned to the crushing chamber for secondary crushing. The qualified particles are conveyed to the air flow separation unit by a screw conveyor. The air flow separation unit uses a cyclone separator to separate unreacted additives and residues (density difference causes residues to be thrown to the wall and discharged) at a tangential wind speed of 10-15 m / s. The purified non-uniform magnetic material flows through the belt conveyor to the magnetic characteristic dynamic detection module. The surface of the belt conveyor is made of anti-static material, and baffles are arranged on both sides. The transmission speed is dynamically adjusted according to the processing capacity of the detection module. The light sensor at the feeding port monitors the material flow in real time and feeds back to the crushing unit, forming a closed-loop control of the crushing particle size.

[0114] After the magnetic feature dynamic detection module receives the material flow, the susceptibility sensor array distributed along the feeding port of the magnetic separation module scans the material susceptibility distribution at a sampling frequency of 100 Hz, generates a two-dimensional susceptibility distribution heat map (high susceptibility area corresponds to magnetite enrichment particles, low susceptibility area corresponds to hematite or impurities), and synchronously deployed X-ray diffraction probe calculates the mass proportion of magnetite and hematite through characteristic diffraction peak intensity analysis, and generates a mineral phase classification matrix. After the above data is denoised by Kalman filtering, it is transmitted to the magnetic field parameter adaptive matching module and the separation path optimization module through industrial Ethernet.

[0115] Based on the susceptibility distribution heat map and the mineral phase classification matrix, the magnetic field parameter adaptive matching module constructs a separation cavity magnetic field gradient requirement model through finite element simulation. In the gradient adjustable permanent magnet-electromagnetic composite magnetic field generator, the permanent magnet array is activated according to the strong magnetic area parameters (corresponding to the high susceptibility area) to generate a low gradient strong magnetic field with an axial gradient of ≤5T / m and a tangential strength of ≥1.2T, which is used to capture magnetite particles; the electromagnetic compensation coil is activated according to the medium-low magnetic area parameters (corresponding to hematite and oxide layer particles) to generate a high gradient medium-strong magnetic field with an axial gradient of ≥8T / m and a tangential strength of 0.8-1.0T. The current intensity of the electromagnetic compensation coil is adjusted in real time by the dynamic magnetic compensation module: the residence time of the material in the cavity is recorded by the RFID marked particle tracking system, the real-time decay amount is predicted by combining the susceptibility decay curve of the historical experiment database, the power amplifier output is adjusted by the PID controller, the reverse magnetic field is injected to offset the magnetic decay effect, and the guide plate angle compensation instruction (CAN bus transmission to the separation path module) is generated synchronously.

[0116] The separation path optimization module integrates the susceptibility distribution heat map, the magnetic field distribution data, and the particle motion image captured by the high-speed CCD camera to construct a multi-scale separation state atlas. The trajectory state observer calculates the theoretical motion trajectory of the particle group based on the balance relationship of magnetic force, gravity, and drag force (magnetic force is determined by susceptibility and magnetic field gradient, gravity is calculated by particle density and cavity inclination, and drag force is derived from material flow rate and fluid viscosity); the multi-objective optimization algorithm (NSGA-II is adopted) takes the concentrate grade ≥85% and the tailings loss rate ≤5% as constraint conditions, and iteratively optimizes the regulation parameters of the guide plate angle (15°-45°) and the flushing water flow rate (0.5-2.0m / s). The regulation parameters are converted into step pulse signals by double encoder servo motors (absolute position encoder calibrates the initial angle, incremental encoder drives rotation, positioning accuracy ±0.5°), which are synchronously sent to the guide plate actuator and the cross-module cooperative control module. The flushing water flow rate parameter is converted into water pressure adjustment instruction by proportional integral valve (valve opening degree is linearly related to flow rate).

[0117] The cross-module cooperative control module receives magnetic field distribution data, regulation parameters and compensation instructions, and then preforms the separation process through discrete element simulation to build a virtual twin model (simulates particle motion trajectory based on material particle density, magnetic susceptibility and cavity geometric parameters), generates virtual separation indicators (concentrate grade, tailings loss rate). When the virtual indicators deviate from the actual detection data (concentrate grade detector, tailings flowmeter) by more than 5%, the magnetic field parameter and separation path synchronous reset instruction is triggered: the magnetic field parameter module recalculates the magnetic field gradient requirement model, adjusts the permanent magnet array and electromagnetic compensation coil activation strategy; the separation path module updates the optimization solution set based on the new magnetic field parameters, generates and executes new regulation parameters. The timestamp alignment module corrects the timing error of the magnetic field regulation instruction and the flow guide plate action instruction (the front segment instruction is delayed by 0.5 seconds, the middle segment is delayed by 1 second, and the rear segment is delayed by 2 seconds) to avoid instruction conflicts.

[0118] The system also includes a dynamic partitioning unit for the separation cavity and a flexible electrode array: the cavity is divided into three segments along the material flow direction, and a photoelectric sensor (RS-485 bus data transmission) is installed in each segment to independently monitor the flow rate (the front segment detects the dispersion state, the middle segment tracks the motion trend under the action of the magnetic field, and the rear segment monitors the distribution before separation); the flexible electrode array (with a spacing of 10-20mm, embedded in the inner wall of the cavity) is connected to a low-frequency signal generator, which applies a 1-5Hz sinusoidal alternating electric field (the electric field strength is dynamically adjusted according to the particle aggregation degree analyzed by the high-speed camera system). The electric field strength data is transmitted to the trajectory state observer to correct the particle aggregation interference term in the magnetic force-gravity-drag force balance equation (quantified through an empirical model of electric field strength and aggregation degree), and the updated trajectory data is fed back to the multi-objective optimization algorithm to iterate the parameters. The electric field regulation instruction and the separation path parameters are synchronized through the timestamp alignment module to ensure consistent action timing.

[0119] Through the closed-loop cooperative control of the above modules, the system realizes dynamic adaptation of the magnetic field distribution to the non-uniform magnetic properties of the material, magnetic attenuation compensation, separation path optimization, and multi-module instruction timing calibration, effectively solving the mismatch problem between static magnetic fields and non-uniform magnetic particles, improving the capture efficiency of medium magnetic particles, and ultimately improving the iron recovery rate and concentrate grade.

[0120] The application solves the problem of static magnetic field adaptation by the following technical solutions: first, the magnetic characteristic dynamic detection module collects the magnetic susceptibility distribution and mineral phase composition data of the material in real time, and generates a magnetic susceptibility distribution heat map and a mineral phase classification matrix. The magnetic field parameter self-adaptive matching module constructs a magnetic field gradient requirement model of the separation cavity based on the above data, and drives the gradient adjustable permanent magnet electromagnetic composite magnetic field generator to generate a dynamically adapted low-gradient magnetic field in the high-magnetic region and a high-gradient magnetic field in the medium-low magnetic region. Through the collaborative regulation of the permanent magnet array and the electromagnetic compensation coil, the real-time matching of the magnetic field distribution and the non-uniform magnetic characteristics of the material is realized, and the insufficient magnetic force of the medium-magnetic particles caused by the mismatch of the static magnetic field strength is avoided.

[0121] The dynamic magnetic compensation module tracks the particle residence time and predicts the magnetic attenuation amount, adjusts the current intensity of the electromagnetic compensation coil to inject a reverse magnetic field to offset the attenuation effect of the particle magnetism. The separation path optimization module integrates the magnetic susceptibility distribution, magnetic field data and particle motion image, and iteratively solves the optimal parameters of the deflector angle and flushing water flow rate through the trajectory state observer and multi-objective optimization algorithm. The calculation results of the magnetic force, gravity and drag force balance equation drive the deflector actuator to dynamically adjust the separation path, maintain the critical balance threshold of the magnetic force and fluid drag force, and prevent the medium-magnetic particles from deviating from the target channel.

[0122] The cross-module collaborative control module preforms the separation process through a virtual twin model, and compares the simulation indicators and actual data in real time. When the deviation exceeds the set threshold, the synchronous reset instruction of the magnetic field parameters and the separation path is triggered, and the time sequence error of the control instruction is corrected through the time stamp alignment module. The dynamic partition unit of the separation cavity and the flexible electrode array further eliminate the particle agglomeration interference and improve the trajectory prediction accuracy. The closed-loop collaborative control of multiple modules realizes the integration of magnetic field dynamic adaptation, separation path optimization and real-time compensation, and finally improves the capture efficiency of medium-magnetic particles and the recovery rate and concentrate grade of iron elements.

Claims

1. Roasting activation magnetic separation combined iron removal system, characterized in that: include: Preprocessing module, magnetic feature dynamic detection module, magnetic field parameter adaptive matching module, dynamic magnetic compensation module, sorting path optimization module and cross-module collaborative control module; The pre-processing module is used to cool and crush the mixed material after roasting and separate the residue, generate a non-uniform magnetic material flow and transmit it to the magnetic characteristic dynamic detection module; The magnetic characteristic dynamic detection module includes a magnetic susceptibility sensor array and an X-ray diffraction probe, which collects the magnetic susceptibility distribution and mineral phase composition data of the material in real time, generates a magnetic susceptibility distribution heat map and a mineral phase classification matrix after filtering, and pushes the heat map and classification matrix to the magnetic field parameter adaptive matching module and the sorting path optimization module; The magnetic field parameter adaptive matching module constructs a magnetic field gradient requirement model for the sorting cavity based on the magnetic susceptibility distribution heat map, drives the gradient-adjustable permanent magnet electromagnetic composite magnetic field generator to switch the magnetic field mode, and generates a low-gradient magnetic field in the strong magnetic region and a high-gradient magnetic field in the medium-low magnetic region corresponding to the magnetic susceptibility distribution; The dynamic magnetic compensation module receives the magnetic field distribution data from the magnetic field parameter adaptive matching module, predicts the magnetic attenuation by tracking the particle retention time, adjusts the current intensity of the electromagnetic compensation coil to inject a reverse magnetic field, and sends a guide plate angle compensation instruction to the sorting path optimization module; The sorting path optimization module integrates the magnetic susceptibility distribution heat map, magnetic field distribution data and particle motion image, solves the control parameters of the guide plate angle and flushing water flow rate through a trajectory state observer and a multi-objective optimization algorithm, and sends the control parameters to the guide plate actuator of the sorting chamber; The cross-module collaborative control module receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, the control parameters of the sorting path optimization module and the compensation instructions of the dynamic magnetic compensation module, and rehearses the sorting process through the virtual twin model. When the deviation between the sorting index of the virtual twin model and the actual sorting index exceeds the set threshold, the synchronous reset of the magnetic field parameters and the sorting path is triggered.

2. The roasting activation magnetic separation combined iron removal system according to claim 1, characterized in that: The pre-processing module comprises: A crushing and screening unit, which is used to receive the roasted mixed material and crush it into a target particle size range through a high-pressure roller crusher; an air flow separation unit, which receives the material output from the crushing and screening unit, removes unreacted additive residues through a cyclone separator, and generates a non-uniform magnetic material flow; The non-uniform magnetic material flows through a belt conveyor and is transported to a magnetic characteristic dynamic detection module.

3. The roasting activation magnetic separation combined iron removal system according to claim 1 is characterized in that: The magnetic feature dynamic detection module is configured as follows: The magnetic susceptibility sensor array is distributed in a ring along the feed port of the magnetic separation module, and scans the magnetic susceptibility distribution of the material in real time and generates a magnetic susceptibility distribution heat map; The X-ray diffraction probe analyzes the mineral phase composition online to generate a mineral phase classification matrix including the proportion of magnetite and hematite; The magnetic susceptibility distribution thermogram and the mineral phase classification matrix are transmitted via industrial Ethernet to a magnetic field gradient requirement model building node of a magnetic field parameter adaptive matching module.

4. The roasting activation magnetic separation combined iron removal system according to claim 1, characterized in that: In the magnetic field parameter adaptive matching module: The gradient-adjustable permanent magnet electromagnetic composite magnetic field generator includes a permanent magnet array and an electromagnetic compensation coil; The permanent magnet array is activated according to the strong magnetic region parameters output by the magnetic field gradient requirement model to generate a low-gradient strong magnetic field; The electromagnetic compensation coil is activated according to the medium and low magnetic zone parameters output by the magnetic field gradient requirement model to generate a high gradient medium strong magnetic field; The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the magnetic attenuation amount predicted by the dynamic magnetic compensation module; The current intensity of the electromagnetic compensation coil is dynamically adjusted according to the magnetic susceptibility attenuation prediction value.

5. The roasting activation magnetic separation combined iron removal system according to claim 1, characterized in that: The sorting path optimization module is configured to: The trajectory state observer calculates the particle group motion trajectory through the magnetic gravity drag balance equation based on the magnetic susceptibility distribution heat map and magnetic field distribution data; The multi-objective optimization algorithm uses the concentrate grade and tailings loss rate as constraints to iteratively optimize the control parameters of the guide plate angle and the flushing water flow rate; The control parameters are converted into step pulse signals of the guide plate actuator through the dual-encoder servo motor and synchronously sent to the cross-module collaborative control module.

6. The roasting activation magnetic separation combined iron removal system according to claim 1, characterized in that: The cross-module collaborative control module is configured to: The virtual twin model receives the magnetic field distribution data of the magnetic field parameter adaptive matching module, the control parameters of the sorting path optimization module, and the compensation instructions of the dynamic magnetic compensation module, and previews the sorting process through discrete element simulation; When the deviation between the sorting index output by the virtual twin model and the actual sorting index exceeds the set threshold, the magnetic field parameters and the sorting path are synchronously reset; The timestamp alignment module corrects the timing error between the control instruction of the magnetic field parameter adaptive matching module and the guide plate action instruction of the sorting path optimization module.

7. The roasting activation magnetic separation combined iron removal system according to claim 1, characterized in that: Also includes: The dynamic partitioning unit of the sorting cavity divides the sorting cavity into three sections: front, middle and back. Photoelectric sensors are deployed in each section to independently monitor the material flow rate. A flexible electrode array is embedded in the inner wall of the sorting chamber to apply a low-frequency alternating electric field to destroy the agglomeration of magnetic particles; The electric field strength data of the flexible electrode array is linked with the trajectory state observer of the sorting path optimization module to correct the particle agglomeration interference term in the magnetic gravity drag balance equation.

Images