Adaptive control system and method for mobile station vehicle-mounted air separation system

Abstract

This invention relates to the field of mobile station vehicle-mounted air separation system technology, and particularly to an adaptive control system and method for a mobile station vehicle-mounted air separation system. The adaptive control method for the mobile station vehicle-mounted air separation system includes: real-time acquisition of environmental wind parameters, material humidity parameters, and finished product cleanliness rate; calculation of feedforward compensation commands based on environmental wind and humidity parameters; calculation of feedback correction commands based on the deviation between the cleanliness rate and the target value; fusion of feedforward and feedback commands to generate the final control command; and control of the variable frequency fan and adjustable air outlets to adjust the air separation airflow. This invention offsets the influence of natural wind and humidity through multi-source heterogeneous disturbance feedforward compensation, directly locks the finished product cleanliness rate through a visual quality two-level closed loop, and configures fault diagnosis and degraded operation strategies; enabling the air separation system to transform from passive defense to active adaptation, maintaining a cleanliness rate of over 95% under combined crosswind and high humidity conditions, saving 15% energy, and achieving unmanned intelligent operation.

Description

Technical Field

[0001] This invention relates to the field of mobile station vehicle-mounted air separation system technology, and in particular to an adaptive control system and method for a mobile station vehicle-mounted air separation system. Background Technology

[0002] Mobile crushing and screening stations are widely used in outdoor sites such as mines, construction waste disposal sites, and temporary engineering projects due to their high flexibility and mobility. Air classification is a key process in its production flow, separating lightweight impurities (such as plastics, wood chips, and paper scraps) and improving the purity of the final product (recycled aggregate). Since the operating environment is usually open-air, air classification systems are inevitably exposed to variable natural weather conditions, posing a significant challenge to their stable operation. Currently, mobile station vehicle-mounted air classification systems mainly employ the following technical solutions to cope with external environmental interference: (1) Physical shielding solution: Install wind deflectors or canvas around the wind classifier to try to reduce the impact of natural wind through physical means. Although this solution has some effect, it cannot completely eliminate airflow turbulence and increases the complexity of the equipment and the difficulty of relocation. It is a passive defense measure.

[0003] (2) Preset redundant air volume scheme: Based on experience, the operator sets the fan speed to a high value to provide sufficient air volume redundancy to suppress the interference of natural wind and cope with high humidity materials. This is a crude open-loop control method with huge energy consumption and unstable effect, and it cannot be dynamically adjusted according to the actual working conditions.

[0004] (3) Manual observation and manual adjustment scheme: Experienced operators continuously observe the sorting effect on site. When they find that the material is affected by external factors (such as seeing the material being blown off course), they manually adjust the inverter knob or the air outlet baffle. This scheme has a slow response, low adjustment accuracy, and relies heavily on manpower, making it difficult to guarantee long-term stable operation.

[0005] (4) Single-variable compensation scheme: Some existing technologies attempt to introduce a single sensor (such as an anemometer) for semi-automatic adjustment. For example, CN113333287B discloses a waste air separation method, which adjusts the parameters according to the material composition by constructing an air separation parameter library and matching historical parameters. However, this scheme is a static parameter call, which cannot sense and compensate for the disturbances of dynamic changes such as natural wind and material humidity in real time, lacks closed-loop verification of the final separation effect, and has limited adaptability.

[0006] (5) Visual feedforward control scheme: CN117415028B discloses a control method for an air classifier, which uses a visual sensor at the feed end to detect the volumetric flow rate and size of the material, and adjusts the air volume accordingly. Although this scheme introduces visual perception, it is only used for feed prediction and belongs to feedforward control. It cannot provide feedback correction for the quality of the sorted finished product. When encountering sudden environmental changes or changes in material characteristics, the control effect is difficult to guarantee.

[0007] (6) Air outlet load balancing scheme: CN119974311B discloses an air separation conveying control method, which uses image recognition of material distribution to adjust the load of each air outlet to achieve balance. This scheme focuses on the load distribution between air outlets, does not involve compensation for external environmental disturbances (such as natural wind and humidity), and cannot take the final product quality as the control target. Summary of the Invention

[0008] The technical problem to be solved by this invention is: in order to solve the problems of low sorting accuracy, high energy consumption and heavy reliance on manual intervention in the existing mobile station vehicle-mounted air sorting system, this invention provides an adaptive control system and method for the mobile station vehicle-mounted air sorting system. By integrating multi-source heterogeneous disturbance perception, feedforward compensation and quality closed-loop control, the air sorting process can be made intelligent and unmanned, ensuring that the target sorting quality can be stably achieved under various harsh working conditions.

[0009] The technical solution adopted by this invention to solve its technical problem is: an adaptive control method for a mobile station vehicle-mounted air separation system, comprising the following steps: S1. Real-time acquisition of at least two heterogeneous disturbance parameters that affect the air separation effect and the quality parameters of the finished product after air separation, wherein the heterogeneous disturbance parameters include ambient wind parameters and material humidity parameters; S2. Based on the heterogeneous disturbance parameters, calculate the first control command for the air classifier actuator through a preset physical model or empirical model to counteract the interference of heterogeneous disturbances on the air classifier process. S3. Based on the deviation between the finished product quality parameters and the preset target quality parameters, calculate the second control command for the air separation actuator to correct the air separation effect; S4. Superimpose or merge the first control command and the second control command to generate the final control command; S5. According to the final control command, control the air separation actuator to adjust the air separation airflow.

[0010] By simultaneously collecting two types of external disturbances with different properties—ambient wind parameters and material humidity parameters—as well as the quality parameters of the finished product after air separation, this approach breaks through the limitations of existing methods that only focus on a single disturbance, comprehensively incorporating key factors affecting air separation performance into the sensing scope. Through a pre-set physical or empirical model, the sensed ambient wind and material humidity are transformed into pre-compensation commands for the air separation actuators, achieving proactive feedforward compensation of disturbances. This allows for advance adjustment of air separation parameters before the disturbances significantly impact the quality of the finished product. By comparing the real-time detected quality parameters of the finished product with preset target values, the deviation is calculated, and correction commands are generated accordingly, establishing a closed-loop feedback control system with the final product quality as the target. By superimposing or intelligently fusing the feedforward compensation commands and feedback correction commands, the final comprehensive control command is generated, achieving coordinated control of feedforward and feedback. This leverages the advantages of rapid feedforward response while utilizing feedback to eliminate steady-state errors. By sending the synthesized commands to actuators such as variable frequency fans and adjustable air outlets, the state of the air separation airflow is actually changed, ultimately improving the air separation effect and achieving a complete closed loop.

[0011] Furthermore, the environmental wind parameters include natural wind speed. And the natural wind direction angle θ, the material moisture parameter is the material moisture content H, Step S2 specifically includes: S21. Based on the material moisture content H and the reference humidity The deviation is used to calculate the target output of the wind turbine. : ,in, As the reference wind turbine frequency, This is the humidity compensation coefficient. For the amplitude limiting function, This is the minimum limit value for the fan frequency. This is the maximum limit value for the fan frequency; S22, Based on natural wind speed Calculate the pre-compensation value for the horizontal deflection angle of the wind outlet, based on the natural wind direction angle θ. : ,in, The measured wind speed at the air outlet. This is the minimum limit value for the horizontal deflection angle. This is the maximum limit value for the horizontal deflection angle; S23. Estimate the combined wind speed along the material's direction of travel. : And calculate the pre-compensation value of the pitch angle of the vent. : ,in, As the reference pitch angle, This is the wind speed compensation coefficient. As the baseline composite wind speed, This is the minimum limit value for the pitch angle. This is the maximum limit value for the pitch angle.

[0012] By linearizing the model This method maps humidity deviation to a compensation amount for fan frequency, and uses a limiting function to ensure safe quantitative compensation for material humidity disturbances. It transforms the previously experience-based assessment of humidity effects into precise mathematical calculations; this is achieved through a vector decomposition model. The lateral component of natural wind is converted into quantitative compensation for the horizontal deflection angle of the air outlet, and a dual-limiting method is used to achieve vector cancellation of natural wind disturbance. Natural wind and system outlet air are modeled as vectors, and the composite lateral velocity is brought close to zero by adjusting the air outlet direction. The composite wind speed along the material's forward direction is calculated. and based on The deviation from the baseline value is linearly corrected for the pitch angle to achieve adaptive adjustment of airflow intensity. The projections of natural wind and system exhaust air in the main direction are superimposed to assess the effective wind energy actually used for separation, and the pitch angle is adjusted accordingly. This is incorporated into various calculation formulas. The limiting function imposes safety boundary constraints on the calculation results, ensuring the engineering safety of the control system and preventing the mechanical / electrical safety limits from being exceeded due to sensor malfunctions, calculation overflows, or extreme operating conditions. Furthermore, the humidity compensation coefficient The following calibration steps are used to obtain the following: While keeping other variables constant, change the moisture content H of the feed. For each moisture content value, adjust the fan frequency to achieve the target cleanliness level of the finished material. ; Record the moisture content H and the corresponding fan frequency F under this operating condition; The humidity compensation coefficient is obtained by performing linear or piecewise linear fitting on the collected data points. Alternatively, a lookup table function F=f(H) can be established.

[0013] By fixing other operating parameters and only changing the feed moisture content H, and recording the fan frequency F required to achieve the target cleanliness rate, a precise correspondence between humidity and frequency is established, quantifying the abstract influence of humidity into a specific compensation coefficient. Least squares fitting is then performed on the collected (H,F) data points to obtain... The compensation model is simplified for engineering purposes. Within the range of humidity variation, the complex nonlinear relationship is approximated as a linear model, which facilitates real-time calculation by PLC. A lookup table function F=f(H) is established as an equivalent replacement for the linear model, providing multiple implementation paths and allowing the selection of the most suitable implementation method based on site conditions and accuracy requirements.

[0014] Furthermore, the wind speed compensation coefficient The following calibration steps are used to obtain the following: While keeping other variables constant, the composite wind speed is changed. ; For each composite wind speed value, adjust the air outlet pitch angle to achieve the target cleanliness rate of the finished material. ; Record the composite wind speed under this operating condition. And the corresponding pitch angle β; The collected data points are subjected to linear or piecewise linear fitting to obtain the wind speed compensation coefficient. Or establish a lookup table function β=f( ).

[0015] Keep other parameters constant and change the composite wind speed Record the pitch angle β required to achieve the target cleanliness rate, establish a precise correspondence between wind speed and angle, making pitch angle compensation quantifiable and reproducible; for the collected ( Fitting the data points (β) together yields... Achieve engineering simplification of the compensation model; establish β=f( The lookup table function provides multiple implementation paths, enhancing the adaptability and flexibility of the solution.

[0016] Furthermore, the finished product quality parameter is the finished product cleanliness rate. The data is acquired in real time by a vision inspection module installed on the finished product conveyor belt; Step S3 specifically includes: S31. Calculate the cleanliness deviation. : ,in, The target cleanliness level is set via the human-machine interface or issued by the process formula. S32, Cleanliness deviation Low-pass filtering and dead-zone handling are implemented to avoid frequent adjustments caused by instantaneous fluctuations in material flow. , , where ε is the dead zone threshold; S33, Based on the cleanliness deviation after treatment The proportional-integral control algorithm is used to calculate the correction amount for the wind classifier actuator: , , ,in, This is the proportional coefficient for wind turbine frequency correction. The integral coefficient for wind turbine frequency correction. The scaling factor for pitch angle correction. The integral coefficients for pitch angle correction. This is the scaling factor for correcting the horizontal deflection angle.

[0017] The cleanliness rate of the sorted finished material can be directly obtained through a vision module installed on the finished product conveyor belt. The control objective is focused on the final product quality, rather than intermediate process quantities. By comparing the real-time cleanliness rate with the target value, a quantitative deviation is obtained to establish an error benchmark for quality control, providing a basis for subsequent corrections. Low-pass filtering removes high-frequency noise, and dead-zone handling ignores minute fluctuations, preventing frequent actuator actions due to instantaneous material flow fluctuations. Proportional-integral control is employed, with the proportional term responding to the current deviation and the integral term eliminating steady-state error, achieving precise and error-free deviation correction. This addresses the problem that simple proportional control cannot eliminate steady-state error, ensuring that the cleanliness rate continues to accurately track the target value after long-term operation. Simultaneously, the fan frequency is calculated. Pitch angle Horizontal deflection angle The correction amount enables collaborative optimization control of multiple variables, solving the problem that the adjustment capability of a single actuator is limited and unable to cope with complex deviations.

[0018] Furthermore, the cleanliness rate of the finished material Obtained through the following visual inspection steps: An industrial camera installed above the finished product conveyor belt captures images of the finished product flow in real time. The acquired image is cropped to extract the effective area of ​​the belt; Convert the cropped image from the RGB color space to the HSV or Lab color space; The suspected lightweight debris area is extracted using a threshold segmentation algorithm, and a binary image is generated. Morphological processing of binary images includes opening operations for noise reduction and closing operations for hole filling; The processed connected components are filtered to remove noise points with an area smaller than a preset threshold and false detection regions whose aspect ratio does not conform to the light and cluttered features. Calculate the number of lightly cluttered pixels remaining after filtering. The cleanliness rate of the finished product is obtained by combining the total number of pixels N of the ROI: .

[0019] By mounting the industrial camera above the finished product conveyor belt, rather than at the feeding end, the problem of visual assessment of the final sorting effect at the feeding end is solved. The effective area of ​​the belt is extracted, eliminating edge interference and improving detection accuracy, while also addressing the interference of irrelevant backgrounds such as belt edges and frames. Color and brightness are decoupled to enhance robustness to changes in lighting, addressing the challenges of adapting to large variations in outdoor lighting and fixed thresholds. Suspected light impurity areas are extracted, and noise reduction and hole filling are performed through opening and closing operations to achieve preliminary identification and image optimization of light impurities, solving the problems of high noise levels and discontinuous targets in the original image. False detections are eliminated based on features such as area and aspect ratio, improving detection accuracy and eliminating interference from reflections and shadows. The cleanliness rate is represented by the pixel area ratio, enabling quantitative output of detection results with simple calculations and good real-time performance.

[0020] Furthermore, the generated final control command is as follows: , , ,in, To ultimately issue control commands to the variable frequency motor, To ultimately issue control commands to the horizontal servo drive, This is the final control command sent to the pitch servo drive.

[0021] By superimposing the feedforward value of humidity compensation with the feedback correction value of the visual closed loop, the advantages of feedforward and feedback are complemented. Feedforward provides a fast response to disturbances, while feedback eliminates steady-state errors, resolving the contradiction that a single control mode cannot balance speed and accuracy. At the same time, the corrections for fan frequency, horizontal angle, and pitch angle are superimposed to achieve coordinated control of three-dimensional airflow, overcoming the limitation of traditional solutions that only adjust a single dimension. Each command is then subjected to SAT limiting again to ensure the absolute safety of the final command and to resolve the issue of exceeding mechanical / electrical limits after superposition.

[0022] Furthermore, it also includes fault diagnosis and degraded operation steps: When the confidence level of the visual detection module is lower than the preset threshold, the visual data is marked as abnormal, the visual correction amount at the current moment is suspended, the correction amount at the previous stable moment is maintained and gradually decayed to zero according to the preset decay function, and the system degenerates into a control mode consisting of feedforward compensation and wind speed closed loop. When the anemometer data is abnormal, the crosswind compensation logic is disabled, the current horizontal deflection angle is kept as the baseline value, and the adjustment is dominated by the quality closed-loop step. When the wind speed sensor data at the air outlet is abnormal, the wind speed closed loop is stopped and the fan frequency open loop control is adopted. At the same time, the maximum fan frequency change rate is limited and the quality closed loop step performs slow correction. When multiple sensors malfunction simultaneously, the system switches to a preset safety baseline condition and issues a manual takeover prompt.

[0023] If the visual confidence level is abnormal, visual correction is paused, and the existing correction amount is gradually decayed to zero to ensure that the system can still operate when vision fails. The correction amount decays gradually according to a preset function rather than instantaneously returning to zero, achieving a smooth transition of control modes and avoiding system oscillations caused by sudden changes in the correction amount. Crosswind compensation is disabled, and the quality closed loop takes over to ensure that basic sorting effects can still be maintained through quality feedback when crosswind perception fails. Wind speed closed loop is disabled, and open-loop control is switched to limit the speed to ensure that the system can still operate when the inner loop feedback fails. The system switches to a preset safety baseline condition to ensure system safety in extreme situations and avoid complete loss of control. An alarm is issued to remind human intervention, realizing human-machine collaboration and seeking human assistance in a timely manner when the system's capabilities are insufficient to solve safety problems that the fully automatic system cannot handle in extreme situations.

[0024] An adaptive control system for a mobile station vehicle-mounted air separation system includes: The environmental sensing module includes an anemometer and a material humidity sensor, which are used to collect environmental wind parameters and material humidity parameters in real time. The quality inspection module includes an industrial camera and vision processing unit mounted on the finished product conveyor belt, used to acquire images of the finished product in real time and calculate the cleanliness rate of the finished product. ; The central control module is connected to the environmental sensing module, the quality detection module, and the air separation actuator. The central control module is configured to perform the following operations: Based on the environmental wind parameters and material humidity parameters, the first control command for the wind separation actuator is calculated using a preset physical model or empirical model. Based on the cleanliness rate of the finished product Compared with the preset target cleanliness rate The deviation is used to calculate the second control command for the air separation actuator; The first control command and the second control command are superimposed or merged to generate the final control command; The final control command is sent to the air separation actuator; The air separation execution module includes a variable frequency fan and an adjustable air outlet, which is used to receive the final control command and adjust the air separation airflow.

[0025] By constructing a multi-source sensing system using anemometers and material humidity sensors, a comprehensive perception of key external disturbances affecting air separation is achieved. Industrial cameras and vision processing units on the finished product conveyor belt acquire cleanliness rates in real time, enabling direct monitoring of final product quality. Integrating feedforward computing, feedback computing, and command fusion, an intelligent control hub is built, enabling the fusion processing and collaborative decision-making of multi-source information. Variable frequency fans and adjustable air outlets achieve vector regulation of airflow, transforming control commands into physical actions, thus solving the problem of traditional actuators having a single adjustment dimension and being unable to accurately reshape the airflow field.

[0026] Furthermore, the adjustable air vent has at least two degrees of freedom: The horizontal deflection degree of freedom is used to adjust the horizontal deflection angle α of the air outlet to counteract the lateral component of the natural wind. Pitch freedom is used to adjust the pitch angle β of the air outlet, thereby adjusting the relative interaction position and separation intensity between the airflow and the material; The variable frequency fan is used to adjust the fan frequency F to change the air velocity at the air outlet.

[0027] The air outlet is oscillated left and right by servo drive, thereby adjusting the lateral direction of airflow and actively counteracting the lateral component of natural wind. The air outlet is also oscillated up and down by servo drive, thereby dynamically adjusting the interaction position between airflow and material to adapt to changes in wind speed and material characteristics. The fan motor speed is adjusted by frequency converter, thereby achieving stepless adjustment of airflow (wind speed) to meet airflow requirements under different working conditions. By adjusting the fan frequency, horizontal deflection angle, and pitch angle, a complete airflow vector adjustment capability is constructed to achieve full-dimensional control of the airflow.

[0028] Furthermore, the central control module is built into an onboard PLC or industrial computer and is configured with a multi-closed-loop control architecture: First-level closed loop: based on the measured value of the wind speed sensor at the air outlet. As a feedback quantity, the PID control algorithm is used to make the fan output track the target wind speed, thereby achieving rapid and stable control of the airflow in the air selection process. Secondary closed loop: based on the cleanliness rate of finished materials As feedback, the correction amount for the fan frequency and air outlet angle is calculated through the proportional-integral control algorithm, and the given value of the first-level closed loop is corrected online to make the cleanliness rate of the finished material approach the target value. Feedforward compensation channel: Based on environmental wind parameters and material humidity parameters, it calculates pre-compensation values ​​for fan frequency and air outlet angle in real time and superimposes them onto the output of the second-level closed loop.

[0029] wind speed sensor For feedback, PID control of the fan frequency enables rapid and precise execution of airflow separation, resulting in a fast response speed; this is achieved by controlling the cleanliness of the finished material. For feedback, proportional-integral control corrects the feedforward value, achieving closed-loop assurance of final product quality and directly locking the control target; based on ambient wind and humidity, pre-compensation values ​​are calculated in real time to actively cancel disturbances, with a response speed faster than feedback; from feedforward to second-level closed loop and then to first-level closed loop, a collaborative system with multiple time scales and multiple control targets is constructed, with feedforward responding quickly to disturbances, second-level closed loop slowly correcting quality, and first-level closed loop quickly tracking commands, achieving decoupling of control loops at the level of hundreds of milliseconds and second-level closed loops at the level of seconds.

[0030] Furthermore, the central control module also includes a fault diagnosis and degradation control unit, specifically including: Real-time monitoring of sensor data and visual inspection confidence level to determine if any anomalies occur; When an abnormality is detected in the vision module, the update of the vision correction amount is automatically paused, and the existing correction amount is gradually reduced to zero according to the preset decay function, and the system switches to the feedforward + wind speed closed-loop control mode. When an anomaly is detected in the wind speed and direction instrument, the crosswind compensation logic is automatically disabled, and the quality closed loop takes the lead in adjustment. When an abnormality is detected in the wind speed sensor, the closed-loop wind speed control is automatically disabled, and the control is switched to open-loop wind frequency control, while limiting the adjustment rate. When multiple sensors are detected to be malfunctioning simultaneously, the system automatically switches to a preset safety baseline condition and issues a manual takeover prompt.

[0031] Continuously monitor the data status and visual confidence level of each sensor to achieve early fault detection; pause correction, reduce attenuation to zero, and degrade operation to ensure that the system does not crash when vision fails; disable crosswind compensation and let the quality closed loop take the lead to ensure that basic sorting can still be maintained when wind direction perception fails; disable wind speed closed-loop, open-loop control and speed limiting to ensure that the outer loop can still work when the inner loop feedback fails; switch safety benchmarks and manual takeover to ensure ultimate safety in extreme cases; different degrade modes correspond to different faults to achieve refined and differentiated fault handling.

[0032] Furthermore, it also includes a human-computer interaction module, used for: Set target cleanliness rate Material type, process formula, and limiting parameters of each actuator; It displays current environmental parameters, material humidity, outlet air velocity, finished product cleanliness rate, control commands, and alarm information in real time. Record and store historical running data for subsequent parameter calibration and optimization.

[0033] Through a touchscreen or host computer interface, operators can set target cleanliness levels, limit parameters, and other human-machine interaction entry points, making the system configurable and adjustable; dynamically display current operating parameters, control commands, and alarm information to achieve transparency of system status, facilitating operator monitoring and intervention; store historical operating data for offline analysis, enabling digital accumulation of experience and providing data support for parameter optimization and fault analysis. Furthermore, the visual detection module includes: An industrial camera is mounted above the finished product conveyor belt, with the lens vertically downwards covering the effective width of the belt. A constant current LED supplemental lighting device is used to provide stable lighting conditions; Protective housing and transparent protective window are used to protect the camera lens; Dust control devices, including air curtains, air knives, or scrapers, are used to keep protective windows clean; The vision processing unit is used to perform image acquisition, threshold segmentation, morphological processing, connected component filtering, and cleanliness calculation.

[0034] The system takes images from above the finished product conveyor belt, enabling direct monitoring of finished product quality; constant current LED supplementary lighting provides stable and controllable illumination, eliminating the impact of natural light variations on image quality; a protective cover and transparent protective window physically isolate dust and water droplets, protecting precision optical components; dustproof devices include air curtains, air knives, or scrapers, actively cleaning the protective window to maintain a clear imaging optical path; and a vision processing unit enables real-time and reliable cleanliness rate calculation.

[0035] A computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the adaptive control method of the mobile station vehicle-mounted air separation system as described above.

[0036] The beneficial effects of this invention are: By simultaneously collecting environmental wind parameters and material humidity parameters, and based on a vector synthesis model and a linearized compensation model, pre-compensation commands for fan frequency, air outlet horizontal deflection angle, and air outlet pitch angle are calculated in real time; simultaneously, the finished material cleanliness rate is used as the basis for calculation. As a feedback quantity, the residual deviation of the feedforward compensation is corrected online through the PI control algorithm, forming a dual guarantee of feedforward and feedback. Regardless of wind, rain, or changes in the dryness or wetness of the material, the system can automatically sense and accurately compensate. Through multi-dimensional disturbance sensing and collaborative compensation, the air separation system is transformed from adding baffles and redundant air volume to precise compensation and closed-loop correction. No matter how the environment changes, the sorting accuracy and the quality of the finished aggregate always remain highly consistent. Through humidity compensation formula Achieving on-demand air supply through a vector compensation model Achieve directional correction; through multi-variable collaborative control, ensure that the fan always operates in the high-efficiency range, only increasing air volume or adjusting angle when necessary; Industrial vision cameras monitor finished product quality in real time, replacing manual visual inspection; through multi-source disturbance feedforward compensation, the central controller calculates the optimal control command in real time based on sensor data, replacing manual experience judgment; through fault diagnosis and degradation strategies, automatic degradation operation and prompts manual intervention are achieved when sensors malfunction, replacing full-time manual monitoring; fully automatic closed-loop control is realized, and with the cooperation of remote central control vehicle, operators can monitor the system operation from a safe central control room, completely eliminating dangerous manual intervention; The system of this invention has the ability to self-diagnose faults and degrade operation. Even if some sensors fail, it can automatically and smoothly degrade operation without immediate shutdown. With the help of remote alarm prompts, it can truly realize an intelligent operation mode with no or few people on duty. Attached Figure Description

[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0038] Figure 1 This is a schematic diagram of the installation of each module of the system of the present invention.

[0039] Figure 2 This is a schematic diagram of the system of the present invention.

[0040] Figure 3 This is a flowchart of the method of the present invention.

[0041] Figure 4 This is a comparison chart of the wind separation control effect of the method of the present invention and the traditional method under different humidity conditions.

[0042] In the diagram: 100, Environmental Sensing Module; 110, Anemometer; 120, Material Humidity Sensor; 200, Quality Inspection Module; 210, Industrial Camera; 220, Vision Processing Unit; 300, Central Control Module; 400, Wind Selection Execution Module; 410, Variable Frequency Fan; 420, Adjustable Air Outlet; 421, Horizontal Servo Drive; 422, Pitch Servo Drive; 430, Wind Speed ​​Sensor; 500, Human-Machine Interaction Module. Detailed Implementation

[0043] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0044] Example 1: like Figure 1 and Figure 2As shown, this embodiment provides an adaptive control system for a mobile station vehicle-mounted wind separation system, including an environmental perception module 100, a quality detection module 200, a central control module 300, a wind separation execution module 400, and a preferably configured human-machine interaction module 500. The environmental perception module 100 includes an anemometer 110 and a material humidity sensor 120. The anemometer 110 is mounted on a telescopic mast on the top of the mobile station. When the mast is raised, it is at least 1 meter above the vehicle's profile to avoid the influence of airflow on the measurement. The anemometer 110 is used to collect natural wind speed in real time. And the natural wind direction angle θ. In this embodiment, the coordinate system is defined as follows: the direction of belt running (i.e., the direction of material movement) is the positive x-axis, the horizontal lateral direction is the y-axis, and the vertical upward direction is the z-axis. The natural wind direction angle θ is defined as the angle between the projection of the natural wind vector in the horizontal plane and the positive x-axis direction. θ is positive when the wind blows from the right.

[0045] The material moisture sensor 120 is a near-infrared (NIR) or microwave non-contact moisture sensor, installed above the feed belt of the vibrating screen with its lens facing the material flow. The material moisture sensor 120 is used to monitor the moisture content H of the feed in real time. The measurement principle is based on the absorption characteristics of water molecules to light of a specific wavelength or the change in the dielectric constant of microwaves. The response time is less than 1 second, meeting the requirements of real-time control.

[0046] The environmental sensing module 100 continuously collects natural wind speed data at a period of 100ms. The data, including the natural wind direction angle θ and the feed moisture content H, are initially filtered and then transmitted to the central control module 300 via industrial Ethernet or CAN bus to provide real-time input for feedforward compensation.

[0047] The quality inspection module 200 includes an industrial camera 210 and a vision processing unit 220. The industrial camera 210 is mounted above the finished product conveyor belt after impurity removal, with its lens pointing vertically downwards, covering the entire effective width of the belt (i.e., the region of interest, ROI). To adapt to harsh outdoor working environments, the industrial camera 210 includes: The constant current LED supplemental lighting device provides stable and controllable lighting conditions, eliminating the impact of natural light variations on image quality; Protective housing and transparent protective window are used to protect camera lenses from dust and water droplets. Dust protection devices, including air curtains, air knives, or scrapers, operate periodically or continuously to keep the protective window clean and ensure clear imaging.

[0048] The vision processing unit 220 is connected to the industrial camera 210 via a high-speed interface and can be implemented using an embedded industrial computer or a high-performance PLC. The vision processing unit 220 performs the following image processing steps: ROI cropping: Extract the effective area of ​​the belt from the original image and exclude irrelevant background interference such as belt edges and frame; Color space conversion: Convert the cropped image from the RGB color space to the HSV or Lab color space to decouple color and brightness, and enhance robustness to changes in lighting. Threshold segmentation: Based on preset brightness, saturation, and chromaticity thresholds, extract regions suspected of being lightweight debris (such as plastic, wood chips, or paper scraps) and generate a binary image; Morphological processing: Opening operation is performed on the binary image to remove small noise, and then closing operation is performed to fill small holes in the clutter area to make the clutter area continuous and complete. Connected component filtering: The processed connected components are filtered to remove noise points with an area smaller than a preset threshold, as well as false detection regions whose aspect ratio does not conform to the characteristics of light noise. Cleanliness calculation: Calculate the number of light impurity pixels remaining after filtration. And the total number of pixels N in the ROI, the cleanliness rate of the finished product is calculated using the following formula. ,in, The value ranges from 0 to 100%, with a larger value indicating a cleaner finished product.

[0049] The vision processing unit 220 outputs a cleanliness rate data every 2 seconds, along with image quality / detection confidence information for subsequent fault diagnosis.

[0050] Working principle: The quality inspection module 200 achieves direct and real-time monitoring of the final effect of air separation by installing an industrial camera on the finished product conveyor belt, providing a reliable feedback signal for quality closed-loop control.

[0051] The central control module 300 is a vehicle-mounted PLC or industrial computer, connected to the environmental sensing module 100, the quality detection module 200, the air separation execution module 400, and the human-machine interface module 500. This module has a built-in control program and performs the following operations: Data acquisition and preprocessing: Data is received from the environmental perception module at a 100ms cycle. The θ and H data are received by the quality detection module at 2-second intervals. And confidence level data, perform preprocessing such as filtering, boundary judgment, and rate of change monitoring; Feedforward compensation calculation: Based on the current environmental wind parameters and material humidity parameters, the first control command to the wind separation actuator is calculated through a preset physical model; Quality closed-loop calculation: based on finished material cleanliness rate Compared with the preset target cleanliness rate The deviation is used to calculate the second control command for the air classifier actuator through the PI control algorithm; Command fusion and output: The first control command and the second control command are superimposed, and after amplitude limiting, the final control command is generated and sent to the wind selection execution module 400; Fault diagnosis and degradation control: Real-time monitoring of the status of each sensor and visual confidence level, and automatic switching to the corresponding degradation operation mode according to the type of anomaly.

[0052] Multi-closed-loop control architecture: such as Figure 2 As shown, the central control module 300 has a three-layer control structure: Feedforward compensation channel: Based on ambient wind and humidity, it calculates the pre-compensation values ​​for fan frequency and air outlet angle in real time, with a fast response speed (second level), and is used to actively cancel known disturbances; First-level closed loop (inner loop): based on the measured value of the wind speed sensor at the air outlet. As a feedback quantity, a PID control algorithm is used to make the fan output track the target wind speed, with a control cycle of 100ms, so as to achieve rapid and stable execution of airflow separation. Secondary closed loop (outer loop): based on the cleanliness rate of finished materials As feedback, the correction amount for the fan frequency and air outlet angle is calculated through the PI control algorithm, and the given value of the first-level closed loop is corrected online. The control cycle is 2 seconds, so that the cleanliness rate of the finished material approaches the target value.

[0053] This multi-closed-loop control architecture enables collaborative optimization across multiple time scales and control objectives. The inner loop executes quickly, the outer loop makes slow corrections, and the feedforward provides proactive compensation. The three loops do not interfere with each other and each performs its own function.

[0054] The air separation execution module 400 includes a variable frequency fan 410 and an adjustable air outlet 420. The variable frequency fan 410 is driven by a frequency converter and receives fan frequency commands from the central control module 300. Adjusting the fan motor speed changes the air velocity at the air outlet. The fan frequency can be adjusted... ~ Continuous adjustment within the range, in this embodiment =20Hz, =50Hz.

[0055] The adjustable vent 420 has at least two degrees of freedom: Horizontal deflection degree of freedom: Controlled by a horizontal servo drive 421, the horizontal deflection angle α of the air outlet can be adjusted to change the direction of airflow in the horizontal plane, thereby counteracting the lateral component of natural wind. Adjustment range ~ In this embodiment =-30°, =30°.

[0056] Pitch freedom: Controlled by a pitch servo drive 422, the pitch angle β of the air outlet can be adjusted to change the relative interaction position and separation intensity between the airflow and the material. Adjustment range ~ In this embodiment =0°, =45°.

[0057] A wind speed sensor 430 is installed at the air outlet to measure the actual air velocity in real time. This signal is then fed back to the central control module 300 as a first-level closed-loop feedback signal.

[0058] Working principle: The air separation execution module 400 receives three-dimensional control commands issued by the central control module 300. , , Through precise execution by frequency converters and servo drives, the wind speed, horizontal direction, and pitch angle of the airflow in the air-separated zone are adjusted in all dimensions, thereby actively generating the required synthetic airflow field in the air-separated zone.

[0059] The human-machine interaction module 500 includes a touch screen or industrial computer display, and is connected to the central control module 300 for: Parameter settings: Allows operators to set target cleanliness levels. Material type, process formula, and limiting parameters of each actuator, etc.; Real-time monitoring: Dynamically displays current environmental parameters and outlet wind speed. Finished material cleanliness rate Control commands ( , , ) and alarm information; Data logging: Records and stores historical operational data. ,θ,H, , , , (etc.), used for subsequent parameter calibration and optimization analysis.

[0060] Example 2: Reference Figure 3 This embodiment provides an adaptive control method for a mobile station vehicle-mounted air separation system, including the following steps: Step 1: Data Acquisition and Baseline Condition Initialization System initialization: Start the fan, air outlet servo mechanism, belt conveyor and vision inspection system, and establish a coordinate system: the direction along the belt / main airflow is the x-axis, the horizontal side is the y-axis, and the angle between the natural wind and the x-axis is θ; Read the preset reference parameters: reference humidity 5%, reference fan frequency 25Hz, reference composite wind speed 18 m / s, reference pitch angle 15°, and the limiting boundary ~ 20-50Hz ~ -30°~30° ~ : 0°~45°.

[0061] Set target cleanliness rate 92%, which can be input through a human-machine interface or automatically distributed through process formula.

[0062] Sensor data acquisition: Obtain natural wind speed from an anemometer. and the natural wind direction angle θ; The moisture content H of the material is obtained from the material humidity sensor; The actual air outlet speed is obtained from the wind speed sensor at the air outlet. ; Cleanliness rate of finished materials obtained from the visual inspection module Cleanliness deviation And image quality / detection confidence information.

[0063] Step 2: Preprocessing and Anomaly Detection Data filtering: Low-pass filtering is performed on the raw data acquired by the sensor to remove high-frequency noise. For example, for... First-order low-pass filtering is used: ; Anomaly detection: Key signals are assessed for out-of-bounds errors, freeze conditions, and rate of change. like If the speed is greater than 35 m / s or remains unchanged for 10 seconds, the anemometer is considered faulty. If H > 35% or H < 0%, the humidity sensor is considered faulty. If | - If the wind speed exceeds 20% and lasts for more than 5 seconds, the wind speed sensor is considered faulty. If the visual confidence level is less than 0.5, the visual module is considered abnormal.

[0064] When an anomaly is detected, the corresponding degradation strategy is triggered based on the anomaly type.

[0065] Step 3: Humidity Compensation Based on the material moisture content H and the reference humidity The deviation is used to calculate the target frequency of the wind turbine. : , Increased humidity makes it easier for lightweight debris to adhere, leading to a higher effective density and a decrease in cleanliness at the same airflow velocity. This formula uses first-order linearization compensation, approximating the frequency increment required to achieve the same separation effect as being proportional to the humidity deviation.

[0066] Humidity compensation coefficient The calibration test revealed that, while keeping other variables constant, varying the feed moisture content H and adjusting the fan frequency for each moisture content value resulted in achieving the desired cleanliness level. Record the (H,F) data points and obtain the results using least squares fitting. In this embodiment, the calibration of construction waste materials is obtained. =1.5.

[0067] This is a limiting function to ensure the output stays within the safe limits. ~ Inside.

[0068] Step 4: Environmental crosswind compensation Vector compensation is performed based on the lateral component of the natural wind. First, the lateral component of the natural wind is calculated: The goal is to minimize the cancellation of the combined lateral velocities in the wind-separated zone, i.e.: Solve for the pre-compensation angle : , When | When |>1, the inner `sat` function restricts the input to [-1,1] to ensure the validity of the `arcsin` operation; the outer `sat` function... The function guarantees output within the mechanical limit. ~ This compensation is a pre-compensation measure, designed to first align the jet direction and reduce the impact of disturbances; whether the final standard is met is then fine-tuned by the visual quality closed-loop system.

[0069] Step 5: Synthetic Wind Speed ​​Compensation Estimate the composite wind speed along the material's direction of travel. : This formula projects and superimposes the jet velocity and the component of natural wind in the main direction to obtain an approximate composite wind speed, which characterizes the effective intensity of the airflow on the material.

[0070] The pitch angle is corrected based on the deviation between the synthesized wind speed and the reference value. : The pitch angle β affects the relative interaction time and position between the jet and the material, when the combined wind speed... When the wind speed is increased, appropriately reducing the pitch angle can prevent excessive blowing from carrying away heavy objects or causing light objects to fall back onto the finished product conveyor belt. In this embodiment, the correction direction is adopted where the higher the wind speed, the smaller β becomes; the sign can be determined by the specific structural calibration.

[0071] Wind speed compensation coefficient The calibration experiment revealed that, while keeping other variables constant, changing the composite wind speed... For each Adjusting the pitch angle allows the cleanliness rate to reach [a certain level]. ,Record( Data points (β) were fitted to obtain... In this embodiment, the calibration is obtained =0.8.

[0072] Step Six: Closed-Loop Airflow Speed ​​at the Air Outlet Actual measured value of wind speed sensor at air outlet As a feedback quantity, a PID control algorithm is used to make the fan output track the target wind speed, which can be determined by... It is obtained by mapping with the characteristic curve of the wind turbine.

[0073] The PID control algorithm is as follows: In this embodiment, the method adopted after on-site adjustment =2.5, =0.3, =0.1.

[0074] This closed loop is used to offset the impact of internal disturbances such as duct blockage, equipment aging, and material load fluctuations on wind speed, ensuring the stable execution of jet intensity.

[0075] Step 7: Visual Cleanliness Level Two-Stage Closed Loop Calculate the cleanliness deviation : ; Filtering and Dead-Time Handling: For Perform low-pass filtering and set a dead zone to avoid frequent adjustments caused by instantaneous fluctuations in material flow: , Where ε is the dead zone threshold, and in this embodiment, ε = 1%; Correction Calculation: Based on the cleanliness deviation after treatment, the correction amount for the air separation actuator is calculated using a PI control algorithm. , , , In this embodiment, the following settings are adopted after on-site adjustment: =6.8, =0.8, =0.15, =0.02, =0.05.

[0076] Command synthesis: The feedforward compensation value and the visual correction value are superimposed to generate the final control command. , , .

[0077] Step 8: Fault Diagnosis and Degraded Operation This step is executed in parallel with the normal control process, and the status of each sensor is monitored in real time: Visual module anomaly degradation: When the visual confidence level falls below a threshold, the visual data is marked as abnormal, the current visual correction is paused, and the correction from the previous stable time is maintained and gradually decays to zero according to an exponential decay function. In this embodiment, τ = 600 seconds is used, and the system degenerates into a feedforward compensation and wind speed closed-loop mode, and an alarm prompt is given for manual inspection.

[0078] Anemometer anomaly degradation: When the anemometer data is abnormal, the crosswind compensation logic is disabled, the current horizontal deflection angle α is maintained as the baseline value, and the adjustment is dominated by the quality closed-loop steps.

[0079] Anomaly-induced degradation of wind speed sensor: When the wind speed sensor data at the air outlet is abnormal, the wind speed closed loop is disabled, and the fan frequency open loop control is adopted. At the same time, the maximum fan frequency change rate is limited, and the quality closed loop step performs slow correction.

[0080] Multi-source anomaly degradation: When multiple sensors malfunction simultaneously, the system switches to a preset safety baseline condition. , , (and issue a manual takeover prompt.)

[0081] Step Nine: Execution and Looping The calculated final control command ( , , The data is sent to the variable frequency fan and servo drive, and updated once every control cycle (100ms).

[0082] The human-machine interface displays the current cleanliness level, environmental parameters, control outputs, and alarm information in real time. The system continuously records operational data. ,θ,H, , (and various instructions), used for subsequent calibration parameter optimization.

[0083] Typical operating condition examples: To more intuitively demonstrate the working process of the method of this invention, the following uses a typical working condition as an example to illustrate the specific calculations of each step: Operating conditions: A construction waste disposal site; the material is a mixture of waste concrete, plastic, and wood chips; Real-time data: =4.2m / s, θ=75°, H=8.7%, =18.3m / s, =87.3%, =92%.

[0084] Control process: Humidity compensation: =25+1.5×(8.7-5)=30.55Hz; Crosswind compensation: =4.2×sin75°=4.06m / s, =arcsin(-4.06 / 20)=-11.7°; Synthetic wind speed compensation: =|20×cos(-11.7°)+4.2×cos75°|=|19.58+1.09|=20.67m / s, =15-0.8×(20.67-18)=12.86°; Visual closed loop: =92% - 87.3% = 4.7%, =6.8×4.7%+0.8×∫4.7%dt=0.32+0.038=0.358Hz, =0.15×4.7%+0.02×∫4.7%dt=0.007+0.00095=0.008°; Instruction synthesis: =30.55 + 0.358 = 30.91 Hz =-11.7°, =12.86 + 0.008 = 12.87°; The final instruction is issued and executed, and the system continuously updates in a loop.

[0085] Comparison examples under different humidity conditions: This embodiment conducted a comparative experiment under different humidity conditions, and the experimental results are shown in Table 1: Table 1: Comparison of Cleanliness Rates between the Embodiment and Traditional Solutions under Different Humidity Conditions

[0086] Experimental data shows that as humidity increases, the cleanliness rate of traditional solutions drops sharply, while the present invention, through humidity compensation and visual closed loop, keeps the cleanliness rate stable at over 83%, with an improvement of up to 32 percentage points under high humidity conditions.

[0087] like Figure 4 As shown, the ambient wind speed is set to zero (ideal situation), the target finished product cleanliness rate is 95%, and the humidity of the test material is gradually increased from 5% to 25%.

[0088] Traditional solution: Operators rely on experience to fix the fan frequency at 40Hz, corresponding to the optimal airflow under dry conditions. When the material moisture content is between 5% and 10%, the cleanliness of the finished product can be maintained at 92% to 95%. However, as the moisture content increases, the adhesion between material particles increases and impurities absorb water and become heavier, making it impossible for the 40Hz airflow to effectively remove light materials. When the moisture content reaches 25%, the cleanliness of the finished product using the traditional solution drops sharply to about 60%, with a large amount of light materials mixed into the finished aggregate.

[0089] In this embodiment, the system's humidity sensor monitors the increase in humidity in real time. The control strategy, based on a humidity-wind speed compensation model, automatically increases the target output of the fan, while the vision system performs fine-tuning. When the humidity rises from 5% to 30%, the system smoothly adjusts the fan frequency from automatic to approximately 50Hz, dynamically increasing the airflow to overcome the increase in material adhesion and specific gravity. The final results show that regardless of humidity changes, the finished product cleanliness rate of this embodiment remains consistently within the target range of 94% to 96%, significantly improving product quality stability.

[0090] Typical harsh working conditions comparison examples Comparative analysis of typical harsh working conditions: open-air work site, after a rain shower (material moisture content H=18%), with an easterly wind blowing (natural wind speed). =6m / s, natural wind direction angle θ=45°). The target cleanliness level is 95%.

[0091] Table 2: Comparison of Traditional Experience-Based Control Schemes, Existing Single-Voltage Variable Frequency Schemes, and This Implementation Scheme under Typical Severe Operating Conditions

[0092] When faced with complex combined conditions of crosswinds and high humidity, simply adjusting the fan frequency cannot change the physical fact that the airflow is deflected by the crosswinds. This embodiment of the solution achieves a true reshaping of a constant sorting airflow field through the coordinated dynamic adjustment of air volume (frequency) and injection direction.

[0093] Visual inspection example: The visual second-order closed-loop convergence process is shown in Table 3: Table 3. Record of Visual Second-Level Loop Convergence Process

[0094] Data shows that it takes about 20 seconds from the occurrence of the disturbance to the recovery to steady state, the overshoot is controlled within 3%, and there is no oscillation, which proves that the second-level closed loop has the characteristics of fast response and accurate convergence.

[0095] It should be noted that, generally, incremental corrections are made to the fan frequency and air outlet angle based on the cleanliness deviation. Only when necessary are minor fine-tuning adjustments made to the horizontal deflection angle. Based on this general principle, Table 3 includes... .

[0096] Example 3: This embodiment provides a computer-readable storage medium storing a computer program thereon. When the program is executed by a processor, it implements the adaptive control method of the mobile station vehicle-mounted air separation system of Embodiment 2. The computer-readable storage medium can be any tangible medium that contains, stores, transmits, propagates, or transmits a program, such as, but not limited to: electronic storage devices, magnetic storage devices (such as hard disks and floppy disks), optical storage devices (such as CD-ROMs and DVDs), magneto-optical storage devices, or any other medium suitable for storing program instructions.

[0097] When the above computer program is run on the processor of an onboard PLC, industrial computer, or embedded system, it causes the device to perform the following operations: Real-time acquisition of environmental wind parameters and material humidity parameters; Real-time acquisition of finished product images and calculation of finished product cleanliness rate; Calculate feedforward compensation instructions based on environmental parameters; Correction instructions are calculated and feedback is given based on the cleanliness deviation; The final control command is generated by integrating feedforward and feedback commands. Control the fan and air outlet to perform adjustments; Perform fault diagnosis and degrade operation.

[0098] In addition, in some cases, wind deflection can be compensated primarily by adjusting the fan speed, or the humidity effect can be compensated primarily by adjusting the adjustable air outlet angle. In other words, the focus of the compensation strategy can be flexibly adjusted according to the actual structure. Material humidity can also be detected indirectly by measuring the material's conductivity or dielectric constant, such as using a capacitive humidity sensor. Besides two-dimensional anemometers, multiple single-axis anemometers can be combined to obtain wind field information, or an ultrasonic anemometer can be used.

[0099] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. An adaptive control method for a mobile station vehicle-mounted air separation system, characterized in that, Includes the following steps: S1. Real-time acquisition of at least two heterogeneous disturbance parameters that affect the air separation effect and the quality parameters of the finished product after air separation, wherein the heterogeneous disturbance parameters include ambient wind parameters and material humidity parameters; S2. Based on the heterogeneous disturbance parameters, calculate the first control command for the air classifier actuator through a preset physical model or empirical model to counteract the interference of heterogeneous disturbances on the air classifier process. S3. Based on the deviation between the finished product quality parameters and the preset target quality parameters, calculate the second control command for the air separation actuator to correct the air separation effect; S4. Superimpose or merge the first control command and the second control command to generate the final control command; S5. According to the final control command, control the air separation actuator to adjust the air separation airflow.

2. The adaptive control method for the mobile station vehicle-mounted wind separation system according to claim 1, characterized in that: The environmental wind parameters include natural wind speed. And the natural wind direction angle θ, the material moisture parameter is the material moisture content H, Step S2 specifically includes: S21. Based on the material moisture content H and the reference humidity The deviation is used to calculate the target output of the wind turbine. : ,in, As the reference wind turbine frequency, This is the humidity compensation coefficient. For the amplitude limiting function, This is the minimum limit value for the fan frequency. This is the maximum limit value for the fan frequency; S22, Based on natural wind speed Calculate the pre-compensation value for the horizontal deflection angle of the wind outlet, based on the natural wind direction angle θ. : ,in, The measured wind speed at the air outlet. This is the minimum limit value for the horizontal deflection angle. This is the maximum limit value for the horizontal deflection angle; S23. Estimate the combined wind speed along the material's direction of travel. : And calculate the pre-compensation value of the pitch angle of the vent. : ,in, As the reference pitch angle, This is the wind speed compensation coefficient. As the baseline composite wind speed, This is the minimum limit value for the pitch angle. This is the maximum limit value for the pitch angle.

3. The adaptive control method for the mobile station vehicle-mounted air separation system according to claim 2, characterized in that: The humidity compensation coefficient The following calibration steps are used to obtain the following: While keeping other variables constant, change the moisture content H of the feed. For each moisture content value, adjust the fan frequency to achieve the target cleanliness level of the finished material. ; Record the moisture content H and the corresponding fan frequency F under this operating condition; The humidity compensation coefficient is obtained by performing linear or piecewise linear fitting on the collected data points. Alternatively, a lookup table function F=f(H) can be established.

4. The adaptive control method for the mobile station vehicle-mounted air separation system according to claim 2, characterized in that: The wind speed compensation coefficient The following calibration steps are used to obtain the following: While keeping other variables constant, the composite wind speed is changed. ; For each composite wind speed value, adjust the air outlet pitch angle to achieve the target cleanliness rate of the finished material. ; Record the composite wind speed under this operating condition. And the corresponding pitch angle β; The collected data points are subjected to linear fitting or piecewise linear fitting to obtain the wind speed compensation coefficient. Or establish a lookup table function β=f( ).

5. The adaptive control method for the mobile station vehicle-mounted wind separation system according to claim 2, characterized in that: The finished product quality parameter is the cleanliness rate of the finished material. The data is acquired in real time by a vision inspection module installed on the finished product conveyor belt; Step S3 specifically includes: S31. Calculate the cleanliness deviation. : ,in, The target cleanliness level is set via the human-machine interface or issued by the process formula. S32, Cleanliness deviation Low-pass filtering and dead-zone handling are implemented to avoid frequent adjustments caused by instantaneous fluctuations in material flow. , , where ε is the dead zone threshold; S33, Based on the cleanliness deviation after treatment The proportional-integral control algorithm is used to calculate the correction amount for the wind classifier actuator: , , ,in, This is the proportional coefficient for wind turbine frequency correction. The integral coefficient for wind turbine frequency correction. The scaling factor for pitch angle correction. The integral coefficients for pitch angle correction. This is the scaling factor for correcting the horizontal deflection angle.

6. The adaptive control method for the mobile station vehicle-mounted wind separation system according to claim 5, characterized in that: The cleanliness rate of the finished material Obtained through the following visual inspection steps: An industrial camera installed above the finished product conveyor belt captures images of the finished product flow in real time. The acquired image is cropped to extract the effective area of ​​the belt; Convert the cropped image from the RGB color space to the HSV or Lab color space; By using a threshold segmentation algorithm, regions suspected to be lightweight debris are extracted and binary images are generated. Morphological processing of binary images includes opening operations for noise reduction and closing operations for hole filling; The processed connected components are filtered to remove noise points with an area smaller than a preset threshold and false detection regions whose aspect ratio does not conform to the light and cluttered features. Calculate the number of lightly cluttered pixels remaining after filtering. The cleanliness rate of the finished product is obtained by combining the total number of pixels N of the ROI: .

7. The adaptive control method for the mobile station vehicle-mounted air separation system according to claim 5, characterized in that, The final control command generated is: , , ,in, To ultimately issue control commands to the variable frequency motor, To ultimately issue control commands to the horizontal servo drive, This is the final control command sent to the pitch servo drive.

8. The adaptive control method for the mobile station vehicle-mounted air separation system according to claim 7, characterized in that, It also includes fault diagnosis and degraded operation steps: When the confidence level of the visual detection module is lower than the preset threshold, the visual data is marked as abnormal, the visual correction amount at the current moment is suspended, the correction amount at the previous stable moment is maintained and gradually decayed to zero according to the preset decay function, and the system degenerates into a control mode consisting of feedforward compensation and wind speed closed loop. When the anemometer data is abnormal, the crosswind compensation logic is disabled, the current horizontal deflection angle is kept as the baseline value, and the adjustment is dominated by the quality closed-loop step. When the wind speed sensor data at the air outlet is abnormal, the wind speed closed loop is stopped and the fan frequency open loop control is adopted. At the same time, the maximum fan frequency change rate is limited and the quality closed loop step performs slow correction. When multiple sensors malfunction simultaneously, the system switches to a preset safety baseline condition and issues a manual takeover prompt.

9. An adaptive control system for a mobile station vehicle-mounted air separation system, characterized in that, include: The environmental sensing module includes an anemometer and a material humidity sensor, which are used to collect environmental wind parameters and material humidity parameters in real time. The quality inspection module includes an industrial camera and vision processing unit mounted on the finished product conveyor belt, used to acquire images of the finished product in real time and calculate the cleanliness rate of the finished product. ; The central control module is connected to the environmental sensing module, the quality detection module, and the air separation actuator. The central control module is configured to perform the following operations: Based on the environmental wind parameters and material humidity parameters, the first control command for the wind separation actuator is calculated using a preset physical model or empirical model. Based on the cleanliness rate of the finished product Compared with the preset target cleanliness rate The deviation is used to calculate the second control command for the air separation actuator; The first control command and the second control command are superimposed or merged to generate the final control command; The final control command is sent to the air separation actuator; The air separation execution module includes a variable frequency fan and an adjustable air outlet, which is used to receive the final control command and adjust the air separation airflow.

10. The adaptive control system of the mobile station vehicle-mounted air separation system according to claim 9, characterized in that: The adjustable air vent has at least two degrees of freedom: The horizontal deflection degree of freedom is used to adjust the horizontal deflection angle α of the air outlet to counteract the lateral component of the natural wind. Pitch freedom is used to adjust the pitch angle β of the air outlet, thereby adjusting the relative interaction position and separation intensity between the airflow and the material; The variable frequency fan is used to adjust the fan frequency F to change the air velocity at the air outlet.

11. The adaptive control system of the mobile station vehicle-mounted wind separation system according to claim 10, characterized in that: The central control module is built into the vehicle-mounted PLC or industrial computer and is configured with a multi-closed-loop control architecture: First-level closed loop: based on the measured value of the wind speed sensor at the air outlet. As a feedback quantity, the PID control algorithm is used to make the fan output track the target wind speed, thereby achieving rapid and stable control of the airflow in the air selection process. Secondary closed loop: based on the cleanliness rate of finished materials As feedback, the correction amount for the fan frequency and air outlet angle is calculated through the proportional-integral control algorithm, and the given value of the first-level closed loop is corrected online to make the cleanliness rate of the finished material approach the target value. Feedforward compensation channel: Based on environmental wind parameters and material humidity parameters, it calculates pre-compensation values ​​for fan frequency and air outlet angle in real time and superimposes them onto the output of the secondary closed loop.

12. The adaptive control system of the mobile station vehicle-mounted wind separation system according to claim 11, characterized in that: The central control module also includes a fault diagnosis and degradation control unit, specifically including: Real-time monitoring of sensor data and visual inspection confidence level to determine if any anomalies occur; When an abnormality is detected in the vision module, the update of the vision correction amount is automatically paused, and the existing correction amount is gradually reduced to zero according to the preset decay function, and the system switches to the feedforward + wind speed closed-loop control mode. When an anomaly is detected in the wind speed and direction instrument, the crosswind compensation logic is automatically disabled, and the quality closed loop takes the lead in adjustment. When an abnormality is detected in the wind speed sensor, the closed-loop wind speed control is automatically disabled, and the control is switched to open-loop wind frequency control, while limiting the adjustment rate. When multiple sensors are detected to be malfunctioning simultaneously, the system automatically switches to a preset safety baseline condition and issues a manual takeover prompt.

13. The adaptive control system of the mobile station vehicle-mounted air separation system according to claim 9, characterized in that: It also includes a human-computer interaction module, used for: Set target cleanliness rate Material type, process formula, and limiting parameters of each actuator; It displays current environmental parameters, material humidity, outlet air velocity, finished product cleanliness rate, control commands, and alarm information in real time. Record and store historical running data for subsequent parameter calibration and optimization.

14. The adaptive control system of the mobile station vehicle-mounted air separation system according to claim 9, characterized in that, The visual detection module includes: An industrial camera is mounted above the finished product conveyor belt, with the lens vertically downwards covering the effective width of the belt. A constant current LED supplemental lighting device is used to provide stable lighting conditions; Protective housing and transparent protective window are used to protect the camera lens; Dust control devices, including air curtains, air knives, or scrapers, are used to keep protective windows clean; The vision processing unit is used to perform image acquisition, threshold segmentation, morphological processing, connected component filtering, and cleanliness calculation.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the adaptive control method of the mobile station vehicle-mounted air separation system as described in any one of claims 1 to 8.

Images