A washing and selecting digital intelligent medium adding system and method based on positive pressure dense phase pneumatic conveying technology

Abstract

The application provides a washing and selecting digital intelligent medium adding system and method based on positive pressure dense phase pneumatic conveying technology, the medium adding system comprises a device end and a digital intelligent control end, bidirectional data interaction is realized through a data acquisition module and an industrial Ethernet communication module, the device end comprises a positive pressure dense phase conveying subsystem, a spiral metering feeding subsystem, a sensing monitoring subsystem, an execution driving subsystem and an industrial control subsystem, the positive pressure dense phase conveying subsystem is used for generating a positive pressure airflow, and conveying medium to the spiral metering feeding subsystem, the spiral metering feeding subsystem conveys the medium to a washing and selecting process end, the sensing monitoring subsystem collects operation parameters of the device end, the execution driving subsystem is used for receiving a control signal and adjusting execution parameters, and the industrial control subsystem receives an instruction of the digital intelligent control end and drives an execution module to operate. Through full-parameter sensing monitoring and remote monitoring, real-time fault early warning and historical data tracing are realized, operation and maintenance cost is reduced, and system operation stability is improved.

Description

Technical Field

[0001] This invention belongs to the field of material conveying and intelligent control technology, specifically relating to a washing and screening intelligent media processing system and method based on positive pressure dense phase pneumatic conveying technology. Background Technology

[0002] In the coal washing and processing industry, heavy media coal preparation is a method of separating coal using a liquid with a density between that of clean coal and gangue (or middlings) as the medium. Clean coal, with a density lower than the medium, floats, while gangue or middlings, with a density higher than the medium, sinks, and are then collected and classified into different products. Heavy media coal preparation is currently a core coal preparation technology with high separation accuracy and a wide range of adjustable separation density. Among its key production steps, the addition of the medium is crucial. A specific medium (currently, a suspension of magnetite powder and water is commonly used as the separation medium both domestically and internationally) must be precisely delivered to a qualified medium tank to ensure the stability of the production process and product quality. The stability of the medium addition system directly affects the separation effect, medium loss, and system cost.

[0003] The main equipment used in the preparation and addition of media in traditional coal preparation plants includes loaders or electromagnetic chucks, media preparation tanks, and thickening pumps. During production, personnel operate loaders or electromagnetic chucks, water supply valves, and air blower valves to prepare high-density media powder (magnetite powder with a true density of 4.5 g / cm³, 90% particle size distribution <45 μm, and moisture content <8%) into a concentrated medium with a density of 1.8-2.0 kg / L. Based on the requirements of the heavy media control system, the thickening pump is activated to add the concentrated medium into the qualified media tank in the sorting workshop. Traditional media addition methods suffer from problems such as inaccurate measurement of media and water consumption, inaccurate density measurement, high manual labor intensity, and poor on-site working environment.

[0004] However, traditional media addition systems still have many shortcomings: First, traditional media preparation is mostly manually controlled or semi-automated. When operators use loaders or electromagnetic chucks to add media to the media tank, they often manually identify the media pile height to select the picking area. In addition, some areas cannot be picked up by the electromagnetic chuck, resulting in uneven media pile height and low media addition efficiency. Furthermore, if the flushing valve or blower valve is not opened in a timely manner by the operator, it will cause media to settle in the tank, affecting the normal operation of the media addition system. Second, the media addition system lacks a closed design, which can easily lead to media loss and dust pollution during the transportation process, failing to meet environmental protection requirements. Third, the existing system has not formed a digital and intelligent closed loop of "perception-decision-execution-feedback", lacks real-time monitoring of the transportation status and equipment operating parameters, has weak fault early warning capabilities, and has high maintenance costs. Fourth, in the existing technology, the coordination between media preparation and media addition processes is insufficient. Fluctuations in media density can easily lead to fluctuations in the amount of media added, and there is a large room for optimization of system energy consumption.

[0005] Positive pressure dense phase pneumatic conveying is a technology that uses the kinetic or pressure energy of airflow to transport solid particles or powders along a planned pipeline route within a closed system. Considering the characteristics of high-density, magnetic, and particle-size-controlled media powders, as well as the need to maintain particle size and the vulnerability of pipelines, dense phase pneumatic conveying technology, with its core features of low air velocity, high solid-to-air ratio, and non-suspended thrombosis, perfectly meets the conveying requirements of media powders. It ensures a production capacity of 45 t / d while maintaining accurate material proportions within the qualified media tanks through 0.5% precision metering control. Therefore, designing a closed, precise, and intelligent positive pressure dense phase pneumatic conveying digital closed-loop media feeding system and method to solve problems such as low media feeding accuracy, severe pollution, and insufficient intelligence in existing technologies has become an urgent need in the coal washing and processing industry. Summary of the Invention

[0006] To address the shortcomings of the existing technologies, this invention provides a washing and beneficiation intelligent media addition system and method based on positive pressure dense phase pneumatic conveying technology. Through the deep integration of positive pressure dense phase conveying technology and intelligent control algorithms, the media powder is continuously, tightly, and precisely added from the media storage silo to multiple qualified media tanks in the coal preparation plant's sorting workshop via a dedicated pipeline.

[0007] The technical solution provided by this invention is a washing and screening intelligent media feeding system based on positive pressure dense phase pneumatic conveying technology, comprising an equipment end and an intelligent control end. The equipment end and the intelligent control end achieve bidirectional data interaction through a data acquisition module and an industrial Ethernet communication module. The intelligent control end is used to receive detection data from the equipment end and control the operation of the equipment end. The equipment end includes a positive pressure dense phase conveying subsystem, a screw metering and feeding subsystem, a sensing and monitoring subsystem, an execution and drive subsystem, and an industrial control subsystem. The positive pressure dense phase conveying subsystem is used to generate a stable positive pressure airflow, conveying the medium in a dense phase state at low speed to the screw metering and feeding subsystem. The screw metering and feeding subsystem is used to convey the medium to the washing and screening process. On the process side, the sensing and monitoring subsystem includes pressure sensors, flow sensors, medium concentration sensors, temperature sensors, and valve position sensors, which collect parameters such as conveying pressure, medium flow rate, added medium amount, and equipment operating temperature in real time. The execution drive subsystem includes a frequency converter, an electromagnetic regulating valve, and a stepper motor, which receives control signals from the digital control terminal and adjusts execution parameters such as high-pressure gas source pressure, conveying flow rate, and feed valve opening. The industrial control subsystem uses an industrial-grade programmable logic controller and an industrial computer to receive instructions from the digital control terminal, drive the execution drive module, store operating parameters and sensor data cache information, and simultaneously collect data from the sensing and monitoring module and upload it to the digital control terminal.

[0008] Furthermore, the digital control terminal includes a database, a digital decision-making module, a parameter optimization module, a data comparison module, and a remote monitoring unit; Database: Stores the identity IDs of each device. kBased on the corresponding basic parameters, historical operating data, and optimized parameter models, establish a device-parameter mapping library ID. k (P0, Q0, T0, M0), where P0 is the rated delivery pressure, Q0 is the rated medium flow rate, T0 is the operating temperature threshold, and M0 is the standard medium supply. The intelligent decision-making module embeds PID control algorithms and machine learning models, receives real-time sensor data uploaded from the device, and generates optimal control parameters by combining process requirements and historical data in the database. Parameter optimization module: Dynamically adjusts control parameters based on real-time operating condition fluctuations to ensure medium addition accuracy and conveying stability; Data comparison module: compares the actual operating parameters uploaded by the device with the optimal parameters generated by the data intelligence decision module to determine whether they exceed the allowable error range. If they do, an early warning is triggered. Remote monitoring unit: Provides a visual operation interface, allowing production and maintenance personnel to view the operating status of the additive system, adjust parameters, detect fault alarms, and trace historical data in real time.

[0009] Furthermore, the PID controller of the intelligent decision-making module is deployed in the local PLC of the device, and a synchronous simulation model is set on the intelligent control terminal for parameter tuning and performance evaluation, which acts on the actuators of the screw metering feed subsystem and the high-pressure air source. The machine learning model is deployed on the server of the intelligent control terminal and adopts a supervised learning framework to predict the optimal control parameter set. The PID and machine learning work together through a setpoint generation-real-time tracking architecture.

[0010] Furthermore, the control loop of the PID is as follows: (1) Flow closed-loop control loop: based on the actual flow rate Q of the medium 动态 As a feedback quantity, the value is set to Q. 目标 The frequency of the inverter is adjusted by controlling the output, thereby changing the speed of the air compressor or the screw feeder, to achieve rapid flow tracking. An incremental PID algorithm is used. u(k) = u(k-1) + K p [e(k)-e(k-1)]+K i e(k)+K d [e(k)-2e(k-1)+e(k-2)] Where, u(k): control output at the k-th sampling time, u(k-1): control output at the (k-1)-th sampling time, K p : Proportional gain, K i Integral gain, K d: Differential gain, e(k): Error value at the k-th sampling time, e(k-1): Error value at the (k-1)-th sampling time, e(k-2): Error value at the (k-2)-th sampling time, e(k)=Q 目标 -Q 动态 (k), K p K i K d The initial values ​​are tuned according to the Ziegler-Nichols method and support online adaptive adjustment; (2) Medium addition closed-loop control loop: based on the cumulative medium addition amount M measured by the weighing sensor in the receiving metering chamber. 动态 For feedback, add medium M to the target. 目标 By adjusting the opening of the pneumatic slide valve or the start-stop cycle of the screw feeder, the error of a single feeding can be ensured to be ≤0.5%.

[0011] Furthermore, the input features and output target of the machine learning model are as follows: The input feature vector x includes: Device Identity ID k The corresponding basic process parameters include P0, Q0, T0, and M0; Real-time sensor data, including current pressure P1, flow rate Q1, temperature T1, medium discharge concentration, and qualified medium tank level; Historical operating status, including average error, energy consumption, and fault records over the past 8 hours; External operating condition variables include the ash content and moisture content of the raw coal to be washed, and the daily production plan load rate; Output target y = (P 目标 Q 目标 M 目标 () is the set value for the next control cycle.

[0012] Furthermore, the positive pressure dense phase conveying subsystem includes a high-pressure gas source, a conveying pipeline, a silo pump, and a receiving metering silo. The high-pressure gas source is connected to the air inlet of the silo pump through the conveying pipeline, the discharge port of the silo pump is connected to the receiving metering silo, and the discharge port of the receiving metering silo is connected to the screw metering feed subsystem.

[0013] Furthermore, the positive pressure dense phase conveying subsystem also includes a dust collector, the inlet of which is connected to the outlet of the silo pump.

[0014] Furthermore, a pneumatic fluidizer is installed at the bottom of the receiving metering chamber, in conjunction with a high-frequency vibrator, to prevent media deposition.

[0015] Furthermore, the spiral metering and feeding subsystem includes a spiral feeder, a precision metering device, and a buffer silo. The medium conveyed by the positive pressure dense phase conveying subsystem is conveyed to the buffer silo via the spiral feeder, and the outlet of the buffer silo is connected to the washing and screening process end.

[0016] Another technical solution provided by this invention: a washing and screening intelligent medium addition method based on positive pressure dense phase pneumatic conveying technology, implemented using the above-mentioned medium addition system, includes the following steps: (1) Initialization phase The initialization phase is conducted under safe operating conditions. Staff input basic parameters P0, Q0, T0, M0, and process requirements for each device via a digital control terminal. The digital control terminal then links this information with the device's identification ID. k Bind the device, store it in the database, and complete device registration and parameter initialization; (2) Digital Intelligence Control Stage After the device is started, the local control unit collects initial operating parameters through the sensor monitoring module, including real-time pressure P1, real-time flow rate Q1, and ambient temperature T1, and uploads the initial parameters along with the device ID to the intelligent control terminal. After receiving the data, the intelligent control terminal matches the device ID with the database. k Based on the corresponding basic parameters and process requirements, the intelligent decision-making module combines the initial parameters with the historical optimization model to generate the first set of optimal control parameters P target 1, Q target 1, M target 1, and sends them to the equipment. (3) Closed-loop conveying and media addition stage After receiving the optimal control parameters, the local control unit drives the execution module to operate: the high-pressure gas source is started and adjusted to the target pressure P target 1; the positive pressure dense phase conveying subsystem fluidizes the medium and delivers it to the screw metering feed subsystem at the target flow rate Q target 1. The spiral metering and feeding subsystem measures the medium through a precision metering device. According to the target amount of medium added, M = target 1, the medium is added to the target equipment in a sealed manner through a sealed feeding valve. At the same time, the sensing and monitoring module collects the dynamic parameters P (dynamic), Q (dynamic), and M (dynamic) during the conveying process in real time and uploads them to the digital control terminal. (4) Feedback optimization stage The data comparison module at the intelligent control terminal compares the received dynamic parameters P_dynamic, Q_dynamic, and M_dynamic with the optimal control parameters, and calculates the error values ​​ΔP = |P_dynamic - P_target1|, ΔQ = |Q_dynamic - Q_target1|, and ΔM = |M_dynamic - M_target1|. If all error values ​​are within the allowable range, the system maintains the current control parameters; if any error value exceeds the threshold, the parameter optimization module dynamically generates corrected control parameters P target2, Q target2, and M target2 based on the error type and operating data, and sends them to the device. After receiving the corrected parameters, the device adjusts the operating status of the actuator in real time until the error between the dynamic parameters and the corrected parameters meets the requirements, thus forming a digital and intelligent closed-loop control.

[0017] The intelligent control terminal stores the parameters, error data, and optimization results of this operation in the database, updates the machine learning model, and provides an optimization basis for subsequent operations.

[0018] The beneficial effects of this invention are: (1) The positive pressure dense phase pneumatic conveying and closed medium addition module are combined to achieve medium leakage-free and dust-free conveying and addition, meet environmental protection requirements, and reduce medium loss by more than 25%.

[0019] (2) By integrating PID control algorithm and machine learning model, a digital and intelligent closed-loop control system is constructed to dynamically adapt to changes in working conditions, and the accuracy of the feed is improved by more than 30% compared with the traditional system.

[0020] (3) Real-time fault warning and historical data traceability can be achieved through full-parameter sensing and monitoring and remote monitoring, thereby reducing operation and maintenance costs and improving system stability.

[0021] (4) The parameter optimization module can dynamically adjust the conveying pressure and flow rate according to the working conditions to avoid ineffective energy consumption and save 15%-25% energy compared with traditional systems.

[0022] (5) The system has strong adaptability and can flexibly adjust the parameter model according to different coal quality, different washing processes and different media characteristics. It is suitable for media addition scenarios in multiple fields such as power coal washing plants and coking coal washing plants.

[0023] This invention is applicable to high-density, fine-particle-size powder materials. Compared with negative pressure conveying or dilute phase conveying, it has advantages such as high conveying efficiency (high solid-to-gas ratio), low pipeline wear, low energy consumption, and no dust pollution. It can meet the triple requirements of "continuous and stable closed feeding + fully intelligent operation + precise metering of media". The research realizes the equipment linkage of the entire media feeding system through a fully intelligent system, which greatly reduces the density fluctuation in the qualified medium tank and the human operation error of media feeding. Attached Figure Description

[0024] Figure 1 is an overall system block diagram of the present invention; Figure 2 is a process flow diagram of Embodiment 1 of the present invention; In the diagram: 1—Positive pressure dense phase conveying subsystem, 2—Screw metering and feeding subsystem, 3—Sensing and monitoring subsystem, 4—Execution and drive subsystem, 5—Industrial control subsystem, 6—Intelligent decision-making module, 7—Parameter optimization module, 8—Data comparison module, 9—Remote monitoring unit, 10—Database. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] Figure 1 shows a digital intelligent conveying system based on positive pressure dense phase pneumatic conveying technology, which includes an equipment end and a digital intelligent control end. The two end achieve bidirectional data interaction through a data acquisition module and an industrial Ethernet communication module.

[0027] (a) Equipment end The equipment includes a positive pressure dense phase conveying subsystem 1, a screw metering and feeding subsystem 2, a sensing and monitoring subsystem 3, an execution and drive subsystem 4, and an industrial control subsystem 5. Each piece of equipment is equipped with a unique identification number (ID). k (k=1, 2, 3...n, corresponding to different target devices).

[0028] 1. Positive pressure dense phase conveying subsystem: It consists of a high-pressure gas source, conveying pipeline, silo pump, receiving metering silo, etc. It is used to generate a stable positive pressure airflow to convey the medium at low speed in a dense phase state, thereby reducing medium loss and pipeline wear.

[0029] (1) Silo pump The selected silo pump has a single silo volume of 1.5m³ (based on a conveying capacity of 3t / h, with a circulation time of approximately 30min per silo).

[0030] Workflow: ① Feeding stage (feed valve open, exhaust valve open, material enters the silo pump by gravity, feed valve closes after full); ② Pressurization stage (air inlet valve open, compressed air is introduced, pressure rises to 0.4-0.5MPa); ③ Conveying stage (discharge valve open, material enters the pipeline under air pressure); ④ Cleaning stage (after conveying, maintain low pressure for 30s to clean residual material in the pipeline).

[0031] Compressed air: It must be dried (dew point ≤ -40℃) and filtered (accuracy ≤ 5μm) to avoid moisture or impurities contaminating the material; the supply pressure is 0.5-0.6MPa and the flow rate is ≥ 3m³ / min.

[0032] (2) Delivery pipeline Material: The main pipeline is made of Φ65×4mm seamless steel pipe, lined with 5mm thick 95% alumina ceramic (Mohs hardness 9, wear resistance is 20 times that of ordinary steel pipe). The ceramic and steel pipe are fixed by interference fit and high temperature resistant adhesive.

[0033] Layout design: A load-bearing support is installed every 5m in the horizontal section (to avoid pipe vibration and wear), and a guide support is installed in the vertical section; ceramic elbows with a radius of curvature R=8D are used for elbows (to reduce resistance and wear).

[0034] Anti-clogging control: Pressure sensors (measurement accuracy ±0.01MPa) are installed every 20m on the pipeline. When the pressure in a certain section suddenly increases (exceeds the set value by 1.2 times), the system automatically opens the backflush valve (purges from downstream to upstream for 10s).

[0035] (3) Receiving metering bin and dust collector Receiving and measuring chamber: 3m³ volume, with a 45° conical bottom, and 3 sets of weighing sensors (accuracy 0.1%FS) on the outside to monitor the weight of materials in the chamber in real time (resolution 0.1kg).

[0036] Arch breaking device: A pneumatic fluidizer (installed at the bottom of the cone, 0.1MPa compressed air is introduced to fluidize the material) is used in conjunction with a high-frequency vibrator (50Hz) to solve the "bridging" problem of ultrafine powder.

[0037] Dust collector: Pulse jet bag filter (handling air volume 3000m³ / h), filter bag material is P84 high temperature resistant filter media (temperature resistance ≤200℃, filtration accuracy 0.3μm), dust cleaning cycle 30min (automatically adjusted according to pressure difference), dust emission concentration ≤10mg / m³.

[0038] 2. Screw metering feeder subsystem: Composed of screw feeder, precision metering device, buffer silo, etc., it realizes the closed addition of media, avoids leakage and dust, and accurately controls the amount of media added at one time.

[0039] (1) Bidirectional screw feeder The weight change (Δm / Δt) of the silo is collected in real time by a weighing sensor. The actual flow rate is calculated and compared with the set flow rate (adjusted according to the needs of the dispersion tank). The screw speed is adjusted by PLC (PID control, response time <0.5s) to ensure that the metering accuracy is ≤0.5%.

[0040] (2) U-shaped screw conveyor The U-shaped channel is made of 304 stainless steel (5mm thick), with a transparent acrylic cover on top (for easy observation of the material status). The spiral shaft has a diameter of 150mm, and the blades are continuous full blades (150mm pitch).

[0041] Discharge port design: Each discharge port is equipped with a pneumatic slide valve (diameter DN100, sealing pressure ≥0.1MPa when closed), and a telescopic chute is installed below the valve port (flexibly connected to the feed port of the dispersion tank to reduce dust); the distance between the two discharge ports is 3m to prevent the material from "crossing over" during the conveying process.

[0042] The bidirectional screw feeds lump coal and fine coal from the weighing hopper, and then the U-shaped screw feeds each qualified medium bucket.

[0043] 3. Sensing and monitoring subsystem: including pressure sensor, flow sensor, medium concentration sensor, temperature sensor, valve position sensor, to collect parameters such as delivery pressure, medium flow rate, added medium amount, and equipment operating temperature in real time.

[0044] 4. Drive subsystem: Composed of frequency converter, electromagnetic regulating valve and stepper motor, it receives control signals. The frequency converter is connected to the high-pressure air source and is used to regulate the pressure of the high-pressure air source (air compressor). The electromagnetic regulating valve is set on the conveying pipeline and is used to regulate the conveying flow rate. The stepper motor is used to regulate the gate opening on the discharge pipe of the U-shaped screw conveyor to the qualified medium bucket.

[0045] 5. Industrial Control Subsystem: Employing an industrial-grade programmable logic controller (PLC) and an industrial computer, this subsystem receives commands from the intelligent control terminal, drives the execution modules, and stores operating parameters, sensor data caches, and other information. It also collects data from the sensor monitoring modules and uploads it to the intelligent control terminal.

[0046] (ii) Digital control terminal The digital control terminal includes a database 10, a digital decision-making module 6, a parameter optimization module 7, a data comparison module 8, and a remote monitoring unit 9.

[0047] Database 10: Stores basic parameters (rated pressure, rated medium supply, process requirement threshold), historical operating data, and optimized parameter models corresponding to the IDK of each device, establishing a device-parameter mapping library ID. k : (P0, Q0, T0, M0), where P0 is the rated delivery pressure, Q0 is the rated medium flow rate, T0 is the operating temperature threshold, and M0 is the standard medium supply.

[0048] Digital Decision Module 6: Embedded with PID control algorithm and machine learning model, it receives real-time sensor data uploaded from the device and combines it with process requirements and historical data in the database to generate optimal control parameters (target pressure P, target flow rate Q, target medium addition M).

[0049] 1. Embedded Implementation of PID Control Algorithm The PID controller is deployed in the local PLC at the equipment end, and a synchronous simulation model is set up at the digital control end for parameter tuning and performance evaluation. Its operating objects are the screw metering feed subsystem and the high-pressure air source actuator, and the specific control loop is as follows: (1) Flow closed-loop control loop: based on the actual flow rate Q of the medium 动态 (Obtained through differential calculation using a high-precision electromagnetic flowmeter or weighing sensor) is used as the feedback quantity, with a set value of Q. 目标 The controller output adjusts the inverter frequency, thereby changing the air compressor speed or screw feeder speed to achieve rapid flow tracking. An incremental PID algorithm is used. u(k) = u(k-1) + K p [e(k)-e(k-1)]+K i e(k)+K d [e(k)-2e(k-1)+e(k-2)] Where, u(k): the control output at the k-th sampling time (current control output), u(k-1): the control output at the (k-1)-th sampling time (previous control output), K p : Proportional gain (proportional coefficient, controls the strength of the proportional term), K i Integral gain (integral coefficient, controlling the strength of the integral term), K d : Differential gain (differential coefficients, controlling the strength of the differential term), e(k): Error value at the k-th sampling time (expected value - actual value), e(k-1): Error value at the (k-1)-th sampling time (previous error), e(k-2): Error value at the (k-2)-th sampling time (error of the time before last), e(k) = Q 目标 -Q 动态 (k), K p K i K d The initial values ​​are tuned according to the Ziegler-Nichols method and support online adaptive adjustment.

[0050] (2) Medium addition closed-loop control loop: based on the cumulative medium addition amount M measured by the weighing sensor in the receiving metering chamber. 动态 For feedback, add medium M to the target. 目标 By adjusting the opening of the pneumatic slide gate valve or the start-stop cycle of the screw feeder, the error of a single feeding can be ensured to be ≤0.5%.

[0051] The PID controller has a response time of less than 0.5 seconds, which meets the requirement for rapid suppression of disturbances (such as changes in medium humidity and fluctuations in pipeline resistance) during dense phase transportation.

[0052] 2. Embedded Implementation of Machine Learning Models The machine learning model is deployed on the intelligent control server and uses a supervised learning framework to predict the optimal set of control parameters (P). 目标 Q 目标 M 目标 Its input features and output target are defined as follows: The input feature vector x includes: Device Identity ID k The corresponding basic process parameters (P0, Q0, T0, M0); Real-time sensor data (current pressure P1, flow rate Q1, temperature T1, medium discharge concentration, qualified medium tank level); Historical operating status (average error, energy consumption, and fault records over the past 8 hours); External operating condition variables (ash content and moisture content of raw coal to be washed, and the planned production load rate for the day).

[0053] Output target y = (P 目标 Q 目标 M 目标 () is the set value for the next control cycle.

[0054] Employing a Random Forest Regressor and Lightweight Gradient Boosting Tree (LightGBM) model, it exhibits strong fitting capabilities for nonlinear and multivariate coupling relationships; it is highly noise-resistant, making it suitable for industrial scenarios where measurement jitter exists; it supports feature importance analysis, facilitating process engineers' understanding of control logic; and it boasts fast inference speed, with a single prediction time of <50ms, meeting real-time scheduling requirements.

[0055] The model training data comes from the system's historical operation database, which contains samples covering at least three months of full operating conditions, including different coal qualities, seasons, and equipment statuses. After each feedback optimization phase, the system automatically labels and adds the dynamic parameters, error data, and correction results of this run to the training set, triggering an online incremental learning mechanism to achieve continuous model evolution. 3. The synergistic mechanism between PID and machine learning The two work together to generate a real-time tracking architecture through setting values: The machine learning model generates a new set of setpoints (P) every 5-10 minutes (or when operating conditions change significantly). 目标 Q 目标 M target ); The PID controller uses this set value as an instruction to complete the fine adjustment of the actuator within a time scale of seconds; If the PID loop exceeds the tolerance multiple times consecutively (e.g., ΔP > threshold for 30 seconds), the machine learning model is triggered to re-predict, forming a digital closed loop of "slow decision-making - fast execution".

[0056] This collaborative mechanism retains the stability and real-time performance of PID control while introducing machine learning's ability to generalize and predict complex operating conditions, effectively overcoming the problem of insufficient adaptability of traditional fixed parameter control in scenarios such as media characteristic fluctuations and equipment aging.

[0057] Parameter optimization module 7: Dynamically adjusts control parameters based on real-time operating condition fluctuations (such as changes in medium density and target equipment load) to ensure medium addition accuracy and conveying stability.

[0058] Data comparison module 8: Compares the actual operating parameters uploaded by the device with the optimal parameters generated by the data-driven decision-making module to determine whether they exceed the allowable error range. If they do, an early warning is triggered.

[0059] Remote monitoring unit 9: Provides a visual operation interface, allowing production and maintenance personnel to view the operating status of the additive system, adjust parameters, detect fault alarms, and trace historical data in real time.

[0060] A digital and intelligent medium addition method based on positive pressure dense phase pneumatic conveying technology includes an initialization stage, a digital and intelligent control stage, a closed-loop conveying medium addition stage, and a feedback optimization stage. The specific steps are as follows: (I) Initialization Phase The initialization phase is conducted under safe operating conditions. Staff input basic parameters (P0, Q0, T0, M0) and process requirements for each device via a digital control terminal. The digital control terminal then links this information with the device's identity ID. k Bind the device, store it in the database, and complete device registration and parameter initialization.

[0061] (II) Digital Intelligence Control Stage After the device is started, the local control unit collects the initial operating parameters (real-time pressure P1, real-time flow Q1, ambient temperature T1) through the sensor monitoring module, and uploads the initial parameters and the device ID to the digital control terminal.

[0062] After receiving the data, the intelligent control terminal matches the device ID with the database. k Based on the corresponding basic parameters and process requirements, the intelligent decision-making module combines the initial parameters with the historical optimization model to generate the first set of optimal control parameters (P target 1, Q target 1, M target 1) and sends them to the equipment.

[0063] (III) Closed-loop conveying and media addition stage After receiving the optimal control parameters, the local control unit drives the execution module to run: the high-pressure gas source is started and adjusted to the target pressure P target 1, and the positive pressure dense phase delivery module fluidizes the medium and delivers it to the closed medium addition module at the target flow rate Q target 1.

[0064] The closed-loop medium addition module measures the medium through a precise metering device, and adds the medium to the target equipment in a closed manner according to the target medium addition amount M target 1 through a sealed feeding valve. At the same time, the sensing and monitoring module collects dynamic parameters (P dynamic, Q dynamic, M dynamic) in real time during the conveying process and uploads them to the digital control terminal.

[0065] (iv) Feedback and Optimization Phase The data comparison module of the intelligent control terminal compares the received dynamic parameters (P dynamic, Q dynamic, M dynamic) with the optimal control parameters and calculates the error values ​​ΔP=|P dynamic - P target 1|, ΔQ=|Q dynamic - Q target 1|, and ΔM=|M dynamic - M target 1|.

[0066] If all error values ​​are within the allowable range (ΔP≤ΔP threshold, ΔQ≤ΔQ threshold, ΔM≤ΔM threshold), the system maintains the current control parameters; if any error value exceeds the threshold, the parameter optimization module dynamically generates corrected control parameters (P target 2, Q target 2, M target 2) based on the error type and operating condition data and sends them to the device.

[0067] After receiving the corrected parameters, the device adjusts the operating status of the actuator in real time until the error between the dynamic parameters and the corrected parameters meets the requirements, thus forming a digital and intelligent closed-loop control.

[0068] The intelligent control terminal stores the parameters, error data, and optimization results of this operation in the database, updates the machine learning model, and provides an optimization basis for subsequent operations. Example 1

[0069] A digital intelligent medium addition system based on positive pressure dense phase pneumatic conveying technology includes an equipment end and a digital intelligent control end. The equipment end includes a positive pressure dense phase conveying subsystem, a screw metering feeding subsystem, a sensing and monitoring subsystem, an execution drive subsystem, an industrial control subsystem, etc. The equipment identification number is ID1 (taking the heavy medium addition scenario in a coal mine coal preparation plant as an example, k=1).

[0070] The intelligent control unit includes a database, an intelligent decision-making module (embedded with PID algorithm and random forest model), a parameter optimization module, a data comparison module, and a remote monitoring unit (built on SCADA system). The database stores the basic parameters corresponding to ID1: rated delivery pressure P0=0.6MPa, rated medium flow rate Q0=5m³ / h, operating temperature threshold T0=50℃, and standard medium supply M0=200kg / h.

[0071] Specific functions of each subsystem module: Positive pressure dense phase conveying subsystem: The high pressure air source adopts a screw air compressor, the silo pump volume is 1.5m³, the fluidization device realizes the fluidization of the medium through the gas distribution plate, and the conveying pipeline diameter is DN65 to ensure that the medium is conveyed at low speed in a dense phase state (solid-to-gas ratio ≥50).

[0072] The screw metering and feeding subsystem feeds only one qualified medium tank at a time. This is because: ① it avoids a decrease in metering accuracy due to simultaneous loading of the bidirectional screws; ② it prevents a pressure difference from forming in the U-shaped trough when two discharge ports are open simultaneously, which would affect conveying stability. The system switching process automatically rotates according to the principle of "lowest material level first" (the material level in the qualified medium tank is monitored by a pressure-type level gauge). Before switching, the current discharge valve is closed, and the next discharge valve is opened after a 5-second delay (to ensure the material in the screw is emptied) to prevent cross-contamination.

[0073] Sensing and monitoring subsystem: pressure sensor range 0~1MPa, accuracy ±0.01MPa; flow sensor range 0~10m³ / h, accuracy ±0.5%; medium concentration sensor accuracy ±1%, providing real-time feedback on medium transport status.

[0074] The drive subsystem consists of a frequency converter that adjusts the air compressor speed to control the pressure, an electromagnetic regulating valve that regulates the pipeline flow, and a stepper motor that controls the opening of the feeding valve. The response time is ≤0.5s.

[0075] The intelligent decision-making module trains a random forest model using historical operating data, which can predict the optimal loading parameters under different operating conditions and combine them with the PID algorithm to achieve rapid parameter adjustment.

[0076] Data comparison module: Set error thresholds ΔP≤0.02MPa, ΔQ≤0.1m³ / h, ΔM≤5kg / h. If the threshold is exceeded, an audible and visual warning will be triggered.

[0077] As shown in Figure 2, the intelligent media addition system for coal washing based on positive pressure dense phase pneumatic conveying technology, taking the conveying and addition of heavy media (magnetite powder) in a coal mine preparation plant as an example, has the following specific execution process: (1) Initialization phase: Staff enter the basic parameters corresponding to ID1 (P0=0.6MPa, Q0=5m³ / h, T0=50℃, M0=200kg / h) through the digital control terminal. The digital control terminal stores the parameters in the database and completes the equipment registration.

[0078] (2) Digital control stage: After the equipment is started, the sensor monitoring module collects the initial parameters (P1=0.55MPa, Q1=4.8m³ / h, T1=25℃) and uploads them to the digital control terminal along with ID1; the digital decision module combines the database parameters and historical optimization model to generate the optimal control parameters (P target1=0.6MPa, Q target1=5m³ / h, M target1=200kg / h) and sends them to the equipment.

[0079] (3) Closed conveying and media addition stage: The PLC controller at the equipment end drives the air compressor to start, and the frequency converter adjusts the air compressor speed to make the conveying pressure reach 0.6MPa; the positive pressure dense phase conveying module fluidizes the magnetite powder and conveys it to the closed media addition module at a flow rate of 5m³ / h; the electromagnetic flowmeter measures the amount of medium, the stepper motor controls the opening of the feeding valve, and the magnetite powder is added to the qualified medium tank in the washing and beneficiation workshop at a rate of 200kg / h; the sensor monitoring module collects dynamic parameters (P dynamic = 0.59MPa, Q dynamic = 4.95m³ / h, M dynamic = 198kg / h) in real time and uploads them.

[0080] (4) Feedback optimization stage: The data comparison module of the digital control terminal calculates the error values ​​ΔP=0.01MPa, ΔQ=0.05m³ / h, and ΔM=2kg / h, all within the allowable threshold, and the system maintains the current parameters. After running for 1 hour, due to a slight increase in the humidity of the medium, the sensor monitoring module collects P_dynamic = 0.63MPa and ΔP=0.03MPa, which exceeds the threshold. The parameter optimization module generates corrected parameters (P_target2=0.58MPa, Q_target2=4.9m³ / h, M_target2=199kg / h) and sends them out. The equipment side adjusts the air compressor speed and pipeline regulating valve to stabilize P_dynamic at 0.58MPa, Q_dynamic at 4.9m³ / h, and M_dynamic at 199kg / h, and the error is restored to the allowable range. The digital control terminal stores the data of this optimization and updates the random forest model.

[0081] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the content of this specification should be included within the protection scope of the present invention.

Claims

1. A washing and screening intelligent media processing system based on positive pressure dense phase pneumatic conveying technology, characterized in that, The system comprises an equipment terminal and a digital control terminal. These two terminals interact bidirectionally via a data acquisition module and an industrial Ethernet communication module. The digital control terminal receives detection data from the equipment terminal and controls its operation. The equipment terminal includes a positive pressure dense phase conveying subsystem, a screw metering and feeding subsystem, a sensing and monitoring subsystem, an execution and drive subsystem, and an industrial control subsystem. The positive pressure dense phase conveying subsystem generates a stable positive pressure airflow, conveying the medium in a dense phase state at low speed to the screw metering and feeding subsystem. The screw metering and feeding subsystem then conveys the medium to the washing and screening process. The sensing and monitoring subsystem includes a pressure sensor. The system includes flow sensors, medium concentration sensors, temperature sensors, and valve position sensors, which collect real-time parameters such as conveying pressure, medium flow rate, added medium amount, and equipment operating temperature. The execution drive subsystem includes a frequency converter, electromagnetic regulating valve, and stepper motor. It receives control signals from the digital control terminal and adjusts execution parameters such as high-pressure gas source pressure, conveying flow rate, and the opening of the screw feeder valve. The industrial control subsystem uses an industrial-grade programmable logic controller and an industrial computer to receive instructions from the digital control terminal, drive the execution drive module, store operating parameters and sensor data cache information, and simultaneously collect data from the sensor monitoring module and upload it to the digital control terminal. The digital control terminal includes a database, a digital decision-making module, a parameter optimization module, a data comparison module, and a remote monitoring unit; Database: Stores the identity IDs of each device. k Based on the corresponding basic parameters, historical operating data, and optimized parameter models, establish a device-parameter mapping library ID. k (P0, Q0, T0, M0), where P0 is the rated delivery pressure, Q0 is the rated medium flow rate, T0 is the operating temperature threshold, and M0 is the standard medium supply. The intelligent decision-making module embeds PID control algorithms and machine learning models, receives real-time sensor data uploaded from the device, and generates optimal control parameters by combining process requirements and historical data in the database. Parameter optimization module: Dynamically adjusts control parameters based on real-time operating condition fluctuations to ensure medium addition accuracy and conveying stability; Data comparison module: compares the actual operating parameters uploaded by the device with the optimal parameters generated by the data intelligence decision module to determine whether they exceed the allowable error range. If they do, an early warning is triggered. Remote monitoring unit: Provides a visual operation interface, allowing production and maintenance personnel to view the operating status of the additive system, adjust parameters, detect fault alarms, and trace historical data in real time.

2. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 1, characterized in that, The PID controller of the intelligent decision-making module is deployed in the local PLC of the device. At the same time, a synchronous simulation model is set on the intelligent control terminal for parameter tuning and performance evaluation. The actuator acts on the screw metering and feeding subsystem and the high-pressure air source. The machine learning model is deployed on the server of the intelligent control terminal and adopts a supervised learning framework to predict the optimal control parameter set. The PID and machine learning work together through the setpoint generation-real-time tracking architecture.

3. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 2, characterized in that, The PID control loop is as follows: (1) Flow closed-loop control loop: based on the actual flow rate Q of the medium 动态 As a feedback quantity, the value is set to Q. 目标 The frequency of the inverter is adjusted by controlling the output, thereby changing the speed of the air compressor or the screw feeder, to achieve rapid flow tracking. An incremental PID algorithm is used. u(k)=u(k-1)+K p [e(k)-e(k-1)]+K i e(k)+K d [e(k)-2e(k-1)+e(k-2)] Where, u(k): control output at the k-th sampling time, u(k-1): control output at the (k-1)-th sampling time, K p : Proportional gain, K i Integral gain, K d : Differential gain, e(k): Error value at the k-th sampling time, e(k-1): Error value at the (k-1)-th sampling time, e(k-2): Error value at the (k-2)-th sampling time, e(k)=Q 目标 -Q 动态 (k), K p K i K d The initial value is tuned according to the Ziegler-Nichols method and supports online adaptive adjustment. Q 目标 For the target traffic, Q 动态 (k) represents the actual flow rate of the medium at the k-th sampling time; (2) Medium addition closed-loop control loop: based on the cumulative medium addition amount M measured by the weighing sensor in the receiving metering chamber. 动态 For feedback, add medium M to the target. 目标 By adjusting the opening of the pneumatic slide valve or the start-stop cycle of the screw feeder, the error of a single feeding can be ensured to be ≤0.5%.

4. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 2, characterized in that, The input features and output target of the machine learning model are as follows: The input feature vector x includes: Device Identity ID k The corresponding basic process parameters include P0, Q0, T0, and M0; Real-time sensor data, including current pressure P1, flow rate Q1, temperature T1, medium discharge concentration, and qualified medium tank level; Historical operating status, including average error, energy consumption, and fault records over the past 8 hours; External operating condition variables include the ash content and moisture content of the raw coal to be washed, and the daily production plan load rate; Output target y = (P 目标 Q 目标 M 目标 () is the set value for the next control cycle.

5. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 1, characterized in that, The positive pressure dense phase conveying subsystem includes a high-pressure gas source, a conveying pipeline, a silo pump, and a receiving metering silo. The high-pressure gas source is connected to the inlet of the silo pump through the conveying pipeline, the outlet of the silo pump is connected to the receiving metering silo, and the outlet of the receiving metering silo is connected to the screw metering feed subsystem.

6. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 5, characterized in that, The positive pressure dense phase conveying subsystem also includes a dust collector, the inlet of which is connected to the outlet of the silo pump.

7. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 5, characterized in that, The bottom of the receiving metering chamber is equipped with a pneumatic fluidizer, which, together with a high-frequency vibrator, prevents media deposition.

8. The intelligent washing and screening system based on positive pressure dense phase pneumatic conveying technology according to claim 1, characterized in that, The spiral metering and feeding subsystem includes a spiral feeder, a precision metering device, and a buffer silo. The medium conveyed by the positive pressure dense phase conveying subsystem is transported to the buffer silo via the spiral feeder, and the outlet of the buffer silo is connected to the washing and screening process end.

9. A washing and screening intelligent media addition method based on positive pressure dense phase pneumatic conveying technology, implemented using the media addition system as described in any one of claims 1-3 or 5-8, characterized in that, Includes the following steps: (1) Initialization phase The initialization phase is conducted under safe operating conditions. Staff input basic parameters P0, Q0, T0, M0, and process requirements for each device via a digital control terminal. The digital control terminal then links this information with the device's identification ID. k Bind the device, store it in the database, and complete device registration and parameter initialization; (2) Digital Intelligence Control Stage After the device is started, the local control unit collects initial operating parameters through the sensor monitoring module, including real-time pressure P1, real-time flow rate Q1, and ambient temperature T1, and uploads the initial parameters along with the device ID to the intelligent control terminal. After receiving the data, the intelligent control terminal matches the device ID with the database. k Based on the corresponding basic parameters and process requirements, the intelligent decision-making module combines the initial parameters with the historical optimization model to generate the first set of optimal control parameters P target 1, Q target 1, and M target 1, and sends them to the equipment. (3) Closed-loop conveying and media addition stage After receiving the optimal control parameters, the local control unit drives the execution module to operate: the high-pressure gas source is started and adjusted to the target pressure P target 1; the positive pressure dense phase conveying subsystem fluidizes the medium and delivers it to the screw metering feed subsystem at the target flow rate Q target 1. The spiral metering and feeding subsystem measures the medium through a precision metering device. According to the target amount of medium added, M = target 1, the medium is added to the target equipment in a sealed manner through a sealed feeding valve. At the same time, the sensing and monitoring module collects the dynamic parameters P (dynamic), Q (dynamic), and M (dynamic) in real time during the conveying process and uploads them to the digital control terminal. (4) Feedback optimization stage The data comparison module at the intelligent control terminal compares the received dynamic parameters P_dynamic, Q_dynamic, and M_dynamic with the optimal control parameters, and calculates the error values ​​ΔP = |P_dynamic - P_target1|, ΔQ = |Q_dynamic - Q_target1|, and ΔM = |M_dynamic - M_target1|. If all error values ​​are within the allowable range, the system maintains the current control parameters; if any error value exceeds the threshold, the parameter optimization module dynamically generates corrected control parameters P_target2, Q_target2, and M_target2 based on the error type and operating data, and sends them to the device. After receiving the correction parameters, the equipment adjusts the operating status of the actuator in real time until the error between the dynamic parameters and the corrected parameters meets the requirements, thus forming a digital and intelligent closed-loop control. The intelligent control terminal stores the parameters, error data, and optimization results of this operation in the database, updates the machine learning model, and provides an optimization basis for subsequent operations.

Images