A method and system for preventing wear in a medium speed coal mill

Abstract

The present application relates to the technical field of medium-speed coal mill wear prevention, and particularly relates to a medium-speed coal mill wear prevention method, which comprises the following steps: collecting wear characteristic data in real time through a nano sensor embedded in a lining plate of the coal mill; after pre-processing and noise reduction of the wear characteristic data, inputting the wear characteristic data and working condition time sequence data into a wear rate prediction model to calculate a current wear rate and a remaining service life, and comparing the current wear rate and the remaining service life with multi-stage wear threshold values to determine a wear grade; based on the wear grade, generating and executing at least two multi-dimensional regulation and control instructions; and returning to S1 to re-collect data after regulation and control, so as to form a closed-loop control. The present application realizes the technical upgrade of the wear prevention and protection of the coal mill from 'passive protection + regular maintenance' to 'active sensing + dynamic regulation and control + predictive maintenance', significantly prolongs the service life of the lining plate, and reduces the number of unplanned shutdowns and operating energy consumption. The present application also discloses a wear prevention system.

Description

Technical Field

[0001] This invention relates to the field of wear prevention technology for medium-speed coal mills, and in particular to a wear prevention method and system for medium-speed coal mills. Background Technology

[0002] Medium-speed coal mills are core equipment in the pulverizing systems of coal-fired power plants, chemical plants, and steel smelters. Their function is to grind raw coal into pulverized coal of a certain fineness, and to complete the drying and conveying process of the pulverized coal, ultimately sending it into the boiler furnace, gasifier, or blast furnace for combustion. The operational stability of the coal mill is directly related to the safety and economy of the entire power generation or production process system.

[0003] During operation, raw coal is fed into the center of the grinding disc by the feeder and moves towards the edge of the disc under centrifugal force. It is then pulverized into coal powder as it passes through the crushing area between the grinding rollers and the disc. Hot primary air enters the mill through the nozzle ring, carrying the coal powder upwards to the separator. Coal powder of the appropriate fineness is discharged from the mill with the airflow, while coarse particles return to the grinding disc for further grinding. During this process, the easily worn areas of the medium-speed coal mill (the frame, separator, nozzle ring, etc.) experience severe wear.

[0004] Currently, the main approach to addressing wear issues in medium-speed coal mills is still the traditional passive protection method combined with periodic shutdowns for inspection / replacement. Although the use of high-performance wear-resistant materials such as high-manganese steel, wear-resistant alloy steel, and metal-ceramic composite liners has improved the basic service life of vulnerable parts, the actual wear state of various components in the coal mill exhibits significant local differences and non-uniformity due to factors such as fluctuations in raw coal quality (changes in ash content, hardness, and particle size), dynamic adjustments to coal mill operating parameters (grinding pressure, speed, and ventilation), and equipment installation deviations. The traditional approach is no longer suitable for the efficient, continuous, and safe operation requirements of modern power plants.

[0005] Of course, existing technologies also indirectly infer wear status by monitoring the operating parameters of coal mills. For example, Chinese patent application CN119334614A discloses a method for monitoring the condition of grinding rollers in a medium-speed coal mill. This method acquires real-time operating parameters of the coal mill (coal feed rate, current, inlet and outlet differential pressure, etc.) and calculates the wear amount of the grinding rollers using a preset coal mill pulverizing output and grinding roller wear algorithm model. Chinese patent application CN117884228B discloses a method for deterioration analysis and fault early warning of a medium-speed coal mill. Based on operating parameters such as coal feed rate, loading force, outlet temperature, current, and inlet and outlet differential pressure, it uses decision trees and deterioration trend lines for deterioration analysis and early warning.

[0006] However, the technical solutions disclosed in the above patents still have drawbacks. The operating parameters are affected by a combination of factors, resulting in insufficient measurement accuracy. The changes in indirect parameters lag behind the actual changes in the wear state, making it impossible to implement precise sensing. At the same time, the indirect parameters reflect the overall comprehensive state of the coal mill and cannot identify abnormal wear in local areas, thus failing to effectively address the problem of non-uniform wear. Summary of the Invention

[0007] To address at least one of the aforementioned technical problems, this invention proposes a wear prevention method and system for medium-speed coal mills, which solves the problems of delayed wear perception, passive protection methods, lack of precision in replacement strategies, difficulty in controlling non-uniform wear, and inability to effectively cope with severe wear caused by fluctuations in coal quality and frequent changes in load in existing technologies.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention provides a method for preventing wear in a medium-speed coal mill, comprising: S1, by using nanosensors embedded in the coal mill liner, wear characteristic data of key vulnerable areas of the coal mill are collected in real time; the wear characteristic data includes the remaining thickness of the liner, micro deformation and vibration signal; S2, after preprocessing and denoising the wear feature data, input the preprocessed wear feature data and working condition time series data into the preset wear rate prediction model, calculate the current wear rate and remaining service life, and compare it with the preset multi-level wear threshold to determine the current wear level; S3, based on the determined wear level, generates multi-dimensional control commands and sends them to the corresponding actuators for execution; the multi-dimensional control commands include at least two of the following: coal feed rate adjustment command, mill disc speed adjustment command, nozzle ring wind speed adjustment command, and permanent magnet motor power adjustment command. S4, after the control command of S3 is executed, returns to step S1 to re-collect wear characteristic data, forming a closed-loop control of "sensing-analysis-control-feedback".

[0009] Preferably, in step S2, the wear rate prediction model is a prediction model constructed based on a time-series neural network; The preprocessing noise reduction includes at least one of the following methods: low-pass filtering, band-pass filtering, wavelet denoising, moving average filtering, and outlier removal; The input features of the time-series neural network include: historical liner thickness time-series data, historical micro-deformation time-series data, historical vibration signal time-series data, and auxiliary operating parameters; the auxiliary operating parameters include material particle size, feed flow rate, equipment operating speed, and running time. The output of the time-series neural network includes: short-term wear rate prediction, remaining service life prediction, and critical wear time prediction.

[0010] Preferably, the multi-level wear threshold determination mechanism in S2 includes: When the current wear rate is lower than the preset normal operating average, it is determined to be a light wear level, and the system maintains the current operating condition. When the current wear rate is higher than the normal operating average but lower than the warning threshold, it is determined to be a moderate wear level and the system issues a warning signal. When the current wear rate exceeds the warning threshold, it is determined to be a severe wear level, and the system automatically triggers multi-dimensional control commands; The warning thresholds include: a remaining thickness of less than 50% of the initial thickness is a warning threshold, and a remaining thickness of less than 30% of the initial thickness is a shutdown recommendation threshold.

[0011] Preferably, the generation of the multi-dimensional control command in S3 adopts a decision logic based on wear-causing factor identification: When the wear mode is identified as scouring wear caused by excessive wind speed, the nozzle ring wind speed adjustment command is executed first. When the wear mode is identified as overload-induced roller wear, the coal feed rate adjustment command and the permanent magnet motor power adjustment command are executed first. When the wear mode is identified as fatigue wear caused by vibration and impact, the grinding disc speed adjustment command and the grinding roller loading force adjustment command are executed first. When the wear pattern is identified as localized wear caused by uneven material distribution, the material drop point adjustment command is executed to balance the wear distribution on the liner surface.

[0012] Preferably, the method further includes a condition-based maintenance scheduling step: When the predicted remaining service life in S2 is lower than the preset replacement threshold, a replacement work order is automatically generated and pushed. When the wear feature data collected in S1 identifies local abnormal wear points, the fault points are automatically located and marked, and targeted maintenance suggestions are generated.

[0013] A second aspect of the present invention provides a wear prevention system for a medium-speed coal mill, used to implement the wear prevention method for a medium-speed coal mill as described in the first aspect, comprising: An embedded nano wear monitoring unit is used to perform step S1, including at least one nano sensor embedded in the coal mill liner for real-time detection of the remaining thickness of the liner, micro deformation and vibration signal; The wear rate prediction and analysis unit is communicatively connected to the embedded nano-wear monitoring unit and is used to execute step S2, calculate the current wear rate and remaining service life based on the preset wear rate prediction model, and determine the wear level. The multidimensional control execution unit is communicatively connected to the wear rate prediction and analysis unit and is used to execute step S3, which generates and executes multidimensional control instructions based on the wear level. The feedback control unit is used to execute step S4, feeding back the collected wear characteristic data after adjustment to the wear rate prediction and analysis unit to form a closed-loop control.

[0014] Preferably, the system further includes a gradient composite wear-resistant liner, which is a three-layer gradient composite structure, consisting of the following layers from the inside out: The base layer is made of heat-resistant alloy steel and has a thickness of 8mm to 15mm; The middle layer is a ceramic-rubber-metal fiber composite buffer layer; The surface layer is a wear-resistant layer, composed of nano-modified zirconia toughened alumina ceramic or high-chromium carbide weld overlay, with a hardness ≥ HRC62. The nanosensor is embedded between the base layer and the intermediate layer of the gradient composite wear-resistant liner, or embedded in a groove on the back of the base layer.

[0015] Preferably, the multidimensional control execution unit includes: The coal feed rate adjustment module is used to automatically reduce the coal feed rate when the wear rate exceeds the limit; The grinding disc speed adjustment module is used to reduce the grinding disc speed when there is severe vibration and impact. The nozzle ring air speed adjustment module is used to control the air speed inside the mill by adjusting the opening of the adjustable nozzle ring; The permanent magnet motor power adjustment module is used to adaptively adjust the drive power according to the wear condition.

[0016] Preferably, the system further includes an adjustable rotating nozzle ring, the outlet cross-sectional area of ​​which can be adjusted in real time; the nozzle ring wind speed adjustment module adjusts the opening of the adjustable rotating nozzle ring according to the control command output by the wear rate prediction and analysis unit.

[0017] Preferably, the embedded nano-wear monitoring unit further includes a high-temperature resistant wireless transmission module, which uses LoRa or industrial ZigBee protocols for wireless data transmission; the nano-sensor and wireless transmission module operate in environments below 400°C, with high dust and weakly corrosive fumes, and have a designed lifespan of ≥2 years.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention introduces a ceramic-rubber-metal fiber composite buffer intermediate layer, which improves the overall impact toughness to more than 25 joules per square centimeter, effectively solving the industry problem of brittle fracture of high-hardness materials and ensuring long-term operational reliability.

[0019] By directly measuring the remaining thickness of the liner plate (measurement accuracy ±10μm to ±20μm), micro-deformation, and vibration signals using embedded nanosensors, the problem of perception lag and insufficient accuracy caused by relying on indirect calculation and downtime detection in existing technologies is solved. Through a nine-grid multi-point arrangement, the spatial distribution perception of wear status is realized, which can accurately locate local abnormal wear points and provide accurate basis for targeted maintenance.

[0020] By constructing a wear rate prediction model through a time-series neural network, based on historical wear time-series data and auxiliary operating parameters, the model outputs short-term wear rate prediction, remaining service life prediction, and critical wear time prediction. This solves the problem that existing technologies cannot achieve predictive maintenance, and upgrades maintenance decisions from passive response to proactive prediction.

[0021] By using decision logic based on wear cause identification, for different wear modes such as scouring wear caused by excessive wind speed, grinding wear caused by overload, fatigue wear caused by vibration and impact, and local wear caused by uneven material distribution, multi-dimensional collaborative control commands are generated, including at least two of the following: coal feed rate adjustment, mill speed adjustment, nozzle ring wind speed adjustment, and permanent magnet motor power adjustment. This realizes the transformation from "single-dimensional open-loop adjustment" to "multi-dimensional collaborative closed-loop optimization". Attached Figure Description

[0022] Figure 1 A flowchart of a wear prevention method for a medium-speed coal mill; Figure 2 This is a structural block diagram of an anti-wear system for a medium-speed coal mill; In the figure: 101, Embedded nano-wear monitoring unit; 102, Wear rate prediction and analysis unit; 103, Multi-dimensional control and execution unit; 104, Feedback control unit. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments of the present invention.

[0024] Example 1 Please refer to the figure for a wear prevention method for a medium-speed coal mill, which includes: S1 uses nanosensors embedded in the liner of the coal mill to collect wear characteristic data of key vulnerable areas of the coal mill in real time.

[0025] It should be noted that the nanosensors used in this embodiment mainly include two types: The first type is the ultrasonic sensor: it emits nanometer-level high-frequency sound waves, which penetrate the liner and are reflected upon reaching the worn surface. The sensor receives the reflected waves and calculates the round-trip time difference between transmission and reception, combined with the propagation speed of the sound waves in the liner material, to determine the remaining thickness of the liner. The ultrasonic sensor has a measurement accuracy of ±10 micrometers to ±20 micrometers and is suitable for liner thickness measurements ranging from 0 mm to 50 mm.

[0026] The second type is the capacitive ranging sensor: This type utilizes the capacitance effect between the sensing surface and the worn surface of the liner. When the liner wears down, causing a reduction in thickness, the distance between the sensing surface and the worn surface changes, thus altering the capacitance value. By calibrating the relationship between the capacitance value and the thickness, the remaining thickness of the liner can be deduced. Capacitive ranging sensors have the advantages of fast response and insensitivity to temperature changes.

[0027] The two types of sensors can be selected according to specific working conditions, or they can be used together in the same coal mill for mutual verification.

[0028] The key vulnerable areas of a medium-speed coal mill mainly include the inner wall area of ​​the nozzle ring, the root area of ​​the separator blades, the edge area of ​​the grinding roller, and the inner wall area of ​​the middle frame.

[0029] Specifically, the nozzle ring is the main channel of the coal mill's heat exchange zone and grinding disc. The high-speed airflow carrying pulverized coal causes continuous erosion and wear in this area. The separator blades are used to separate qualified pulverized coal from coarse particles. The pulverized coal airflow scours the root of the blades. The grinding rollers and grinding disc grind the raw coal. The edges of the grinding rollers bear high contact stress and shear stress, making them prone to fatigue wear and abrasive wear. The middle frame is the main structure of the coal mill. The pulverized coal airflow scours the inner wall of the frame during its upward movement.

[0030] To facilitate effective monitoring of the aforementioned critical vulnerable areas by the sensor, the nanosensor in this embodiment is installed using the following embedded method: The sensor installation points are arranged in a nine-square grid, specifically, with nine backing plates forming a group, and the backing plate in the central area selected as the sensor installation target. This arrangement can represent the overall wear condition of the area while reducing the number of sensors and lowering system costs.

[0031] A groove is machined on the back of the liner (i.e., the non-wear surface). Specifically, in this embodiment, a milling or drilling process is used to machine a shallow groove with a depth of 2 mm to 5 mm on the back of the liner. The shape and size of the groove match the shape of the sensor. At the same time, the groove depth should not be too deep to avoid weakening the structural strength of the liner; nor should it be too shallow, otherwise the sensor cannot be fully embedded.

[0032] The nanosensor is embedded in the groove, specifically with the sensor probe facing the wear surface of the liner, to ensure that the acoustic wave or capacitance signal can effectively penetrate the liner.

[0033] High-temperature resistant epoxy or ceramic adhesive is used to fill and fix the sensor. It can be understood that the filling adhesive not only serves to fix the sensor, but also isolates the sensor from the high temperature and high dust environment, protecting the sensor from corrosion and mechanical damage.

[0034] The sensor signal cable is led out along the groove and protected. Specifically, the cable is sheathed with a stainless steel corrugated pipe or a high-temperature resistant ceramic sleeve and runs along the back of the liner to the outside of the coal mill housing, where it is connected to the signal acquisition module.

[0035] It should be noted that the wear characteristic data in this embodiment includes the remaining thickness of the liner, micro-deformation, and vibration signals. Specifically, the remaining thickness of the liner is the most intuitive wear characterization parameter, which can be obtained through the ultrasonic time-of-flight method or capacitance ranging method mentioned above. The liner undergoes minute elastic or plastic deformation under stress and temperature changes, reflecting the stress state and cumulative damage degree of the liner, and can be collected by the strain-sensitive element in the nanosensor. Simultaneously, during the operation of the coal mill, vibrations are generated by the grinding of the grinding rollers and grinding disc, the impact of coal powder particles, and the dynamic balance deviation of rotating parts. The vibration signal includes multiple dimensions such as vibration frequency, peak impact value, and root mean square value of vibration, and can be collected by the piezoelectric sensitive element in the nanosensor.

[0036] Understandably, when the collected wear characteristic data identifies localized abnormal wear points—for example, if the thickness reading of a certain sensor is significantly smaller than that of adjacent sensors—the system automatically locates and marks the fault location. The positioning accuracy can reach the level of a single liner plate. Simultaneously, the system generates targeted maintenance suggestions based on the characteristics of the abnormal wear (e.g., whether it is erosion-type or impact-type), such as checking the flow field distribution in the area and detecting any installation deviations.

[0037] Considering the rotating components (such as grinding discs and rollers) inside the coal mill, wired wiring poses risks of entanglement and wear. To facilitate data transmission, this embodiment employs a high-temperature resistant wireless transmission module. Specifically, the wireless transmission module uses either the LoRa protocol or the industrial ZigBee protocol. The LoRa protocol offers advantages such as long transmission distance, strong penetration capability, and low power consumption, making it suitable for scenarios where signals need to penetrate the metal casing of the coal mill. The industrial ZigBee protocol offers advantages such as flexible networking and strong anti-interference capabilities, making it suitable for scenarios involving multi-sensor network monitoring.

[0038] The wireless transmission module and the nanosensor are integrated and packaged together, and installed in a recessed area of ​​the substrate. The entire sensor module is configured to operate stably for a long time in environments below 400°C, with high dust concentrations, and in weakly corrosive fumes, with a designed service life of no less than 2 years.

[0039] S2. After preprocessing and denoising the wear characteristic data, the preprocessed wear characteristic data and operating condition time series data are input into the preset wear rate prediction model to calculate the current wear rate and remaining service life, and compare them with the preset multi-level wear thresholds to determine the current wear level.

[0040] It is understandable that there are various noise sources in the coal mill operating environment that can easily interfere with the original sensor signals, including but not limited to: power frequency interference, mechanical shock noise, high-frequency random noise, and temperature drift.

[0041] To eliminate noise interference, this embodiment employs at least one preprocessing noise reduction method, including but not limited to: low-pass filtering, band-pass filtering, wavelet denoising, moving average filtering, and outlier removal. After preprocessing noise reduction, the signal-to-noise ratio of the original signal is significantly improved, providing reliable input data for the accurate calculation of the subsequent prediction model.

[0042] It should be noted that the wear rate prediction model in this embodiment is a prediction model based on a time-series neural network. Specifically, a time-series neural network is a type of neural network model specifically designed for processing time-series data. Its characteristic is that it can capture the dependencies and patterns of change in data over time.

[0043] Preferred temporal neural networks of this invention include, but are not limited to, the following: The first type is Long Short-Term Memory Neural Network (LSTM): LSTM can effectively solve the gradient vanishing and gradient explosion problems of traditional recurrent neural networks when processing long sequences by introducing three gating mechanisms: input gate, forget gate and output gate. It is suitable for learning long-term dependencies of wear state evolution.

[0044] The second type is the Gated Recurrent Unit (GRU): GRU is a simplified variant of LSTM, which combines the input gate and forget gate into an update gate. It has fewer parameters and higher computational efficiency, and can achieve prediction accuracy similar to LSTM when there is sufficient data.

[0045] The third type is the Transformer model: Based on the self-attention mechanism, the Transformer can process the entire time series in parallel, which is advantageous for capturing long-term trends and periodic patterns in wear and tear.

[0046] The main input features of the prediction model in this embodiment include: Historical liner thickness time-series data, specifically, is a sequence of liner thickness measurements collected at fixed time intervals (e.g., once per minute) over a past period (e.g., the past 72 hours). This sequence reflects the cumulative effect of wear.

[0047] Historical micro-deformation time-series data, specifically, a sequence of micro-deformation measurements of the liner collected concurrently. The deformation sequence reflects the stress state and elastic recovery of the liner.

[0048] Historical vibration signal time-series data, specifically, is a sequence of vibration characteristic parameters such as vibration frequency, peak impact value, and root mean square value collected concurrently. The vibration sequence reflects the dynamic stability and impact load conditions of the equipment.

[0049] Auxiliary operating parameters, specifically, include material particle size (the particle size distribution of raw coal), feed flow rate (the amount of raw coal entering the coal mill per unit time), equipment operating speed (the rotational speed of the grinding disc), and operating time (the cumulative operating time since the last liner replacement or major overhaul). These operating parameters are important external factors affecting the wear rate.

[0050] The output of the prediction model in this embodiment mainly includes: Short-term wear rate forecast, specifically, refers to the wear thickness per unit time over a future period (e.g., the next 24 hours), measured in millimeters per thousand hours. This value reflects the recent trend in wear evolution.

[0051] The projected remaining useful life (RUL) is, specifically, the estimated remaining operating time, measured in months or hours, from the current moment until the liner thickness reaches its end-of-life threshold. This value is a core basis for developing maintenance plans.

[0052] The critical wear moment prediction value, specifically, refers to the specific time point at which the wear state is expected to change from light or moderate to severe. This value is used for early warning and scheduling intervention.

[0053] To improve the accuracy and robustness of the prediction, this embodiment performs supplementary calculations and verifications on the neural network output values.

[0054] Specifically, this embodiment uses instantaneous wear rate to reflect the rate of wear at the current moment. The calculation formula is as follows:

[0055] In the formula: The thickness wear amount is the amount of reduction in the thickness of the liner plate within a time interval, expressed in millimeters. The time interval is in hours.

[0056] Meanwhile, the remaining useful life (RUL) can be calculated using the following formula:

[0057] In the formula: This represents the current remaining thickness of the liner. The threshold thickness for scrapping the liner (usually taken as 30% of the initial thickness); This represents the current instantaneous wear rate.

[0058] The above formula is based on the linear wear assumption and is applicable to the steady-state wear stage. Of course, in practical applications, the neural network prediction value and the analytical formula calculation value can be fused, for example, by using weighted averaging or Kalman filtering. Combining the advantages of the two methods can effectively improve the accuracy and robustness of the prediction.

[0059] Based on the predicted wear rate and remaining thickness values ​​mentioned above, they are compared with preset multi-level thresholds to determine the current wear level. Specifically: The mild wear level is determined by the following criteria: the current wear rate is lower than the preset average value for normal operation. For example, the average value for normal operation is 0.08 mm per thousand hours. When the actual wear rate is lower than this value, it indicates that the liner is in a normal wear stage and the wear condition is good. The system response is to maintain the current operating condition, without performing any intervention operations, and only continuously monitor.

[0060] The moderate wear level is determined by the following criteria: the current wear rate is higher than the normal operating average but lower than the warning threshold. The warning threshold is set as follows: the remaining thickness is less than 50% of the initial thickness. For example, if the initial thickness is 30 mm, a warning is triggered when the remaining thickness is less than 15 mm. When the wear level is determined to be moderate, the system response is to issue a warning signal and suggest that operators conduct on-site inspections, but it does not automatically adjust operating parameters.

[0061] The severe wear level is determined by the following criteria: the current wear rate exceeds the warning threshold and the wear rate continues to rise. When the remaining thickness is less than 30% of the initial thickness (e.g., 9 mm), the system issues a shutdown recommendation. When the wear level is determined to be severe, the system response is to automatically trigger multi-dimensional control commands to actively adjust the operating parameters of the coal mill to reduce the wear rate.

[0062] In summary, mild wear levels maintain operation, moderate wear levels provide early warnings, and severe wear levels require proactive intervention. This wear level classification system avoids excessive intervention that could impact production efficiency while ensuring timely intervention when wear becomes severe.

[0063] Understandably, when the predicted remaining service life is lower than the preset replacement threshold, a replacement work order is automatically generated and pushed out. Specifically, the replacement work order includes the following information: the location of the liner to be replaced, the estimated replacement time window, the required spare parts model and quantity, and the estimated downtime. The work order can be pushed to maintenance personnel's mobile phones via SMS, or to the maintenance team via WeChat or email.

[0064] S3 generates multi-dimensional control commands based on the determined wear level and sends them to the corresponding actuators for execution.

[0065] Understandably, different causes of wear require different control strategies. To ensure the accuracy of the control strategies, it is necessary to identify the wear patterns of the wear-causing factors.

[0066] Specifically, when the wear mode is identified as erosion wear caused by excessive wind speed, the nozzle ring wind speed adjustment command is executed first. Specifically, erosion wear is characterized by the wear area aligning with the airflow direction, exhibiting a streamlined or grooved morphology, and the vibration signal primarily consisting of high-frequency, small-amplitude vibrations. When this mode is identified, the system prioritizes executing the nozzle ring wind speed adjustment command, reducing the nozzle ring opening or closing the airflow damper to lower the wind speed inside the mill and reduce the erosion force of coal dust on the lining.

[0067] When the wear mode is identified as overload-induced grinding wear, the system prioritizes executing the coal feed rate adjustment command and the permanent magnet motor power adjustment command. Specifically, grinding wear is characterized by indentations and scratches on the wear surface, a positive correlation between the wear rate and the coal feed rate, and predominantly medium-frequency periodic vibration in the vibration signal. When this mode is identified, the system prioritizes executing the coal feed rate adjustment command and the permanent magnet motor power adjustment command to reduce the coal feed rate per unit time, while simultaneously reducing the motor output power to alleviate the grinding pressure between the grinding roller and the liner.

[0068] When the wear mode is identified as fatigue wear caused by vibration and impact, the system prioritizes executing the grinding disc speed adjustment command and the grinding roller loading force adjustment command. Specifically, fatigue wear is characterized by flaking or cracking on the wear surface, and the impact peaks in the vibration signal appear intermittently with large amplitudes. When this mode is identified, the system prioritizes executing the grinding disc speed adjustment command and the grinding roller loading force adjustment command, reducing the grinding disc speed to decrease the impact frequency, and simultaneously reducing the hydraulic loading force of the grinding roller to decrease the impact amplitude.

[0069] When the wear pattern is identified as localized wear caused by uneven material distribution, a material drop point adjustment command is executed to balance the wear distribution on the liner surface. Specifically, localized wear is characterized by significant differences in wear thickness between different areas, with sensor readings in some areas deviating significantly from the average value. When this pattern is identified, the system executes a material drop point adjustment command, adjusting the position of the coal feeder's drop point or adding a material equalization device to make the raw coal distribution on the grinding disc more uniform, thereby balancing the wear distribution on the liner surface.

[0070] Given the multidimensional and complex nature of wear-causing factors, the multidimensional control commands in this embodiment include at least two of the following: coal feed rate adjustment command, mill disc speed adjustment command, nozzle ring air speed adjustment command, and permanent magnet motor power adjustment command.

[0071] Specifically, the coal feed rate adjustment command controls the coal feeder's feeding rate. When the wear rate exceeds the limit, the coal feed rate per unit time is automatically reduced to decrease the scouring and grinding effect of the material on the liner. Once the wear rate returns to normal, the feed rate is gradually restored to the rated feed rate. The coal feed rate adjustment typically employs a proportional-integral-derivative control algorithm to achieve a smooth transition.

[0072] Grinding disc speed adjustment command: This command controls the speed of the main motor of the coal mill. When vibration and impact are severe or micro-deformation is large, the grinding disc speed is reduced to decrease the relative speed between the grinding rollers and the liner, thereby reducing friction and wear. During the low-wear stage, a high-efficiency speed is maintained to ensure pulverizing output.

[0073] Nozzle ring air velocity adjustment command: This command controls the opening of the nozzle ring or the position of the airflow damper. By adjusting the outlet cross-sectional area of ​​the adjustable nozzle ring, the air velocity inside the mill is controlled within a preset optimized range. Excessive air velocity will exacerbate erosion and wear, while insufficient air velocity will affect pulverized coal conveying and drying efficiency; therefore, a balance must be struck between the two.

[0074] Permanent magnet motor power adjustment command: The permanent magnet motor, as the drive unit of the coal mill, replaces the traditional asynchronous motor and reducer system. Permanent magnet motors have advantages such as high efficiency, wide speed range, and fast response. This command adaptively adjusts the output power of the permanent magnet motor according to the wear condition, avoiding impact wear caused by hard starts and sudden load changes.

[0075] For example, when the load on the coal mill suddenly increases: A single instruction scheme (such as simply increasing the coal feed rate) may actually exacerbate wear due to mismatched wind speed and rotation speed.

[0076] Multi-dimensional coordination solution: The system will simultaneously issue linkage adjustment commands for coal feed rate, rotation speed and wind speed, so that the three can reach the optimal matching state again under the new load, ensuring output while minimizing wear increase.

[0077] S4, after the control command of S3 is executed, returns to step S1 to re-collect wear characteristic data, forming a closed-loop control of "sensing-analysis-control-feedback".

[0078] Specifically, the basic principle of closed-loop control in this embodiment is as follows: the sensor collects wear state data, which is processed and analyzed to generate control commands. After the actuator executes the commands, the wear state changes, and the sensor collects new data again, and so on in a continuous cycle.

[0079] Example 2 Please refer to Figure 2 This embodiment provides a wear prevention system for a medium-speed coal mill, used to implement the wear prevention method for a medium-speed coal mill as described in Embodiment 1, including: An embedded nano-wear monitoring unit 101 is used to perform step S1, including at least one nano-sensor embedded in the coal mill liner for real-time detection of the remaining thickness of the liner, micro-deformation and vibration signal.

[0080] To facilitate signal transmission, the unit also includes a high-temperature resistant wireless transmission module. The wireless transmission module uses either the LoRa protocol or the industrial ZigBee protocol for wireless data transmission. The LoRa protocol operates in the 433 MHz or 868 MHz frequency band and features strong penetration and long transmission distance. The industrial ZigBee protocol operates in the 2.4 GHz frequency band and supports star, tree, and mesh network topologies, making it suitable for multi-sensor network monitoring.

[0081] Meanwhile, the nanosensor and wireless transmission module are packaged into one unit, achieving an overall protection level of IP68. This allows for long-term stable operation in environments below 400 degrees Celsius, with high dust concentrations (up to 1000 mg / m³), and weakly corrosive fumes. The designed service life is no less than 2 years.

[0082] It should be noted that, to avoid signal interference, this unit also includes a signal preprocessing module, which can be deployed on an edge computing gateway or integrated inside the nanosensor. This module amplifies and performs analog-to-digital conversion on the original sensor signal, converting the analog signal into a digital signal; it executes denoising algorithms such as low-pass filtering, band-pass filtering, wavelet denoising, moving average filtering, and outlier removal; and it extracts features from the denoised signal, such as calculating statistical features like the thickness change rate and root mean square value of vibration within a sliding window.

[0083] Wear rate prediction and analysis unit 102 is communicatively connected to embedded nano-wear monitoring unit and is used to execute step S2, calculate the current wear rate and remaining service life based on preset wear rate prediction model, and determine the wear level.

[0084] Specifically, this unit can be deployed on an edge computing gateway, a local server in the factory, or a cloud server. Its core components include: a temporal neural network prediction model (using LSTM, GRU, or Transformer architecture), a model training and update module, and a multi-level wear threshold comparison module.

[0085] This unit receives feature data from the signal preprocessing unit, loads a pre-trained prediction model, calculates the current wear rate and remaining service life, compares it with a preset threshold, and outputs the wear level determination result. The output frequency of the calculation result is configurable, with a typical value of once every 5 minutes.

[0086] The multidimensional control execution unit 103 is communicatively connected to the wear rate prediction and analysis unit and is used to execute step S3, which generates and executes multidimensional control instructions based on the wear level.

[0087] Specifically, the multi-dimensional control execution unit includes: The coal feed rate adjustment module is used to automatically reduce the coal feed rate when the wear rate exceeds the limit. Specifically, this module is connected to the frequency converter or actuator of the coal feeder and sends the coal feed rate setting value through analog signals (4 to 20 mA current or 0 to 10 volts voltage) or digital communication protocols (such as Modbus, Profibus). When the wear rate exceeds the limit, the coal feed rate is automatically reduced; when the wear rate returns to normal, it is gradually restored to the rated coal feed rate.

[0088] The grinding disc speed regulation module is used to reduce the grinding disc speed when subjected to severe vibration and impact. Specifically, this module is connected to the frequency converter driver of the permanent magnet motor. When the vibration and impact are severe or the micro-deformation is large, a deceleration command is sent to the frequency converter driver to reduce the grinding disc speed.

[0089] The nozzle ring airflow adjustment module is used to control the airflow speed inside the mill by adjusting the opening of the adjustable nozzle ring. Specifically, this module is connected to the electric actuator of the adjustable nozzle ring or the drive mechanism of the airflow damper. When erosion wear is detected, the airflow speed inside the mill is controlled within a preset optimization range by adjusting the opening of the adjustable nozzle ring or the opening of the airflow damper.

[0090] The permanent magnet motor power regulation module is used to adaptively adjust the drive power according to the wear condition. Specifically, this module is connected to the permanent magnet motor controller and adaptively adjusts the output power of the permanent magnet motor according to the wear condition to achieve soft start and soft stop, avoiding mechanical shock.

[0091] The feedback control unit 104 is used to execute step S4, feeding back the wear characteristic data collected after adjustment to the wear rate prediction and analysis unit to form closed-loop control. Specifically, this unit is connected to the multi-dimensional adjustment execution unit and the embedded nano-wear monitoring unit respectively, feeding back the new wear characteristic data collected after adjustment to the wear rate prediction and analysis unit to form closed-loop control.

[0092] The feedback control unit is essentially a data flow management module responsible for coordinating the timing of data transmission between various units. Specifically, its core functions include: detecting the completion status of control commands; triggering new data acquisition cycles; ensuring that the predictive model receives the latest data after control; and recording complete historical data of closed-loop control for subsequent analysis.

[0093] To improve wear resistance, the liner in this embodiment uses a gradient composite wear-resistant liner. The gradient composite wear-resistant liner has a three-layer gradient composite structure, consisting of the following layers from the inside out: The base layer, made of heat-resistant alloy steel, can be T-28CrMoCuB alloy or 12Cr1MoV alloy steel, with a thickness of 8mm to 15mm. The function of the base layer is to provide overall structural strength and a connection interface with the coal mill shell.

[0094] The middle layer is a ceramic-rubber-metal fiber composite buffer layer. Specifically, this layer is composed of alumina particles, nitrile rubber, and stainless steel wire mesh. The rubber component absorbs the energy of the particle impact, the steel wire mesh provides tear resistance, and the ceramic particles increase wear resistance. The function of the middle layer is to buffer impact loads and prevent the hard and brittle ceramic surface from shattering under impact.

[0095] The surface layer, a wear-resistant layer, consists of nano-modified zirconia-toughened alumina ceramic or a high-chromium carbide weld overlay. The zirconia-toughened alumina ceramic improves the fracture toughness of the ceramic through the martensitic phase transformation toughening mechanism of zirconia. The chromium carbide (chemical formula Cr7C3) in the high-chromium carbide weld overlay has a hardness of HRC62 or higher. The thickness of the surface layer is typically 6 mm to 12 mm.

[0096] Tests showed that the improved liner wear layer has a hardness 150-260 times that of traditional high manganese steel. Under typical medium-speed coal mill operating conditions (coal ash content >30%, temperature 300-450℃), the liner replacement cycle is extended from 3-6 months to 18-24 months, significantly reducing the frequency of downtime.

[0097] In this embodiment, the nanosensor is embedded between the base layer and the intermediate layer of the gradient composite wear-resistant liner, or embedded in a groove on the back of the base layer. Thus, the liner serves as both a wear-protecting component and a wear-sensing carrier.

[0098] It should be noted that, to facilitate the adjustment of primary air parameters, this system also includes an adjustable rotating nozzle ring, the outlet cross-sectional area of ​​which can be adjusted in real time; the nozzle ring wind speed adjustment module adjusts the opening of the adjustable rotating nozzle ring according to the control command output by the wear rate prediction and analysis unit. The structure and working principle of the nozzle ring are existing technologies and will not be described in detail here.

[0099] The closed-loop anti-wear system provided in this embodiment operates according to the following timing sequence: The first step involves using nanosensors to collect raw signals such as the remaining thickness of the liner, minute deformations, and vibration amplitude at a millisecond-level sampling frequency.

[0100] The second step involves the edge computing unit preprocessing the original signal to reduce noise and extract features.

[0101] The third step is to input the extracted feature data into the temporal neural network prediction model to calculate the current wear rate and remaining lifespan.

[0102] The fourth step is to compare the wear rate with preset multi-level thresholds to determine the wear level.

[0103] The fifth step involves the decision engine generating multi-dimensional control commands based on the wear level and the identified wear causes, specifying the target values ​​for adjusting the coal feed rate, mill speed, nozzle ring opening, and permanent magnet motor power.

[0104] The sixth step is for the implementing agency to execute the control instructions.

[0105] In the seventh step, the sensor continues to collect new wear data.

[0106] Step 8: Go back to step 3 and recalculate the wear rate.

[0107] The single closed-loop response time is controlled within seconds to ensure that the system can respond promptly to rapid changes in wear conditions.

[0108] In this embodiment, when the system detects accelerated wear, it actively reduces the wear rate by adjusting operating parameters to prevent further wear deterioration, thus effectively suppressing the wear rate. Simultaneously, as closed-loop operation continues, the system accumulates a large amount of historical data, allowing the predictive model to be continuously updated and optimized, gradually improving the accuracy of control decisions and continuously optimizing control precision.

[0109] This embodiment uses the ZGM type medium-speed coal mill as an example to compare the parameters before and after optimization. The specific parameters are shown in the table below:

[0110] As shown in the table above, by improving the liner material and monitoring the wear status in real time based on wear characteristic data collected by sensors, and then analyzing the wear level and wear mode of the coal mill using a time-series prediction model, and employing differentiated strategies for multi-dimensional closed-loop control based on different causes, the wear process of the optimized coal mill liner is actively intervened in. This results in a 40%-50% reduction in the wear rate of the optimized coal mill liner, a 60%-80% extension in the liner replacement cycle, a reduction of more than 70% in unplanned downtime, and an 8%-12% reduction in coal mill power consumption.

[0111] The above description is a specific implementation of the embodiments of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for preventing wear in a medium-speed coal mill, characterized in that, include: S1, by using nanosensors embedded in the coal mill liner, wear characteristic data of key vulnerable areas of the coal mill are collected in real time; the wear characteristic data includes the remaining thickness of the liner, micro deformation and vibration signal; S2, after preprocessing and denoising the wear feature data, input the preprocessed wear feature data and working condition time series data into the preset wear rate prediction model, calculate the current wear rate and remaining service life, and compare it with the preset multi-level wear threshold to determine the current wear level; S3, based on the determined wear level, generates multi-dimensional control commands and sends them to the corresponding actuators for execution; The multi-dimensional control commands include at least two of the following: coal feed rate adjustment command, mill disc speed adjustment command, nozzle ring air speed adjustment command, and permanent magnet motor power adjustment command. S4, after the control command of S3 is executed, returns to step S1 to re-collect wear characteristic data, forming a closed-loop control of "sensing-analysis-control-feedback".

2. The wear prevention method for medium-speed coal mills according to claim 1, characterized in that, In S2, the wear rate prediction model is a prediction model constructed based on a time-series neural network; The preprocessing noise reduction includes at least one of the following methods: low-pass filtering, band-pass filtering, wavelet denoising, moving average filtering, and outlier removal; The input features of the time-series neural network include: historical liner thickness time-series data, historical micro-deformation time-series data, historical vibration signal time-series data, and auxiliary operating parameters; the auxiliary operating parameters include material particle size, feed flow rate, equipment operating speed, and running time. The output of the time-series neural network includes: short-term wear rate prediction, remaining service life prediction, and critical wear time prediction.

3. The wear prevention method for medium-speed coal mills according to claim 1, characterized in that, The multi-level wear threshold determination mechanism in S2 includes: When the current wear rate is lower than the preset normal operating average, it is determined to be a light wear level, and the system maintains the current operating condition. When the current wear rate is higher than the normal operating average but lower than the warning threshold, it is determined to be a moderate wear level and the system issues a warning signal. When the current wear rate exceeds the warning threshold, it is determined to be a severe wear level, and the system automatically triggers multi-dimensional control commands; The warning thresholds include: a remaining thickness of less than 50% of the initial thickness is a warning threshold, and a remaining thickness of less than 30% of the initial thickness is a shutdown recommendation threshold.

4. The wear prevention method for medium-speed coal mills according to claim 1, characterized in that, The generation of the multi-dimensional control command in S3 adopts a decision logic based on wear-cause identification: When the wear mode is identified as scouring wear caused by excessive wind speed, the nozzle ring wind speed adjustment command is executed first. When the wear mode is identified as overload-induced roller wear, the coal feed rate adjustment command and the permanent magnet motor power adjustment command are executed first. When the wear mode is identified as fatigue wear caused by vibration and impact, the grinding disc speed adjustment command and the grinding roller loading force adjustment command are executed first. When the wear pattern is identified as localized wear caused by uneven material distribution, the material drop point adjustment command is executed to balance the wear distribution on the liner surface.

5. The wear prevention method for medium-speed coal mills according to claim 1, characterized in that, The method also includes a state maintenance scheduling step: When the predicted remaining service life in S2 is lower than the preset replacement threshold, a replacement work order is automatically generated and pushed. When the wear feature data collected in S1 identifies local abnormal wear points, the fault points are automatically located and marked, and targeted maintenance suggestions are generated.

6. A wear prevention system for a medium-speed coal mill, used to implement the wear prevention method for a medium-speed coal mill as described in any one of claims 1-5, characterized in that, include: An embedded nano wear monitoring unit (101) is used to perform step S1, including at least one nano sensor embedded in the coal mill liner for real-time detection of the remaining thickness of the liner, micro deformation and vibration signal; Wear rate prediction and analysis unit (102) is communicatively connected to the embedded nano wear monitoring unit and is used to execute step S2, calculate the current wear rate and remaining service life based on the preset wear rate prediction model, and determine the wear level; The multidimensional control execution unit (103) is communicatively connected to the wear rate prediction and analysis unit and is used to execute step S3, generating and executing multidimensional control instructions according to the wear level; The feedback control unit (104) is used to execute step S4, feeding back the wear characteristic data collected after adjustment to the wear rate prediction and analysis unit to form a closed-loop control.

7. The anti-wear system for medium-speed coal mills according to claim 6, characterized in that, The system also includes a gradient composite wear-resistant liner, which has a three-layer gradient composite structure, consisting of the following layers from the inside out: The base layer is made of heat-resistant alloy steel and has a thickness of 8mm to 15mm; The middle layer is a ceramic-rubber-metal fiber composite buffer layer; The surface layer is a wear-resistant layer, composed of nano-modified zirconia toughened alumina ceramic or high-chromium carbide weld overlay, with a hardness ≥ HRC62. The nanosensor is embedded between the base layer and the intermediate layer of the gradient composite wear-resistant liner, or embedded in a groove on the back of the base layer.

8. The anti-wear system for medium-speed coal mills according to claim 6, characterized in that, The multidimensional control execution unit includes: The coal feed rate adjustment module is used to automatically reduce the coal feed rate when the wear rate exceeds the limit; The grinding disc speed adjustment module is used to reduce the grinding disc speed when there is severe vibration and impact. The nozzle ring air speed adjustment module is used to control the air speed inside the mill by adjusting the opening of the adjustable nozzle ring; The permanent magnet motor power adjustment module is used to adaptively adjust the drive power according to the wear condition.

9. The anti-wear system for medium-speed coal mills according to claim 8, characterized in that, The system also includes an adjustable rotating nozzle ring, the outlet cross-sectional area of ​​which can be adjusted in real time; the nozzle ring wind speed adjustment module adjusts the opening of the adjustable rotating nozzle ring according to the control command output by the wear rate prediction and analysis unit.

10. The anti-wear system for medium-speed coal mills according to claim 6, characterized in that, The embedded nano wear monitoring unit also includes a high-temperature resistant wireless transmission module, which uses LoRa or industrial ZigBee protocols for wireless data transmission; the nano sensor and wireless transmission module operate in environments below 400°C, with high dust and weakly corrosive fumes, and have a designed lifespan of ≥2 years.

Images