A dynamic coal powder metering system and device

Abstract

The present application belongs to the technical field of powder flow metering, and provides a coal powder dynamic metering system and device, which comprises the following steps: collecting real-time data of the weight of a buffer bin, the derivative of the gravity change rate, the opening degree of the previous stage of discharging, and the flow variance of the metering scale, determining the stability through multiple conditions, and superimposing time length verification to open a stable calibration window without material supplement, extracting the real delivery capacity of the buffer bin by using sliding average filtering within the calibration window, synchronously integrating to obtain the effective cumulative capacity of the metering scale and calculating the initial calibration coefficient, setting a static interval and a dynamic trend double verification mechanism, identifying abnormal calibration data in combination with a historical correction coefficient prediction model, weighting the compliance coefficient to iteratively update the correction coefficient, and finally issuing the deviation compensation after dead zone verification to improve the anti-interference ability and fault tolerance of coal powder metering calibration under harsh working conditions, avoid parameter mutation out of control, and effectively guarantee the coal powder metering precision and system operation stability.

Description

Technical Field

[0001] This invention belongs to the field of powder flow metering technology, specifically a dynamic coal powder metering system and device. Background Technology

[0002] High-precision and stable metering of pulverized coal is a core prerequisite for ensuring gasification reaction efficiency, product conversion rate and safe and stable operation of the equipment. Metering accuracy directly determines the accuracy of process proportions and the stability of production conditions.

[0003] Currently, industrial applications mainly use continuous metering equipment such as rotor scales, Coriolis scales, and quantitative feeders to achieve dynamic metering of pulverized coal. However, pulverized coal has a small particle size and is easily fluidized. The closed conveying process is a complex gas-solid two-phase flow state, which is highly susceptible to the influence of material humidity, air pressure pulsation, and fluidization air volume fluctuations. This frequently leads to abnormal operating conditions such as material slugging, bridging, and air leakage, resulting in instantaneous force disorder of the metering scale, large flow fluctuations, and significant dynamic metering errors.

[0004] Flow metering equipment operating under high loads for extended periods will experience a continuous decline in accuracy due to mechanical wear, material property drift, and sensor aging. Traditional calibration methods such as hanging codes, chain codes, and physical calibration are all static calibration methods that require shutdown. These methods not only interrupt continuous production and incur high maintenance costs, but also fail to simulate the dynamic operating conditions of real gas-solid two-phase flow, resulting in large calibration deviations and an inability to correct real-time operating errors of the equipment.

[0005] Existing technologies, in order to address the shortcomings of shutdown calibration, generally utilize the comparison between the weight loss of the buffer bin and the cumulative weight of the weighing scale to achieve simple online calibration. However, this approach suffers from three fatal technical defects under harsh industrial conditions: First, the bin weight signal is superimposed with high-frequency interference from multiple sources, such as mechanical vibration, air pressure pulsation, and material replenishment impact, making it impossible to extract the true and effective weight loss, resulting in low benchmark accuracy. Second, there is a lack of accurate steady-state condition identification and boundary judgment mechanisms, making it easy to collect data during periods of pseudo-steady-state, fluctuation, and material replenishment interference, causing serious distortion of the calibration benchmark. Third, there is no confidence verification of calibration results or historical weighted smoothing mechanisms, and most methods adopt a direct overlay correction method. A single abnormal disturbance can cause a sudden change in the calibration coefficient, leading to divergence in metrological control, drastic fluctuations in operating conditions, and significant production safety hazards.

[0006] Therefore, there is an urgent need in the field for a dynamic metering and online self-calibration technology for pulverized coal that can accurately identify the true steady-state calibration window, filter out multi-source operating condition interference, and possess an anti-divergence dynamic correction mechanism. To this end, the present invention provides a dynamic metering system and device for pulverized coal. Summary of the Invention

[0007] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0008] The technical solution adopted by this invention to solve its technical problem is: One objective of this invention is to provide a dynamic coal powder metering system, comprising: Data acquisition unit: collects instantaneous weight data of the weighing buffer bin, derivative characteristics of the rate of change of gravity, physical opening signal of the front feeding mechanism, and instantaneous flow variance of the continuous weighing scale; Steady-state boundary determination unit: When it is detected that the front feeding mechanism is completely closed, the instantaneous silo weight data is within the safe fluidization range, the absolute value of the derivative characteristic of the gravity change rate is less than the silo weight disturbance threshold, and the instantaneous flow variance is less than the feeding fluctuation threshold, a candidate steady state is generated. The duration of the candidate steady state is compared with the set time. If the duration exceeds the set time, the no-feeding steady calibration window is opened. Synchronous integration unit: Within the no-replenishment stable calibration window, the reference actual conveying volume of the weighing buffer bin is calculated using a moving average filtering algorithm, and the effective cumulative volume of the continuous weighing scale within the no-replenishment stable calibration window is obtained through synchronous integration. Dynamic correction unit: Based on the reference actual conveying volume and effective cumulative volume, it calculates the initial calibration coefficient for the current calibration cycle and performs the first-level static interval verification. It uses historical correction coefficients to build a prediction model and outputs the predicted correction coefficient. It then uses the predicted correction coefficient to perform the second-level dynamic trend verification on the initial calibration coefficient. After both the first-level static interval verification and the second-level dynamic trend verification pass, it performs a weighted iterative calculation of the initial calibration coefficient and the correction coefficient of the previous calibration cycle to generate the correction coefficient for the current calibration cycle. After the dead zone verification of the correction coefficient for the current calibration cycle passes, it is sent to the continuous weighing scale for deviation compensation.

[0009] As a further improvement of the present invention, the specific process of collecting the instantaneous weight data of the weighing buffer bin, the derivative characteristics of the rate of change of gravity, the physical opening signal of the front-stage feeding mechanism, and the instantaneous flow variance of the continuous weighing scale is as follows: The gross weight analog electrical signal of the weighing buffer bin is collected by the weighing sensor, and the gross weight analog electrical signal is transmitted to the controller for analog-to-digital conversion to obtain the instantaneous bin weight data. Based on the instantaneous weight data, the controller uses a differential algorithm in the discrete time domain to calculate the weight change between adjacent sampling periods, obtains the initial gravity change rate, and uses a low-pass filter to smooth the initial gravity change rate to obtain the derivative characteristics of the gravity change rate. The system acquires the electrical signals fed back from the flow valve or pneumatic gate valve and the valve positioner of the non-standard feeding mechanism, converts the feedback electrical signals into a 0-100% opening percentage value, and acquires the physical opening signal. When digital control is used, the physical opening signal is acquired directly. The instantaneous linear load of pulverized coal is obtained through the weighing bridge inside the continuous weighing scale, and the instantaneous operating speed of the weighing scale is obtained through the speed sensor. The instantaneous linear load and the instantaneous operating speed are multiplied to obtain the instantaneous feed flow rate. A sliding time window containing a preset number of sampling points is set in the controller, and the mathematical variance of the instantaneous feed flow rate within the sliding time window is calculated in real time. The calculation result is used as the instantaneous flow rate variance.

[0010] As a further improvement of the present invention, the specific process of opening the no-feed stable calibration window is as follows: Determine if the physical opening signal of the front-stage feeding mechanism is zero; if it is, then condition one is met. Determine if the instantaneous bin weight data is within the safe fluidization range; if it is, then condition two is met. Determine if the absolute value of the derivative characteristic of the gravity change rate is less than the bin weight disturbance threshold; if it is less than the bin weight disturbance threshold, then condition three is met. Determine if the instantaneous flow variance is less than the feeding fluctuation threshold; if it is less than the feeding fluctuation threshold, then condition four is met. When all four judgment conditions are met, the system is in a candidate stable operating state. The timer inside the controller is started to keep track of the duration. If any of the four judgment conditions is no longer met during the timer, the timer is reset to zero. If the duration exceeds the set time, the no-feeding stable calibration window is opened.

[0011] As a further improvement of the present invention, the specific process of obtaining the effective cumulative amount of the continuous weighing scale within the stable calibration window without material replenishment through synchronous integration is as follows: At the start of the no-replenishment stabilization calibration window, acquire the first instantaneous bin weight data sequence containing multiple consecutive sampling points; at the end of the no-replenishment stabilization calibration window, acquire the second instantaneous bin weight data sequence containing multiple consecutive sampling points. The first and second instantaneous bin weight data sequences are processed by a moving average filtering algorithm to obtain the initial effective bin weight and the final effective bin weight. The initial effective bin weight is subtracted from the final effective bin weight to obtain the actual weight reduction value of the weighing buffer bin within the no-replenishment stable calibration window. The actual weight reduction value is used as a reference for the actual conveying volume. At the start of the no-feeding stable calibration window, the continuous weighing scale records the initial cumulative value. During the duration of the no-feeding stable calibration window, the instantaneous feed flow rate output by the continuous weighing scale is continuously integrally calculated in the time dimension. When the no-feeding stable calibration window ends, the calculation is stopped and the ending cumulative value is recorded. The ending cumulative value is subtracted from the initial cumulative value to obtain the effective cumulative amount.

[0012] As a further improvement of the present invention, the specific process of calculating the initial calibration coefficient of the current calibration period and performing the first-level static interval verification is as follows: Obtain the reference actual delivery volume and the effective cumulative volume, and divide the reference actual delivery volume by the effective cumulative volume to obtain the initial calibration coefficient for the current calibration cycle; The calibration coefficient confidence interval is set in the controller. If the initial calibration coefficient falls within the calibration coefficient confidence interval, the first-level static interval verification is considered to have passed.

[0013] As a further improvement of the present invention, the specific process of constructing a prediction model and outputting prediction correction coefficients using historical correction coefficients is as follows: The controller extracts the actual issuance time points and corresponding historical correction coefficients for N consecutive historical calibration cycles. It constructs a time variable sequence with the actual issuance time points as the independent variable sequence and a correction coefficient sequence with the corresponding historical correction coefficients as the dependent variable sequence. Using the Pearson correlation coefficient algorithm, the correlation coefficient between the independent variable series and the dependent variable series is calculated. If the absolute value of the correlation coefficient is greater than or equal to the correlation threshold, the least squares method is used to linearly fit the independent variable series and the dependent variable series to construct a prediction model. If the absolute value of the correlation coefficient is less than the correlation threshold, a nonlinear fitting algorithm is used to perform nonlinear mapping on the independent variable sequence and the dependent variable sequence to construct a prediction model. Obtain the target time point at which the current calibration cycle is expected to be sent to the continuous weighing scale, input the target time point into the prediction model, and output the prediction correction coefficient.

[0014] As a further improvement of the present invention, the specific process of performing a second-level dynamic trend verification of the initial calibration coefficient using the prediction correction coefficient is as follows: Calculate the absolute value of the difference between the predicted correction coefficient and the initial calibration coefficient that passed the first-level static interval verification, and compare the absolute value of the difference with the dynamic confidence threshold; If the absolute value of the difference is greater than or equal to the dynamic confidence threshold, the second-level dynamic trend verification fails. If the absolute value of the difference is less than the dynamic confidence threshold, then the second-level dynamic trend verification passes.

[0015] As a further improvement of the present invention, the specific process of generating the correction coefficient for the current calibration period by weighted iterative summation of the initial calibration coefficient and the correction coefficient of the previous calibration period is as follows: A smoothing weight coefficient is introduced, with a value between 0 and 1. The initial calibration coefficient, which has passed the first-level static interval verification and the second-level dynamic trend verification, is multiplied by the smoothing weight coefficient to obtain the current correction component. The correction coefficient of the previous period is multiplied by the value obtained by subtracting the smoothing weight coefficient from 1 to calculate the historical retained component. The current correction component and the historical retained component are algebraically added to generate the correction coefficient of the current calibration period.

[0016] As a further improvement of the present invention, the specific process of sending the dead-zone verification of the current calibration cycle correction coefficient to the continuous weighing scale for deviation compensation is as follows: Calculate the absolute value of the change between the correction coefficient of the current calibration cycle and the actual correction coefficient issued in the previous calibration cycle, and compare the absolute value of the change with the adjustment dead zone threshold. If the absolute value of the change is greater than or equal to the dead zone threshold, the dead zone check is passed. The controller then sends the current calibration cycle correction coefficient to the underlying controller of the continuous weighing scale in real time via the industrial communication bus. The underlying control logic of the continuous weighing scale uses the current calibration cycle correction coefficient as a multiplication factor in the flow calculation formula or the target set value of the PID controller to adjust the drive motor speed and compensate for static errors in real time.

[0017] The second objective of this invention is to provide a dynamic coal powder metering device, comprising: The weighing buffer chamber is equipped with a weighing sensor for real-time acquisition of chamber weight data. The pre-feeding mechanism is located at the inlet end of the weighing buffer bin; A constant flow feeder is installed at the bottom discharge port of the weighing buffer silo; Continuous metering scales receive the discharge from constant flow feeders and are used for continuous dynamic metering of pulverized coal. The controller is communicatively connected to the upstream feeding mechanism, the weighing sensor, and the continuous metering scale; the controller is applied to the control logic of each unit in a dynamic coal powder metering system.

[0018] The beneficial effects of this invention are as follows: This invention simultaneously collects four core state variables: instantaneous bin weight, derivative of bin weight-gravity change rate, front-stage discharge opening, and metering scale flow variance. It uses a combination of four conditions—discharge closure, bin position range, disturbance threshold, and flow stability—to determine the true steady state, and incorporates a duration filtering mechanism to distinguish between true and false steady states. This effectively avoids calibration under adverse conditions such as feed impact, bin collapse, gas-solid two-phase flow pressure pulsation, and drastic flow fluctuations. The calibration window is only opened under true, stable, no-feed conditions, improving the effectiveness and reliability of the calibration benchmark and overcoming the shortcomings of traditional online calibration methods, such as poor condition identification and susceptibility to benchmark distortion.

[0019] This invention abandons the traditional single numerical verification mode and innovatively sets up a two-level linkage verification mechanism. The first level completes static numerical screening through a fixed confidence interval, eliminating abnormal calibration coefficients that exceed the range. The second level, based on historical correction coefficient time-series data, constructs a prediction model through correlation discrimination and adaptive fitting algorithms, and uses the predicted correction coefficients to complete dynamic trend verification. It identifies abnormal data from both the magnitude and trend of changes, effectively avoiding erroneous calibration results caused by instantaneous sensor noise and extreme operating condition disturbances, and preventing measurement deviations caused by single abnormal data.

[0020] This invention employs a weighted iterative update strategy using historical coefficients and current effective coefficients to avoid abrupt changes in correction coefficients, ensuring smooth transitions in measurement parameters and stable, unfluctuating operating conditions. Simultaneously, a final dead-zone verification mechanism is added, executing updates only when coefficient changes exceed a threshold, effectively preventing frequent parameter refreshes and equipment adjustments caused by minor numerical disturbances. This addresses the industry pain points of traditional online calibration, which are prone to parameter jumps, control divergence, operating condition oscillations, and even production accidents. Attached Figure Description

[0021] The invention will now be further described with reference to the accompanying drawings.

[0022] Figure 1 This is a system module diagram of a dynamic coal powder metering system according to the present invention; Figure 2 This is a schematic diagram of a dynamic coal powder metering device according to the present invention; Figure 3 This is a comparison chart of simulation data showing the dynamic evolution of the correction coefficient, the two-level verification, and the dead zone control effect in the embodiments of the present invention.

[0023] Figure 2 The meanings of each symbol in the text are as follows: 1. Foreground raw material silo; 2. Pneumatic gate valve and non-standard discharge system; 3. Discharge chute or horizontal screw conveyor with flow valve; 4. Dust collector; 5. Weighing buffer silo; 5.1 Weighing sensor; 6. Dual-drive flow stabilizer; 7. Continuous weighing scale; 8. Screw conveyor; 9. Airlock valve; 10. Pneumatic gate valve; 11. Kneader; 12. Gas phase balance pipe; 13. Controller. Detailed Implementation

[0024] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments. Example

[0025] like Figure 1 and Figure 2As shown, the underlying physical hardware architecture upon which this invention is based, from top to bottom, includes: a front-stage raw material silo 1 connected to the inlet end of a weighing buffer silo 5 via a discharge chute with a flow valve and a pneumatic gate valve and a non-standard discharge 2, or a horizontal screw conveyor 3; a dust collector 4 connected to the top of the weighing buffer silo 5, and a dual-drive flow stabilizer 6 (i.e., a flow stabilizer feeding device) connected to its bottom outlet; a continuous metering scale 7 receiving the discharge from the dual-drive flow stabilizer 6 for continuous dynamic metering; and metered coal powder passing sequentially through a screw conveyor 8, an airlock valve 9, and a pneumatic gate valve 10, finally entering a kneader 11, a gas phase balance pipe 12, and an inlet pipe to maintain gas phase pressure balance, and a controller 13. Based on the above hardware architecture, the coal powder dynamic metering system described in this invention includes: The data acquisition unit is used to collect in real time the instantaneous weight data of the weighing buffer bin, the derivative characteristics of the rate of change of gravity, the physical opening signal of the front feeding mechanism, and the instantaneous flow variance of the continuous weighing scale. In this embodiment, the data acquisition unit consists of a weighing sensor 5.1 installed on the side wall or bottom of the weighing buffer chamber 5, a valve position feedback end of the front-stage feeding mechanism (i.e., a feeding chute or horizontal screw conveyor 3 with a flow valve and a pneumatic gate valve and a non-standard feeding 2), a speed measuring and weighing sensor inside the continuous weighing scale 7, and a controller 13 that is communicatively connected to the above hardware.

[0026] In some embodiments, the specific process of real-time acquisition of instantaneous weight data of the weighing buffer bin in the data acquisition unit is as follows: The weighing sensor 5.1 continuously collects the gross weight analog electrical signal of the weighing buffer 5 at a preset sampling frequency; the gross weight analog electrical signal is transmitted to the controller 13 for analog-to-digital conversion, and the fixed tare weight of the weighing buffer 5 is deducted; at the same time, combined with the micro-negative pressure state parameters maintained by the gas phase balance pipe 12 and the dust collector 4, the converted weight data is compensated for gas phase pressure to eliminate false weighing fluctuations caused by airflow and pressure pulsation in the system, thereby obtaining high-precision instantaneous hopper weight data.

[0027] In some embodiments, the specific process of real-time acquisition of the derivative characteristics of the rate of change of gravity in the data acquisition unit is as follows: Based on the instantaneous sludge weight data, the controller 13 uses a differential algorithm in the discrete time domain to calculate the change in sludge weight between adjacent sampling periods, and obtains the initial gravity change rate (i.e., the first derivative of sludge weight with respect to time). In order to eliminate the transient mechanical impact noise generated during pulverized coal feeding and the random disturbance of the gas-solid two-phase flow, a low-pass filter is used to smooth the initial gravity change rate, and a smoothed first derivative value that can truly reflect the pulverized coal outflow rate is obtained. The smoothed first derivative value is used as the derivative feature of the gravity change rate.

[0028] In some embodiments, the specific process of real-time acquisition of the physical opening signal of the front-stage feeding mechanism in the data acquisition unit is as follows: The system acquires the electrical signals fed back from the flow valve or pneumatic gate valve and the valve positioner of the non-standard feeder 2 in the pre-feeding mechanism; when analog control is used, the feedback current or voltage signal is converted into an opening percentage value of 0-100%; when digital control is used, the closed / open state of the limit switch is directly read to obtain the accurate physical opening signal. The physical opening signal is used to accurately characterize whether the pre-raw material bin 1 is replenishing the weighing buffer bin 5.

[0029] In some embodiments, the specific process of real-time acquisition of the instantaneous flow variance of the continuous weighing scale in the data acquisition unit is as follows: The instantaneous linear load of pulverized coal is obtained through the weighing bridge inside the continuous weighing scale 7, and the instantaneous operating speed of the weighing scale is obtained simultaneously through the speed sensor. The two are multiplied to obtain the instantaneous feed flow rate. A sliding time window containing a preset number of sampling points is set in the controller 13, and the mathematical variance of the instantaneous feed flow rate within the sliding time window is calculated in real time. The calculation result is used as the instantaneous flow rate variance. The instantaneous flow rate variance is used to quantitatively evaluate the actual feeding stability of pulverized coal after the dual-drive flow stabilizer 6 stirs and stabilizes the flow.

[0030] The steady-state boundary determination unit is used to determine the state of the data acquired by the data acquisition unit. When it is simultaneously detected that the front-end feeding mechanism is completely closed, the instantaneous hopper weight data is within the preset safe fluidization range, the absolute value of the derivative characteristic of the gravity change rate is less than the preset hopper weight disturbance threshold, and the instantaneous flow variance is less than the preset feeding fluctuation threshold, a candidate steady state is generated, and the duration of the candidate steady state is compared with the set time. If the duration exceeds the set time, the no-feeding steady calibration window is triggered and opened. In this embodiment, the steady-state boundary determination unit consists of a microprocessor (such as a PLC or DSP chip) and a memory module inside the controller 13. It has a state discrimination and logic control algorithm program embedded inside, which is used to perform logical operations and state comparisons on the received data.

[0031] In some embodiments, the specific process of state discrimination of the data acquired by the data acquisition unit in the steady-state boundary determination unit is as follows: The controller 13 reads the various data transmitted by the data acquisition unit in real time and performs multi-condition concurrent logic judgments: Judgment condition 1: Determine whether the physical opening signal of the front feeding mechanism is zero. If the physical opening signal is zero, then judgment condition 1 is satisfied. Judgment condition 1 is to confirm that the front feeding mechanism is completely closed and to ensure that no external materials are mixed into the weighing buffer chamber 5. This is the physical premise for calibration using natural weightlessness. Judgment Condition Two: Determine whether the instantaneous silo weight data is within the preset safe fluidization range. If the instantaneous silo weight data is within the preset safe fluidization range, then Judgment Condition Two is satisfied. Satisfying Judgment Condition Two can effectively avoid the risk of pulverized coal rushing caused by excessively high silo levels, and the risk of air leakage or material shortage caused by excessively low silo levels, ensuring that the pulverized coal is in a stable fluidization state within the silo. For example, the safe fluidization range can be set to 30% to 80% of the full silo capacity. Judgment Condition 3: Determine whether the absolute value of the derivative characteristic of the rate of change of gravity is less than the preset weight disturbance threshold. If the absolute value of the derivative characteristic of the rate of change of gravity is less than the preset weight disturbance threshold, then Judgment Condition 3 is satisfied. The weight disturbance threshold is set according to the normal feeding characteristics of the weighing buffer bin 5 and is used to characterize that no sudden physical disturbances such as material collapse, bridging or crust shedding have occurred in the bin, so as to ensure that the weight loss curve is smooth. Judgment condition four: Determine whether the instantaneous flow variance is less than the preset feeding fluctuation threshold. If the instantaneous flow variance is less than the preset feeding fluctuation threshold, then judgment condition four is satisfied. The feeding fluctuation threshold is used to confirm that the mixing layer and the flow stabilizing layer of the dual-drive flow stabilizer 6 have good homogenization effect, and that the feeding of the continuous metering scale 7 is not subjected to transient impact of the gas-solid two-phase flow and is in a stable feeding state.

[0032] When all four of the above conditions are met, the determination system is currently in a stable operating state.

[0033] In some embodiments, the specific process of triggering and opening the unreplenished stability calibration window in the steady-state boundary determination unit is as follows: When the controller 13 determines that the system has entered the candidate stable operating state, it starts an internal timer to accumulate the duration. During the timing process, if any of the four determination conditions mentioned above are no longer met, the timer is reset and the system waits again. The accumulated duration is compared in real time with a preset time, which is used to filter out transient pseudo-steady-state phenomena (such as brief flow stagnation or sensor transient noise) in the gas-solid two-phase flow system. If the duration exceeds the preset time, the system is confirmed to have entered a true physical steady state. At this time, the controller 13 generates a trigger command to officially open the feedless stable calibration window, providing a reliable time interval for subsequent high-precision integration and calibration.

[0034] The synchronous integration unit is used to extract the effective weight loss of the weighing buffer as a reference actual conveying amount within the no-replenishment stable calibration window using a moving average filtering algorithm, and synchronously integrates to obtain the effective cumulative amount of the continuous weighing scale within the no-replenishment stable calibration window. In this embodiment, the synchronization integration unit consists of a digital signal processing module and a high-precision clock synchronization module inside the controller 13, which achieves high-speed data interaction with the weighing sensor 5.1 of the weighing buffer chamber 5 and the bottom integrator of the continuous weighing scale 7.

[0035] In some embodiments, within the synchronous integration unit, the specific process of extracting the effective weight loss of the weighing buffer as a reference actual conveying volume using a moving average filtering algorithm within the unreplenished stable calibration window is as follows: At the start of the no-replenishment stable calibration window, a first segment of instantaneous silo weight data sequence containing multiple continuous sampling points is acquired; at the end of the no-replenishment stable calibration window, a second segment of instantaneous silo weight data sequence containing multiple continuous sampling points is acquired; for the first and second segments of instantaneous silo weight data sequences, a moving average filtering algorithm is used respectively, that is, the arithmetic mean of the weight data of multiple sampling points within a preset data window length is calculated to filter out the high-frequency mechanical vibration interference of the motor during the operation of the dual-drive flow stabilizer 6 and the high-frequency noise caused by the gas phase pressure pulsation inside the system, thereby calculating the smoothed initial effective silo weight and the final effective silo weight respectively; the initial effective silo weight is subtracted from the final effective silo weight to calculate the actual weight reduction value of the weighing buffer silo 5 within the no-replenishment stable calibration window period, and the actual weight reduction value is used as the effective weight loss, and the effective weight loss is used as the reference actual conveying volume for subsequent calibration.

[0036] In some embodiments, the specific process of synchronous integration in the synchronous integration unit to obtain the effective cumulative amount of the continuous weighing scale within the stable calibration window without material replenishment is as follows: According to the high-precision clock synchronization module inside the controller 13, at the start of the no-feeding stable calibration window, the underlying integrator of the continuous weighing scale 7 is synchronously triggered to record the initial cumulative value (or the integrator is cleared and accumulation begins); during the duration of the no-feeding stable calibration window, the instantaneous feed flow rate output by the continuous weighing scale 7 is subjected to continuous definite integral calculation in the time dimension; when the no-feeding stable calibration window ends, the integration calculation is synchronously stopped and the end cumulative value is recorded; the end cumulative value is subtracted from the initial cumulative value (or the total integral value after clearing is directly read) to obtain the total cumulative integral value within the no-feeding stable calibration window, and the total cumulative integral value is used as the effective cumulative amount. Through timestamp alignment, it is ensured that the effective cumulative amount and the effective weight loss correspond completely in the physical time scale, eliminating time asynchronous errors caused by communication delays or mechanical response lags.

[0037] The dynamic correction unit is used to calculate the initial calibration coefficient for the current calibration cycle and perform the first-level static interval verification; it uses historical correction coefficients to build a prediction model to output predicted correction coefficients, and uses the predicted correction coefficients to perform the second-level dynamic trend verification on the initial calibration coefficients; after both levels of verification pass, the initial calibration coefficients and the correction coefficients of the previous cycle are weighted and iteratively calculated to generate the correction coefficients for the current cycle; finally, after performing adaptive dead-zone verification on the correction coefficients for the current cycle, the process is triggered to send the data to the continuous weighing scale for deviation compensation.

[0038] In this embodiment, the dynamic correction unit consists of a closed-loop control calculation module and a communication sending module inside the controller 13. Its output terminal is connected to the frequency converter driver or the underlying feeding control board of the continuous weighing scale 7, serving as the final execution center of the entire online self-calibration system.

[0039] In some embodiments, the specific process of calculating the initial calibration coefficient for the current calibration period and performing the first-level static interval verification in the dynamic correction unit is as follows: Obtain the reference actual conveying amount (i.e., effective weight loss) and the effective cumulative amount output by the synchronous integration unit; divide the reference actual conveying amount by the effective cumulative amount and calculate the ratio between the two; the ratio directly reflects the degree of deviation between the actual measured value of the continuous weighing scale 7 and the actual physical weight loss of the weighing buffer chamber 5 within the current calibration cycle, and use the ratio as the initial calibration coefficient for the current calibration cycle.

[0040] The specific process for performing the first-level static interval verification on the initial calibration coefficients is as follows: A reasonable calibration coefficient confidence interval (e.g., 0.95 to 1.05) is pre-set in the controller 13. The calibration coefficient confidence interval is used to define the maximum allowable physical drift range for a single calibration. The calculated initial calibration coefficient is compared with the calibration coefficient confidence interval. If the initial calibration coefficient falls within the calibration coefficient confidence interval, it is determined that it has passed the first-level static interval verification and the initial calibration coefficient is retained. If it exceeds the calibration coefficient confidence interval, it is determined that the initial calibration coefficient is unreliable due to extreme abnormal interference and is directly discarded while keeping the original system parameters unchanged.

[0041] In some embodiments, the specific process of constructing a prediction model and outputting prediction correction coefficients using historical correction coefficients in the dynamic correction unit is as follows: The controller 13 extracts the actual issuance time points and corresponding historical correction coefficients for N consecutive historical calibration cycles, constructs a time variable sequence using the actual issuance time points as the independent variable sequence, and constructs a correction coefficient sequence using the corresponding historical correction coefficients as the dependent variable sequence. The correlation coefficient between the independent variable sequence and the dependent variable sequence is calculated using the Pearson correlation coefficient algorithm, and the absolute value of the calculated correlation coefficient is compared with a preset correlation threshold (preferably 0.8). If the absolute value of the correlation coefficient is greater than or equal to the correlation threshold, it indicates that the measurement parameters of the historical cycle show a significant linear drift trend over time. In this case, the least squares method is used to linearly fit the independent variable sequence and the dependent variable sequence to construct a linear prediction model. If the absolute value of the correlation coefficient is less than the correlation threshold, it indicates that the measurement parameters of the historical cycle exhibit complex nonlinear fluctuations over time. In this case, a nonlinear fitting algorithm (such as polynomial fitting or exponential fitting) is used to perform nonlinear mapping on the independent variable sequence and the dependent variable sequence to construct a nonlinear prediction model. Obtain the target time point at which the current calibration cycle is expected to be sent to the continuous weighing scale 7, input the target time point into the constructed prediction model (linear prediction model or nonlinear prediction model), and output the prediction correction coefficient.

[0042] In some embodiments, the specific process of performing a second-level dynamic trend verification of the initial calibration coefficients using the predicted correction coefficients in the dynamic correction unit is as follows: Calculate the absolute value of the difference between the predicted correction coefficient and the initial calibration coefficient that passed the first-level verification; compare the absolute value of the difference with a preset dynamic confidence threshold; If the absolute value of the difference is greater than or equal to the dynamic confidence threshold, it is determined that the current initial calibration coefficient deviates significantly from the normal evolution trend of the system (such as encountering coal powder internal collapse or sensor transient failure), the second-level dynamic trend verification fails, and the current initial calibration coefficient is discarded. If the absolute value of the difference is less than the dynamic confidence threshold, it is determined that the current initial calibration coefficient conforms to the historical evolution pattern, and the second-level dynamic trend verification is passed; It should be noted that the dynamic confidence threshold is an empirical value determined based on the statistical standard deviation of historical prediction errors or a preset tolerance band width. Its physical meaning lies in defining the reasonable fluctuation range of model prediction errors. If the difference exceeds the dynamic confidence threshold, it indicates that the current deviation cannot be explained by normal mechanical wear or gradual changes in material properties, and must indicate a sudden physical anomaly (such as a sensor being subjected to a transient impact from a hard object, or a large-scale collapse of pulverized coal), thus requiring isolation. For example, the dynamic confidence threshold can be set to 0.02, allowing a maximum reasonable deviation fluctuation of 2% between the predicted correction coefficient and the initial calibration coefficient.

[0043] In some embodiments, the specific process of generating the current period's correction coefficient by weighted iterative summation of the initial calibration coefficient and the correction coefficient of the previous calibration period in the dynamic correction unit is as follows: After the initial calibration coefficients have passed the first-level static interval verification and the second-level dynamic trend verification, a preset smoothing weight coefficient (within the range of 0 to 1) is introduced. The verified initial calibration coefficients are multiplied by the smoothing weight coefficient to obtain the current correction component. At the same time, the correction coefficient of the previous calibration cycle is multiplied by the remaining smoothing weight coefficient (1 minus the smoothing weight coefficient) to obtain the historical retained component. The current correction component and the historical retained component are algebraically added to smoothly update the measurement parameters and generate the correction coefficient for the current calibration cycle.

[0044] In some embodiments, the specific process by which the dynamic correction unit performs adaptive dead-zone verification on the current cycle correction coefficient and triggers the dispatch process to the continuous weighing scale for deviation compensation is as follows: Calculate the absolute value of the change between the correction coefficient of the current calibration cycle and the correction coefficient actually issued in the previous calibration cycle, and compare the absolute value of the change with the preset adjustment dead zone threshold. If the absolute value of the change is less than the adjustment dead zone threshold, the dead zone check fails. It is determined that the current system deviation is within the tolerable mechanical fault tolerance range, and the issuance of this calibration command is temporarily suspended to avoid high-frequency vibration and mechanical fatigue of the drive motor of the continuous weighing scale 7 due to frequent fine-tuning. If the absolute value of the change is greater than or equal to the adjustment dead zone threshold, the sending process is triggered, the dead zone verification is passed, and the current calibration cycle correction coefficient is sent in real time to the underlying controller of the continuous weighing scale 7 via the industrial communication bus through the controller 13. The underlying control logic of the continuous weighing scale 7 uses it as a multiplication factor and directly applies it to the internal flow calculation formula or the target set value of the PID controller, thereby dynamically adjusting the speed of the drive motor to compensate for static errors in real time.

[0045] It should be noted that the adjustment dead zone threshold is an empirical or dynamically calculated value determined based on the minimum control resolution of the inverter of the continuous weighing scale 7 drive motor and the gear backlash of the mechanical transmission mechanism. This is because when the change in the correction coefficient causes the motor speed adjustment command to be lower than the inverter's minimum resolution, or when the mechanical gear backlash cannot be overcome, the actual physical feed rate will not change effectively. Instead, it will lead to the control system frequently outputting invalid commands and increasing the electromagnetic heat loss of the motor. For example, the adjustment dead zone threshold can be set to 0.005 (corresponding to 0.5% control deviation). The dead zone mechanism can effectively filter out invalid, minute control commands, maximizing the service life of the electromechanical equipment while ensuring metering accuracy.

[0046] To further verify the effectiveness of the two-level verification and adaptive dead-time mechanism in the dynamic correction unit, such as Figure 3 As shown, Figure 3 The evolution trajectory of the correction coefficient over multiple consecutive calibration cycles of this invention is shown.

[0047] The horizontal axis in the graph represents the continuous calibration cycle number, and the vertical axis represents the correction factor value, such as... Figure 3 As shown, the historical correction coefficients (dots) exhibit a slow linear drift trend. A linear prediction model (dashed line) was constructed, and a dynamic confidence interval (shaded band) was generated by floating up and down based on the prediction model. In the kth period, the calculated initial calibration coefficients (asterisks) were severely offset due to sudden interference, exceeding the dynamic confidence interval. This triggered the second-level dynamic trend verification to fail, and the coefficients were directly discarded, thus successfully isolating the anomaly.

[0048] Furthermore, for the initial calibration coefficients that have passed verification, the current period correction coefficient curve (solid line) generated after weighted iteration is smoother than the original calculated value, especially in some periods. Figure 3 The solid line in the middle represents a horizontal stepped area. Since the change in the correction coefficient between adjacent cycles is less than the preset dead zone threshold (e.g., 0.005), the dead zone interception mechanism is triggered, keeping the underlying control parameters unchanged. Macroscopically, it presents a stable stepped control trajectory, effectively filtering out invalid high-frequency fine-tuning commands. This intuitively demonstrates the protective effect of this invention on electromechanical equipment while ensuring measurement accuracy. Example

[0049] Based on the specific implementation process of Example 1, such as Figure 2 As shown, the present invention also provides a dynamic coal powder metering device, which includes the following physical structure: The weighing buffer chamber 5 is equipped with a weighing sensor 5.1 for real-time acquisition of chamber weight data; The pre-feeding mechanism is located at the feed end of the weighing buffer chamber 5; A constant flow feeder (i.e., a dual-drive constant flow machine 6) is installed at the bottom discharge port of the weighing buffer hopper 5; The continuous metering scale 7 receives the discharge from the constant flow feeder and is used for continuous dynamic metering of pulverized coal. The controller 13 is communicatively connected to the front-stage feeding mechanism, the weighing sensor 5.1, and the continuous weighing scale 7, respectively. The controller 13 is configured to execute the control logic of each unit in the pulverized coal dynamic metering system described in Embodiment 1.

[0050] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic coal powder metering system, characterized in that: include: Data acquisition unit: collects instantaneous weight data of the weighing buffer bin, derivative characteristics of the rate of change of gravity, physical opening signal of the front feeding mechanism, and instantaneous flow variance of the continuous weighing scale; Steady-state boundary determination unit: When it is detected that the front feeding mechanism is completely closed, the instantaneous silo weight data is within the safe fluidization range, the absolute value of the derivative characteristic of the gravity change rate is less than the silo weight disturbance threshold, and the instantaneous flow variance is less than the feeding fluctuation threshold, a candidate steady state is generated. The duration of the candidate steady state is compared with the set time. If the duration exceeds the set time, the no-feeding steady calibration window is opened. Synchronous integration unit: Within the no-replenishment stable calibration window, the reference actual conveying volume of the weighing buffer bin is calculated using a moving average filtering algorithm, and the effective cumulative volume of the continuous weighing scale within the no-replenishment stable calibration window is obtained through synchronous integration. Dynamic correction unit: Based on the reference actual conveying volume and effective cumulative volume, it calculates the initial calibration coefficient for the current calibration cycle and performs the first-level static interval verification. It uses historical correction coefficients to build a prediction model and outputs the predicted correction coefficient. It then uses the predicted correction coefficient to perform the second-level dynamic trend verification on the initial calibration coefficient. After both the first-level static interval verification and the second-level dynamic trend verification pass, it performs a weighted iterative calculation of the initial calibration coefficient and the correction coefficient of the previous calibration cycle to generate the correction coefficient for the current calibration cycle. After the dead zone verification of the correction coefficient for the current calibration cycle passes, it is sent to the continuous weighing scale for deviation compensation.

2. The pulverized coal dynamic metering system according to claim 1, characterized in that, The specific process for collecting the instantaneous weight data of the weighing buffer bin, the derivative characteristics of the rate of change of gravity, the physical opening signal of the front-stage feeding mechanism, and the instantaneous flow variance of the continuous weighing scale is as follows: The gross weight analog electrical signal of the weighing buffer bin is collected by the weighing sensor, and the gross weight analog electrical signal is transmitted to the controller for analog-to-digital conversion to obtain the instantaneous bin weight data. Based on the instantaneous weight data, the controller uses a differential algorithm in the discrete time domain to calculate the weight change between adjacent sampling periods, obtains the initial gravity change rate, and uses a low-pass filter to smooth the initial gravity change rate to obtain the derivative characteristics of the gravity change rate. The system acquires the electrical signals fed back from the flow valve or pneumatic gate valve and the valve positioner of the non-standard feeding mechanism, converts the feedback electrical signals into a 0-100% opening percentage value, and acquires the physical opening signal. When digital control is used, the physical opening signal is acquired directly. The instantaneous linear load of pulverized coal is obtained through the weighing bridge inside the continuous weighing scale, and the instantaneous operating speed of the weighing scale is obtained through the speed sensor. The instantaneous linear load and the instantaneous operating speed are multiplied to obtain the instantaneous feed flow rate. A sliding time window containing a preset number of sampling points is set in the controller, and the mathematical variance of the instantaneous feed flow rate within the sliding time window is calculated in real time. The calculation result is used as the instantaneous flow rate variance.

3. The dynamic coal powder metering system according to claim 1, characterized in that, The specific process for opening the no-feeding stable calibration window is as follows: Determine if the physical opening signal of the front-stage feeding mechanism is zero; if it is, then condition one is met. Determine if the instantaneous bin weight data is within the safe fluidization range; if it is, then condition two is met. Determine if the absolute value of the derivative characteristic of the gravity change rate is less than the bin weight disturbance threshold; if it is less than the bin weight disturbance threshold, then condition three is met. Determine if the instantaneous flow variance is less than the feeding fluctuation threshold; if it is less than the feeding fluctuation threshold, then condition four is met. When all four judgment conditions are met, the system is in a candidate stable operating state. The timer inside the controller is started to keep track of the duration. If any of the four judgment conditions is no longer met during the timer, the timer is reset to zero. If the duration exceeds the set time, the no-feeding stable calibration window is opened.

4. The dynamic coal powder metering system according to claim 1, characterized in that, The specific process of obtaining the effective cumulative amount of the continuous weighing scale within the unreplenished stable calibration window through synchronous integration is as follows: At the start of the no-replenishment stabilization calibration window, acquire the first instantaneous bin weight data sequence containing multiple consecutive sampling points; at the end of the no-replenishment stabilization calibration window, acquire the second instantaneous bin weight data sequence containing multiple consecutive sampling points. The first and second instantaneous bin weight data sequences are processed by a moving average filtering algorithm to obtain the initial effective bin weight and the final effective bin weight. The initial effective bin weight is subtracted from the final effective bin weight to obtain the actual weight reduction value of the weighing buffer bin within the no-replenishment stable calibration window. The actual weight reduction value is used as a reference for the actual conveying volume. At the start of the no-feeding stable calibration window, the continuous weighing scale records the initial cumulative value. During the duration of the no-feeding stable calibration window, the instantaneous feed flow rate output by the continuous weighing scale is continuously integrally calculated in the time dimension. When the no-feeding stable calibration window ends, the calculation is stopped and the ending cumulative value is recorded. The ending cumulative value is subtracted from the initial cumulative value to obtain the effective cumulative amount.

5. The dynamic coal powder metering system according to claim 1, characterized in that, The specific process for calculating the initial calibration coefficients for the current calibration period and performing the first-level static interval verification is as follows: Obtain the reference actual delivery volume and the effective cumulative volume, and divide the reference actual delivery volume by the effective cumulative volume to obtain the initial calibration coefficient for the current calibration cycle; The calibration coefficient confidence interval is set in the controller. If the initial calibration coefficient falls within the calibration coefficient confidence interval, the first-level static interval verification is considered to have passed.

6. The dynamic coal powder metering system according to claim 1, characterized in that, The specific process of constructing a prediction model using historical correction coefficients and outputting prediction correction coefficients is as follows: The controller extracts the actual issuance time points and corresponding historical correction coefficients for N consecutive historical calibration cycles. It constructs a time variable sequence with the actual issuance time points as the independent variable sequence and a correction coefficient sequence with the corresponding historical correction coefficients as the dependent variable sequence. Using the Pearson correlation coefficient algorithm, the correlation coefficient between the independent variable series and the dependent variable series is calculated. If the absolute value of the correlation coefficient is greater than or equal to the correlation threshold, the least squares method is used to linearly fit the independent variable series and the dependent variable series to construct a prediction model. If the absolute value of the correlation coefficient is less than the correlation threshold, a nonlinear fitting algorithm is used to perform nonlinear mapping on the independent variable sequence and the dependent variable sequence to construct a prediction model. Obtain the target time point at which the current calibration cycle is expected to be sent to the continuous weighing scale, input the target time point into the prediction model, and output the prediction correction coefficient.

7. The dynamic coal powder metering system according to claim 1, characterized in that, The specific process of performing a second-level dynamic trend verification of the initial calibration coefficients using the predicted correction coefficients is as follows: Calculate the absolute value of the difference between the predicted correction coefficient and the initial calibration coefficient that passed the first-level static interval verification, and compare the absolute value of the difference with the dynamic confidence threshold; If the absolute value of the difference is greater than or equal to the dynamic confidence threshold, the second-level dynamic trend verification fails. If the absolute value of the difference is less than the dynamic confidence threshold, then the second-level dynamic trend verification passes.

8. The dynamic coal powder metering system according to claim 1, characterized in that, The specific process of generating the correction coefficient for the current calibration period by weighted iteratively averaging the initial calibration coefficient and the correction coefficient from the previous calibration period is as follows: A smoothing weight coefficient is introduced, with a value between 0 and 1. The initial calibration coefficient, which has passed the first-level static interval verification and the second-level dynamic trend verification, is multiplied by the smoothing weight coefficient to obtain the current correction component. The correction coefficient of the previous period is multiplied by the value obtained by subtracting the smoothing weight coefficient from 1 to calculate the historical retained component. The current correction component and the historical retained component are algebraically added to generate the correction coefficient of the current calibration period.

9. A dynamic coal powder metering system according to claim 1, characterized in that, The specific process of sending the dead-zone verification of the correction coefficient for the current calibration cycle to the continuous weighing scale for deviation compensation is as follows: Calculate the absolute value of the change between the correction coefficient of the current calibration cycle and the actual correction coefficient issued in the previous calibration cycle, and compare the absolute value of the change with the adjustment dead zone threshold. If the absolute value of the change is greater than or equal to the dead zone threshold, the dead zone check is passed. The controller then sends the current calibration cycle correction coefficient to the underlying controller of the continuous weighing scale in real time via the industrial communication bus. The underlying control logic of the continuous weighing scale uses the current calibration cycle correction coefficient as a multiplication factor in the flow calculation formula or the target set value of the PID controller to adjust the drive motor speed and compensate for static errors in real time.

10. A dynamic coal powder metering device, characterized in that: include: The weighing buffer chamber is equipped with a weighing sensor for real-time acquisition of chamber weight data. The pre-feeding mechanism is located at the inlet end of the weighing buffer bin; A constant flow feeder is installed at the bottom discharge port of the weighing buffer silo; Continuous metering scales receive the discharge from constant flow feeders and are used for continuous dynamic metering of pulverized coal. The controller is communicatively connected to the front-stage feeding mechanism, the weighing sensor, and the continuous metering scale; the controller is used to execute the control logic of each unit in the pulverized coal dynamic metering system according to any one of claims 1 to 9.

Images