Storage bin discharging metering redundancy detection and correction control method

Abstract

The invention relates to a bin discharging metering redundancy detection and correction control method, and belongs to the technical field of industrial automatic material conveying control. In order to solve the problems that in the prior art, due to the fact that the bin weight change rate conflicts with the actual discharging amount dimension, the control direction is misjudged, and the fixed weight cannot adapt to the complex working condition, the method comprises the steps that firstly, the bin weight change rate is obtained in real time through a bin weight sensor to serve as a first detection signal; meanwhile, the actual discharging amount accumulated value is collected through the disc feeder metering device to serve as a second detection signal; dynamically constructing a collaborative deviation model, calculating a first deviation between the bin weight change rate and the actual blanking amount and a second deviation between the planned blanking amount and the actual blanking amount, and generating a fusion deviation according to a preset weight coefficient; and finally outputting the control quantity to drive a disk feeder frequency converter. The method is mainly used for accurate metering and closed-loop control of raw material bin discharging in the industries of steel, cement and the like, dimensional conflict accidents are effectively solved, and the adaptability to complex working conditions is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial automation material conveying control, and in particular relates to a silo unloading metering redundancy detection and correction control method. Background Art

[0002] In the field of industrial material conveying control, silo unloading metering usually adopts a dual-sensor redundant detection mechanism. However, the existing technology has the following inherent defects: 1. Dimensional conflict leads to misjudgment of control direction The silo weight sensor outputs an instantaneous rate of change (unit: tons / hour), while the metering device outputs a cumulative value (unit: tons). This dimensional difference between the two makes direct comparisons physically inconsistent. For example, when the silo weight drops suddenly due to vibration, the system mistakenly interprets this as a surge in the material discharge rate, erroneously triggering a deceleration command, which actually results in insufficient material discharge. The root cause lies in the lack of a unified unit mechanism. The direct superposition of the instantaneous differential signal and the cumulative integral signal causes logical confusion. This problem has long plagued production lines, with control errors caused by dimensional conflicts accounting for an average of 34% of annual failures.

[0003] 2. Fixed weights are difficult to adapt to dynamic working conditions Traditional solutions use a fixed weighting (usually 1:1) for the silo weight and flow signals. However, in actual operation, when the disc feeder blades wear (flow error +12% after two years of operation), the distorted signal is still used at a fixed ratio. During the rainy season, when the moisture content of the material increases and sticks to the silo wall, the silo weight sensor fails (output returns to zero), but the system still uses a 50% weighting. When bearing wear causes metering device anomalies, the fixed weighting continues to integrate the fault signal. Due to the lack of a dynamic weighting adjustment mechanism, control accuracy deteriorates to ±25% under complex operating conditions. Manual weighting has been attempted, but the response lag exceeds 5 minutes and relies on operator experience.

[0004] 3. Linear Cumulative Amplification of Historical Errors Existing technology permanently stores historical deviation values and accumulates them linearly. Early abnormal data (e.g., a single sensor failure resulting in a +300% deviation) continuously contaminates the system, causing the correction factor K to be excessively amplified during subsequent normal operation due to historical abnormal data. This causes output frequency overshoot and oscillation (fluctuation amplitude ±15%), requiring a manual reset of the historical database, a process that takes an average of seven minutes. The underlying issue is that historical data processing lacks an attenuation mechanism, allowing abnormal values to continuously dominate control logic.

[0005] 4. Equipment status is disconnected from the control system Changes in mechanical health status were not fed back to the control layer. The system continued to operate according to a fixed strategy even when bearings showed early signs of wear (vibration increase >15%). The inertia effect of the feeder during startup and shutdown (iron ore experienced a lag of up to 6.7 seconds) was not compensated. When a material blockage occurred, the position signal jumped to physically impossible values (such as negative values or exceeding the device's upper limit). Because the control unit lacked access to equipment status monitoring signals, the misjudgment rate during mechanical anomalies surged to 41%, and there was no parameter out-of-bounds protection mechanism.

[0006] 5. Transient process control inaccuracy The startup and shutdown phases account for 5% of the operating time, but the resulting metering errors account for 38% of the total error: actual material delivery lags behind the frequency command during startup and acceleration; residual material is not accounted for during shutdown and deceleration; and the inertia differences between different materials (iron concentrate and bentonite) are not differentiated. Traditional methods use a fixed hysteresis compensation value, but the compensation error during material switching can reach 12%.

[0007] These issues stem from the inability to coordinate the processing of heterogeneous data from multiple sources (equipment status signals, material characteristic parameters, and real-time measurement values). This results in an average of 7.2 process interruptions annually at industrial sites due to control inaccuracies. The challenges in addressing these issues lie in establishing a unified dimensional model, designing a dynamic weight reconstruction mechanism, intelligently attenuating historical errors, and establishing a closed loop between equipment status and control—requirements that have long been unmet by existing technologies. Summary of the Invention

[0008] One purpose of the present invention is to solve the problem of misjudgment of control direction caused by the dimensional conflict between the bin weight change rate (instantaneous differential signal) and the actual material discharge amount (cumulative integral signal), as well as the defect that fixed weights cannot adapt to dynamic working conditions such as equipment wear and changes in material properties.

[0009] Solve the problem of correction out of control caused by historical abnormal data pollution, and equipment overload caused by signal jumps under extreme working conditions (such as negative position value during material blockage).

[0010] Solve the problem of flow meter error drift (+12%) caused by blade wear after long-term operation of the equipment, but continuous batching deviation caused by the weight coefficient being fixed for life.

[0011] This solves the problem that the operator's experience in adjusting weights for specific working conditions (such as adhesion in the rainy season) cannot be recorded and reused by the system.

[0012] This solves the problem of the control system continuing to use distorted flow signals (error +18%), which leads to process interruption when early bearing wear is not identified.

[0013] Solve the problem that sudden changes in material viscosity (such as a sudden increase in moisture content) cause the silo weight signal to fail and the system accelerates incorrectly, causing the chain conveyor to break.

[0014] This solves the problem that human experience cannot be embedded in the control system in real time and emergency conditions rely on delayed responses.

[0015] Solve the problem of imbalance in the weight distribution between machine autonomous control and human intervention decision-making, which leads to failure in responding to unknown working conditions.

[0016] Solve the problem of unbalanced flow rates between adjacent silos (difference > 15 tons / hour) causing a mixing rate of up to 12%.

[0017] Solve the problem of instantaneous metering deviation accounting for 38% of the total amount due to material inertia differences (iron ore lag 6.7 seconds / bentonite 3.2 seconds) during the start-up and shutdown stages.

[0018] In order to achieve these purposes and other advantages according to the present invention, the present invention provides a method for detecting and correcting redundant material metering in a silo, comprising the following steps: a) Using a bin weight sensor to detect the bin weight change rate in real time as a first detection signal; b) obtaining the accumulated value of the actual feeding amount through the disc feeder metering device as the second detection signal; c) Dynamically build collaborative bias models: Calculate the first deviation ΔW: ΔW = | Bin weight change rate − actual unloading amount |, Calculate the second deviation ΔQ: ΔQ = |Planned cutting quantity − Actual cutting quantity|, Generate fusion deviation Δ according to the preset weight coefficient (α, β) 融合 : , and α+β=1 d) When the fusion deviation Δ 融合 When the threshold is exceeded, based on the historical deviation sequence Calculate the adaptive correction coefficient K: ; e) Output control quantity U out : , U 当前 is the current operating frequency of the disc feeder; The output control quantity U out As the frequency converter input frequency of the disc feeder.

[0019] Preferably, the silo unloading metering redundancy detection and correction control method further includes: When the system is first run, the initial setting value is modified, or the device is triggered by malfunction, the control unit performs a reset operation, which includes clearing the historical accumulated deviation value and resetting the initial setting value to zero; The control unit sets upper and lower thresholds for the reduction in bin weight and position, actual material discharge, and historical cumulative deviation values. When the detection value exceeds the upper or lower threshold range, the value is forcibly locked as the corresponding threshold boundary value.

[0020] Preferably, the silo unloading metering redundancy detection and correction control method further includes: The control unit automatically adjusts the reference values of the weight coefficients α and β according to the cumulative production output. The reference value of α increases by a step size every time the cumulative output increases by the set threshold. , β baseline value reduction step size .

[0021] Preferably, the silo unloading metering redundancy detection and correction control method further includes: When the operator manually modifies the weight coefficient, the system records the modification amplitude Δα, Δβ and the corresponding working conditions, and automatically preloads the modified weight coefficient when the same working conditions occur again.

[0022] Preferably, the silo unloading metering redundancy detection and correction control method further includes: Collect the vibration spectrum of the disc feeder in real time and extract the characteristic frequency amplitude Af; When the Af of a specific frequency band of 10-100Hz increases by more than 15% for three consecutive samples, a bearing wear warning is generated; When an early warning is triggered, the weight coefficient α is automatically increased to the upper limit of 1.0, and the correction response speed of the historical cumulative deviation is increased by 50%.

[0023] Preferably, the silo unloading metering redundancy detection and correction control method further includes: Real-time calculation of the correlation coefficient ρ between the bin weight change rate and the feeder current; When ρ is less than 0.6 for 5 consecutive minutes, it is determined that the material viscosity has increased abnormally. When the viscosity has increased abnormally, the weight coefficient α is reduced to 0.3, β is increased to 0.7, and the high-frequency vibrator is activated to assist in material unloading.

[0024] Preferably, the silo unloading metering redundancy detection and correction control method further includes the step of manual experience quantitative input: Operators use a dedicated interface on the control room touch screen to perform real-time confidence scoring for two key sensors: Bin weight sensor confidence value: Slide bar setting 0-100 points, where 0 = completely distorted and 100 = absolutely reliable; Flow meter confidence value: independent slider setting 0-100 points; The score takes effect immediately after submission, and the system automatically records the operator's number and timestamp.

[0025] Preferably, the silo unloading metering redundancy detection and correction control method further includes a human-machine fusion decision-making step: Machine weight reception: obtain automatically calculated weight coefficients α and β; Manual weight generation: Convert the confidence value of the bin weight sensor to a weight factor: Manual bin weight weight = confidence value / 100; Convert the confidence value of the flow meter to a weight factor: Manual flow meter weight = confidence value / 100 Fusion execution: Final bin weight fusion weight = machine weight α × 70% + manual bin weight × 30%; Final flow meter weight = machine weight β × 70% + manual flow meter weight × 30%.

[0026] Preferably, the silo unloading metering redundancy detection and correction control method further includes: Silo coupling effect perception: Real-time calculation of the difference in unloading rates between adjacent silos ; When the unloading rate difference ΔS continues to be greater than the set coupling threshold, it is determined that the silo is unbalanced; Dynamic Trim Control: In the unbalanced state, the weight coefficient of the high-speed silo is automatically reduced , improve the weight coefficient of low-speed silo ; Synchronously adjust the disc feeder frequency: reduce the frequency by K·ΔS for high-speed silos and increase the frequency by K·ΔS for low-speed silos; Material transfer warning: When the imbalance duration > T max When mixing materials, a risk alert is generated and a transfer plan is recommended.

[0027] T max It is the maximum time threshold that the system allows the silo to be out of balance. Its value is dynamically calculated based on two key factors: Material safe transfer time: the time required to empty all the materials in the silo (for example: silo capacity 50 tons ÷ belt conveyor speed 10 tons / hour = 5 hours); Safety factor: Add an additional 20% buffer time (to prevent unexpected delays).

[0028] Final formula: T max = Material safe transfer time × 1.2.

[0029] Preferably, the redundant detection and correction control method for silo unloading metering performs transient compensation control during the start-up and shutdown phases of the disc feeder: Start-stop transient recognition: When the feeder accelerates from standstill to 90% of the set frequency, it is marked as startup acceleration transient; When the feeder decelerates from the running state to 10% of the frequency, it is marked as the shutdown deceleration transient; By real-time monitoring of the frequency change rate, it is determined that: during the acceleration phase, the frequency increases by more than 5 Hz per second and exceeds 90% of the set value; during the deceleration phase, the frequency decreases by more than 3 Hz per second and is less than 10%; Dynamic inertia compensation: The preset lag time parameters are called according to the material type: 6.7 seconds for iron ore and 3.2 seconds for bentonite. The compensation increment is calculated as follows: lag time multiplied by acceleration and then multiplied by the adjustment coefficient 0.8. The final output frequency is the normal control value plus the compensation increment.

[0030] The present invention has at least the following beneficial effects: By unifying dimensions (the silo weight change rate and the actual material discharge volume are converted into tons / hour), misjudgment of control direction is eliminated. The fusion deviation model combined with dynamic weight distribution improves the control accuracy under complex working conditions to ±0.4% (traditional solution ±6.2%), and the average annual dimensional conflict accidents are reduced to 0.

[0031] The historical data clearing mechanism eliminates early abnormal value contamination (reset time ≤ 200 milliseconds), and the dynamic threshold constraint limits the position signal jump during material blockage to a physically feasible range (such as negative values returning to zero), avoiding equipment overload and damage, and improving system stability by 40%.

[0032] The cumulative output drives the weight self-tuning (α+0.03 for every 50,000 tons), which automatically reduces the flow meter weight after equipment wear. The control accuracy of the production line remains at ±1.8% after three years of operation (the traditional solution deteriorates to ±9.5%).

[0033] Manually adjusted weights are bound to and stored with operating condition characteristics (e.g., α=0.7 when humidity is >85%), and empirical parameters are automatically reused for the same operating conditions. The response lag is compressed from 5 minutes to 200 milliseconds, reducing raw material waste by 17%.

[0034] Bearing wear warning linkage control (vibration increase > 15% triggers α = 1.0) shields fault signals and accelerates correction responses. The control accuracy during the mechanical abnormality period remains at ±2.1%, and the fault downtime rate is reduced by 89%.

[0035] Abnormal viscosity perception (ρ < 0.6 for 5 minutes) triggers weight switching (β = 0.7) and vibrator activation, breaking adhesion within 8 minutes and preventing chain conveyor breakage, with a control accuracy of ±3.2% (traditional solutions ±25%).

[0036] The real-time scoring mechanism of the slider (0-100 points) digitizes manual experience and automatically issues warnings when the scoring deviation exceeds the limit (such as a red border when the score is less than 60 points), reducing the operational error rate by 52%.

[0037] The machine weight of 70% is dominant to ensure basic stability, while the manual weight of 30% retains the ability to intervene in unknown working conditions (such as sudden blockage). The error rate of human-machine collaborative decision-making is only 1 / 5 of that of the purely manual solution.

[0038] The flow rate differences between adjacent silos are dynamically balanced (when ΔS>15 tons / hour, the speed bin is reduced and the frequency is increased by K·ΔS), reducing the mixing rate from 12% to 0.8%. If the imbalance lasts for more than 45 minutes, a silo transfer plan is automatically generated, and the decision-making time takes 3.2 minutes (traditional 47 minutes).

[0039] The start-stop transients are matched with hysteresis parameters according to the material type (6.7 seconds for iron ore and 3.2 seconds for bentonite). After compensation, the instantaneous metering error is reduced from 12% to 1.8%, which can reduce the process start-stop deviation by 38% annually. DETAILED DESCRIPTION

[0040] The present invention is described in further detail below so that those skilled in the art can implement the invention with reference to the description.

[0041] Example 1: A high-precision weighing sensor (range 0-100 tons, resolution ±10kg) was installed at the bottom of the raw material silo to collect real-time silo weight data and calculate the silo weight change rate to generate a first detection signal. Simultaneously, a nuclear scale (accuracy ±0.5%) was used on the disc feeder to obtain the accumulated actual material discharge volume to generate a second detection signal. The sampling period for both was 1 second.

[0042] The control unit performs deviation fusion: calculates the instantaneous metering error ΔW: ΔW=|bin weight change rate-actual unloading quantity|; calculates the set tracking error ΔQ: ΔQ=|planned unloading quantity-actual unloading quantity|; among them, the actual unloading quantity specifically refers to the converted instantaneous rate value, that is, the actual unloading rate (unit: tons / hour), that is, the actual unloading quantity = (current cumulative value-previous cycle cumulative value) / adopting cycle length, the planned unloading quantity specifically refers to the planned unloading rate (unit: tons / hour), that is, the planned unloading quantity = planned total quantity / planned completion time.

[0043] Dynamic weighting: When the vibration sensor detects a silo amplitude greater than 5 mm / s (risk of arching), the silo weight α is set to 0.2, and the flow meter weight β is set to 0.8. If the flow meter's historical error exceeds 5% for three consecutive times, α is set to 0.7 and β to 0.3. Under normal operating conditions, α = β = 0.5, where α and β are preset weighting coefficients (α + β = 1).

[0044] Generate fusion deviation: Δ 融合 =α×ΔW+β×ΔQ.

[0045] Call the historical database (store the latest 100 deviation records) and perform nonlinear processing on the fusion deviation: take the square root of the sum of squares of historical deviations (√∑Δ²) as the benchmark denominator, and divide the current Δ 融合 Divide by (0.8 × reference value + ) to obtain the correction coefficient K. When a blocking trend is detected (ΔW suddenly increases by 300%), the denominator coefficient 0.8 is automatically adjusted to 0.5 to enhance the correction strength.

[0046] Compare the planned feeding amount with the actual feeding amount in real time: if the actual feeding amount is lower than the planned feeding amount, mark the direction sign sign = +1 (i.e. sign (ΔQ)); if the actual feeding amount is higher than the planned feeding amount, mark sign = -1; Directional output control quantity: read the current frequency U of the disc feeder 当前 (such as 45Hz) Calculate the new frequency: U 当前 ×(1+K×sign). When sign = +1, the output is a speed-up signal (e.g., 45×(1+0.1)=49.5Hz), and when sign = -1, the output is a speed-down signal (e.g., 45×(1-0.1)=40.5Hz).

[0047] This embodiment solves the technical problem of loss of control of the correction direction caused by dimensional conflict. In the prior art, the reduction in bin position (instantaneous differential signal) and the planned total material quantity (cumulative integral signal) are directly superimposed, and the dimensions of the two are tons / hour and tons, respectively. The units are confusing, and the system is prone to misjudgment when the bin position decreases. This embodiment dynamically converts the reduction in bin position into the bin weight change rate (tons / hour), and the planned total material quantity into the planned unloading quantity (tons / hour), thereby achieving dimensional unification. The weights are dynamically assigned through the working condition perception module: when the bin vibrates greatly, the flow meter signal is trusted (weight 0.8); when the flow meter is abnormal, the bin weight change rate (weight 0.7) is trusted. Effect: Dimension conflict accidents on the production line are reduced.

[0048] This embodiment addresses the technical issue of overshoot oscillation caused by the linear accumulation of historical errors. In existing technologies, simply accumulating historical deviations results in the continuous amplification of earlier abnormal data. This embodiment introduces a time decay mechanism to address historical deviations: the current deviation is weighted 100%, the previous deviation is weighted 80%, and earlier deviations are exponentially decayed. The strength of the reference denominator is dynamically adjusted based on the current operating conditions, increasing the correction force when material blockage occurs and smoothing the suppression during normal conditions.

[0049] This embodiment addresses the technical issue of a surge in false positives during equipment failure. In existing technologies, dual sensor signals are weighted at a fixed 1:1 ratio. This embodiment establishes a mechanism for self-diagnosis of failure modes and self-reconstruction of weights: the vibration sensor detects arching risk, and the flowmeter self-diagnosis module assesses metering reliability. Weight switching is completed within 200 milliseconds (for example, the flowmeter weight increases to 0.8 in the event of an arch). The result: control accuracy remains within ±2.5% even in the event of a sensor failure, far exceeding the ±15% of conventional solutions.

[0050] This invention utilizes a four-level progressive mechanism: dimensional unification, error attenuation, intelligent weighting, and directional orientation. It upgrades redundancy detection from simple signal superposition to a condition-adaptive system. Error correction capabilities evolve from passive signal reception to active condition sensing. Historical data utilization evolves from linear amplification to intelligent forgetting. Control logic evolves from open-loop compensation to directional closed-loop. This resolves the technical challenges of dimensional conflict, error amplification, and fault vulnerability in industrial material handling control.

[0051] According to one embodiment of the present invention, a metering compensation reset mechanism is provided. When the system is first started, the initial setting value is manually modified, or a device malfunction is triggered (such as a feeder jam alarm), the control unit automatically performs the following: 1. Clear historical accumulated deviation: delete all historical deviation data stored in the PLC register; 2. Reset the initial setting value: restore the planned material quantity to the default zero value; 3. Fault sign clearing: reset diagnostic signs such as vibration exceeding limit and flow meter abnormality; Execution time: ≤200 milliseconds from trigger to completion (traditional solutions require manual operation >5 minutes); A dynamic constraint mechanism is also included, imposing hard threshold limits on key parameters: 1. Reduction of heavy positions: limited to [0, Q max ](Q max is the theoretical maximum unloading rate of the equipment); For example, ball mill silo Q max =80\ tons / hour. When vibration causes the signal to jump to 120 tons / hour, it is forced to lock at 80 tons / hour.

[0052] 2. Actual feeding quantity: limited to [0, R cap ](R cap is the rated capacity of the disc feeder); 3. Historical cumulative deviation: limited to [-S, +S] (S is the safety margin, which is 20% of the planned material cutting amount); This implementation addresses the issue of uncontrolled corrections caused by historical data contamination. While existing technologies permanently store historical deviations, this implementation utilizes a metering compensation reset mechanism to automatically clear historical data upon system initialization or after a fault. Correction errors caused by historical data contamination are reset to zero.

[0053] This implementation also addresses signal jumps caused by extreme operating conditions. Existing technologies lack parameter constraints. When a material blockage occurs, the bin position signal plummets to a negative value, causing the system to erroneously accelerate material discharge and cause overflow. This implementation utilizes a dynamic constraint mechanism to constrain key parameters within physically feasible ranges: bin weight reduction ≥ 0 (eliminating negative values), actual discharge volume ≤ equipment rated capacity (preventing over-range), and historical cumulative deviation fluctuations are limited to prevent sudden K-factor abrupt changes.

[0054] For example, if ΔW suddenly increases by 500% during material blockage, it should be calculated as 300% after constraint to avoid overload and burning of the inverter.

[0055] This implementation addresses deficiencies in system robustness by automatically resetting and clearing dirty data and combining dynamic thresholds to protect safety boundaries: from passive to active: traditional manual reset is upgraded to triggered self-healing; from unbounded to bounded: parameter constraints eliminate physically impossible values.

[0056] According to another embodiment of the present invention, an operating parameter self-tuning mechanism is included; 1. Cumulative output monitoring: Real-time statistics of the total amount of material discharged by the disc feeder (e.g. updated every time 10,000 tons are completed); When the cumulative amount reaches the set threshold (such as 50,000 tons), the weight self-tuning is triggered.

[0057] 2. Dynamic weight adjustment: α baseline value increased (Default 0.03), β baseline value is reduced (Default 0.03); Example: Initially, α=β=0.5. After a cumulative output of 150,000 tons, α=0.95 and β=0.05.

[0058] 3. Clamp protection: The upper limit of α is 1.0 and the lower limit is 0.2; the upper limit of β is 0.8 and the lower limit is 0 (to avoid excessive deviation).

[0059] This implementation also includes a manual intervention coordination mechanism: 1. Manual operation record: When the operator manually modifies the weight (e.g., setting α=0.7 when material is blocked), the system captures the current working condition characteristics: vibration frequency>10Hz, sudden increase in bin weight change rate>200%, ambient humidity>85%; 2. Working condition feature storage: Bind the modified weights and working condition features and store them in the knowledge base; 3. Intelligent preloading: Real-time monitoring of working conditions. When the same characteristics are matched again (vibration > 10 Hz and humidity > 85%), α is automatically switched to 0.7 and a prompt "Historical tuning parameters have been loaded" is displayed.

[0060] This embodiment solves the technical problem of inaccurate control caused by long-term wear of equipment. In the prior art, the weight coefficient is fixed for life. Due to the wear of the disc feeder blades (the flow error is +12% after 2 years of operation), the system still uses the flow signal according to β=0.5, resulting in a continuous deviation of 8% in the batching. The present invention uses a production-driven self-tuning method to make the weight adaptive with equipment loss: for every 50,000 tons increase in cumulative production, α+0.03 (trust bin weight sensor) corresponds to β-0.03 (reducing the weight of the wear flow meter). Effect: After 3 years of operation of the production line, the control accuracy remains at ±1.8% (the traditional solution deteriorates to ±9.5%) This implementation solves the technical problem of expert experience being lost. An industry pain point: Engineers increased α to 0.8 to account for material adhesion during the rainy season, but this experience wasn't recorded in the system, leading new employees to repeat mistakes and waste raw materials. This invention digitizes this experience through a collaborative human intervention mechanism.

[0061] According to another embodiment of the present invention, a vibration sensor is installed on the bearing seat of a disc feeder to collect real-time vibration spectra in the 10-100 Hz frequency range. If the amplitude of a specific characteristic frequency (e.g., 65 Hz) increases by more than 15% over three consecutive sampling periods, a bearing wear warning signal is automatically generated.

[0062] After the warning is triggered, the system performs a two-fold response: 1. Increase the weight coefficient α to the upper limit of 1.0, so that the control is completely dependent on the bin weight sensor to avoid the risk of flow meter signal distortion; 2. Increase the correction response speed of historical cumulative deviations by 50% to accelerate the suppression of material feeding fluctuations.

[0063] Existing technologies lack equipment status monitoring and only respond passively after a fault occurs. Even in the event of a mechanical anomaly, the fixed-weight fusion algorithm is still executed mechanically, resulting in persistent distortion of the control signal. While traditional solutions require a complete bearing seizure to be detected, the present invention proactively identifies early signs of wear by detecting a vibration spectrum increase greater than 15%. Traditional solutions continue to fuse signals at an α value of 0.5 in the event of a mechanical anomaly, while the present invention automatically switches to a safe control mode at an α value of 1.0.

[0064] The invention implements a closed-loop linkage of prediction and control: from vibration symptom identification to real-time adjustment of control parameters and coordination of maintenance instructions, a two-level protection mechanism is implemented: a first-level defense (α = 1.0 shields faulty equipment signals) combined with a second-level acceleration (historical deviation response is accelerated by 50%).

[0065] Deeply embed equipment health management into the control system to achieve: capturing early signs of mechanical degradation through vibration spectrum; dynamically reconstructing the fusion model based on the warning level; and directly triggering pre-emptive maintenance work orders through warning signals.

[0066] According to another embodiment of the present invention, the invention comprises: Step 1: Real-time perception of material flow A current transmitter is installed on the disc feeder drive motor to synchronously collect the bin weight sensor data. The correlation coefficient ρ between the bin weight change rate and the motor current is calculated once every second: When the material flowability is good, the decrease in bin weight is strongly correlated with the increase in current (ρ>0.85); When the material viscosity increases (such as a sudden increase in moisture content), the bin weight change stagnates and the current fluctuates, causing the ρ value to drop. When ρ is less than 0.6 for 5 consecutive minutes, it is determined that the material viscosity is abnormal.

[0067] Step 2: Dual-mode control reconstruction After the viscosity abnormality warning is triggered, the system executes: 1. Dynamic switching of weight coefficients: Reduce the α of weight 1 from the baseline value of 0.5 to 0.3 (reduce the weight of the failed bin signal), and increase β from 0.5 to 0.7 (increase the weight of the flow meter signal); 2. Physical mobility intervention: Activate the high-frequency vibrator on the silo wall (50Hz amplitude continuous output) until the ρ value returns to >0.75 and then turn it off.

[0068] Existing technology ignores changes in material properties. Increased humidity during the rainy season causes increased material viscosity, causing the bin weight signal to remain at zero. The system misjudged the bin empty and rapidly increased speed, causing the chain conveyor to overload and break. Without physical intervention, the only way to address the arching is to manually tap the bin.

[0069] The key problems solved by the present invention are: 1. No perception of viscosity mutation Traditional solutions rely on visual observation of material fluidity, with a response lag of >30 minutes. The present invention uses the current-bin weight correlation coefficient ρ to capture viscosity changes in real time and provide an early warning within 5 minutes. For example, during the rainy season, if the ρ value drops sharply from 0.92 to 0.53, the system activates the vibrator 12 minutes in advance.

[0070] 2. Control strategy is out of line with material characteristics Conventional solutions still fuse signals at α = 0.5 during viscosity anomalies, resulting in 50% of the zero-bin weight signal being adopted. This new approach dynamically switches to a dominant mode at β = 0.7, avoiding contamination from invalid signals. The result: Control accuracy during viscosity anomalies remains within ±3.5% (compared to ±25% for conventional solutions).

[0071] 3. Lack of proactive liquidity intervention The traditional solution requires manual climbing up the silo to knock, which is inefficient and dangerous. The present invention automatically activates the high-frequency vibrator and directly connects the mechanical execution from the electrical control. For example, the arch processing time is shortened from 45 minutes to 8 minutes.

[0072] According to another embodiment of the present invention, a dedicated human-machine collaborative touch screen is deployed in the central control room, providing two independent control channels: Channel A (bin weight sensor credibility): This dynamically displays a comparison chart of the current bin weight signal waveform and historical curves. The operator drags a slider to set the confidence value (0-100 points). A red warning border is automatically drawn if the confidence value falls below 60 points. Channel B (flow meter credibility): This displays a real-time heat map of the correlation between flow meter speed and discharge volume. The slider is set with vibration feedback, and a buzzer alert is triggered if the confidence value falls below 50 points.

[0073] Permission management mechanism: Standard operators: Adjustment range ±20 points (e.g., from 70 to 50 points). Engineers: Unlimited adjustment, which can override automatic weighting. Each adjustment generates an electronic log: "Operator Zhang xx reduced the flow meter confidence value from 75 to 60 at 2:28 PM on March 12, 2025."

[0074] Human-machine fusion decision engine: The system performs triple weight fusion every 2 seconds: 1. Machine autonomous weight: Obtain the α and β values calculated in real time (e.g., for vibration conditions, α=0.2, β=0.8); 2. Artificial experience weight: The confidence value C of the bin weight sensor set by the operator w , Flowmeter confidence value C q The linear mapping is the weight factor C w / 100, C q / 100.

[0075] 3. Authority Balance and Integration: Final position weight = machine α × 70% + manual factor × 30%; Final flow meter weight = machine β × 70% + manual factor × 30%; The 70 / 30 ratio ensures the safety of machine-led control while retaining the ability of human experts to intervene in response to unknown working conditions.

[0076] According to another embodiment of the present invention, a silo topology monitoring system is deployed in the central control room to map the physical position relationship between each silo and the mixing conveyor in real time. The system scans the discharge rate of adjacent silos every minute. If it detects that the flow rate difference between silo A and silo B exceeds 15 tons / hour for 5 minutes, it automatically triggers an orange imbalance alarm. At this time, the dynamic balancing mechanism is immediately activated: 1. The weight of high-speed silo flow meter is reduced from 0.5 to 0.3; 2. The weight of the low-speed silo is increased from 0.5 to 0.7; 3. The frequency of high-speed silo is reduced by 1Hz for every 5-ton flow rate difference (for example, if the flow rate difference is 20 tons / hour, the frequency is reduced by 4Hz); 4. The frequency of low-speed silo increases in the same proportion; If the imbalance persists for 45 minutes without relief, the system generates a transfer plan (e.g., transferring 15% of the material from silo A to a backup silo 35 meters away) and dynamically plans and maps an avoidance path through a 3D visualization interface. If the imbalance lasts for more than 60 minutes, and the probability of material mixing reaches 89%, the system forcibly activates the transfer robot to execute the plan, reducing the mixing conveyor speed to 70%.

[0077] Existing technologies use independent silo control strategies, with no data exchange between silo control cabinets. This invention's real-time flow rate differential monitoring reduces the mixing rate from 12% to 0.8%. A 45-minute imbalance warning prevents 89% of losses 40 minutes in advance. The intelligent transfer system reduces decision-making time from 47 minutes to 3.2 minutes, and route planning avoids production disruptions.

[0078] According to another embodiment of the present invention, during the start-up and shutdown phases of the disc feeder in the pelletizing production line, the system automatically executes a transient compensation control process. When the equipment starts from a standstill, the frequency change rate is monitored in real time. If the frequency exceeds 90% of the set value within 3 seconds (e.g., accelerating from 50 Hz to above 45 Hz), the system is immediately marked as in the startup acceleration transient phase. At this point, the preset material database is called up—with a matching lag time of 6.7 seconds for iron ore and 3.2 seconds for bentonite. For example, using iron ore acceleration as an example: if the acceleration is detected at 4 Hz per second, the compensation increment is calculated as: 6.7 seconds × 4 Hz / s × adjustment factor 0.8 = 21.44 Hz, which is added to the normal control frequency output. The adjustment factor 0.8 is the empirical attenuation factor for inertia compensation, derived from 200 iron ore / bentonite start-up and shutdown experiments. Its physical meaning is the effective proportion of the compensation amount. When the equipment stops and slows down, the frequency drops to 5 Hz (10% of the set value) and the bin weight sensor detects a sudden increase in the remaining material (e.g., 2.3 tons). The weighing device's reading of 98.5 tons is corrected to: 98.5 + 2.3 × 0.85 = 100.455 tons. This entire process is dynamically switched by the material type adaptive module, eliminating the need for manual intervention.

[0079] Example 2: 1. Convert the reduction in storage space into the rate of change in storage weight (tons / hour), and convert the planned total material quantity into the planned material discharge rate (tons / hour) to achieve unit unification.

[0080] 2. Real-time detection: Bin weight change rate: 48.2 tons / hour; Actual flow meter value: 52.1 tons / hour; Planned unloading capacity: 50 tons / hour.

[0081] 3. Dynamic weight allocation (equipment has processed 100,000 tons of materials cumulatively): Trust warehouses have more weight on sensors (60% weight), and flow meters have a weight of 40%; Fusion bias = 60% × |48.2-52.1| + 40% × |50-52.1| = 3.18.

[0082] 4. Output control: The correction factor is calculated to be 0.46. Since the actual material discharge exceeds the plan, the feeder frequency is reduced from 45Hz to 24.3Hz.

[0083] Results: The material feeding rate is stabilized to 49.8 tons / hour (error 0.4%). Compared with the traditional method, the material feeding fluctuation is ±6.2 tons / hour when the fixed weight is 50%-50%.

[0084] Example 3: 1. The vibration sensor detects that the 65Hz amplitude increases by more than 15% three times in a row (12μm, 14μm, 16.5μm), triggering a bearing wear warning.

[0085] 2. The system responds immediately: Increase the weight of the bin weight sensor to 100% (totally avoid the risk of flow meter distortion); The speed of historical deviation correction is increased by 50% (accelerated suppression of fluctuations).

[0086] 3. Results: During the period of flow meter signal distortion (error +18%), the control accuracy remained at ±2.1%, and the production line continued to operate.

[0087] Example 4: 1. Real-time calculation of the correlation coefficient between the warehouse weight change rate and the motor current: The correlation coefficient is > 0.85 for normal liquidity; When the viscosity is <0.6 (actually measured 0.52) for 5 consecutive minutes, an abnormal viscosity alarm is triggered.

[0088] 2. Automatic execution: Reduce the weight of the bin to 30% and increase the weight of the flow meter to 70%; Activate high-frequency vibrators on the silo wall to break up buildup.

[0089] 3. Results: Fluidity recovered after 8 minutes (correlation coefficient > 0.75), and the control accuracy during the abnormal period was ±3.2%, avoiding equipment damage.

[0090] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and embodiments shown and described herein.

Claims

1. A method for detecting and correcting redundant material metering in a silo, characterized in that: The following steps are involved: a) Using a bin weight sensor to detect the bin weight change rate in real time as a first detection signal; b) obtaining the accumulated value of the actual feeding amount through the disc feeder metering device as the second detection signal; c) Dynamically build collaborative bias models: Calculate the first deviation ΔW: ΔW = | Bin weight change rate − actual unloading amount |, Calculate the second deviation ΔQ: ΔQ = |Planned cutting quantity − Actual cutting quantity|, Generate fusion deviation Δ according to the preset weight coefficient (α, β) 融合 : , and α+β=1; d) When the fusion deviation Δ 融合 When the threshold is exceeded, based on the historical deviation sequence Calculate the adaptive correction coefficient K: ; e) Output control quantity U out : , U 当前 is the current operating frequency of the disc feeder; The output control quantity U out As the inverter input frequency of the disc feeder.

2. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: Also includes: When the system is first run, the initial setting value is modified, or the device malfunctions, the control unit performs a reset operation, which includes clearing the historical accumulated deviation value and resetting the initial setting value to zero; The control unit sets upper and lower thresholds for the reduction in bin weight and position, actual material discharge, and historical cumulative deviation values. When the detection value exceeds the upper or lower threshold range, the value is forcibly locked as the corresponding threshold boundary value.

3. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: Also includes: The control unit automatically adjusts the reference values of the weight coefficients α and β according to the cumulative production output. The reference value of α increases by a step size every time the cumulative output increases by the set threshold. , β baseline value reduction step size .

4. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: Also includes: When the operator manually modifies the weight coefficient, the system records the modification amplitude Δα, Δβ and the corresponding working conditions, and automatically preloads the modified weight coefficient when the same working conditions occur again.

5. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: Also includes: Collect the vibration spectrum of the disc feeder in real time and extract the characteristic frequency amplitude Af; When the Af of a specific frequency band of 10-100Hz increases by more than 15% for three consecutive samples, a bearing wear warning is generated; When an early warning is triggered, the weight coefficient α is automatically increased to the upper limit of 1.0, and the correction response speed of the historical cumulative deviation is increased by 50%.

6. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: Also includes: Real-time calculation of the correlation coefficient ρ between the bin weight change rate and the feeder current; When ρ is less than 0.6 for 5 consecutive minutes, it is determined that the material viscosity has increased abnormally. When the viscosity has increased abnormally, the weight coefficient α is reduced to 0.3, β is increased to 0.7, and the high-frequency vibrator is activated to assist in material unloading.

7. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: It also includes the steps of manual experience quantification input: Operators use a dedicated interface on the control room touch screen to perform real-time confidence scoring for two key sensors: Bin weight sensor confidence value: Slide bar setting 0-100 points, where 0 = completely distorted and 100 = absolutely reliable; Flow meter confidence value: independent slider setting 0-100 points; The score takes effect immediately after submission, and the system automatically records the operator's number and timestamp.

8. The silo unloading metering redundancy detection and correction control method according to claim 7 is characterized in that: It also includes the steps of human-machine fusion decision-making: Machine weight reception: obtain automatically calculated weight coefficients α and β; Manual weight generation: Convert the bin weight sensor confidence value into a weight factor: Manual bin weight weight = confidence value / 100; Convert the flow meter confidence value into a weight factor: Manual flow meter weight = confidence value / 100 Fusion execution: Final bin weight fusion weight = machine weight α × 70% + manual bin weight weight × 30%; Final flow meter weight = machine weight β × 70% + manual flow meter weight × 30%.

9. The silo unloading metering redundancy detection and correction control method according to claim 1 is characterized in that: Also includes: Silo coupling effect perception: Real-time calculation of the difference in unloading rates between adjacent silos ; When the unloading rate difference ΔS continuously exceeds the set coupling threshold, it is determined that the silo is unbalanced; Dynamic Trim Control: In the unbalanced state, the weight coefficient of the high-speed silo is automatically reduced , improve the weight coefficient of low-speed silo ; Synchronously adjust the disc feeder frequency: reduce the frequency by K·ΔS for high-speed silos and increase the frequency by K·ΔS for low-speed silos; Material transfer warning: When the imbalance duration > T max When mixing materials, a risk alert is generated and a transfer plan is recommended.

10. The silo unloading metering redundancy detection and correction control method according to claim 1, characterized in that: Execute transient compensation control during the start-up and shutdown phases of the disc feeder: Start-stop transient recognition: When the feeder accelerates from standstill to 90% of the set frequency, it is marked as startup acceleration transient; When the feeder decelerates from the running state to 10% of the frequency, it is marked as the shutdown deceleration transient; By real-time monitoring of the frequency change rate, it is determined that: during the acceleration phase, the frequency increases by more than 5 Hz per second and exceeds 90% of the set value; during the deceleration phase, the frequency decreases by more than 3 Hz per second and is less than 10% of the set value; Dynamic inertia compensation: The preset lag time parameters are called according to the material type: 6.7 seconds for iron ore and 3.2 seconds for bentonite. The compensation increment is calculated as follows: lag time × acceleration × adjustment coefficient 0.

8. The final output frequency is the normal control value plus the compensation increment.