Method for detecting and automatically unblocking a conveyor transfer point clogging

Abstract

This invention discloses a method for detecting and automatically clearing blockages at conveyor transfer drop points, specifically relating to the field of machine testing and inspection technology. The method includes: S1: deployment and parameter calibration of the detection system; S2: acquisition of baseline data under normal operating conditions; S3: real-time multi-dimensional detection data acquisition; S4: multi-parameter fusion blockage judgment and level classification; S5: adaptive matching of clearing parameters; S6: precise automatic clearing execution; S7: verification of clearing effect and secondary processing; and S8: data recording and model optimization. This invention employs multi-dimensional detection unit collaboration and data fusion analysis to effectively capture complex state changes. The multi-parameter cross-validation mechanism significantly improves the comprehensiveness of feature recognition and the accuracy of state judgment under different materials and diverse blockage conditions, resulting in higher accuracy in detecting blockages at conveyor transfer drop points. This solves the problem of poor accuracy in detecting blockages at conveyor transfer drop points.

Description

Technical Field

[0001] This invention relates to the field of machine testing and inspection technology, and more specifically, to a method for detecting and automatically clearing blockages at conveyor transfer drop points. Background Technology

[0002] In the static and dynamic balance tests of machine components, the blockage detection and automatic unblocking method at the material drop point of the conveyor effectively avoids abnormal equipment vibration and load imbalance caused by blockage by sensing the material accumulation status in real time and automatically triggering the unblocking action. This ensures the dynamic balance and stability of the conveying system and the overall equipment, and guarantees the accuracy of blockage detection during the testing process.

[0003] Among existing publicly available documents, patent publication number CN102213648A discloses an online simulation test machine for the solid conveying section of a screw extruder and a method for measuring output and pressure. This technology uses a cone sleeve discharge port exposed outside the barrel bushing. Under the thrust of the screw ribs and the guiding effect of the conical surface of the guide cone, the high-pressure material accumulated at the screw head is steadily extruded along the cone sleeve discharge port. This overcomes the defects of existing methods and equipment for detecting the output and pressure of the solid conveying section of a screw extruder, such as material movement trajectory distortion, discharge port blockage, and pressure fluctuations. However, this technology still has the following drawbacks.

[0004] The blockage is static at first, but becomes dynamic during unblocking, with complex state transitions and diverse blockage conditions, resulting in a large amount of data and increasing the difficulty of detection. At the same time, it is questionable whether the specific unblocking effect after the blockage test can be accurately tested, and there is a lack of precise quantitative evaluation of the unblocking effect, making it difficult to determine whether the detection and unblocking measures are truly effective. These factors are intertwined, making it difficult for existing testing methods to comprehensively and accurately capture the characteristics and changes of blockage, resulting in poor accuracy of blockage testing at the conveyor transfer drop point. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for detecting and automatically clearing blockages at conveyor transfer material drop points, comprising the following specific steps:

[0006] S1: Detection system deployment and parameter calibration. Detection units are deployed in the downstream section of the feeding belt at the material transfer point of the conveyor, the inner circumferential direction of the material discharge port, and the upstream section of the receiving belt.

[0007] S2: Normal operating condition baseline data acquisition, control the conveyor to run according to normal operating parameters, continuously collect data from each detection unit for 30-50 minutes, and store it as baseline data;

[0008] S3: Real-time multi-dimensional detection data acquisition. Each detection unit collects the material coverage area and accumulation height of the feeding belt and receiving belt in real time at a sampling frequency of 10-50Hz, the pressure on the inner wall of the discharge port, the material falling trajectory and accumulation height data inside the discharge port, and calculates the real-time relative flow index RFI, the average pressure P_avg, and the real-time material falling speed v_real.

[0009] S4: Multi-parameter fusion blockage judgment and level classification. It performs fusion analysis on real-time detection data, judges whether blockage has occurred based on preset conditions, and classifies the blockage into three levels: mild, moderate and severe according to the degree of blockage.

[0010] S5: Adaptive matching of dredging parameters. Based on the blockage level and real-time detection parameters, the corresponding dredging parameters are matched for the dual-stage adaptive dredging mechanism. The dual-stage adaptive dredging mechanism includes a first-stage high-frequency vibration unit and a second-stage spiral pushing unit.

[0011] S6: Precise automatic dredging execution. Based on the matched dredging parameters, the dredging device is driven to perform dredging operations, and the dredging parameters are updated in real time during the dredging process.

[0012] S7: Verification of dredging effect and secondary treatment. After dredging is completed, verify whether the detection parameters have returned to the normal range, determine the dredging effect, and perform secondary dredging if necessary.

[0013] S8: Data recording and model optimization, storing complete blockage handling data, and regularly optimizing the blockage level judgment threshold and dredging parameter matching model.

[0014] In a preferred embodiment, the detection unit in S1 includes an industrial AI camera, a fiber Bragg grating pressure sensor, and a lidar. Parameters of each detection unit are calibrated, setting the material coverage area recognition error to 1%–2.5%, the pressure measurement error to 0.5%–1%, and the spatial positioning error to 1.8–3 cm. The industrial AI cameras on the downstream section of the feeding belt and the upstream section of the receiving belt are installed at a height of 1.2–1.5 m above the belt surface, with an installation angle of 45–65° to the belt surface. 4–10 sets of fiber Bragg grating pressure sensors are evenly deployed circumferentially inside the discharge port, with the spacing between adjacent sensors being 1 / 4 of the discharge port's circumference. The sensors are embedded in the wear-resistant liner on the inner wall of the discharge port, with the sensing surface flush with the liner surface. The lidar is deployed at the top of the discharge port, with the laser emission direction perpendicularly pointing towards the material falling area inside the discharge port.

[0015] In a preferred embodiment, the normal operating parameter of the conveyor in S2 is the feed belt speed. =1.5m / s, conveyor belt speed =1.8m / s, material flow rate =500t / h, calculate the relative flow index under normal operating conditions. Normal pressure value of the inner wall of the material discharge port Average speed of material falling and the range of fluctuations, including the relative flow index The normal fluctuation range is 0.95-1.05, among which... This refers to the material coverage area of ​​the feed belt under normal operating conditions. This represents the material accumulation height on the feed belt under normal operating conditions. This refers to the material coverage area of ​​the receiving conveyor belt under normal operating conditions. The material accumulation height of the receiving conveyor belt and the normal pressure value of the inner wall of the discharge port under normal operating conditions. Fluctuation ranges from 0.5 to 2 kN, with an average falling speed of the material. =3m / s, with a fluctuation range of 2.5-3.5m / s.

[0016] In a preferred embodiment, the fusion analysis condition in S4 is the Real-Time Relative Flow Index (RFI). A blockage is defined as follows: a pressure of 0.95-1.05 m / s for 3-5 seconds, an average pressure of P_avg of 2-4 kN or a maximum pressure of P_max of 5-10 kN, a material falling velocity of v_real of 1.5-2.5 m / s for 2-6 seconds, and a material accumulation height of H of 0.5-0.8 m.

[0017] In a preferred embodiment, in step S4, mild blockage is defined as meeting one blockage determination condition, with P_avg = 1-3 kN, H = 0.6-1 m, and RFI ∈ [0.8-0.95) ∪ (1.05-1.2]; moderate blockage is defined as meeting two blockage determination conditions, with P_avg = 3-8 kN, H = 1-2 m, and RFI ∈ [0.6-0.8) ∪ (1.2-1.5]; and severe blockage is defined as meeting three or more blockage determination conditions, with P_avg = 8-10 kN, H = 2-5 m, and RFI ∈ [0-0.6) ∪ (1.5-+∞).

[0018] In a preferred embodiment, the unblocking parameters for mild blockage in S5 are: only the first-level vibration unit is activated, with a vibration frequency of 30-50Hz, an amplitude of 8-15mm, and a vibration duration of 10-35s; the unblocking parameters for moderate blockage are: both the first-level and second-level unblocking units are activated, with a vibration frequency of 40-80Hz, an amplitude of 15-25mm, a spiral pushing speed of 30-50r / min, and a pushing depth of 1.5-1.8m; the unblocking sequence is: after 5-10s of vibration, the spiral pushing is activated simultaneously.

[0019] In a preferred embodiment, the unblocking parameters for S5 for severe blockage are as follows: activating the primary and secondary unblocking units, with a vibration frequency of 50-80Hz, an amplitude of 20-30mm, a spiral pushing speed of 50-80r / min, and a pushing depth of 3-3.8m, while simultaneously performing a jetting operation with a jetting pressure of 0.6-1.2MPa and a jetting flow rate of 0.8-1.5m³ / min.

[0020] In a preferred embodiment, during the unblocking process in S6, if the average pressure P_avg decreases by 10%-15% within 3-5 seconds, the vibration frequency increases by 5-75Hz, the screw speed increases by 5-15r / min, and the material accumulation height H decreases at a rate of 0.2-0.8m / s, the pushing depth increases by 0.5-0.9m.

[0021] In a preferred embodiment, the verification standard for the unblocking effect in S7 is that all detection parameters return to the normal operating condition reference range, i.e., RFI∈[0.95-1.05], P_avg∈[0.5-2kN], v_real∈[2.5-3.5m / s], H0.3-0.5m. If the standard is not met, secondary unblocking is performed. The maximum number of secondary unblocking is 3-15 times. If the standard is not met after 3-15 times, the conveyor is controlled to stop and an emergency alarm is issued, with both buzzer and light alarms.

[0022] In a preferred embodiment, in step S8, material characteristics, detection data, and blockage levels are extracted from the case library, and a gradient boosting tree twin model is constructed. If the false positive rate of a certain type of material with mild blockage is 1%-3%, then the RFI is fine-tuned 5-10 times and the P_avg judgment threshold is fine-tuned 3-8 times. If the level unblocking success rate is 90%-95%, then the corresponding initial unblocking parameters are optimized. After optimization, the standard is verified by 3-10 sets of working conditions.

[0023] The technical effects and advantages of this invention are as follows:

[0024] 1. This invention employs multi-dimensional detection unit collaboration and data fusion analysis to effectively capture complex state changes and achieve high detection accuracy. Addressing the complexity of blockage transitioning from static to dynamic, multiple points are simultaneously collected, including an industrial AI camera downstream of the feeding belt, a circumferential fiber optic pressure sensor inside the discharge port, and an industrial AI camera upstream of the receiving belt. These points collect multiple parameters such as material coverage area, accumulation height, inner wall pressure, and material trajectory speed. The relative flow index (RFI) is calculated and combined with pressure and speed thresholds for fusion judgment. Simultaneously, the blockage is categorized into mild, moderate, and severe levels, solving the problems of missed detection with single signals and difficulty in capturing dynamic blockages with static features. The multi-parameter cross-validation mechanism significantly improves the comprehensiveness of feature recognition and the accuracy of state judgment under different materials and diverse blockage conditions, resulting in higher accuracy in blockage testing at conveyor transfer discharge points.

[0025] 2. This invention employs a system that automatically matches the initial parameters such as the starting combination, frequency, and amplitude of the primary vibration unit, secondary spiral pushing unit, and jet unit based on the blockage level. During the unblocking process, the parameters are dynamically adjusted based on feedback from real-time monitoring data such as the average pressure P_avg and the accumulation height H, increasing the vibration frequency and pushing depth. After unblocking, the effect is precisely quantified and evaluated by verifying whether RFI, P_avg, v_real, and H have returned to the normal range. A secondary processing flow is also included to ensure the effectiveness of the "testing and detection - unblocking - verification" closed loop, resulting in higher accuracy in the blockage testing and detection at the conveyor transfer drop point.

[0026] 3. This invention employs a case-based continuous learning optimization model, which enhances the system's adaptability and long-term reliability to different materials and operating conditions. All data from each blockage handling process, including material characteristics, detection data, blockage level, unblocking parameters, and effects, are recorded and incorporated into the case library. The judgment threshold and unblocking strategy are periodically optimized using models such as gradient boosting trees. In the example, after optimization, the false judgment rate was reduced to approximately 2%, and the unblocking success rate was increased to over 93%. This self-learning mechanism enables the system to accumulate and adapt to the blockage characteristics of different materials such as coal, iron ore, and wheat, effectively addressing the challenge of massive data volumes. Verified under multiple operating conditions, it has achieved continuous improvement in detection accuracy and unblocking success rate. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the operation of the conveyor transfer drop point blockage detection and automatic unblocking method of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] like Figure 1 The conveyor transfer drop point blockage detection and automatic unblocking method shown is illustrated in the following three sets of embodiments:

[0030] Example 1: Applicable to coal materials.

[0031] S1: Deployment and parameter calibration of the detection system. Detection units are deployed in the downstream section of the feeding belt, the inner circumferential direction of the discharge port, and the upstream section of the receiving belt. The industrial AI cameras in the upstream sections of the feeding belt and receiving belt are installed at a height of 1.3m above the belt surface and at an angle of 55° to the belt surface. Six sets of fiber optic pressure sensors are evenly deployed in the inner circumferential direction of the discharge port, with the spacing between adjacent sensors being 1 / 4 of the circumference of the discharge port. The sensors are embedded in the wear-resistant liner of the inner wall of the discharge port and the sensing surface is flush with the surface of the liner. The lidar is deployed at the top of the discharge port, with the laser emission direction pointing vertically to the material falling area inside the discharge port. The parameters are calibrated and set with a material coverage area recognition error of 1.8%, a pressure measurement error of 0.7%, and a spatial positioning error of 2.4cm.

[0032] S2: Normal operating condition baseline data acquisition, controlling the conveyor to operate according to normal operating parameters, feed belt speed =1.5m / s, conveyor belt speed =1.8m / s, material flow rate =500t / h, continuously collect data from each detection unit for 40min, store as baseline data, and calculate the baseline parameters for normal operating conditions: =0.8m² material coverage area of ​​the feed conveyor belt =0.3m feed belt material accumulation height, =0.9m² material coverage area of ​​the receiving conveyor belt =0.28m material accumulation height on the receiving conveyor belt, relative flow rate index =(0.8×0.3×1.5) / (0.9×0.28×1.8)≈1.02, Normal pressure value of the inner wall of the discharge port. Fluctuation range 0.5kN, average falling speed of material =3m / s, fluctuation range 3m / s.

[0033] S3: Real-time multi-dimensional detection data acquisition. Each detection unit collects data in real time at a sampling frequency of 30Hz. The material coverage area of ​​the feeding belt is 0.92m², and the stacking height is 0.36m. The material coverage area of ​​the receiving belt is 0.75m², and the stacking height is 0.22m. The pressure on the inner wall of the discharge port is 2.2kN. The material falling trajectory deviates from the normal trajectory by 5cm. The stacking height is 0.7m. Real-time parameters are calculated: relative flow index RFI≈1.18, average pressure P_avg=2.2kN, and real-time material falling speed v_real=2.4m / s.

[0034] S4: Multi-parameter fusion for congestion judgment and level classification. The real-time data is fused and analyzed. The real-time relative flow index RFI is 1.18 and the duration is 4s, which meets one congestion judgment condition. At the same time, P_avg=2.2kN, H=0.7m, and RFI∈(1.05-1.2], so it is judged as a mild congestion.

[0035] S5: Adaptive matching of unblocking parameters. Based on the level of mild blockage, the unblocking parameters are matched, and only the first-level vibration unit is activated. The vibration frequency is 40Hz, the amplitude is 12mm, and the vibration duration is 25s.

[0036] S6: Precise automatic unblocking execution, driving the first-level vibration unit to perform unblocking operations according to matching parameters, monitoring data in real time during the unblocking process, and increasing the vibration frequency by 10Hz according to the rule when the average pressure P_avg decreases by 12% within 4 seconds.

[0037] S7: Verification of dredging effect and secondary treatment. After dredging is completed, the following parameters are verified: RFI=1.01, P_avg=1.2kN, v_real=2.9m / s, H=0.4m. All parameters return to the normal range, indicating that the dredging is effective and no secondary dredging is required.

[0038] S8: Data recording and model optimization. Store all data from this blockage handling, extract coal material characteristics, detection data, and mild blockage level information and include them in the case library. Optimize based on gradient boosting tree twin model. The false judgment rate for mild blockage of this type of coal is 2%. Fine-tune the RFI judgment threshold 7 times and the P_avg judgment threshold 5 times. After optimization, the success rate of clearing mild blockage is 94%, which has been verified to meet the standard under 6 working conditions.

[0039] Example 2: Applicable to iron ore materials.

[0040] S1: Deployment and parameter calibration of the detection system, deployment of the detection unit, installation height of the industrial AI camera is 1.4m, installation angle is 60° with the surface of the belt, 8 sets of fiber optic pressure sensors are evenly deployed on the inner side of the material inlet, and the deployment position of the lidar is the same as in Example 1. Parameter calibration settings: material coverage area recognition error is 2.2%, pressure measurement error is 0.9%, and spatial positioning error is 2.8cm.

[0041] S2: Normal operating condition baseline data acquisition. The normal operating parameters of the conveyor are the same as in Example 1. Data is continuously collected for 45 minutes and stored as baseline data. Baseline parameters are then calculated. =0.9m² =0.35m =1.0m² =0.32m, =(0.9×0.35×1.5) / (1.0×0.32×1.8)≈0.98; Normal pressure value of the inner wall of the material discharge port. Fluctuation of 1kN, average velocity of material falling =3m / s.

[0042] S3: Real-time multi-dimensional detection data acquisition, sampling frequency 40Hz, real-time data acquisition, material coverage area of ​​the feeding belt is 1.05m², stacking height is 0.42m, material coverage area of ​​the receiving belt is 0.68m², stacking height is 0.18m, inner wall pressure of the discharge port is 5kN, material falling speed is 2.1m / s and lasts for 4s, stacking height is 1.5m, real-time parameters are calculated: RFI≈0.72, P_avg=5kN, v_real=2.1m / s.

[0043] S4: Multi-parameter fusion blockage judgment and level classification. Fusion analysis: RFI=0.72 and P_avg=5kN, which meet two blockage judgment conditions. At the same time, H=1.5m and RFI∈(0.6-0.8), it is judged as moderate blockage.

[0044] S5: Adaptive matching of unblocking parameters. Matching the unblocking parameters for moderate blockage, starting the primary vibration unit and the secondary spiral pushing unit. The vibration frequency is 60Hz, the amplitude is 20mm, the spiral pushing speed is 40r / min, and the pushing depth is 1.6m. The unblocking sequence is to start the spiral pushing synchronously after 8s of vibration.

[0045] S6: Precise and automatic dredging execution, performing dredging operations according to parameters, monitoring during dredging: material accumulation height H decreases at a rate of 0.5m / s, and increasing the pushing depth by 0.7m according to rules.

[0046] S7: Verification of dredging effect and secondary treatment. After the first dredging, the verification parameters were: RFI=0.92, close to the normal range; P_avg=2.5kN, slightly exceeding the normal range; v_real=2.6m / s; H=0.6m, slightly exceeding the normal range, failing to meet the verification standard. A second dredging was performed. After the second dredging, the parameters returned to normal: RFI=0.98, P_avg=1.6kN, v_real=2.8m / s; H=0.42m, indicating that the dredging was effective.

[0047] S8: Data recording and model optimization. Store the data processed this time and include it in the iron ore material case library. After model optimization, the misjudgment rate of moderate blockage is 2.5%, and the unblocking success rate is 93%. It has been verified to meet the standards through 8 sets of working conditions.

[0048] Example 3: Applicable to wheat materials.

[0049] S1: Deployment and parameter calibration of the detection system, deployment of the detection unit, installation height of the industrial AI camera is 1.2m, installation angle is 50° with the surface of the belt, 10 sets of fiber optic pressure sensors are evenly deployed on the inner side of the material inlet, and the deployment position of the lidar is the same as in the previous embodiment 1. Parameter calibration settings: material coverage area recognition error is 1.2%, pressure measurement error is 0.6%, and spatial positioning error is 2.0cm.

[0050] S2: Normal operating condition baseline data acquisition. The normal operating parameters of the conveyor are the same as in the previous Example 1. Data is continuously collected for 38 minutes and stored as baseline data. Baseline parameters are then calculated. =1.0m² =0.25m =1.1m² =0.22m, =(1.0×0.25×1.5) / (1.1×0.22×1.8)≈1.01; Normal pressure value of the inner wall of the material discharge port. Fluctuation of 1kN, average velocity of material falling =3m / s.

[0051] S3: Real-time multi-dimensional detection data acquisition, sampling frequency 50Hz, real-time data acquisition, material coverage area of ​​the feeding belt is 1.2m², stacking height is 0.5m, material coverage area of ​​the receiving belt is 0.55m², stacking height is 0.12m, inner wall pressure of the discharge port is 9kN, material falling speed is 2.0m / s for 5s, stacking height is 3.5m, real-time parameters are calculated: RFI≈0.55, P_avg=9kN, v_real=2.0m / s.

[0052] S4: Multi-parameter fusion blockage judgment and level classification. Fusion analysis: RFI=0.55, P_avg=9kN, v_real=2.0m / s and lasts for 5s, which meets the three blockage judgment conditions. At the same time, H=3.5m and RFI∈(0-0.6), it is judged as severe blockage.

[0053] S5: Adaptive matching of unblocking parameters, matching parameters for severe blockage: Start the primary vibration unit and the secondary spiral pushing unit, and simultaneously perform air jet operation, vibration frequency 65Hz, amplitude 25mm, spiral pushing speed 65r / min, pushing depth 3.5m, air jet pressure 0.9MPa, air jet flow rate 1.2m³ / min.

[0054] S6: Precise automatic unblocking execution, executes unblocking operations according to parameters, monitors during unblocking process, the average pressure P_avg decreases by 14% within 3 seconds, increases the vibration frequency by 10Hz to 75Hz, and increases the screw speed by 10r / min.

[0055] S7: Verification of dredging effect and secondary treatment. After the first dredging, the verification parameters were: RFI=0.78, P_avg=4.2kN, v_real=2.4m / s, H=1.8m, which did not meet the standard. After the second dredging, the parameters were: RFI=0.96, P_avg=1.8kN, v_real=2.9m / s, H=0.4m. All parameters returned to the normal range, and the dredging was deemed effective.

[0056] S8: Data recording and model optimization. Store the data processed this time and include it in the wheat material case library. After model optimization, the misjudgment rate of severe blockage is 2.8%, and the unblocking success rate is 95%. It has been verified to meet the standards through 10 sets of working conditions.

[0057] The following data table is derived from the above three sets of embodiments:

[0058]

[0059] Indicator Explanation: 1. Detection accuracy improvement rate = (Detection error before optimization - Detection error after optimization) / Detection error before optimization × 100%, calculated based on parameter calibration error and misjudgment rate after model optimization in the embodiment.

[0060] 2. State transition detection accuracy = (Number of times data is accurately captured during state transition / Total number of transitions) × 100%, reflecting the detection stability during static-dynamic transitions.

[0061] 3. Data processing accuracy = effective data volume / total data volume collected × 100%, solving the problem of processing deviation caused by large data volume.

[0062] 4. The unblocking effect compliance rate = number of successful unblocking attempts / total number of blockages × 100%, which is the optimized data clearly given in the example.

[0063] 5. Parameter regression rate = Number of test groups whose parameters returned to the normal range after unblocking / Total number of test groups × 100%.

[0064] 6. Validation condition compliance rate = Number of validation condition groups / Total number of validation condition groups × 100%.

[0065] In summary, the advantages include the use of multi-dimensional detection units in collaboration and data fusion analysis, which effectively captures complex state changes and achieves high detection accuracy. Addressing the complexity of blockage transitioning from static to dynamic, multiple parameters such as material coverage area, accumulation height, inner wall pressure, and material trajectory speed are simultaneously collected from multiple points, including an industrial AI camera downstream of the feeding belt, a circumferential fiber optic pressure sensor inside the discharge port, and an industrial AI camera upstream of the receiving belt. By calculating the relative flow index (RFI) and combining it with pressure and speed thresholds, a fusion judgment and severity classification (mild, moderate, severe) is performed. This solves the problems of easy missed detection with single signals and difficulty in capturing dynamic blockages with static features. The multi-parameter cross-validation mechanism significantly improves the comprehensiveness of feature recognition and the accuracy of state judgment under different materials and diverse blockage conditions, resulting in higher accuracy in blockage testing at the conveyor transfer discharge point.

[0066] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting and automatically clearing blockages at the material drop point of a conveyor, characterized in that: The specific steps are as follows: S1: Detection system deployment and parameter calibration. Detection units are deployed in the downstream section of the feeding belt at the material transfer point of the conveyor, the inner circumferential direction of the material discharge port, and the upstream section of the receiving belt. S2: Normal operating condition baseline data acquisition, control the conveyor to run according to normal operating parameters, continuously collect data from each detection unit for 30-50 minutes, and store it as baseline data; S3: Real-time multi-dimensional detection data acquisition. Each detection unit collects the material coverage area and accumulation height of the feeding belt and receiving belt in real time at a sampling frequency of 10-50Hz, the pressure on the inner wall of the discharge port, the material falling trajectory and accumulation height data inside the discharge port, and calculates the real-time relative flow index RFI, the average pressure P_avg, and the real-time material falling speed v_real. S4: Multi-parameter fusion blockage judgment and level classification. Real-time detection data is fused and analyzed to determine whether blockage has occurred based on preset conditions, and the blockage is classified into three levels: mild, moderate, and severe. The fusion analysis conditions are: real-time relative flow index (RFI) ∉ 0.95-1.05 and duration 3-5s; average pressure P_avg 2-4kN or maximum pressure P_max 5-10kN; real-time material falling velocity v_real 1.5-2.5m / s and duration 2-6s; material accumulation height H 0.5-0.8m. A blockage is determined by meeting any one of the following conditions: mild blockage is defined as meeting one blockage condition, with P_avg = 1-3 kN, H = 0.6-1 m, and RFI ∈ [0.8-0.95)∪(1.05-1.2]; moderate blockage is defined as meeting two blockage conditions, with P_avg = 3-8 kN, H = 1-2 m, and RFI ∈ [0.6-0.8)∪(1.2-1.5]; and severe blockage is defined as meeting three or more blockage conditions, with P_avg = 8-10 kN, H = 2-5 m, and RFI ∈ [0-0.6)∪(1.5-+∞). S5: Adaptive matching of unblocking parameters. Based on the blockage level and real-time detection parameters, corresponding unblocking parameters are matched for the dual-stage adaptive unblocking mechanism. The dual-stage adaptive unblocking mechanism includes a primary high-frequency vibration unit and a secondary spiral pushing unit. For mild blockage, the unblocking parameters are: only the primary high-frequency vibration unit is activated, with a vibration frequency of 30-50Hz, an amplitude of 8-15mm, and a vibration duration of 10-35s. For moderate blockage, the unblocking parameters are: both the primary high-frequency vibration unit and the secondary spiral pushing unit are activated, with a vibration frequency of 40-80Hz and an amplitude of 15-15mm. 25mm, spiral pushing speed 30-50r / min, pushing depth 1.5-1.8m, the unblocking sequence is to start the spiral pushing synchronously after vibration for 5-10s. The unblocking parameters for severe blockage are: start the first-stage high-frequency vibration unit and the second-stage spiral pushing unit, vibration frequency 50-80Hz, amplitude 20-30mm, spiral pushing speed 50-80r / min, pushing depth 3-3.8m, and simultaneously perform air jet operation, air pressure 0.6-1.2MPa, air flow rate 0.8-1.5m³ / min; S6: Precise automatic dredging execution. Based on the matched dredging parameters, the dredging device is driven to perform dredging operations, and the dredging parameters are updated in real time during the dredging process. S7: Verification of dredging effect and secondary treatment. After dredging is completed, verify whether the detection parameters return to the normal working condition reference range, determine the dredging effect, and perform secondary dredging if the standard is not met. S8: Data recording and model optimization. Store complete blockage handling data and regularly optimize the blockage level judgment threshold and dredging parameter matching model by extracting case library.

2. The method for detecting and automatically clearing blockages at the conveyor transfer drop point according to claim 1, characterized in that: The detection unit in S1 includes an industrial AI camera, a fiber optic pressure sensor, and a lidar. Parameters of each detection unit are calibrated, with material coverage area recognition error set at 1%–2.5%, pressure measurement error at 0.5%–1%, and spatial positioning error at 1.8–3 cm. The industrial AI cameras on the downstream section of the feeding belt and the upstream section of the receiving belt are installed at a height of 1.2–1.5 m above the belt surface, with an installation angle of 45–65° to the belt surface. 4–10 sets of fiber optic pressure sensors are evenly deployed circumferentially inside the discharge port, with the spacing between adjacent sensors being 1 / 4 of the discharge port's circumference. The sensors are embedded in the wear-resistant liner on the inner wall of the discharge port, with the sensing surface flush with the liner surface. The lidar is deployed at the top of the discharge port, with the laser emission direction perpendicularly pointing towards the material falling area inside the discharge port.

3. The method for detecting and automatically clearing blockages at the conveyor transfer drop point according to claim 1, characterized in that: The normal operating parameters of the conveyor in S2 are the feed belt speed. Feeding belt speed Material flow Calculate the relative flow index under normal operating conditions. Normal pressure value of the inner wall of the material discharge port Average speed of material falling and the range of fluctuations, including the relative flow index The normal fluctuation range is 0.95-1.05, among which... This refers to the material coverage area of ​​the feed conveyor belt under normal operating conditions. This represents the material accumulation height on the feed belt under normal operating conditions. This refers to the material coverage area of ​​the receiving conveyor belt under normal operating conditions. The material accumulation height of the receiving conveyor belt and the normal pressure value of the inner wall of the discharge port under normal operating conditions. Fluctuation ranges from 0.5 to 2 kN, with an average falling speed of the material. The fluctuation range is 2.5-3.5 m / s.

4. The method for detecting and automatically clearing blockages at the conveyor transfer drop point according to claim 1, characterized in that: During the unblocking process in S6, if the average pressure P_avg decreases by 10%-15% within 3-5 seconds, the vibration frequency increases by 5-75Hz, the screw speed increases by 5-15r / min, and the material accumulation height H decreases at a rate of 0.2-0.8m / s, the pushing depth increases by 0.5-0.9m.

5. The method for detecting and automatically clearing blockages at the conveyor transfer drop point according to claim 1, characterized in that: The verification standard for the unblocking effect in S7 is that all detection parameters return to the normal operating condition reference range, namely RFI∈[0.95-1.05], P_avg∈[0.5-2kN], v_real∈[2.5-3.5m / s], H∈0.3-0.5m. If the standard is not met, secondary unblocking is performed. The maximum number of secondary unblocking is 3-15 times. If the standard is not met after 3-15 times, the conveyor is controlled to stop and an emergency alarm is issued, with both buzzer and light alarms.

6. The method for detecting and automatically clearing blockages at the material drop point of a conveyor as described in claim 1, characterized in that: In step S8, material characteristics, detection data, and blockage levels are extracted from the case library to construct a gradient boosting tree twin model. If the false positive rate of a certain type of material with mild blockage is 1%-3%, then the RFI is fine-tuned 5-10 times and the P_avg judgment threshold is fine-tuned 3-8 times. If the level of unblocking success rate is 90%-95%, then the corresponding initial unblocking parameters are optimized. After optimization, the standard is verified by 3-10 sets of working conditions.

Images