Coal conveying belt fault early warning method and system

Abstract

The invention discloses a coal conveying belt fault early warning method and system, relates to the field of safety monitoring of coal conveying equipment, and solves the problems of lagging and high false alarm rate of traditional single-parameter early warning. The system comprises a multi-dimensional data acquisition layer, a dynamic feature extraction layer, a belt-material coupling analysis module, a fault judgment layer and an early warning output layer, wherein the acquisition layer acquires parameters such as belt tension, speed, material granularity and weight; noise reduction is carried out on the extraction layer, parameters are normalized, and features are extracted; the coupling module builds a dynamic coupling model to calculate a coupling coefficient, matches a fault feature library and outputs an abnormal signal; determining a layer fault type and a risk level; and the output layer gives early warning information and suggestions. According to the method, the system is applied, belt and material characteristic correlation early warning is achieved through sub-model construction and coupling coefficient calculation, the early warning accuracy and timeliness are improved, and the method is suitable for multiple scenes and low in operation and maintenance cost.

Description

Technical Field

[0001] This invention relates to the field of safety monitoring of coal conveying equipment, specifically to a method and system for early warning of coal conveyor belt failures. Background Technology

[0002] Coal conveyor belts are core equipment for transporting coal in thermal power plants, coal mines, and other enterprises, and their operational stability directly affects production efficiency and operational safety. Currently, early warning systems for coal conveyor belt failures mostly rely on single-parameter monitoring methods, such as using tension sensors to monitor belt tension, using cameras to identify surface damage on the belt, or using speed sensors to determine belt slippage.

[0003] However, traditional early warning methods have significant drawbacks: Faults in coal conveyor belts (such as tearing or misalignment) are often the result of the combined effects of belt operation and material characteristics. For example, when the instantaneous weight of the material is excessive and the particle size is uneven, it can cause a sudden increase in local belt tension, leading to tearing. However, traditional single-parameter monitoring can only capture the result of "sudden increase in tension" and cannot correlate it with the underlying causes of "material weight and particle size," easily resulting in delayed early warnings. Furthermore, when belt parameter fluctuations are caused by normal changes in material characteristics (such as a slight increase in moisture content), traditional methods are prone to misinterpreting them as faults, leading to a high false alarm rate. Therefore, a technical solution that can correlate belt and material characteristics and improve the accuracy of early warnings is urgently needed. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for early warning of coal conveyor belt failures, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a coal conveyor belt fault early warning system, comprising: A multi-dimensional data acquisition layer is used to collect belt parameters and material parameters during the operation of the coal conveyor belt. The belt parameters include belt tension, running speed, and surface temperature, while the material parameters include particle size distribution, instantaneous weight, and moisture content. The dynamic feature extraction layer, connected to the multi-dimensional data acquisition layer, is used to perform noise reduction and normalization processing on the acquired belt parameters and material parameters, and to extract the real-time change features of each parameter. The belt-material coupling analysis module is connected to the dynamic feature extraction layer. It is used to construct a dynamic coupling model between the belt running state and material characteristics, calculate the coupling coefficient between the two, and output coupling abnormal signals based on the coupling coefficient and a preset fault feature library. The fault determination layer, connected to the belt-material coupling analysis module, is used to receive coupling abnormal signals and, in combination with the real-time change characteristics of various parameters, determine the fault type and risk level. The early warning output layer, connected to the fault determination layer, is used to output corresponding early warning information and handling suggestions based on the fault type and risk level.

[0006] Preferably, the multi-dimensional data acquisition layer includes a tension sensor, a speed sensor, an infrared temperature sensor, a particle size analyzer, a weight sensor, and a moisture content detector, which respectively collect belt tension, running speed, surface temperature, material particle size distribution, instantaneous weight, and moisture content.

[0007] Preferably, in the belt-material coupling analysis module, the construction of the dynamic coupling model is based on the belt material characteristic parameters and the material physical characteristic parameters. The belt material characteristic parameters include elastic modulus and friction coefficient, and the material physical characteristic parameters include bulk density and internal friction angle.

[0008] Preferably, the specific working logic of the fault determination layer is as follows: Signal reception and verification: Receive structured coupling anomaly signals output by the coupling analysis module, including coupling coefficient values, anomaly direction (above / below the normal range), preliminary suspected fault markers, and timestamps; verify signal integrity (no missing fields, timestamp deviation from system time ≤ 1s); if an anomaly is found, request retransmission to ensure the validity of the input data; Parameter feature correlation: Based on the signal timestamp, retrieve the core features of the belt and the material correlation features within 5 seconds before and after. The core features of the belt are tension change rate, tension difference between the two sides, and speed fluctuation amplitude. The material correlation features are weight deviation, off-center load, and moisture content deviation. Align the parameter features with the abnormal signal by time to form the parameter change curve before and after the abnormality, and clarify the correlation between the two. Fault type determination: Call the preset fault determination rule library (indexed by suspected fault markers, including core parameter combinations and thresholds). For example, "suspected tear" corresponds to "tension change rate > 8% / s + weight deviation > 15% + idler speed deviation < -10r / min". Verify the thresholds one by one for the core features of the belt and the material-related features: if all / most features are met, the fault type is confirmed (if all 3 are met, the fault is determined to be "confirmed tear fault"). If it is a combination of suspected markers (such as "tear + deviation"), the corresponding rules are verified separately to determine whether it is a single or compound fault. Risk level assessment and result output: The comprehensive score is calculated based on three-dimensional indicators: "fault impact range (percentage of belt length involved), development speed (parameter deterioration rate), and downtime loss," with weights accounting for 40%, development speed accounting for 35%, and downtime loss accounting for 25%. Then, based on preset range thresholds, high, medium, and low risks are classified. The structured judgment results are output to the early warning output layer. The specific judgment structure includes fault type, risk level, core basis, and fault location, and is associated with the stored logs to support subsequent traceability.

[0009] Preferably, a method for early warning of coal conveyor belt failure, applied to a coal conveyor belt failure early warning system, includes the following steps: S1: The multi-dimensional data acquisition layer collects the belt parameters and material parameters of the coal conveyor belt in real time and transmits them to the dynamic feature extraction layer; S2: The dynamic feature extraction layer performs noise reduction and normalization on the collected parameters, extracts the real-time change features of each parameter, and transmits them to the belt-material coupling analysis module. S3: The belt-material coupling analysis module executes a dynamic coupling analysis process and outputs coupling anomaly signals to the fault determination layer; S4: The fault determination layer combines the coupled abnormal signal with the real-time change characteristics of each parameter to determine the fault type and risk level, and transmits it to the early warning output layer. S5: The early warning output layer outputs early warning information and handling suggestions based on the fault type and risk level.

[0010] Preferably, in step S3, the dynamic coupling analysis process of the belt-material coupling analysis module includes: S31: Import the preset belt material characteristic parameters, combine them with the real-time change characteristics of belt parameters extracted from the rated operating speed S2, establish a belt operating state sub-model, and obtain the belt operating state score through the belt operating state sub-model. ; S32: Import the preset material physical property parameters, combine them with the real-time change characteristics of the material parameters extracted in step S2, establish a material property sub-model, and obtain the material property fit score through the material property sub-model. ; S33: Based on the belt conveyor running state sub-model and the material characteristic sub-model, construct a dynamic coupling model and define the coupling coefficient calculation function: ,in The coupling coefficient is... These are the weighting coefficients, and , For real-time belt tension, Rated belt tension, For real-time running speed, For the rated operating speed, This represents the instantaneous weight of the material in real time. The instantaneous weight of the rated material. This is the real-time internal friction angle of the material. This refers to the internal friction angle of standard materials. S34: Real-time calculation of coupling coefficient Set the normal coupling coefficient range ,like If so, it is marked as a coupling anomaly; S35: Call a preset fault feature library, which stores the abnormal range of coupling coefficients and parameter change feature combinations corresponding to different fault types, and then selects the currently abnormal coupling coefficients. The values ​​are matched with the characteristics of parameter changes, and the output includes coupled abnormal signals containing suspected fault types.

[0011] Preferably, the specific construction content of the belt running state sub-model is as follows: based on the belt material property parameters (elastic modulus, friction coefficient), inputting three core features of the dynamic feature extraction layer—tension change rate, speed fluctuation amplitude, and surface temperature deviation—and establishing the model using a multiple linear regression algorithm: ,in Rate the belt's operating status (0-10 points, the higher the score, the more stable it is). The regression coefficients (obtained through training with historical normal operation data) ), The tension change rate (normalized). The amplitude of velocity fluctuation (after normalization). For surface temperature deviation (after normalization); The specific construction content of the material property sub-model is as follows: based on the material's physical property parameters (bulk density, standard internal friction angle), three core features of the dynamic feature extraction layer—weight deviation, particle size distribution uniformity, and moisture content deviation—are input, and the same multiple linear regression algorithm is used to establish the model. ,in The material characteristics are rated for compatibility (0-10 points, the higher the score, the better the compatibility with belt operation). For regression coefficients (obtained through training) ), This is the weight deviation value (after normalization). This represents the uniformity of particle size distribution (after normalization, the smaller the value, the less uniform the distribution). This represents the moisture content deviation (after normalization).

[0012] Compared with the prior art, the beneficial effects of the present invention are: This invention constructs a dynamic coupling model through a belt-material coupling analysis module, linking the belt's operating state (tension, speed, etc.) with material characteristics (weight, particle size, moisture content, etc.). The coupling coefficient quantifies the degree of matching between the two, and combined with a fault feature database, it achieves "causal correlation"-based early warning. For example, when overweight and uneven particle size of the material cause abnormal belt tension, the model can identify the "material cause" in advance through abnormal coupling coefficients, avoiding misjudging simple material fluctuations as faults.

[0013] This invention calculates the fit between the belt and the material in real time using a coupling model. When the belt parameters are below the threshold but the coupling relationship is abnormal (such as the coupling coefficient deviating from the normal range due to uneven material loading), it can output a coupling abnormality signal. Taking belt tearing failure as an example, traditional early warning systems can only alarm after a sudden increase in tension, while this invention can identify potential problems 3-5 minutes in advance through the combination of "high weight deviation - low coupling coefficient" features, allowing maintenance personnel time to adjust the feed rate and check the belt condition, effectively preventing the failure from escalating. This invention's multi-dimensional data acquisition layer supports coal conveyor belts with different bandwidths and speeds. Only the coupling model parameters need to be adjusted based on the belt material characteristics (such as elastic modulus and friction coefficient) and rated parameters (such as rated tension and rated weight), without requiring hardware replacement. It is adaptable to coal conveying equipment in different enterprises such as thermal power plants and coal mines, reducing enterprise modification costs. The tiered early warning system aligns with actual production: the fault judgment layer classifies risks into high, medium, and low risks based on three-dimensional indicators: "impact range, development speed, and downtime loss." The early warning output layer outputs differentiated responses accordingly (high risk triggers emergency shutdown and audible / visual alarms; medium risk prompts adjustment operations; low risk only records logs), avoiding the interference of traditional "one-size-fits-all" early warnings on normal production and balancing safety monitoring with production efficiency. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the system structure of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0015] 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.

[0016] Please see Figure 1-2 This invention provides a technical solution: a coal conveyor belt fault early warning system, comprising: A multi-dimensional data acquisition layer is used to collect belt parameters and material parameters during the operation of the coal conveyor belt. The belt parameters include belt tension, running speed, and surface temperature, while the material parameters include particle size distribution, instantaneous weight, and moisture content. The dynamic feature extraction layer, connected to the multi-dimensional data acquisition layer, is used to perform noise reduction and normalization processing on the acquired belt parameters and material parameters, and to extract the real-time change features of each parameter. The belt-material coupling analysis module is connected to the dynamic feature extraction layer. It is used to construct a dynamic coupling model between the belt running state and material characteristics, calculate the coupling coefficient between the two, and output coupling abnormal signals based on the coupling coefficient and a preset fault feature library. The fault determination layer, connected to the belt-material coupling analysis module, is used to receive coupling abnormal signals and, in combination with the real-time change characteristics of various parameters, determine the fault type and risk level. The early warning output layer, connected to the fault determination layer, is used to output corresponding early warning information and handling suggestions based on the fault type and risk level.

[0017] Furthermore, the multi-dimensional data acquisition layer includes a tension sensor, a speed sensor, an infrared temperature sensor, a particle size analyzer, a weight sensor, and a moisture content detector, which respectively collect data on belt tension, running speed, surface temperature, material particle size distribution, instantaneous weight, and moisture content. For example, the tension sensor collects the current belt tension as 25kN (rated tension 20kN), the weight sensor collects the instantaneous weight of the material as 1.2t (rated weight 1.0t), and the moisture content detector collects the material moisture content as 8% (standard moisture content 5%), and transmits this raw data to the dynamic feature extraction layer in real time.

[0018] Furthermore, in the belt-material coupling analysis module, the dynamic coupling model is constructed based on the belt material properties and the material physical properties. The belt material properties include the elastic modulus and the coefficient of friction, while the material physical properties include the bulk density and the angle of internal friction. Furthermore, the specific working logic of the fault determination layer is as follows: Signal reception and verification: Receive structured coupling anomaly signals output by the coupling analysis module, including coupling coefficient values, anomaly direction (above / below the normal range), preliminary suspected fault markers, and timestamps; verify signal integrity (no missing fields, timestamp deviation from system time ≤ 1s); if an anomaly is found, request retransmission to ensure the validity of the input data; Parameter feature correlation: Based on the signal timestamp, retrieve the core features of the belt and the material correlation features within 5 seconds before and after. The core features of the belt are tension change rate, tension difference between the two sides, and speed fluctuation amplitude. The material correlation features are weight deviation, off-center load, and moisture content deviation. Align the parameter features with the abnormal signal by time to form the parameter change curve before and after the abnormality, and clarify the correlation between the two. Fault type determination: Call the preset fault determination rule library (indexed by suspected fault markers, including core parameter combinations and thresholds). For example, "suspected tear" corresponds to "tension change rate > 8% / s + weight deviation > 15% + idler speed deviation < -10r / min". Verify the thresholds one by one for the core features of the belt and the material-related features: if all / most features are met, the fault type is confirmed (if all 3 are met, the fault is determined to be "confirmed tear fault"). If it is a combination of suspected markers (such as "tear + deviation"), the corresponding rules are verified separately to determine whether it is a single or compound fault. Risk level assessment and result output: A comprehensive score is calculated based on three-dimensional indicators: "fault impact range (percentage of belt length involved), development speed (parameter deterioration rate), and downtime loss," with weights assigned as follows: fault impact range accounts for 40%, development speed accounts for 35%, and downtime loss accounts for 25%. Based on preset range thresholds, risks are categorized as high, medium, and low. Structured judgment results are output to the early warning output layer, with the specific judgment structure including fault type, risk level, core basis, and fault location, and associated with stored logs for subsequent traceability. Specifically, each dimension corresponds to specific quantitative indicators and weights: Fault impact range (weight 40%): The quantitative indicator is "percentage of belt length affected by the fault (unit: %)", which is divided into three levels: "local (<10%), regional (10%-30%), and overall (>30%)". Fault development speed (weight 35%): The quantitative indicator is "parameter characteristic deterioration rate (unit: % / s)", such as the growth rate of tension change rate in tearing failure, which is divided into three levels: "slow (<2% / s), medium (2%-5% / s), and fast (>5% / s)". Downtime loss (weight 25%): The quantitative indicator is "estimated hourly downtime loss due to failure (unit: RMB 10,000)", which is calculated based on the company's historical production data and is divided into three levels: "low (< RMB 50,000), medium (RMB 50,000-150,000), and high (> RMB 150,000)". The system normalizes the quantitative indicators for each assessment dimension (mapping them to 0-100 points), and calculates the comprehensive risk score according to the weights. The formula is as follows: Overall risk score = Fault impact range score × 40% + Fault development speed score × 35% + Downtime loss score × 25%; Risk is categorized into three levels based on the overall risk score: High risk: Overall score > 80 points, corresponding to "rapid development of failure, involving the entire belt, and high estimated downtime losses" (e.g., if a tearing failure is confirmed, the tension change rate increases by 6% / s, involving 40% of the belt length, and the estimated downtime loss is 200,000 yuan). Medium risk: Overall score of 50-80 points, corresponding to "the fault is developing at a moderate speed, the belts in the affected area are affected, and the estimated downtime loss is moderate" (if the belt misalignment fault is confirmed, the tension difference between the two sides increases by 1kN / s, the belt length is affected by 20%, and the estimated downtime loss is 80,000 yuan). Low risk: Overall score < 50 points, corresponding to "slow development of fault, involving local belt, and low estimated downtime loss" (e.g., if wear fault is confirmed, temperature deviation increases by 0.5℃ / s, involving 5% of belt length, and estimated downtime loss of 30,000 yuan).

[0019] Furthermore, a method for early warning of coal conveyor belt failure, applied to a coal conveyor belt failure early warning system, includes the following steps: S1: The multi-dimensional data acquisition layer collects the belt parameters and material parameters of the coal conveyor belt in real time and transmits them to the dynamic feature extraction layer; S2: The dynamic feature extraction layer performs noise reduction and normalization on the collected parameters, extracts the real-time change features of each parameter, and transmits them to the belt-material coupling analysis module. For example, the dynamic feature extraction layer uses a sliding window filter (window size set to 5 seconds) to remove instantaneous interference in the original data, and then uses Min-Max normalization to map the tension of 25kN to 0.8 ((25-15) / (25-15)=1? Correction here: Assuming the rated tension is 20kN and the normal range is 15-25kN, then 25kN normalized is (25-15) / (25-15)=1), with a weight of 1.2t. The mapping is 0.6 ((1.2-0.8) / (1.2-0.8)=1? Correction: Rated weight 1.0t, normal range 0.8-1.2t, then 1.2t normalized is (1.2-0.8) / (1.2-0.8)=1); then the tension change rate of 10% / s (compared to the tension of 22.5kN in the previous second) and the weight deviation value of 20% are extracted, and these feature data are transmitted to the belt-material coupling analysis module; S3: The belt-material coupling analysis module executes a dynamic coupling analysis process and outputs coupling anomaly signals to the fault determination layer; S4: The fault determination layer combines the coupled abnormal signal with the real-time change characteristics of each parameter to determine the fault type and risk level, and transmits it to the early warning output layer. S5: The early warning output layer outputs early warning information and handling suggestions based on the fault type and risk level. Furthermore, in step S3, the dynamic coupling analysis process of the belt-material coupling analysis module includes: S31: Import the preset belt material characteristic parameters, combine them with the real-time change characteristics of belt parameters extracted from the rated operating speed S2, establish a belt operating state sub-model, and obtain the belt operating state score through the belt operating state sub-model. ; S32: Import the preset material physical property parameters, combine them with the real-time change characteristics of the material parameters extracted in step S2, establish a material property sub-model, and obtain the material property fit score through the material property sub-model. ; The specific construction content of the belt operation state sub-model is as follows: Based on the belt material property parameters (elastic modulus, friction coefficient), three core features of the dynamic feature extraction layer are input: tension change rate, speed fluctuation amplitude, and surface temperature deviation. A multiple linear regression algorithm is used to establish the model. ,in Rate the belt's operating status (0-10 points, the higher the score, the more stable it is). The regression coefficients (obtained through training with historical normal operation data) ), The tension change rate (normalized). The amplitude of velocity fluctuation (after normalization). For surface temperature deviation (after normalization); for example, when , , hour, The belt's operating condition was determined to be unstable.

[0020] The specific construction content of the material property sub-model is as follows: based on the material's physical property parameters (bulk density, standard internal friction angle), three core features of the dynamic feature extraction layer—weight deviation, particle size distribution uniformity, and moisture content deviation—are input, and the same multiple linear regression algorithm is used to establish the model. ,in The material characteristics are rated for compatibility (0-10 points, the higher the score, the better the compatibility with belt operation). For regression coefficients (obtained through training) ), This is the weight deviation value (after normalization). This represents the uniformity of particle size distribution (after normalization, the smaller the value, the less uniform the distribution). This represents the moisture content deviation (after normalization). For example, when... , , hour, The results indicate that the material characteristics are poorly matched with the belt.

[0021] S33: Based on the belt conveyor running state sub-model and the material characteristic sub-model, construct a dynamic coupling model and define the coupling coefficient calculation function: ,in The coupling coefficient is... These are the weighting coefficients, and , For real-time belt tension, Rated belt tension, For real-time running speed, For the rated operating speed, This represents the instantaneous weight of the material in real time. The instantaneous weight of the rated material. This is the real-time internal friction angle of the material. This refers to the internal friction angle of standard materials. S34: Real-time calculation of coupling coefficient Set the normal coupling coefficient range ,like If so, it is marked as a coupling anomaly; combining the above example, let's assume... , , , , ,but The normal coupling coefficient range is set as follows: (Based on historical normal data statistics) Currently This was determined to be a coupling anomaly.

[0022] S35: Call the preset fault feature library. The fault feature library stores the abnormal range of coupling coefficients and parameter change feature combinations corresponding to different fault types. The current abnormal coupling coefficients will be... The values ​​are matched with the characteristics of parameter changes, and the output includes coupled abnormal signals containing suspected fault types.

[0023] This invention discloses a method and system for early warning of coal conveyor belt faults, relating to the field of safety monitoring of coal conveying equipment, aiming to solve the problems of delayed early warning and high false alarm rate in traditional single-parameter early warning systems. The system includes a multi-dimensional data acquisition layer, a dynamic feature extraction layer, a belt-material coupling analysis module, a fault determination layer, and an early warning output layer: the multi-dimensional data acquisition layer collects belt parameters such as belt tension and speed, and material parameters such as particle size and weight; the dynamic feature extraction layer reduces noise and normalizes the parameters and extracts real-time change features; the belt-material coupling analysis module constructs a dynamic coupling model, calculates the coupling coefficient, and matches it with a fault feature library to output coupled abnormal signals; the fault determination layer combines abnormal signals and parameter features to determine the fault type and risk level; and the early warning output layer outputs corresponding early warning information and handling suggestions.

[0024] The method utilizes the aforementioned system, and the steps are as follows: collecting and transmitting parameters; processing parameters and extracting features; performing coupling analysis to output abnormal signals; determining the fault and risk level; and outputting early warnings and suggestions. Specifically, the coupling analysis establishes a belt conveyor operating state sub-model and a material characteristic sub-model, and uses a coupling coefficient calculation function to determine anomalies, achieving early warnings that correlate belt and material characteristics. This improves the accuracy and timeliness of early warnings, is adaptable to multiple scenarios, and has low operation and maintenance costs.

[0025] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A coal conveyor belt fault early warning system, characterized in that, include: A multi-dimensional data acquisition layer is used to collect belt parameters and material parameters during the operation of the coal conveyor belt. The belt parameters include belt tension, running speed, and surface temperature, while the material parameters include particle size distribution, instantaneous weight, and moisture content. The dynamic feature extraction layer, connected to the multi-dimensional data acquisition layer, is used to perform noise reduction and normalization processing on the acquired belt parameters and material parameters, and to extract the real-time change features of each parameter. The belt-material coupling analysis module is connected to the dynamic feature extraction layer. It is used to construct a dynamic coupling model between the belt running state and material characteristics, calculate the coupling coefficient between the two, and output coupling abnormal signals based on the coupling coefficient and a preset fault feature library. The fault determination layer, connected to the belt-material coupling analysis module, is used to receive coupling abnormal signals and, in combination with the real-time change characteristics of various parameters, determine the fault type and risk level. The early warning output layer, connected to the fault determination layer, is used to output corresponding early warning information and handling suggestions based on the fault type and risk level.

2. The coal conveyor belt fault early warning system according to claim 1, characterized in that: The multi-dimensional data acquisition layer includes a tension sensor, a speed sensor, an infrared temperature sensor, a particle size analyzer, a weight sensor, and a moisture content detector, which respectively collect belt tension, running speed, surface temperature, material particle size distribution, instantaneous weight, and moisture content.

3. The coal conveyor belt fault early warning system according to claim 1, characterized in that: In the belt-material coupling analysis module, the dynamic coupling model is constructed based on the belt material characteristic parameters and the material physical characteristic parameters. The belt material characteristic parameters include elastic modulus and friction coefficient, and the material physical characteristic parameters include bulk density and internal friction angle.

4. The coal conveyor belt fault early warning system according to claim 1, characterized in that: The specific working logic of the fault determination layer is as follows: Signal reception and verification: Receive the structured coupling anomaly signal output by the coupling analysis module, including coupling coefficient value, anomaly direction, preliminary suspected fault marker, and timestamp; Verify signal integrity; if an anomaly is detected, request a retransmission to ensure the validity of the input data. Parameter feature correlation: Based on the signal timestamp, retrieve the core features of the belt and the material-related features within 5 seconds before and after the signal. The core features of the belt are the tension change rate, the tension difference between the two sides, and the speed fluctuation amplitude. The material-related features are the weight deviation, the off-center load, and the moisture content deviation. Align the parameter features with the abnormal signal by time to form the parameter change curve before and after the abnormality, and clarify the correlation between the two. Fault type determination: Call the preset fault determination rule library; check the threshold one by one for the core features of the belt and the material association features: if all / most features are met, the fault type is confirmed; If the suspected flag is a combination of flags, then the corresponding rules are checked separately to determine whether it is a single or compound fault. Risk level assessment and result output: The comprehensive score is calculated based on the three-dimensional indicators of "fault impact range, development speed, and downtime loss" and their weights. The fault impact range accounts for 40% of the weight, the development speed accounts for 35%, and the downtime loss accounts for 25%. Then, based on the preset range threshold, the risk is divided into high, medium, and low risk. The structured judgment results are output to the early warning output layer. The specific judgment structure includes the fault type, risk level, core basis, and fault location, and is associated with the storage logs to support subsequent traceability.

5. A method for early warning of coal conveyor belt failure, applied to the coal conveyor belt failure early warning system according to any one of claims 1-4, characterized in that, Includes the following steps: S1: The multi-dimensional data acquisition layer collects the belt parameters and material parameters of the coal conveyor belt in real time and transmits them to the dynamic feature extraction layer; S2: The dynamic feature extraction layer performs noise reduction and normalization on the collected parameters, extracts the real-time change features of each parameter, and transmits them to the belt-material coupling analysis module. S3: The belt-material coupling analysis module executes a dynamic coupling analysis process and outputs coupling anomaly signals to the fault determination layer; S4: The fault determination layer combines the coupled abnormal signal with the real-time change characteristics of each parameter to determine the fault type and risk level, and transmits it to the early warning output layer. S5: The early warning output layer outputs early warning information and handling suggestions based on the fault type and risk level.

6. The method for early warning of coal conveyor belt failure according to claim 5, characterized in that: In step S3, the dynamic coupling analysis process of the belt-material coupling analysis module includes: S31: Import the preset belt material characteristic parameters, combine them with the real-time change characteristics of belt parameters extracted from the rated operating speed S2, establish a belt operating state sub-model, and obtain the belt operating state score through the belt operating state sub-model. ; S32: Import the preset material physical property parameters, combine them with the real-time change characteristics of the material parameters extracted in step S2, establish a material property sub-model, and obtain the material property fit score through the material property sub-model. ; S33: Based on the belt conveyor running state sub-model and the material characteristic sub-model, construct a dynamic coupling model and define the coupling coefficient calculation function: ,in The coupling coefficient is... These are the weighting coefficients, and , For real-time belt tension, Rated belt tension, For real-time running speed, For the rated operating speed, This represents the instantaneous weight of the material in real time. The instantaneous weight of the rated material. This is the real-time internal friction angle of the material. This refers to the internal friction angle of standard materials. S34: Real-time calculation of coupling coefficient Set the normal coupling coefficient range ,like If so, it is marked as a coupling anomaly; S35: Call a preset fault feature library, which stores the abnormal range of coupling coefficients and parameter change feature combinations corresponding to different fault types, and then selects the currently abnormal coupling coefficients. The values ​​are matched with the characteristics of parameter changes, and the output includes coupled abnormal signals containing suspected fault types.

7. The method for early warning of coal conveyor belt failure according to claim 6, characterized in that: The specific construction content of the belt operation state sub-model is as follows: based on the belt material characteristic parameters, three core features of the dynamic feature extraction layer are input: tension change rate, speed fluctuation amplitude, and surface temperature deviation. A multiple linear regression algorithm is used to establish the model. ,in The belt operating condition is scored. For regression coefficients, The rate of change of tension. For the speed fluctuation range, For surface temperature deviation; The specific construction content of the material characteristic sub-model is as follows: based on the physical property parameters of the material, three core features are input into the dynamic feature extraction layer: weight deviation value, particle size distribution uniformity, and moisture content deviation. The same multiple linear regression algorithm is used to establish the model. ,in Scoring the material property suitability. For regression coefficients, This is the weight deviation value. For particle size distribution uniformity, This is due to moisture content deviation.

Images