Wharf belt conveyor steel wire leakage early warning method, device and equipment and storage medium

Abstract

The invention provides a wharf belt conveyor steel wire leakage early warning method, device and equipment and a storage medium, and the method comprises the steps: carrying out the edge detection and texture comparison through collecting a belt surface image, and generating a steel wire texture definition degradation rate; meanwhile, infrared radiation data of a joint area are collected, and the thermal radiation intensity of the joint is obtained through normalization processing and thermal radiation energy calculation. And calculating wear, stress concentration and environmental attenuation coefficients in combination with the standardized belt state characteristic matrix, and inputting a belt health degradation model to generate a health degradation index. And determining a curve gradient coefficient and a risk offset according to the service life of the belt, the environmental corrosivity parameter and the historical missing report event. And finally, through product calculation and Sigmoid function conversion, a steel wire exposure risk index is generated, dynamic quantitative evaluation, accurate pre-judgment and advanced early warning are realized, safe operation of the belt conveyor is guaranteed, and fault loss is reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of conveyors, and particularly relates to a port belt conveyor wire leakage early warning method, device, equipment and storage medium. BACKGROUND

[0002] As the core equipment for dry bulk cargo transfer, the belt conveyor bears the heavy responsibility of continuous and efficient conveying. However, under the long-term high load and harsh environmental conditions, the conveyor belt is prone to a series of problems such as wear, stress concentration and aging, among which the wire exposure is particularly critical. It is not only a potential cause of belt rupture and shutdown maintenance, but also may cause serious safety accidents, directly affecting the operation efficiency and economic benefits of the port.

[0003] Traditionally, the condition monitoring of the belt conveyor mainly relies on manual inspection or data collection of a single type of sensor, such as monitoring the belt thickness with a laser thickness gauge, collecting the running tension with a tension sensor, and detecting the local temperature with an infrared thermal imager, and making fault judgments based on experience and setting fixed thresholds. Although this method can find faults to some extent, it has significant limitations: first, the data collection dimension is single, and it cannot fully reflect the health status of the belt; second, the risk assessment model is simple, and it mainly uses linear weighting or fixed formula calculation, which is difficult to accurately quantify the potential risks under complex working conditions; third, the early warning mechanism has poor adaptability, and the static early warning threshold does not consider the cumulative influence of working condition changes and historical data, resulting in frequent false positives and false negatives.

[0004] Especially for the serious fault of wire exposure, the traditional method is difficult to accurately predict. Wire exposure is often accompanied by a gradual degradation process caused by belt wear, stress concentration and environmental factors, which involves multi-dimensional and nonlinear interaction of influencing factors.

[0005] Therefore, how to effectively integrate and dynamically quantify the multi-source heterogeneous data becomes the key to improving the accuracy of wire exposure risk early warning. SUMMARY

[0006] The application aims to overcome the defects in the prior art and provide a port belt conveyor wire leakage early warning method, device, equipment and storage medium.

[0007] The application provides a port belt conveyor wire leakage early warning method, which comprises the following steps:

[0008] Collecting the image of the belt surface of the belt conveyor, performing edge detection and texture contrast on the image of the belt surface, and generating a wire texture definition degradation rate;

[0009] Collecting infrared radiation data of the belt joint area, performing normalization processing and thermal radiation energy calculation operation on the infrared radiation data to generate joint thermal radiation intensity;

[0010] Obtaining a standardized belt state feature matrix, performing wear feature coefficient calculation, stress concentration coefficient calculation and environmental attenuation coefficient calculation on the standardized belt state feature matrix, inputting the calculation results into a belt health degradation model to generate a belt health degradation index;

[0011] Based on the service life of the belt and the environmental corrosive parameters, the curve steepness coefficient is determined;

[0012] Based on the historical false alarm events and the environmental corrosive parameters, the risk offset is determined;

[0013] According to the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient and the risk offset, the product calculation and Sigmoid function conversion operation are performed to generate a steel wire exposure risk index.

[0014] Optionally, the belt surface image is subjected to edge detection and texture contrast to generate a steel wire texture definition degradation rate, comprising:

[0015] Performing edge detection operation on the belt surface image to extract steel wire texture edge;

[0016] Comparing the extracted texture edge with a standard texture template to calculate texture edge blurring degree and contrast attenuation amount;

[0017] According to the texture edge blurring degree and contrast attenuation amount, the steel wire texture definition degradation rate is generated.

[0018] Optionally, the infrared radiation data is subjected to normalization processing and thermal radiation energy calculation operation to generate joint thermal radiation intensity, comprising:

[0019] Performing median filtering operation on the infrared radiation data to eliminate environmental heat source interference;

[0020] According to the filtered data, the total value of joint area thermal radiation is calculated;

[0021] According to the total value of thermal radiation and joint area, the unit area thermal radiation intensity is generated.

[0022] Optionally, the calculation results are input into a belt health degradation model to generate a belt health degradation index, comprising:

[0023] Obtaining scratch density and real-time belt speed;

[0024] when the first proportion of the scratch density is greater than the critical density of the scratch, updating a wear risk weight according to the scratch density and a real-time belt speed;

[0025] inputting the updated weight into a belt health degradation model to generate a belt health degradation index.

[0026] Optionally, based on historical false alarm events and environmental corrosive parameters, a risk offset is determined, including:

[0027] determining a baseline offset according to the environmental corrosive parameters;

[0028] when a false alarm event occurs, reducing the risk offset by a fixed value.

[0029] Optionally, after the steel wire exposure risk index is generated, the method further includes:

[0030] when the steel wire exposure risk index reaches an emergency threshold, fusing a tension anomaly area and a thermal imaging coordinate;

[0031] outputting a steel wire exposure position coordinate according to the fusion result.

[0032] Optionally, based on a belt service life and environmental corrosive parameters, a curve steepness coefficient is determined, including:

[0033] determining an initial curve steepness coefficient according to the belt service life;

[0034] when the environmental corrosive parameters indicate a high salt mist area, increasing the curve steepness coefficient by a fixed proportion.

[0035] The application also provides a port belt conveyor steel wire leakage early warning device, including:

[0036] an image module that collects a belt surface image of the belt conveyor, performs edge detection and texture contrast on the belt surface image, and generates a steel wire texture definition degradation rate;

[0037] an infrared module that collects infrared radiation data of a belt joint area, performs normalization processing and thermal radiation energy calculation operations on the infrared radiation data, and generates joint thermal radiation intensity;

[0038] a model module that acquires a standardized belt state feature matrix, performs wear feature coefficient calculation, stress concentration coefficient calculation, and environmental attenuation coefficient calculation on the standardized belt state feature matrix, inputs the calculation results into a belt health degradation model, and generates a belt health degradation index;

[0039] a curve module that determines a curve steepness coefficient based on a belt service life and environmental corrosive parameters;

[0040] an offset module configured to determine a risk offset based on historical false alarm events and environmental corrosive parameters;

[0041] an index module configured to perform a product calculation and a Sigmoid function conversion operation according to the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient and the risk offset, and generate a steel wire exposure risk index.

[0042] The application also provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0043] The application also provides a computer readable storage medium, which stores a computer program, and the computer program, when executed in a computer, causes the computer to execute the method described above.

[0044] The application has the following beneficial effects:

[0045] The application provides a wharf belt conveyor steel wire leakage early warning method, which comprises the following steps: collecting a belt surface image of a belt conveyor, performing edge detection and texture contrast on the belt surface image, and generating a steel wire texture definition degradation rate; collecting infrared radiation data of a belt joint area, performing normalization processing and thermal radiation energy calculation operation on the infrared radiation data, and generating a joint thermal radiation intensity; obtaining a standardized belt state feature matrix, performing wear feature coefficient calculation, stress concentration coefficient calculation and environmental attenuation coefficient calculation on the standardized belt state feature matrix, inputting the calculation results into a belt health degradation model, and generating a belt health degradation index; determining a curve steepness coefficient based on a belt service life and environmental corrosive parameters; determining a risk offset based on historical false alarm events and environmental corrosive parameters; and performing a product calculation and a Sigmoid function conversion operation according to the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient and the risk offset, and generating a steel wire exposure risk index. The application integrates multiple source data through AI technology, dynamically quantitatively evaluates the belt health state, accurately predicts the steel wire exposure risk, reduces false alarms and missed alarms, realizes advanced early warning and graded response, guarantees the safe and stable operation of the wharf belt conveyor, and reduces fault loss. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 is a wharf belt conveyor steel wire leakage early warning process schematic diagram in the application;

[0047] Figure 2 is a wharf belt conveyor steel wire leakage early warning device schematic diagram in the application. DETAILED DESCRIPTION

[0048] Exemplary embodiments of the present disclosure will be provided in greater detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it is understood that various forms implement the present disclosure and should not be limited by the embodiments set forth herein. Rather, the embodiments are provided to enable a more thorough understanding of the present disclosure and to convey the full scope of the present disclosure to those skilled in the art.

[0049] The present application provides a wharf belt conveyor wire leakage early warning method, which is applied to the field of belt conveyor state monitoring and used for real-time early warning of the risk of belt conveyor belt wire exposure, thereby ensuring safe operation of the equipment.

[0050] As shown in Figure 1 , the wharf belt conveyor wire leakage early warning method provided by the present application comprises:

[0051] S101, collecting a belt surface image of the belt conveyor, performing edge detection and texture contrast on the belt surface image, and generating a wire texture clarity degradation rate;

[0052] At key monitoring points of the conveyor running path, such as the bearing section and the backhaul section clean area, a high-frame-rate industrial camera is deployed, and a ring-shaped LED fill light is used to ensure image clarity. The camera focuses on the area where the wire exposure risk may occur, and real-time microscopic images of the belt surface and the shallow layer are taken. When collecting, the joint and the repair area are avoided, and an effective detection area of 300x300 pixels is selected.

[0053] Through image processing techniques such as edge detection algorithms, the real-time collected wire texture image is compared with a standard wire texture template (i.e., a wire texture image under the perfect state of the belt), and indexes such as texture edge blur and contrast attenuation are calculated. Finally, the degradation proportion of the texture clarity, i.e., the wire texture clarity degradation rate, is quantified. .

[0054] The edge detection operation is used to extract the wire texture edge. After the image is collected by the high-frame-rate industrial camera, an image processing algorithm is applied to identify the texture contour. The texture contrast involves comparing the extracted edge data with the standard template to calculate the blur and attenuation. For example, when the number of scratches on the belt surface increases, the edge blur will increase, and the contrast attenuation will rise, reflecting the degree of material degradation. The generation of the wire texture clarity degradation rate is based on quantitative indicators, such as when the texture edge blur exceeds 10%, it is increased by 0.15 to match the actual risk. The texture edge blur is calculated based on pixel gradient changes, and the contrast attenuation is generated through color contrast analysis; a 10% increase in blur corresponds to a 0.15 increase, reflecting the degree of material degradation.

[0055] ​​​S102, collect infrared radiation data of the belt joint area, perform normalization processing and thermal radiation energy calculation operation on the infrared radiation data to generate joint thermal radiation intensity;

[0056] In the conveyor joint area, such as the belt lap joint, the infrared thermal imager is deployed at the vulcanization joint position to monitor the temperature field distribution of the joint area in real time at a high frequency of ≥1 Hz. The infrared thermal imager captures the infrared radiation signal of the joint position and converts it into a temperature distribution image, and extracts the thermal radiation energy value of the joint area. The collected thermal radiation energy value is normalized, and the thermal conduction characteristics of the joint material are combined to calculate the thermal radiation intensity per unit area, i.e. the joint thermal radiation intensity .

[0057] The normalization processing operation includes performing median filtering operation on the infrared radiation data to eliminate environmental heat source interference; for example, applying median filtering to remove abnormal points and retain true thermal radiation characteristics. The thermal radiation energy calculation operation calculates the total thermal radiation value of the joint area based on the filtered data, and then generates the thermal radiation intensity per unit area based on the total thermal radiation value and the joint area ; the joint area is a fixed value (such as 0.5 m²), and the total thermal radiation value is divided by the area . The median filtering operation is used to remove abnormal points of environmental heat source interference and retain true thermal radiation characteristics; for example, applying median filtering to remove temperature abnormal values caused by environmental heat sources in high temperature environments.

[0058] S103, obtain a standardized belt state feature matrix, perform wear feature coefficient calculation, stress concentration coefficient calculation and environmental attenuation coefficient calculation on the standardized belt state feature matrix, input the calculation results into a belt health degradation model to generate a belt health degradation index.

[0059] The standardized belt state feature matrix is generated by a data fusion preprocessing module, which performs time-space alignment, noise filtering and missing value interpolation on the belt health state time series data set collected by the multi-source heterogeneous data collection module.

[0060] The time-space alignment is based on the system unified clock in time, and all data are stamped with accurate time stamps and calibrated for delay to ensure that the data set at the same time corresponds to the same running state of the belt, such as synchronizing longitudinal temperature difference gradient, scratch density, real-time tension, etc. to the same time axis; according to the conveyor running speed, equipment installation position and belt motion trajectory, the spatial mapping relationship of each collection point is established, such as calculating the real-time position of the belt by the driving drum speed, associating different position sensor data with specific spatial areas of the belt, such as the load section and the joint area, to realize the unity of data in time and space dimensions.

[0061] The noise filtering, for sensor data such as real-time tension, relative humidity, real-time belt speed, adopts sliding average or Kalman filtering to process high-frequency random noise, separates and eliminates periodic interference through Fourier transform; for image data such as scratch density, RGB color vector, steel wire texture definition degradation rate, uses Gaussian filtering or morphological filtering to remove image interference and reduce the influence of motion blur caused by belt running; for thermal radiation data such as joint thermal radiation intensity, adopts median filtering to remove abnormal points of environmental heat source interference and retains the real thermal radiation characteristics.

[0062] The missing value interpolation, when data is missing due to equipment failure, transmission interruption or harsh environment, is interpolated based on relevance: time series data such as real-time belt speed and real-time tension adopts linear or polynomial interpolation, and if it has periodicity, it is filled with historical data of the same period; data with strong correlation with other features, such as thickness difference and scratch density, are predicted by known features using a trained regression model such as random forest; short-term missing of key features is filled forward, and long-term missing is marked as low confidence data and reduced in weight in subsequent analysis.

[0063] After time and space alignment, noise filtering and missing value interpolation, the features of different dimensions, such as thickness difference in mm and belt speed in m / s, are normalized by min-max to map them to the interval [0, 1], and then a matrix is constructed with time as the row index and each feature, such as the absolute difference of nominal thickness △h, real-time belt speed v, scratch density , longitudinal temperature difference gradient △T, as the column index.

[0064] The wear feature coefficient calculation is performed on the normalized belt state feature matrix to generate the wear feature coefficient K w , and the calculation formula is:

[0065]

[0066] Where k is the scratch sensitivity coefficient, λ is the material attenuation factor, △h is the absolute difference between the current thickness and the nominal thickness, h0 is the thickness damage threshold, v is the real-time belt speed, v max is the safe belt speed threshold, S d is the scratch density, is the critical scratch density, and e is the natural constant.

[0067] The scratch sensitivity coefficient is used to quantify the influence of scratch density on the wear feature coefficient, and its setting is based on the comprehensive results of material properties, historical failure data and experimental calibration. Specifically, the value of k is determined by analyzing the correlation between scratch density S0 and steel wire exposure risk in historical belt failure cases:

[0068] When the scratch density approaches the critical value , such as a certain type of belt = 2 cm 2 When the failure rate shows a steep increase trend, the k value is fitted to meet the Sigmoid function:

[0069]

[0070] Material properties directly affect the value of k. For example, the k value of a rubber cover layer with a thick belt is low due to strong scratch resistance, and the typical range is 0.5-0.8, while the k value of a fabric core or thin layer belt needs to be increased to 1.2-1.8 due to high scratch sensitivity. Further laboratory calibration simulates different scratch densities in a controlled environment and measures the belt strength decay rate, and the k value is optimized using regression analysis to ensure accurate matching of the actual wear response curve. When initially deployed, k is preset with an initial value based on the material safety data sheet provided by the supplier, such as the wear resistance index, and is dynamically updated based on real-time scratch monitoring data during operation. The material attenuation factor λ is used to represent the change in belt thickness: And the ratio of belt speed to safety belt speed threshold: The composite attenuation effect of the wear characteristic coefficient K w is set depending on material fatigue tests, working environment adaptation, and dynamic belt speed correction. Material fatigue tests measure the quantitative relationship between belt thickness loss and residual strength through accelerated aging tests such as cyclic tension and thermal oxidative aging: when the thickness loss reaches a certain threshold, such as: = 10%, if the measured strength decay rate is 30%, then λ ≈ 1.2 is fitted to meet the exponential decay model:

[0071] ×

[0072] The working environment has a significant impact on the value of λ. In high wear scenarios such as iron ore transportation, a larger λ value is required, typically in the range of 1.3-1.8 to strengthen thickness sensitivity, while in mild environments such as coal transportation, it can be reduced to 0.5-0.9. At the same time, λ needs to be dynamically adapted to the belt speed, and when the real-time belt speed v approaches the safety threshold, λ is scaled up to highlight the risk of thickness loss under high-speed working conditions. At the same time, λ is continuously calibrated through real-time thickness monitoring data and is re-calibrated based on the fatigue life data provided by the supplier when the belt batch is changed.

[0073] Perform stress concentration coefficient calculation on the standardized belt state feature matrix to generate the stress concentration coefficient . The calculation formula is:

[0074]

[0075] Where α is the thermal stress sensitivity factor, β is the deformation amplification coefficient, and γ is the tension fluctuation gain coefficient, is a longitudinal temperature difference gradient, is a reference temperature, is an edge waviness standard deviation, is a real-time tension, is an average running tension, is a hyperbolic tangent function, is a natural logarithm.

[0076] The thermal stress sensitivity factor a is used to quantify the longitudinal temperature difference gradient:

[0077]

[0078] The nonlinear influence strength of the belt thermal stress is set based on the material thermal expansion characteristics, infrared thermal imaging calibration, and working condition adaptation. Specifically, the thermal expansion coefficients δ of each layer of the material, such as the rubber covering layer and the steel wire rope core, are determined through material thermodynamic experiments, combined with the Young's modulus E, to establish a temperature difference-stress conversion relationship:

[0079]

[0080] The fitting is:

[0081]

[0082] In the working condition adaptation stage, the infrared thermal imager is used to collect longitudinal temperature distribution data under typical loads, and the a value is optimized through regression analysis: for high-throughput iron ore wharfs that are frequently subjected to thermal shock, a is taken as 1.8-2.5 to strengthen the temperature difference response; for mild working conditions, such as coal transportation, a is reduced to 0.7-1.2; the initial value of the deformation amplification coefficient β is set as 0.15 mm^(-2) and can be adjusted, such as calculating the real-time σ w based on the transverse profile data collected by the laser thickness gauge. w When σ t exceeds the deformation risk critical value, β is proportionally increased by 20%-50% to match the high deformation risk scenario; the deformation risk critical value is set as 0.3 mm; the value of the tension fluctuation gain coefficient γ is set in the range of 0.8-1.2, wherein for high-impact load working conditions, such as iron ore wharfs, γ=1.2 is taken to strengthen the tension sudden change sensitivity, and for mild working conditions, such as coal transportation, γ=0.8 is taken to reduce the false alarm rate, when |F avg |>0.2F avg , γ is forced to increase to 1.5-2.0 to trigger a strong warning response.

[0083] An environmental attenuation coefficient calculation is performed on the standardized belt state feature matrix to generate an environmental attenuation coefficient The calculation formula is:

[0084]

[0085] wherein w is an environmental reference coefficient, η is a humidity gain coefficient, RH is a relative humidity, is a current RGB color vector, is a standard working condition color vector, is a color difference tolerance, max is a maximum function.

[0086] The environmental reference coefficient w sets an initial range according to the grade of the belt cover material: w = 0.9-1.0 for high-weather-resistant materials, and w = 0.7-0.8 for ordinary natural rubber; secondly, combined with historical environmental factors such as acid rain and high temperature leading to belt aging cases, a reference attenuation rate is fitted to calibrate the initial value; finally, dynamic adjustment is made according to the deployment environment: the upper limit w = 1.0 for mild dry environments, and the lower limit w = 0.75 for high-corrosion environments such as coastal iron ore terminals, to strengthen the influence of environmental degradation; the humidity gain coefficient η, in high-humidity environments, i.e. RH > 80%, takes η = 1.8-2.2 to strengthen the humidity penalty term, and in dry environments, i.e. RH < 30%, takes η = 0.3-0.5 to weaken the influence, and a dynamic adjustment mechanism is introduced during operation: if RH > 70% is continuously monitored for more than 2 hours, the sensitivity is immediately increased by η new = η × (1+0.2 × RH), the weight is increased by 20% in the rainy season and decreased by 10% in the dry season, to realize real-time response to humidity risk.

[0087] The wear characteristic coefficient K w , the stress concentration coefficient K s and the environmental degradation coefficient K e are input into the belt health degradation model to generate the belt health degradation index P. The calculation formula is:

[0088]

[0089] wherein a is a wear risk weight, b is a stress risk weight, c is an environmental risk weight, a+b+c = 1 and a, b, c ∈ [0, 1], P is a belt health degradation index, K w is a wear characteristic coefficient, K s is a stress concentration coefficient, K e is an environmental degradation coefficient, arctan is an inverse tangent function, π is a circular constant, and e is a natural constant.

[0090] The generation of the belt health degradation index P includes obtaining the scratch density S d and the real-time belt speed v. When the scratch density S d is greater than 80% of the critical scratch density S0, the wear risk weight a is updated according to S d and v.

[0091] The initial value ranges for the wear risk weight 'a', stress risk weight 'b', and environmental risk weight 'c' are set as a = 0.5~0.7, b = 0.2~0.3, and c = 0.1~0.2, respectively; these values ​​will be adjusted later through a dynamic adjustment mechanism, specifically:

[0092] When S d When >0.8×S0:

[0093]

[0094] Where f(v) is the band speed correction factor, specifically:

[0095]

[0096] when or hour, ;

[0097] When RH>80% or >0.8× hour, ;

[0098] Forced reallocation after each adjustment:

[0099]

[0100]

[0101]

[0102] in, As the final wear and tear risk weight, For the final stress risk weight, This represents the final environmental risk weight.

[0103] Increased scratch density enhances the impact of wear risk; for example... / When = 0.9, a increases by 0.072.

[0104] S104. Based on the service life of the belt and environmental corrosivity parameters, determine the curve steepness coefficient.

[0105] Curve steepness coefficient η r The value is determined based on the belt's service life and environmental corrosivity parameters. For every two years increase in service life, [the value is adjusted accordingly]. Initial value increased by 20%; new belt =3.5, 5-year belt =4.2.

[0106] When environmental corrosion parameters indicate a high salt spray area, increase a fixed proportion of 20%~25%; for example, coastal environment from 4.2 to 5.0.

[0107] the steepness coefficient η of the curve r Typical value range is 3.5~6.5, specific setting logic is: when value increases, such as 4.0~6.0, the curve will become steep, making the R value jump sharply when the input quantity P× × approaches the risk threshold, thereby strengthening the early warning sensitivity of high-risk working conditions, such as the iron ore wharf, to ensure that minor risk changes are captured in time; conversely, in a mild scenario with high tolerance for false positives, such as coal transportation, need to be reduced to 3.0~4.0 to flatten the response curve and avoid frequent false triggers, and need to be dynamically adjusted according to the belt state, new belts have stable material properties, and the initial value is set to a lower value, such as 3.5, to reduce oversensitivity; as the service life increases, when the belt enters the aging stage, such as more than 5 years, need to be increased by 20% to compensate for the implicit risk response delay caused by material fatigue.

[0108] In addition, in the scenario of multi-source sensor data conflict, such as the inconsistency between the steel wire texture degradation rate and the joint thermal radiation intensity monitoring results, temporarily reduce to 80% of the baseline value to suppress false positives by weakening the curve steepness. The initial curve steepness coefficient is determined according to the service life of the belt, including: new belt =3.5, linearly increasing with age. When the environmental corrosive parameter indicates a high salt fog area, increase the fixed proportion; logically, high salt fog accelerates material degradation, increasing the response sensitivity. For example, when in service for 3 years =3.8, increase to 4.56 under high salt fog.

[0109] S105, based on historical false alarm events and environmental corrosive parameters, determine the risk offset.

[0110] The risk offset θ is determined according to historical false alarm events and environmental corrosive parameters.

[0111] The environmental corrosive parameter is used to set the baseline offset: increase θ by 0.15~0.25 in coastal high salt fog areas, and decrease θ by 0.1 in dry environments. New belt θ=0.8, 5-year belt θ=1.0. When a historical false alarm event occurs, reduce θ by a fixed value of 0.05; for example, after a false alarm, θ decreases from 1.0 to 0.95.

[0112] The risk offset θ establishes a baseline by quantifying the cumulative effect of service life on risk: for every 2 years of service life, θ needs to be increased by 0.1 to offset the implicit risk increase due to material aging, with new belts θ = 0.8 and 5-year belts θ = 1.0; at the same time, environmental corrosiveness is indispensable to the modification of θ: in coastal high-salt-mist or acid rain areas, θ needs to be additionally increased by 0.15-0.25, such as from 1.0 to 1.25, to cover the accelerated degradation of humidity and chemical corrosion on the belt structure; while in dry and moderate environments, θ can be reduced by 0.1 to match the actual risk level; historical operation data provides the basis for dynamic calibration: if the system has a missed event, such as actual exposure of steel wire but no early warning, θ needs to be reduced by 0.05 to increase sensitivity; if the false alarm rate exceeds the standard, θ is increased by 0.03 to enhance stability. The determination of the baseline offset according to the environmental corrosiveness parameter includes: the baseline is based on service life, and θ is increased by 0.1 every 2 years. When a missed event occurs, a fixed value is reduced; logically, a missed event indicates that the system is not sensitive enough, and reducing θ increases the response. For example, θ = 1.25 under high-salt-mist, and reduces to 1.20 after a missed event.

[0113] S106, according to the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient and the risk offset, a product calculation and a Sigmoid function conversion operation are performed to generate a steel wire exposure risk index.

[0114] The product calculation is performed: the belt health degradation index P, the steel wire texture definition degradation rate and the joint thermal radiation intensity are multiplied to obtain an intermediate value.

[0115] The Sigmoid function conversion operation is performed: based on the product result, the curve steepness coefficient and the risk offset θ, a Sigmoid function is applied to generate a steel wire exposure risk index R.

[0116] The calculation formula is:

[0117]

[0118] wherein, is the curve steepness coefficient, P is the belt health degradation index, is the steel wire texture definition degradation rate, is the joint thermal radiation intensity, θ is the risk offset, and e is the natural constant.

[0119] After the steel wire exposure risk index R is generated, when the steel wire exposure risk index R reaches an emergency threshold, the tension abnormal area and the thermal imaging coordinates are fused.

[0120] Specifically, the early warning threshold is obtained by comparing R with a dynamic threshold in real time through an embedded comparator.

[0121]

[0122]

[0123]

[0124] wherein T1, T2 and T3 are early warning thresholds, k1, k2 and k3 are reference threshold coefficients, ω1, ω2 and ω3 are historical data weight coefficients, ΔT is a longitudinal temperature difference gradient, T ref is a reference temperature, v is a real-time belt speed, v max is a safe belt speed threshold, R h is a historical steel wire exposure risk index sequence, σ h is a historical steel wire exposure risk index variance, is a historical steel wire exposure risk index mean, and max is a maximum value function.

[0125] When R≥T1, a fusion positioning algorithm based on the tension abnormal area, thermal imaging coordinates (x, y) and visual damage area outputs the exposure position coordinates and a quantitative damage report containing the exposure area and crack width. The hierarchical response mechanism compares the steel wire exposure risk index R with a dynamic threshold in real time through an embedded comparator. When R≥T1, a shutdown code 0xE1 is sent to the conveyor PLC through the PROFIBUS bus and the hydraulic brake is activated. At the same time, a fusion positioning algorithm based on the tension abnormal area, thermal imaging coordinates (x, y) and visual damage area outputs the exposure position coordinates and a quantitative damage report containing the exposure area and crack width. When , the dynamic speed is reduced according to the formula:

[0126]

[0127] , the infrared monitoring frequency of the joint area is increased to 5Hz, wherein v new is the target belt speed after the risk is triggered, η v is a speed reduction coefficient, the range is 0.15~0.5, the default value is 0.3, and it can be adjusted; when , the visual sampling rate is increased by 4 times and specific maintenance instructions are output, such as completing the joint vulcanization state detection within 72 hours.

[0128] The steel wire exposure position coordinates are output according to the fusion result; the coordinates are used to guide maintenance, such as a report containing the exposure area.

[0129] As shown in Figure 2 , the application also provides a port belt conveyor wire leakage early warning device, which comprises:

[0130] The image module 201 collects the image of the belt surface of the belt conveyor, performs edge detection and texture contrast on the image of the belt surface, and generates a steel wire texture definition degradation rate.

[0131] The infrared module 202 collects infrared radiation data of the belt joint area, performs normalization processing and thermal radiation energy calculation operations on the infrared radiation data, and generates joint thermal radiation intensity.

[0132] The model module 203 obtains a standardized belt state feature matrix, performs wear feature coefficient calculation, stress concentration coefficient calculation, and environmental attenuation coefficient calculation on the standardized belt state feature matrix, inputs the calculation results into a belt health degradation model, and generates a belt health degradation index.

[0133] The curve module 204 determines a curve steepness coefficient based on the service life of the belt and the environmental corrosiveness parameter.

[0134] The offset module 205 determines a risk offset based on historical false alarm events and the environmental corrosiveness parameter.

[0135] The index module 206 performs product calculation and Sigmoid function conversion operations on the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient, and the risk offset to generate a steel wire exposure risk index.

[0136] The application also provides an electronic device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described above.

[0137] The application also provides a computer-readable storage medium storing a computer program, which, when executed in a computer, causes the computer to execute the method described above.

[0138] The above embodiments are provided to facilitate the understanding and application of the application by those skilled in the art. Those skilled in the art can easily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without inventive labor. Therefore, the application is not limited to the above embodiments, and any improvements and modifications made to the application by those skilled in the art based on the disclosure of the application should be within the scope of protection of the application.

Claims

1. A method for early warning of broken wire of a port belt conveyor, characterized in that, The method comprises the following steps: Collecting the image of the belt surface of the belt conveyor, performing edge detection and texture contrast on the image of the belt surface, and generating the steel wire texture definition degradation rate; Collecting infrared radiation data of the belt joint area, performing normalization processing and thermal radiation energy calculation operation on the infrared radiation data, and generating joint thermal radiation intensity; Obtaining a standardized belt state feature matrix, performing wear feature coefficient calculation, stress concentration coefficient calculation and environmental attenuation coefficient calculation on the standardized belt state feature matrix, inputting the calculation results into a belt health degradation model, and generating a belt health degradation index; Determine the curve steepness coefficient based on the service life of the belt and the environmental corrosivity parameter; Determine the risk offset based on the historical false alarm events and the environmental corrosivity parameter; According to the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient and the risk offset, the product calculation and Sigmoid function conversion operation are performed to generate the steel wire exposure risk index.

2. The method of claim 1, wherein, Performing edge detection and texture contrast on the image of the belt surface to generate the steel wire texture definition degradation rate, comprising: Performing edge detection operation on the image of the belt surface to extract the steel wire texture edge; Comparing the extracted texture edge with the standard texture template to calculate the texture edge blur and contrast attenuation; According to the texture edge blur and contrast attenuation, the steel wire texture definition degradation rate is generated.

3. The method of claim 1, wherein, Performing normalization processing and thermal radiation energy calculation operation on the infrared radiation data to generate joint thermal radiation intensity, comprising: Performing median filtering operation on the infrared radiation data to eliminate the interference of environmental heat source; According to the filtered data, the total value of the thermal radiation of the joint area is calculated; According to the total value of the thermal radiation and the joint area, the unit area thermal radiation intensity is generated.

4. The method of claim 1, wherein, Input the calculation results into the belt health degradation model to generate the belt health degradation index, comprising: Obtaining the scratch density and real-time belt speed; When the first proportion of the scratch density is greater than the critical density of the scratch, the wear risk weight is updated according to the scratch density and real-time belt speed; The updated weight is input into the belt health degradation model to generate the belt health degradation index.

5. The method of claim 1, wherein, Determine the risk offset based on the historical false alarm events and the environmental corrosivity parameter, comprising: Determine the reference offset according to the environmental corrosivity parameter; When a false alarm event occurs, the fixed value of the risk offset is reduced.

6. The method of claim 1, wherein, After generating the steel wire exposure risk index, it further comprises: When the steel wire exposure risk index reaches the emergency threshold, the tension abnormal area and the thermal imaging coordinates are fused; According to the fusion result, the steel wire exposure position coordinates are output.

7. The method of claim 1, wherein, Determine the curve steepness coefficient based on the service life of the belt and the environmental corrosivity parameter, comprising: Determine the initial curve steepness coefficient according to the service life of the belt; When the environmental corrosivity parameter indicates a high salt mist area, the fixed proportion of the curve steepness coefficient is increased.

8. A port belt conveyor broken wire early warning device, characterized in that, It comprises: An image module collects the image of the belt surface of the belt conveyor, performs edge detection and texture contrast on the image of the belt surface, and generates the steel wire texture definition degradation rate; An infrared module acquires infrared radiation data of the belt joint area, performs normalization processing and thermal radiation energy calculation operations on the infrared radiation data, and generates joint thermal radiation intensity; A model module acquires a standardized belt state feature matrix, performs wear feature coefficient calculation, stress concentration coefficient calculation, and environmental attenuation coefficient calculation on the standardized belt state feature matrix, inputs the calculation results into a belt health degradation model, and generates a belt health degradation index; A curve module determines a curve steepness coefficient based on the belt service life and the environmental corrosiveness parameter; An offset module determines a risk offset based on historical false alarm events and the environmental corrosiveness parameter; An index module performs product calculation and Sigmoid function conversion operations on the belt health degradation index, the steel wire texture definition degradation rate, the joint thermal radiation intensity, the curve steepness coefficient, and the risk offset to generate a steel wire exposure risk index.

9. An electronic device, comprising: A computer program product, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program product, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method of any one of claims 1-7.

Images