Intelligent security system for conveyors

Abstract

This invention relates to the field of conveyor equipment technology, and discloses an intelligent safety system for conveyors. The system includes dividing the safety cycle into sub-periods for collecting comprehensive operating parameters and operating status values ​​of the conveyor during each sub-period. It trains a machine learning model to predict the operating status value of the next sub-period, determines whether to enter a shutdown safety mode, marks adjustment parameters from the comprehensive operating parameters, and formulates conveyor optimization instructions. Compared to existing technologies, this invention, by dividing the conveyor's safety cycle into sub-periods, allows for refined control of the conveyor's safety control duration. Based on a machine learning model, it can predict the conveyor's operating status in the next sub-period, thereby accurately predicting future anomalies and formulating corresponding data optimization instructions. This avoids the lag caused by emergency shutdowns due to real-time assessment of operating status, effectively improving the safety performance of the conveyor.

Description

Technical Field

[0001] This invention relates to the field of conveying equipment technology, and more specifically, to an intelligent safety system for conveyors. Background Technology

[0002] Conveyors are widely used in many fields such as coal mining, mining, and logistics. During operation, a series of problems may occur, such as excessive speed, coal piling, deviation, and tearing, which can lead to serious accidents. At the same time, if the conveyor experiences malfunctions such as jamming or motor overheating, it may also cause equipment damage and production stoppage. Therefore, it is necessary to accurately monitor the operating status of the conveyor and develop a suitable safety system to protect the conveyor's operating status.

[0003] Reference patent application CN116280987A discloses a belt conveyor fault monitoring and early warning system. This system uses an image acquisition system to capture and monitor key parts of the belt conveyor during operation, and a data acquisition system to collect and monitor critical data during operation. This enables real-time monitoring of the belt conveyor's operation. An image data recognition system processes the information collected by the image and data acquisition systems, and a fault early warning system further identifies abnormal information in the images and data to confirm the anomaly. The system can then issue timely warnings to address the fault promptly, eliminating potential safety hazards caused by undetected conveyor malfunctions.

[0004] The existing technology has the following shortcomings:

[0005] Existing conveyor safety systems collect real-time operating parameters of the conveyor during operation and combine this data with calculation and analysis to assess the conveyor's real-time operating status and implement safety measures based on this status. However, since the conveyor operates continuously, this real-time analysis and assessment method can only evaluate the current operating status of the conveyor and cannot predict future anomalies. Consequently, when anomalies or malfunctions occur in the future, the system cannot control the conveyor to stop in time, leading to a lag in subsequent safety measures and reducing the overall safety effectiveness of the conveyor.

[0006] In view of this, the present invention proposes an intelligent safety system for conveyors to solve the above problems. Summary of the Invention

[0007] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: an intelligent safety system for conveyors, applied to a conveyor control terminal, comprising:

[0008] The sub-period segmentation module is used to obtain the security cycle of the conveyor and divide the security cycle into sub-periods based on the segmentation criteria. The segmentation criteria are: the start and end points of any two adjacent sub-periods coincide on the timeline.

[0009] The data acquisition module is used to collect the historical training data set of the conveyor in the sub-period. The historical training data set includes comprehensive operating parameters and operating status values. The comprehensive operating parameters include bearing temperature value, smoke concentration value, speed over-limit value, coal pile height value, belt tear degree and belt deviation degree. The operating status values ​​include normal status and abnormal status.

[0010] The model training module is used to train a machine learning model that predicts the operating state value of the next sub-period based on a set of historical training data.

[0011] The prediction and judgment module is used to collect real-time comprehensive operating parameters, input them into the trained machine learning model, predict the operating status value of the next sub-period, and determine whether to enter the shutdown security mode.

[0012] The instruction formulation module is used to mark adjustment parameters from the comprehensive operating parameters and formulate delivery optimization instructions based on the adjustment parameters.

[0013] Furthermore, methods for dividing sub-time periods include:

[0014] A point is randomly marked on the conveyor belt of the conveyor and recorded as the detection node;

[0015] Start the conveyor and continuously increase its operating power until it reaches the rated power, then stop increasing it. Record the moment when the conveyor's operating power first reaches the rated power as the zero point.

[0016] After midnight, the camera captures two images of the detection node passing the same position on the conveyor belt, obtaining the first image and the second image.

[0017] The timestamps are used to query the times corresponding to the first and second images one by one, and these timestamps are recorded as the start time and end time of the cycle. The duration between the start time and the end time of the cycle is recorded as the security cycle.

[0018] Based on the preset division duration, mark i+1 division points within the security cycle;

[0019] Based on the (i+1)th dividing point, the security cycle is divided into i consecutive sub-time periods, and the time corresponding to the (i+1)th dividing point is recorded as the start time of the (i+1)th sub-time period and the end time of the (i)th sub-time period.

[0020] Furthermore, methods for obtaining the coal pile height value include:

[0021] Using the preset height detection duration as a standard, mark p height detection moments one by one within i sub-time periods;

[0022] Along the conveyor's conveying direction, a camera captures images of the conveyor head at p height detection moments, obtaining p height detection images;

[0023] Using computer vision technology, the regions corresponding to the conveyor belt and coal in p height detection images are identified one by one, and the belt region and coal pile region are obtained respectively.

[0024] Draw a horizontal line along the upper boundary of the tape area to obtain the horizontal line of the tape;

[0025] Draw a line along the outer boundary line of the coal pile area to obtain the coal pile boundary line, and mark the pixels on the coal pile boundary line one by one to obtain the boundary pixels.

[0026] Measure the vertical distance from each boundary pixel to the horizontal line of the tape, and record the maximum vertical distance as the sub-height value to obtain p sub-height values;

[0027] The average of the p sub-height values ​​of the i sub-time periods is calculated to obtain the i coal pile height values.

[0028] The expression for the coal pile height value is:

[0029]

[0030] In the formula, DM gdi Let GD be the coal pile height value for the i-th sub-period. zia It represents the a-th sub-height value of the i-th sub-time period.

[0031] Further methods for obtaining the tear strength of the tape include:

[0032] Using the preset tear detection duration as a standard, mark s non-adjacent tear detection times in i sub-time periods;

[0033] By measuring the tension value of the conveyor belt at s tear detection moments one by one using a tension sensor installed on the roller and in rolling contact with the conveyor belt, s real-time tension values ​​are obtained.

[0034] The difference between the maximum value and the minimum value of the i real-time tension values ​​is used to obtain the i tension change values.

[0035] The expression for the change in tension is:

[0036] ZL bhi =ZL zdi -ZL zxi ;

[0037] In the formula, ZL bhi ZL represents the tension change value in the i-th sub-period. zdi ZL represents the maximum real-time tension value in the i-th sub-period. zxi This represents the minimum real-time tension value in the i-th sub-period.

[0038] The duration between the tear detection time corresponding to the maximum value of the i real-time tension values ​​and the tear detection time corresponding to the minimum value of the i real-time tension values ​​is recorded as the tear change duration, and i tear change durations are obtained.

[0039] By comparing the i tension change values ​​with the i tear change durations in sequence, the i tape tear degrees are obtained;

[0040] The expression for tape tear strength is:

[0041]

[0042] In the formula, JD sli Let SL be the tape tear strength in the i-th sub-time period. bhi Let be the duration of the tearing change in the i-th sub-period.

[0043] Further methods for obtaining tape deviation include:

[0044] Record the midpoint of each of the i sub-time periods as the deviation detection time, and obtain i deviation detection times;

[0045] i top-view images of the conveyor belt are obtained by capturing top-view images of the conveyor belt at i deviation detection moments using a camera installed above the conveyor belt.

[0046] Using computer vision technology, the conveyor belt in i top-view deviation images is identified one by one, and two long side lines and two short side lines are drawn along the outer boundary of the conveyor belt. The area within the two long side lines and the two short side lines is recorded as the real-time area, and i real-time areas are obtained.

[0047] Mark the midpoints of two long side lines in i real-time regions, and draw dividing lines through the two midpoints. Record the real-time regions located below the dividing lines as target regions, and obtain i target regions.

[0048] Mark the pixels in the i target regions one by one, count the number of pixels, and obtain the total value of the i pixels;

[0049] After querying the database, a standard region is obtained, and i target regions are superimposed on the standard region in sequence. The region located outside the standard region is recorded as the deviation region, and i deviation regions are obtained.

[0050] Mark the pixels in the i deviation regions one by one, count the number of pixels, and obtain the i deviation values;

[0051] Compare each of the i deviation values ​​with the total value of the i pixels to obtain the i tape deviation degrees.

[0052] The expression for tape deviation is:

[0053]

[0054] In the formula, JD ppi Let be the tape deviation in the i-th sub-time period, PP lzi Let XS be the deviation value for the i-th sub-time period. zli Let be the total number of pixels in the i-th sub-time period.

[0055] Furthermore, training methods for machine learning models that predict the running state value of the next sub-period include:

[0056] The comprehensive operating parameters are transformed into multiple feature vectors using the sliding window method. The operating state value is converted into a label corresponding to the comprehensive operating parameters according to the sliding step size. The normal state is converted into 0 and the abnormal state is converted into 1. Each feature vector corresponds to a label and constitutes a set of training data. Multiple sets of training data constitute a training set. The comprehensive operating parameters are arranged in the order of collection time. The prediction time step size Z, the sliding step size Q, and the sliding window length N are preset.

[0057] The feature vector is used as the input to the machine learning model, the predicted running state value of the next sub-period after time step Z is used as the output, the subsequent running state value of each training set is used as the prediction target, and the sum of prediction errors is minimized as the training objective. The machine learning model is trained to generate a machine learning model that predicts the running state value of the next sub-period based on the comprehensive running parameters of the previous sub-period.

[0058] Furthermore, the methods for determining whether to enter the shutdown security mode include:

[0059] When the output of the machine learning model is 0, the running status value of the next sub-period is normal.

[0060] When the output of the machine learning model is 1, the running status value of the next sub-period is an abnormal state.

[0061] If the current sub-period running status value is normal, it is determined that the system will not enter the shutdown security mode;

[0062] If the operating status value of the next sub-period is abnormal, it is determined that the system will enter the shutdown security mode.

[0063] Furthermore, the methods for marking the adjustment parameters include:

[0064] The bearing temperature value is compared with the preset safe temperature value. When the bearing temperature value is greater than the preset safe temperature value, the bearing temperature value is marked as an adjustment parameter.

[0065] The smoke concentration value is compared with the preset safe concentration value. When the smoke concentration value is greater than the preset safe concentration value, the smoke concentration value is marked as an adjustment parameter.

[0066] The speed limit value is compared with the preset safety value. When the speed limit value is greater than the preset safety value, the speed limit value is marked as an adjustment parameter.

[0067] Compare the coal pile height value with the preset safety height value. When the coal pile height value is greater than the preset safety height value, mark the coal pile height value as an adjustment parameter.

[0068] Compare the tape tear strength with the preset tear safety value. When the tape tear strength is greater than the preset tear safety value, mark the tape tear strength as an adjustment parameter.

[0069] The tape deviation is compared with the preset deviation safety value. When the tape deviation exceeds the preset deviation safety value, the tape deviation is marked as an adjustment parameter.

[0070] Furthermore, the conveying optimization instructions include instructions to reduce bearing temperature, reduce smoke concentration, reduce speed exceeding limits, reduce coal pile height, reduce belt tearing, and reduce belt misalignment.

[0071] Furthermore, the methods for formulating instructions to reduce bearing temperature, reduce smoke concentration, reduce speed exceeding limits, reduce coal pile height, reduce conveyor belt tearing, and reduce conveyor belt misalignment include:

[0072] When the parameter is adjusted to the bearing temperature value, a command to reduce the bearing temperature value is generated;

[0073] When the parameter is adjusted to the smoke concentration value, an instruction to reduce the smoke concentration value is generated;

[0074] When the adjusted parameter is set to exceed the speed limit, an instruction to reduce the speed exceeding the speed limit is generated;

[0075] When the parameter is adjusted to the coal pile height value, an instruction to reduce the coal pile height value is issued;

[0076] When the parameter is adjusted to tape tear strength, an instruction to reduce tape tear strength is generated;

[0077] When the parameter is adjusted to tape deviation, an instruction to reduce tape deviation is generated.

[0078] The technical effects and advantages of the intelligent safety system for conveyors of this invention are as follows:

[0079] This invention obtains the safety cycle of a conveyor and divides it into sub-periods based on a segmentation criterion. It then collects historical training data sets of the conveyor within these sub-periods, including comprehensive operating parameters and operating status values. Based on this historical training data set, a machine learning model is trained to predict the operating status value of the next sub-period. Real-time comprehensive operating parameters are collected and input into the trained machine learning model to predict the operating status value of the next sub-period and determine whether to enter a shutdown safety mode. Adjustment parameters are marked from the comprehensive operating parameters, and based on these adjustment parameters, conveyor optimization instructions are formulated. Compared to existing technologies, this invention improves upon existing technologies by... By dividing the security cycle into sub-periods, the long-duration security cycle can be further subdivided, allowing for refined control of the conveyor's security control duration. Combined with comprehensive operating parameters from past sub-periods, a machine learning model can predict the conveyor's operating status for the next sub-period. This enables accurate prediction of impending anomalies and allows for early shutdown of the security system. Furthermore, appropriate data optimization instructions are developed for any abnormal data, improving the safety of the conveyor's operation and avoiding the lag caused by emergency shutdowns based on real-time operational status assessments. This effectively enhances the conveyor's security performance. Attached Figure Description

[0080] Figure 1 This is a schematic diagram of the intelligent security system for conveyors provided in Embodiment 1 of the present invention;

[0081] Figure 2 This is a flowchart illustrating the intelligent security method for conveyors provided in Embodiment 2 of the present invention. Detailed Implementation

[0082] 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.

[0083] Example 1: Please refer to Figure 1 As shown, the intelligent safety system for the conveyor described in this embodiment is applied to the conveyor control terminal and includes:

[0084] The sub-period segmentation module is used to obtain the security cycle of the conveyor and divide the security cycle into sub-periods based on the segmentation criteria;

[0085] The security cycle refers to the time required for the conveyor to perform a complete set of security data monitoring and security command formulation. It can uniformly limit the time used for the security operation of the conveyor and serve as the basis for subsequent sub-time periods. The sub-time period is the smallest unit of length that constitutes the security cycle and serves as the standard duration for subsequent data collection. Since the security cycle is long and contains a lot of data, in order to break down the data within the security cycle into smaller parts, it is necessary to divide the security cycle into multiple consecutive and equal sub-time periods through the division criteria.

[0086] The dividing criterion is that the start and end points of any two adjacent sub-time periods coincide on the timeline.

[0087] Methods for dividing sub-time periods include:

[0088] A point is randomly marked on the conveyor belt of the conveyor and recorded as the detection node;

[0089] Start the conveyor and continuously increase its operating power until it reaches the rated power, then stop increasing it. Record the moment when the conveyor's operating power first reaches the rated power as the zero point. Rated power refers to the operating power of the conveyor when it reaches normal operation, ensuring that the conveyor can maintain normal operation. Rated power is obtained by consulting the conveyor's technical parameter table.

[0090] After midnight, the camera captures two images of the detection node passing the same position on the conveyor belt, obtaining the first image and the second image.

[0091] The timestamps are used to query the times corresponding to the first and second images one by one, and these timestamps are recorded as the start time and end time of the cycle. The duration between the start time and the end time of the cycle is recorded as the security cycle.

[0092] Using a preset division duration as the standard, i+1 division points are marked within the security cycle. The preset division duration is the minimum time required for a complete and sufficient change in the security data within the conveyor to be collected, ensuring that the security data changes within each sub-period, thereby enabling accurate data collection of the security status within each sub-period. The preset division duration is set according to the actual length of the conveyor's security cycle; for example, if the security cycle is 40 seconds, then the sub-period is 2 seconds.

[0093] Based on the (i+1)th dividing point, the security cycle is divided into i consecutive sub-time periods, and the time corresponding to the (i+1)th dividing point is recorded as the start time of the (i+1)th sub-time period and the end time of the (i)th sub-time period.

[0094] It should be noted that by dividing the data into sub-periods, the overall security cycle with a large time span can be further subdivided, allowing a large amount of security data to be distinguished by different time periods. This facilitates the subsequent collection and analysis of security data, thereby avoiding data collection errors.

[0095] The data acquisition module is used to collect the historical training data set of the conveyor in the sub-period. The historical training data set includes comprehensive operating parameters and operating status values. The comprehensive operating parameters include bearing temperature value, smoke concentration value, speed over-limit value, coal pile height value, belt tear degree and belt deviation degree.

[0096] Comprehensive operating parameters refer to the internal data of the conveyor itself and the external data of the environment in which the conveyor is located that may change during the process of conveying coal. These data can comprehensively represent the safety of the conveyor during the coal conveying process, thereby specifically representing the operating status of the conveyor and facilitating the determination of whether the operating status of the conveyor is safe.

[0097] Comprehensive operating parameters include bearing temperature, smoke concentration, speed exceeding limit, coal pile height, conveyor belt tear degree, and conveyor belt deviation degree;

[0098] The bearing temperature value refers to the local temperature value generated at the connection between the drum and the conveyor belt during the conveyor's coal conveying process due to the rotation of the bearing. When the bearing temperature value is higher, the local temperature between the drum and the conveyor belt on the conveyor will be higher, which will cause excessive wear, failure, or even fire of the drum and the conveyor belt due to high temperature, thus causing the conveyor's operating condition to deteriorate. The bearing temperature value is obtained by temperature sensor installed between the conveyor drum and the conveyor belt.

[0099] Smoke concentration value refers to the smoke concentration value in the mine environment where the conveyor is located. It can represent the smoke situation in the environment where the conveyor is located. When the smoke concentration value is higher, the smoke concentration value in the environment where the conveyor is located will be higher, the probability of fire in the environment where the conveyor is located will be higher, and the worse the operating condition of the conveyor will be. The smoke concentration value is detected and obtained by a smoke concentration sensor installed above the drive unit of the conveyor and located downwind.

[0100] Speed ​​exceeding the limit refers to the extent to which the real-time conveying speed of the conveyor exceeds the safe speed when transporting coal. The larger the speed exceeding the limit, the greater the extent to which the real-time conveying speed exceeds the safe speed, the greater the probability of coal spillage or conveyor malfunction, and the worse the conveyor's operating condition. The speed exceeding the limit is obtained by subtracting the real-time conveying speed from the safe speed after detecting the real-time conveying speed by a speed sensor installed on the roller. The safe speed refers to the rated conveying speed of the conveyor when transporting coal under safe operating conditions, and is obtained by consulting the conveyor's technical parameter table.

[0101] The coal pile height value refers to the height of the coal piled on the conveyor belt during the coal conveying process. It can be used to represent the height of the top of the coal pile on the conveyor. The larger the coal pile height value, the higher the height of the coal piled on the conveyor belt, which leads to a greater probability of the coal collapsing and spilling on the conveyor belt, and thus a worse operating condition of the conveyor.

[0102] Methods for obtaining the coal pile height value include:

[0103] Using a preset height detection duration as the standard, p height detection moments are marked sequentially within i sub-time periods. The preset height detection duration refers to the time interval between capturing two adjacent height detection images, thus limiting the time span between two adjacent height detection images. This ensures that the coal pile on the conveyor belt can be transported a certain distance, improving the stability of the coal pile and preventing collapse. The preset height detection duration is set according to the actual operating speed of the conveyor; the higher the operating speed, the shorter the preset height detection duration. The specific value of the preset height detection duration is not unique, as long as it meets the requirements.

[0104] Along the conveyor's conveying direction, a camera captures images of the conveyor head at p height detection moments, obtaining p height detection images;

[0105] Using computer vision technology, the regions corresponding to the conveyor belt and coal in p height detection images are identified one by one, and the belt region and coal pile region are obtained respectively.

[0106] Draw a horizontal line along the upper boundary of the conveyor belt area to obtain the conveyor belt horizontal line; the conveyor belt horizontal line serves as the dividing line between the conveyor belt and the coal pile, and can provide a reference basis for subsequent data measurement of the coal pile height to ensure the accuracy of the coal pile height calculation;

[0107] Draw a line along the outer boundary line of the coal pile area to obtain the coal pile boundary line, and mark the pixels on the coal pile boundary line one by one to obtain the boundary pixels.

[0108] Measure the vertical distance from each boundary pixel to the horizontal line of the tape, and record the maximum vertical distance as the sub-height value to obtain p sub-height values;

[0109] The average of the p sub-height values ​​of the i sub-time periods is calculated to obtain the i coal pile height values.

[0110] The expression for the coal pile height value is:

[0111]

[0112] In the formula, DM gdi Let GD be the coal pile height value for the i-th sub-period. zia It represents the a-th sub-height value of the i-th sub-time period.

[0113] Belt tear degree refers to the severity of longitudinal tearing of the conveyor belt during the conveying of coal. It can be numerically expressed to indicate the tearing condition of the conveyor belt. The greater the belt tear degree, the more severe the longitudinal tearing of the conveyor belt, the higher the probability of the conveyor belt breaking and causing operational failure, and the worse the operating condition of the conveyor.

[0114] Methods for obtaining the tear strength of adhesive tape include:

[0115] Using a preset tear detection duration as a standard, s non-adjacent tear detection times are marked within i sub-time periods. The preset tear detection duration refers to the time interval between detecting two adjacent real-time tension values, which limits the time span between two adjacent real-time tension values ​​to ensure that the longitudinal positions of the conveyor belts corresponding to two adjacent real-time tension values ​​do not overlap. The preset tear detection duration is specifically set according to the actual requirements of the conveyor belt for tension measurement length.

[0116] By measuring the tension value of the conveyor belt at s tear detection moments one by one using a tension sensor installed on the roller and in rolling contact with the conveyor belt, s real-time tension values ​​are obtained.

[0117] The difference between the maximum value and the minimum value of the i real-time tension values ​​is used to obtain the i tension change values.

[0118] The expression for the change in tension is:

[0119] ZL bhi =ZL zdi -ZL zxi ;

[0120] In the formula, ZL bhi ZL represents the tension change value in the i-th sub-period. zdi ZL represents the maximum real-time tension value in the i-th sub-period. zxiThis represents the minimum real-time tension value in the i-th sub-period.

[0121] The duration between the tear detection time corresponding to the maximum value of the i real-time tension values ​​and the tear detection time corresponding to the minimum value of the i real-time tension values ​​is recorded as the tear change duration, and i tear change durations are obtained.

[0122] By comparing the i tension change values ​​with the i tear change durations in sequence, the i tape tear degrees are obtained;

[0123] The expression for tape tear strength is:

[0124]

[0125] In the formula, JD sli Let SL be the tape tear strength in the i-th sub-time period. bhi Let be the duration of the tearing change in the i-th sub-period.

[0126] Belt misalignment refers to the degree to which the conveyor belt deviates to one side during the conveying of coal, causing the conveyor belt to tilt. It can be numerically expressed as the degree of belt deviation to both sides. The greater the belt misalignment, the greater the degree of belt deviation to one side, the more serious the belt misalignment, and the worse the operating condition of the conveyor.

[0127] Methods for obtaining tape deviation include:

[0128] Record the midpoint of each of the i sub-time periods as the deviation detection time, and obtain i deviation detection times;

[0129] i top-view images of the conveyor belt are obtained by capturing top-view images of the conveyor belt at i deviation detection moments using a camera installed above the conveyor belt.

[0130] Using computer vision technology, the conveyor belt in i top-view deviation images is identified one by one, and two long side lines and two short side lines are drawn along the outer boundary of the conveyor belt. The area within the two long side lines and the two short side lines is recorded as the real-time area, and i real-time areas are obtained.

[0131] Mark the midpoints of the two long side lines in the i real-time regions, and draw the dividing lines through the two midpoints. Record the real-time regions below the dividing lines as target regions to obtain i target regions. When the conveyor belt deviates, the regions below the dividing lines will shift, so that the target regions can accurately represent the areas where the conveyor belt deviates.

[0132] Mark the pixels in the i target regions one by one, count the number of pixels, and obtain the total value of the i pixels;

[0133] After querying the database, a standard area is obtained, and i target areas are superimposed on the standard area in sequence. The area outside the standard area is recorded as the deviation area, and i deviation areas are obtained. The standard area refers to the image of the standard area obtained by the above method and stored in the database when the conveyor belt does not deviate. It serves as the basis for subsequent deviation area identification of the target area.

[0134] Mark the pixels in the i deviation regions one by one, count the number of pixels, and obtain the i deviation values;

[0135] Compare each of the i deviation values ​​with the total value of the i pixels to obtain the i tape deviation degrees.

[0136] The expression for tape deviation is:

[0137]

[0138] In the formula, JD ppi Let be the tape deviation in the i-th sub-time period, PP lzi Let XS be the deviation value for the i-th sub-time period. zli Let be the total number of pixels in the i-th sub-time period.

[0139] Operating status values ​​refer to the overall operating condition of the conveyor when transporting coal, which can be used to represent the safety of the conveyor during operation. Operating status values ​​include normal and abnormal states. When the operating status value is normal, it indicates that the safety of the conveyor during operation is good, and when the operating status value is abnormal, it indicates that the safety of the conveyor during operation is poor. Operating status values ​​are obtained by collecting bearing temperature values, smoke concentration values, speed over-limit values, coal pile height values, belt tear degree, and belt misalignment degree corresponding to the normal and abnormal states.

[0140] The model training module is used to train a machine learning model that predicts the operating state value of the next sub-period based on a set of historical training data.

[0141] Once the historical training data set is obtained, a machine learning model that can accurately predict the operating status value of the next sub-period can be trained based on the combined operating parameters and operating status values, thereby enabling advance prediction of the future situation of the conveyor.

[0142] Training methods for machine learning models that predict the running state value of the next sub-period include:

[0143] The comprehensive operating parameters are transformed into multiple feature vectors using the sliding window method. The operating state value is converted into a label corresponding to the comprehensive operating parameters according to the sliding step size. The normal state is converted into 0 and the abnormal state is converted into 1. Each feature vector corresponds to a label and constitutes a set of training data. Multiple sets of training data constitute a training set. The comprehensive operating parameters are arranged in the order of collection time. The prediction time step size Z, the sliding step size Q, and the sliding window length N are preset.

[0144] The feature vector is used as the input to the machine learning model, the predicted running state value of the next sub-period after time step Z is used as the output, the subsequent running state value of each training set is used as the prediction target, and the sum of prediction errors is minimized as the training objective. The machine learning model is trained to generate a machine learning model that predicts the running state value of the next sub-period based on the comprehensive running parameters of the previous sub-period.

[0145] For example, the machine learning model is either a CNN neural network model or AlexNet;

[0146] The formula for calculating prediction error is:

[0147] zk=(ak-wk) 2

[0148] In the formula, zk is the prediction error, k is the group number of the feature vector; ak is the predicted state value corresponding to the k-th feature vector, and wk is the actual state value corresponding to the k-th training data.

[0149] In machine learning models, feature vectors represent comprehensive operating parameters, and state values ​​represent operating state values.

[0150] Other model parameters of the machine learning model, such as the target loss value, optimization algorithm, ratio of training set to test set to validation set, and optimization of loss function, are all obtained through actual engineering implementation and continuous experimental tuning.

[0151] Specifically, a simple example of the sliding window method is as follows: Suppose we want to train a machine learning model using comprehensive running parameters (Q1, Q2, Q3, Q4, Q5, Q6) to predict the value of one future time step. We can use a sliding window of length 4 and a sliding step size of length 1 to generate a training set and a prediction target for the future. For example, the training set is (Q1, Q2, Q3, Q4) and (Q2, Q3, Q4, Q5), and the prediction targets are (E5) and (E6). E5 is the running state value corresponding to the comprehensive running parameters of the next sub-time period after the sub-time period in which Q4 is located, and E6 is the running state value corresponding to the comprehensive running parameters of the next sub-time period in which Q5 is located.

[0152] The sliding window method can accurately and in advance predict the future operating status value based on the existing historical data of the conveyor during the coal conveying process. This allows for the early detection of changes in the operating status value before any abnormal changes occur, and provides the information to the conveying control terminal as a basis for safety assessment and emergency stop operations. This helps to avoid serious malfunctions caused by drastic changes in various parameters during the conveying process.

[0153] The prediction and judgment module is used to collect real-time comprehensive operating parameters, input them into the trained machine learning model, predict the operating status value of the next sub-period, and determine whether to enter the shutdown security mode.

[0154] Once the machine learning model is trained, real-time comprehensive operating parameters can be collected and input into the machine learning model. Based on the output of the machine learning model, the operating status value of the conveyor can be obtained, which serves as the basis for subsequent determination of whether to enter the shutdown safety mode. When the shutdown safety mode is entered, it indicates that the operating status of the conveyor is poor and an emergency shutdown of the conveyor is required, and instructions are given for the subsequent adjustment of operating parameters.

[0155] The methods for determining whether to enter the shutdown security mode include:

[0156] When the output of the machine learning model is 0, the running status value of the next sub-period is normal.

[0157] When the output of the machine learning model is 1, the running status value of the next sub-period is an abnormal state.

[0158] If the operating status value of the next sub-period is normal, and the operating status of the conveyor is good, then it is determined that the shutdown safety mode will not be entered.

[0159] If the operating status value of the next sub-period is abnormal, and the operating status of the conveyor is poor, it will be determined to enter the shutdown safety mode.

[0160] When entering the shutdown safety mode, the conveyor will first shut down in an emergency to prevent malfunctions and improve the safety of the coal conveying process. It also provides time for subsequent adjustments to the relevant parameters of the conveyor's operating status, avoiding the lag caused by stopping the conveyor only when a malfunction occurs.

[0161] The instruction formulation module is used to mark adjustment parameters from the comprehensive operating parameters and formulate delivery optimization instructions based on the adjustment parameters;

[0162] Adjustment parameters refer to one or more of the following parameters in the comprehensive operating parameters that may cause abnormal operating conditions of the conveyor: bearing temperature, smoke concentration, speed exceeding limit, coal pile height, belt tearing degree, and belt misalignment degree. These adjustment parameters serve as the basis for optimizing and adjusting the operating condition of the conveyor.

[0163] The methods for marking adjustment parameters include:

[0164] The bearing temperature value is compared with the preset temperature safety value. The preset temperature safety value refers to the maximum bearing temperature value when the conveyor is in normal operating condition. In other words, the maximum bearing temperature value in normal operating condition is limited. The preset temperature safety value is obtained by collecting the maximum bearing temperature values ​​of a large number of historical conveyor operating conditions when the bearing temperature is in normal operating condition and then calculating the average value.

[0165] When the bearing temperature exceeds the preset safe temperature value, the local temperature between the conveyor drum and the conveyor belt is too high, which will have a significant negative impact on the operation of the conveyor. Therefore, the bearing temperature value is marked as an adjustment parameter.

[0166] The smoke concentration value is compared with the preset safe concentration value. The preset safe concentration value refers to the maximum smoke concentration value when the conveyor is in normal operating condition. In other words, the maximum smoke concentration value in normal operating condition is limited. The preset safe concentration value is obtained by collecting the maximum smoke concentration values ​​of a large number of historical conveyor operating conditions when the smoke concentration value is in normal operating condition and then calculating the average value.

[0167] When the smoke concentration value exceeds the preset safe concentration value, the smoke concentration in the mine environment where the conveyor is located is too high, which will have a significant negative impact on the operation of the conveyor. In this case, the smoke concentration value is marked as an adjustment parameter.

[0168] The speed over-limit value is compared with the preset over-limit safety value. The preset over-limit safety value refers to the maximum speed over-limit value when the conveyor's operating state is normal, which limits the maximum speed over-limit value in the normal state. The preset over-limit safety value is obtained by collecting the average value of the maximum speed over-limit values ​​of a large number of historical conveyor operating states when the operating state is normal.

[0169] When the speed exceeds the preset safety value, the real-time conveying speed of the conveyor exceeds the safe speed by too much, which will have a significant negative impact on the operation of the conveyor. In this case, the speed exceeds the limit and is marked as an adjustment parameter.

[0170] The coal pile height is compared with the preset safety height value. The preset safety height value refers to the maximum coal pile height value when the conveyor is in normal operating condition. This limits the maximum coal pile height value in normal operating condition. The preset safety height value is obtained by averaging the maximum coal pile height values ​​collected from a large number of historical conveyor operating conditions when they are in normal operating condition.

[0171] When the coal pile height is greater than the preset safety height value, the coal pile height on the conveyor belt is too high during the conveyor conveying process, which will have a significant negative impact on the operation of the conveyor. Therefore, the coal pile height value is marked as an adjustment parameter.

[0172] The belt tear strength is compared with the preset tear safety value. The preset tear safety value refers to the maximum value of the belt tear strength when the conveyor is in normal operating condition, which limits the maximum value of the belt tear strength in normal operating condition. The preset tear safety value is obtained by collecting the maximum value of the belt tear strength when the conveyor is in normal operating condition from a large number of historical data and then calculating the average value.

[0173] When the belt tear degree is greater than the preset tear safety value, the more severe the longitudinal tearing of the conveyor belt is during the conveying of coal, the greater the negative impact on the operation of the conveyor. Therefore, the belt tear degree is marked as an adjustment parameter.

[0174] The belt deviation is compared with the preset safety deviation value. The preset safety deviation value refers to the maximum value of belt deviation when the conveyor is in normal operating condition. This limits the maximum value of belt deviation in normal operating condition. The preset safety deviation value is obtained by averaging the maximum value of belt deviation when a large number of historical conveyor operating conditions are in normal operating condition.

[0175] When the belt deviation exceeds the preset safety value, the greater the belt tilt during the conveyor's coal conveying process, the more negative the impact on the conveyor's operation. Therefore, the belt deviation is marked as an adjustment parameter.

[0176] Once the adjustment parameters are marked, it indicates that there are parameters with abnormally large values ​​in the overall operating parameters of the conveyor. In this case, it is necessary to formulate a conveying optimization command based on the specific type of adjustment parameter. This will enable the conveying optimization command to optimize and adjust various adjustment parameters during the coal conveying process, ensuring that the conveyor can maintain a safe operating state and avoid more serious failures in the subsequent operation of the conveyor.

[0177] The conveying optimization instructions include instructions to reduce bearing temperature, reduce smoke concentration, reduce speed exceeding limits, reduce coal pile height, reduce belt tearing, and reduce belt misalignment.

[0178] The methods for formulating instructions to reduce bearing temperature, reduce smoke concentration, reduce excessive speed, reduce coal pile height, reduce conveyor belt tearing, and reduce conveyor belt misalignment include:

[0179] When the parameter is adjusted to the bearing temperature value, it is necessary to cool down the local temperature between the conveyor drum and the conveyor belt, so an instruction to reduce the bearing temperature value is formulated.

[0180] When the parameter is adjusted to the smoke concentration value, it is necessary to adsorb the smoke concentration value in the mine environment where the conveyor is located, and then formulate an instruction to reduce the smoke concentration value.

[0181] When the adjusted parameter is the speed exceeding the limit, it is necessary to reduce the real-time conveying speed of the conveyor. Therefore, a speed reduction instruction is formulated.

[0182] When the parameter is adjusted to the coal pile height value, it is necessary to sort out the coal piled on the conveyor belt during the coal conveying process, and then formulate an instruction to reduce the coal pile height value.

[0183] When the parameter is adjusted to the belt tear strength, the conveyor belt needs to be replaced during the coal conveying process, and an instruction to reduce the belt tear strength is formulated.

[0184] When the adjustment parameter is belt deviation, it is necessary to adjust the deviation of the conveyor belt during the coal conveying process, and thus formulate an instruction to reduce belt deviation.

[0185] It should be noted that after issuing commands to reduce bearing temperature, reduce smoke concentration, reduce speed over-limit, reduce coal pile height, reduce belt tear, and reduce belt deviation, the conveyor control terminal needs to control the corresponding mechanisms and equipment to execute the conveyor optimization commands based on the actual conveyor optimization commands. This will enable the conveyor to contact the abnormal state and ensure that the subsequent conveyors can operate normally and safely.

[0186] Specifically, when an instruction to reduce bearing temperature is issued, the automatic sprinkler system will automatically spray water at the connection between the roller and the conveyor belt to achieve a cooling effect. When an instruction to reduce smoke concentration is issued, the automatic sprinkler system and the exhaust fan will work simultaneously to achieve the effect of spraying water to extinguish the fire while simultaneously exhausting air to reduce smoke concentration. When an instruction to reduce speed exceeding the limit is issued, the drive motor of the conveyor will be controlled to reduce its speed to achieve a deceleration effect. When an instruction to reduce coal pile height is issued, the automatic coal pusher will push and comb the coal on the conveyor belt to reduce the coal height. When an instruction to reduce belt tearing is issued, the conveyor belt needs to be stopped and replaced. When an instruction to reduce belt misalignment is issued, the conveyor needs to be stopped, and the belt misalignment will be adjusted by adjusting the offset of the roller.

[0187] In this embodiment, the security cycle of the conveyor is obtained and divided into sub-periods based on a division criterion. Historical training data sets of the conveyor within these sub-periods are collected. These historical training data sets include comprehensive operating parameters and operating status values. Based on these historical training data sets, a machine learning model is trained to predict the operating status value of the next sub-period. Real-time comprehensive operating parameters are collected and input into the trained machine learning model to predict the operating status value of the next sub-period and determine whether to enter a shutdown security mode. Adjustment parameters are marked from the comprehensive operating parameters, and based on these adjustment parameters, conveyor optimization instructions are formulated. Compared to existing technologies, by using the conveyor... The security cycle is divided into sub-periods, which can further subdivide security cycles with large durations, thereby enabling refined control of the conveyor's security control duration. By combining comprehensive operating parameters from past sub-periods, a machine learning model can predict the conveyor's operating status for the next sub-period, thus accurately predicting potential anomalies in the future and initiating a shutdown security mode in advance. Furthermore, appropriate data optimization instructions are formulated for abnormal data to improve the safety of the conveyor's operation. This avoids the lag caused by emergency shutdowns that rely on real-time assessment of the operating status, effectively enhancing the conveyor's security performance.

[0188] Example 2: Please refer to Figure 2 As shown, parts not described in detail in this embodiment are described in Embodiment 1. An intelligent safety method for conveyors is provided, applied to a conveyor control terminal, and implemented based on an intelligent safety system for conveyors, including:

[0189] S1: Obtain the security cycle of the conveyor and divide the security cycle into sub-time periods based on the division criteria. The division criteria are: the start and end points of any two adjacent sub-time periods coincide on the timeline.

[0190] S2: Collect the historical training data set of the conveyor in the sub-period. The historical training data set includes comprehensive operating parameters and operating status values. The comprehensive operating parameters include bearing temperature value, smoke concentration value, speed over-limit value, coal pile height value, belt tear degree and belt deviation degree. The operating status values ​​include normal status and abnormal status.

[0191] S3: Based on the historical training data set, train a machine learning model to predict the running state value of the next sub-period;

[0192] S4: Collect real-time comprehensive operating parameters, input them into the trained machine learning model, predict the operating status value of the next sub-period, and determine whether to enter the shutdown security mode; if the shutdown security mode is entered, execute S5; if the shutdown security mode is not entered, repeat S4.

[0193] S5: Mark the adjustment parameters from the comprehensive operating parameters, and formulate delivery optimization instructions based on the adjustment parameters.

[0194] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A smart safety system for conveyors, applied to conveyor control terminals, characterized in that: include: The sub-period segmentation module is used to obtain the security cycle of the conveyor and divide the security cycle into sub-periods based on the segmentation criteria. The segmentation criteria are: the start and end points of any two adjacent sub-periods coincide on the timeline. The data acquisition module is used to collect the historical training data set of the conveyor in the sub-period. The historical training data set includes comprehensive operating parameters and operating status values. The comprehensive operating parameters include bearing temperature value, smoke concentration value, speed over-limit value, coal pile height value, belt tear degree and belt deviation degree. The operating status values ​​include normal status and abnormal status. Methods for obtaining the coal pile height value include: Based on the preset height detection time, in Mark each sub-time period individually A high-altitude detection moment; Along the conveyor's conveying direction, the camera captures images of the conveyor head section by section. Images at each height detection time are obtained. One height detection image; Identify one by one using computer vision technology The areas corresponding to the conveyor belt and coal in the height detection images are obtained respectively, and the belt area and the coal pile area are obtained respectively; Draw a horizontal line along the upper boundary of the tape area to obtain the horizontal line of the tape; Draw a line along the outer boundary line of the coal pile area to obtain the coal pile boundary line, and mark the pixels on the coal pile boundary line one by one to obtain the boundary pixels. Measure the vertical distance from each boundary pixel to the horizontal line of the tape, and record the maximum vertical distance as the sub-height value to obtain... Height value; Will Individual time period The average of the summed height values ​​is obtained. Individual coal pile height values; The expression for the coal pile height value is: ； In the formula, For the first Coal pile height values ​​for each sub-period. For the first The first sub-period Height value; Methods for obtaining the tear strength of adhesive tape include: Based on the preset tear detection time, in Within each sub-time period, mark s non-adjacent tear detection times; By measuring the tension value of the conveyor belt at s tear detection moments one by one using a tension sensor installed on the roller and in rolling contact with the conveyor belt, s real-time tension values ​​are obtained. Will The maximum value of each real-time tension value and After subtracting the minimum value of each real-time tension value, we obtain Each tension change value; The expression for the change in tension is: ； In the formula, For the first The tension change value for each sub-period. For the first The maximum real-time tension value for each sub-period. For the first The minimum real-time tension value for each sub-period; Will The tear detection time corresponding to the maximum value of each real-time tension value up to The duration between the tear detection moments corresponding to the minimum real-time tension values ​​is recorded as the tear change duration. The duration of each tearing change; Will The tension change values ​​are sequentially related to By comparing the duration of each tear change, we can obtain... The tear strength of the tape; The expression for tape tear strength is: ； In the formula, For the first The tear strength of the tape during each time period For the first Duration of tearing changes in individual time periods; The model training module is used to train a machine learning model that predicts the operating state value of the next sub-period based on a set of historical training data. The prediction and judgment module is used to collect real-time comprehensive operating parameters, input them into the trained machine learning model, predict the operating status value of the next sub-period, and determine whether to enter the shutdown security mode. The instruction formulation module is used to mark adjustment parameters from the comprehensive operating parameters and formulate delivery optimization instructions based on the adjustment parameters.

2. The intelligent security system for conveyors according to claim 1, characterized in that, The method for dividing the sub-time periods includes: A point is randomly marked on the conveyor belt of the conveyor and recorded as the detection node; Start the conveyor and continuously increase its operating power until it reaches the rated power, then stop increasing it. Record the moment when the conveyor's operating power first reaches the rated power as the zero point. After midnight, the camera captures two images of the detection node passing the same position on the conveyor belt, obtaining the first image and the second image. The timestamps are used to query the times corresponding to the first and second images one by one, and these timestamps are recorded as the start time and end time of the cycle. The duration between the start time and the end time of the cycle is recorded as the security cycle. Based on the preset time interval, mark the period within the security cycle. One dividing point; by Based on these dividing points, the security cycle is divided into... The number of consecutive sub-time periods, and the number of... The time corresponding to the first division point is denoted as the i-th... The start point of each sub-period and the first The end point of each sub-period.

3. The intelligent security system for conveyors according to claim 2, characterized in that, The training method for the machine learning model that predicts the operating state value of the next sub-period includes: The comprehensive operating parameters are transformed into multiple feature vectors using the sliding window method. The operating state value is converted into a label corresponding to the comprehensive operating parameters according to the sliding step size. The normal state is converted into 0 and the abnormal state is converted into 1. Each feature vector corresponds to a label and constitutes a set of training data. Multiple sets of training data constitute a training set. The comprehensive operating parameters are arranged in the order of collection time. The prediction time step size Z, the sliding step size Q, and the sliding window length N are preset. The feature vector is used as the input to the machine learning model, the predicted running state value of the next sub-period after time step Z is used as the output, the subsequent running state value of each training set is used as the prediction target, and the sum of prediction errors is minimized as the training objective. The machine learning model is trained to generate a machine learning model that predicts the running state value of the next sub-period based on the comprehensive running parameters of the previous sub-period.

4. The intelligent safety system for conveyors according to claim 3, characterized in that, The method for determining whether to enter the shutdown security mode includes: When the output of the machine learning model is 0, the running status value of the next sub-period is normal. When the output of the machine learning model is 1, the running status value of the next sub-period is an abnormal state. If the current sub-period running status value is normal, it is determined that the system will not enter the shutdown security mode; If the operating status value of the next sub-period is abnormal, it is determined that the system will enter the shutdown security mode.

5. The intelligent safety system for conveyors according to claim 4, characterized in that, The method for marking the adjustment parameters includes: The bearing temperature value is compared with the preset safe temperature value. When the bearing temperature value is greater than the preset safe temperature value, the bearing temperature value is marked as an adjustment parameter. The smoke concentration value is compared with the preset safe concentration value. When the smoke concentration value is greater than the preset safe concentration value, the smoke concentration value is marked as an adjustment parameter. The speed limit value is compared with the preset safety value. When the speed limit value is greater than the preset safety value, the speed limit value is marked as an adjustment parameter. Compare the coal pile height value with the preset safety height value. When the coal pile height value is greater than the preset safety height value, mark the coal pile height value as an adjustment parameter. Compare the tape tear strength with the preset tear safety value. When the tape tear strength is greater than the preset tear safety value, mark the tape tear strength as an adjustment parameter. The tape deviation is compared with the preset deviation safety value. When the tape deviation exceeds the preset deviation safety value, the tape deviation is marked as an adjustment parameter.

6. The intelligent safety system for conveyors according to claim 5, characterized in that, The conveying optimization commands include commands to reduce bearing temperature, reduce smoke concentration, reduce speed exceeding limits, reduce coal pile height, reduce belt tearing, and reduce belt misalignment.

7. The intelligent safety system for conveyors according to claim 6, characterized in that, The methods for formulating the commands to reduce bearing temperature, reduce smoke concentration, reduce excessive speed, reduce coal pile height, reduce conveyor belt tearing, and reduce conveyor belt misalignment include: When the parameter is adjusted to the bearing temperature value, a command to reduce the bearing temperature value is generated; When the parameter is adjusted to the smoke concentration value, an instruction to reduce the smoke concentration value is generated; When the adjusted parameter is set to exceed the speed limit, an instruction to reduce the speed exceeding the speed limit is generated; When the parameter is adjusted to the coal pile height value, an instruction to reduce the coal pile height value is issued; When the parameter is adjusted to tape tear strength, an instruction to reduce tape tear strength is generated; When the parameter is adjusted to tape deviation, an instruction to reduce tape deviation is generated.

Images