A belt tear detection method and system for a mining conveyor

Abstract

The present invention relates to the field of mining conveying detection technology, and in particular to a belt tear detection method and system for mining conveyors. First, an adaptive echo sampling method is used to collect ultrasonic signals, a drift compensation algorithm is used to correct the drift of the baseline signal, and overlapping echo separation technology is used to separate overlapping echoes; then, an environmental adaptation algorithm is introduced to build a machine learning model; micro-tear recognition technology is used to identify ultrasonic signals processed by the machine learning-based environmental adaptation algorithm, and a mutation point detection algorithm and a pattern recognition and classification algorithm are introduced to perform tear feature recognition and tear pattern recognition; finally, a tear event confirmation algorithm is introduced, logical rules are set, a tear position positioning technology is introduced to locate the tear position, and a tear degree assessment algorithm is introduced to assess the tear degree. The technical problems of inaccurate data processing, poor accuracy in belt longitudinal tear detection, and low sensitivity in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of mining conveyor detection, and in particular to a belt tear detection method and system for a mining conveyor. Background Art

[0002] In the mining industry, conveyor systems are key material handling equipment used to transport a variety of materials, including coal, ore, and other mined materials, in and out of mines. These systems typically rely on long-distance conveyor belts to achieve efficient material handling. However, conveyor belts are prone to wear and tear during operation, especially in harsh mining environments. Damage to conveyor belts not only leads to downtime and reduced production efficiency, but may also cause safety accidents. At present, belt damage detection in mining conveyor systems mainly relies on regular manual inspections or monitoring methods based on simple mechanical devices. Existing longitudinal tear detection systems often have the following technical problems: Line laser-based image recognition technology is easily affected by factors such as lighting conditions, conveyor belt movement, and equipment vibration. The actual detection effect is not ideal and cannot adapt well to the harsh working environment and continuous operation requirements of coal mines; the existing longitudinal tear detection system cannot accurately measure the longitudinal tear length; the longitudinal tear detection system using a single wheel-type ultrasonic transducer has poor adaptability and is difficult to cope with different industrial scenarios, especially when the vibration of the on-site equipment is large, and the ultrasonic signal is difficult to recognize at the receiving end.

[0003] There are many methods for mining conveyor detection. Zhang Haitao et al. proposed a Chinese invention patent "A Ferromagnetic Material Detection System for Mining Material Conveyors", application number: "CN202310934146.2", which mainly includes: a conveyor belt for transporting crushed coal; an iron remover, which is arranged above the conveyor belt and used to absorb metal substances mixed in the coal; a sensor component, which is arranged on the conveyor belt and located on the side of the iron remover close to the conveyor belt feed port, and the sensor component is used to detect the metal information mixed in the coal, and is used to control the working state of the iron remover and the conveyor belt; a leveling baffle, which is fixedly arranged above the conveyor belt and located on the side of the iron remover close to the conveyor belt feed port, and a preset distance is left between the bottom end of the leveling baffle and the conveyor belt; a scattering component, which is arranged on the leveling baffle and used to scatter the coal transported on the conveyor belt, and the scattering component adjusts the degree of scattering of the coal based on the pressure between the coal and the leveling baffle. This application can improve the transportation efficiency of coal.

[0004] However, the above technology has at least the following technical problems: insufficiently accurate data processing, poor accuracy in detecting longitudinal tearing of the belt, and low sensitivity. Summary of the invention

[0005] The present invention solves the technical problems of inaccurate data processing, poor accuracy and low sensitivity in belt longitudinal tear detection in the prior art by providing a belt tear detection method and system for a mining conveyor, thereby achieving the technical effects of accurate data processing and high-precision and high-sensitivity detection of belt longitudinal tear.

[0006] The present invention provides a belt tear detection method and system for a mining conveyor, which specifically includes the following technical solutions:

[0007] A belt tear detection system for a mining conveyor includes the following parts:

[0008] Ultrasonic detection module, environmental adaptation module, micro-tear identification module, modular integration module, self-diagnosis and maintenance module;

[0009] The ultrasonic detection module selects and deploys ultrasonic sensors, and uses an adaptive echo sampling method to collect ultrasonic signals. It uses a drift compensation algorithm to correct the drift of the baseline signal and uses overlapping echo separation technology to separate overlapping echoes to obtain ultrasonic signal data, and then transmits the collected ultrasonic signal data to the environmental adaptation module.

[0010] The environmental adaptation module collects ultrasonic signal data under different environmental conditions, records environmental parameters under different environmental conditions, builds a machine learning model, and applies the machine learning model to the ultrasonic detection module;

[0011] The micro-tear recognition module uses the micro-tear recognition technology to recognize the ultrasonic signal data processed by the environmental adaptation module, further determines the tearing result based on the recognition result, and generates a report;

[0012] Modular integration module, which forms integration scheme and implementation guide according to the report of micro-tear identification module and design parameters of conveyor system;

[0013] The self-diagnosis and maintenance module detects the health status of the system, performs self-diagnosis, triggers the maintenance process, obtains maintenance notifications, system status reports and preventive maintenance suggestions, and feeds back the maintenance notifications and system status reports to the ultrasonic detection module and the environmental adaptation module to optimize the detection parameters; the detection data is the data collected by modules other than the self-diagnosis and maintenance module and the modular integration module, including the data provided by the ultrasonic detection module, the data processed by the environmental adaptation module, and the data provided by the micro-tear identification module.

[0014] A method for detecting belt tearing of a mining conveyor comprises the following steps:

[0015] S1. Adopt adaptive echo sampling method to collect ultrasonic signals, use drift compensation algorithm to correct drift of baseline signal and separate overlapping echoes by overlapping echo separation technology;

[0016] S2. Introduce the environment adaptation algorithm, build a machine learning model, and integrate the machine learning-based environment adaptation algorithm into the belt longitudinal tear detection implementation;

[0017] S3. Use micro-tear recognition technology to identify the ultrasonic signal processed by the machine learning-based environment adaptation algorithm, introduce the mutation point detection algorithm and pattern recognition and classification algorithm to perform tear feature recognition and tear pattern recognition;

[0018] S4. Based on the results obtained by the mutation point detection algorithm and the pattern recognition and classification algorithm, a tear event confirmation algorithm is introduced, logical rules are set, a tear position positioning technology is introduced to locate the tear position, and a tear degree assessment algorithm is introduced to assess the tear degree.

[0019] Preferably, S1 specifically includes:

[0020] Select the ultrasonic sensor according to the width and thickness of the conveyor belt; set the baseline signal level of the ultrasonic sensor; adjust the sensitivity of the ultrasonic sensor to collect ultrasonic signals.

[0021] Preferably, in S1, it further includes:

[0022] When the conveyor belt starts running, the adaptive echo sampling method is used to collect data.

[0023] Preferably, in S1, it further includes:

[0024] During data acquisition, a drift compensation algorithm is applied in real time to correct for drift of the baseline signal.

[0025] Preferably, in S1, it further includes:

[0026] When the data acquisition process is running, the overlapping echo separation technology is introduced to separate the overlapping echoes.

[0027] Preferably, S2 specifically includes:

[0028] Collect ultrasonic signal data under different environmental conditions, record environmental parameters under different environmental conditions, apply signal processing techniques, including fast Fourier transform and wavelet transform, to extract features from ultrasonic signals; build a machine learning model, use the extracted features and environmental parameters as input, output the signal quality score, and adjust the model parameters.

[0029] Preferably, S3 specifically includes:

[0030] Firstly, an adaptive filtering algorithm is used to preprocess the ultrasonic signal after being processed by the environment adaptation algorithm based on machine learning. Then, the tearing feature is identified using the mutation point detection algorithm. The mutation point indicates the potential tearing of the conveyor belt, and is determined by integrating the influence of the first-order derivative and the second-order derivative of the ultrasonic signal through the concept of the modulus length of the vector.

[0031] Preferably, in S3, it further includes:

[0032] The extreme point analysis method is used to extract the peak and valley values ​​of the signal. The extreme point is an indicator of the tearing feature. The extreme point analysis method locates the extreme point by calculating the derivative of the signal and finding its zero point. Then, a parameter adaptive adjustment strategy is executed to dynamically adjust the parameters of the ultrasonic sensor, including sensitivity and sampling frequency, according to the detected features. Pattern recognition and classification algorithms are introduced to perform tearing pattern recognition.

[0033] Preferably, S4 specifically includes:

[0034] Based on the results obtained by the mutation point detection algorithm and the pattern recognition and classification algorithm, a tear event confirmation algorithm is introduced, and logical rules are set to confirm the tear event. After confirming the tear event, the tear position positioning technology is introduced to locate the tear position. Then the tear degree assessment algorithm is introduced. The tear degree assessment algorithm assesses the severity of the tear according to the depth and length of the tear feature, and provides a quantitative assessment of the tear degree by analyzing the geometric shape and change trend of the tear feature to obtain the tear result. Finally, the tear response decision protocol is implemented according to the tear result. The tear response decision protocol decides whether to shut down immediately or continue running according to the location and severity of the tear, monitors the development of the tear, and forms a report.

[0035] The beneficial effects of the technical solution of the present invention are:

[0036] 1. The present invention uses an adaptive echo sampling method to allow the system to dynamically adjust the sampling frequency according to the operating status of the conveyor belt, accurately collect ultrasonic signals, and thus improve the accuracy of detection. The drift compensation algorithm effectively corrects the drift of the baseline signal and ensures the stability and reliability of the ultrasonic data. The overlapping echo separation technology enables each echo to be accurately analyzed and located even under high-speed operating conditions when echo signals overlap, thereby improving the accuracy of tear location.

[0037] 2. The present invention introduces an adaptive filtering algorithm to dynamically adjust the filter coefficients to optimize the clarity of the signal, thereby effectively eliminating noise and interference, improving the signal-to-noise ratio of the signal, and providing a clearer signal basis for the identification of tearing features. The concept of vector modulus is used to integrate the influence of the first-order and second-order derivatives of the signal, so that changes in the signal can be detected more sensitively, thereby improving the sensitivity and accuracy of tearing detection. The parameter adaptive adjustment strategy dynamically adjusts the parameters of the ultrasonic sensor, such as sensitivity and sampling frequency, according to the detected features. In this way, the detection settings can be optimized according to actual conditions to improve the accuracy and efficiency of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a module diagram of the belt tear detection system of the mining conveyor of the present invention;

[0039] Figure 2 The present invention is a flow chart of a method for detecting belt tearing of a mining conveyor. DETAILED DESCRIPTION

[0040] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0042] The specific scheme of a belt tear detection method and system for a mining conveyor provided by the present invention is described in detail below with reference to the accompanying drawings.

[0043] Refer to the attached Figure 1 , a belt tear detection system for a mining conveyor, comprising the following parts:

[0044] Ultrasonic detection module, environmental adaptation module, micro-tear identification module, modular integration module, self-diagnosis and maintenance module;

[0045] The ultrasonic detection module selects and deploys ultrasonic sensors, and uses an adaptive echo sampling method to collect ultrasonic signals. It uses a drift compensation algorithm to correct the drift of the baseline signal and uses overlapping echo separation technology to separate overlapping echoes to obtain ultrasonic signal data, and then transmits the collected ultrasonic signal data to the environmental adaptation module.

[0046] Environmental adaptation module, which collects a large amount of ultrasonic signal data under different environmental conditions, records the corresponding environmental parameters, builds a machine learning model to ensure the stable operation of the mining conveyor in different mining environments, and applies it to the ultrasonic detection module;

[0047] The micro-tear recognition module uses the micro-tear recognition technology to recognize the ultrasonic signal data processed by the environmental adaptation module, further determines the tearing result based on the recognition result, and generates a report;

[0048] Modular integration module, based on the report of micro-tear identification module and the design parameters of conveyor system, ensures that the detection system can be seamlessly integrated into the existing conveyor, forms an integration plan and implementation guide to ensure the compatibility of the detection system with the conveyor;

[0049] The self-diagnosis and maintenance module detects the health status of the system according to the integrated solution and detection data, performs self-diagnosis, triggers the maintenance process, obtains maintenance notifications, system status reports and preventive maintenance suggestions, and feeds back the maintenance notifications and system status reports to the ultrasonic detection module and the environmental adaptation module to optimize the detection parameters; the detection data is the data collected by other modules, including the data provided by the ultrasonic detection module, the data processed by the environmental adaptation module, and the data provided by the micro-tear recognition technology.

[0050] Refer to the attached Figure 2 , a belt tear detection method for a mining conveyor, comprising the following steps:

[0051] S1. Adopt adaptive echo sampling method to collect ultrasonic signals, use drift compensation algorithm to correct drift of baseline signal and separate overlapping echoes by overlapping echo separation technology;

[0052] Select an ultrasonic sensor with the appropriate frequency based on the width and thickness of the conveyor belt to ensure penetration and resolution, and install the selected ultrasonic sensor at key locations on the conveyor belt, which are locations prone to tearing, such as joints and high-stress areas; at the same time, ensure that the installation angle of the ultrasonic sensor can maximize the coverage of the conveyor belt surface to capture all possible tear signals; set the baseline signal level of the sensor based on an undamaged conveyor belt; adjust the sensitivity of the sensor to ensure that it can detect tiny signal changes, which can quickly capture conveyor tears.

[0053] When the conveyor belt starts running, in order to obtain more accurate ultrasonic data, the present invention adopts an adaptive echo sampling method for data collection. The adaptive echo sampling method dynamically adjusts the sampling frequency according to the running state of the conveyor belt to obtain more accurate data. The specific implementation process is as follows:

[0054] First, the initial sampling setting; when the conveyor belt starts, the initial sampling frequency f is set based on the design parameters of the conveyor belt, such as maximum speed and maximum load. init ; Then, in the initial sampling stage, the collected signal S(t) is subjected to Fourier transform FFT to obtain the signal spectrum and analyze its frequency components and amplitude information; Finally, in order to obtain more accurate ultrasonic data, the sampling frequency is dynamically adjusted; According to the signal stability index Δs(t), the sampling frequency f is dynamically adjusted sample (t) to adapt to the changes in the running state of the conveyor belt; the adjustment of the sampling frequency follows the following formula:

[0055] f sample (t) = f init ·e -α·Δs(t) +f min

[0056] Among them, f min It is the minimum frequency to ensure signal sampling, ensuring that sufficient data can be obtained even in the case of low signal changes; α is the adjustment coefficient, which is used to control the attenuation rate of the sampling frequency to adapt to signal changes; Δs(t) is the stability index of the signal, which reflects the degree of change of the signal over time and is a key parameter for dynamically adjusting the sampling frequency.

[0057] Furthermore, in order to prevent the baseline signal of the ultrasonic sensor from drifting due to environmental changes and affecting the accuracy of the data, the present invention applies a drift compensation algorithm in real time to correct the drift of the baseline signal during the data acquisition process; specifically, while adaptively sampling the echo, the baseline signal S is monitored in real time. baseline (t), and calculate its moving average μ baseline (t) and standard deviation σ baseline , to determine the drift of the signal and correct the signal according to the drift:

[0058] S corrected (t) = S raw (t)-μ baseline (t)+β·σ baseline (t)

[0059] Where β is the adjustment factor, which is used to compensate for the signal fluctuation caused by environmental changes and ensure the accuracy of signal correction; μ baseline (t)+β·σ baseline (t) represents the drift; S raw (t) represents the original signal, which is the data directly captured by the sensor and contains drift information; S corrected (t) represents the corrected signal, where the drift is subtracted from the original signal;

[0060] Furthermore, when the above implementation process is running at high speed, ultrasonic echoes may overlap, resulting in difficulty in tearing detection. The present invention introduces an overlapping echo separation technology to separate overlapping echoes so as to analyze each echo separately. First, a signal analysis method is used to detect whether the echoes overlap. If echo overlap is detected, a high-resolution time-frequency analysis is immediately performed on the signal S(t) using a short-time Fourier transform (STFT) to distinguish the time and frequency characteristics of different echoes. Then, an echo feature extraction method is used to extract the center frequency, bandwidth, amplitude and phase of each echo from the STFT result. Finally, a reconstruction formula is used to absorb and separate and reconstruct the overlapping echoes so as to analyze each echo separately. The echo signal reconstruction formula is as follows:

[0061]

[0062] Among them, E separated (t) is the reconstructed echo signal; A i is the amplitude of the ith echo, reflecting the strength of the echo, which is related to the size of the tear; j is an imaginary number; φ i is the phase of the ith echo, which is related to the temporal characteristics of the echo; τ i is the arrival time of the ith echo, which is the key parameter for determining the echo position; Δτ i is the time width of the ith echo, which determines the duration of the echo; rect is the rectangular function; n is the number of echoes; t represents continuous time, which is used to describe the change of ultrasonic signal over time;

[0063] After the above process, the corrected and separated ultrasonic signals are obtained.

[0064] S2. Introduce the environment adaptation algorithm, build a machine learning model, and integrate the machine learning-based environment adaptation algorithm into the belt longitudinal tear detection implementation;

[0065] The specific implementation of the environment adaptation algorithm is as follows:

[0066] First, a large amount of ultrasonic signal data under different environmental conditions is collected, and the corresponding environmental parameters, such as temperature, humidity and dust level, are recorded at the same time; the environmental parameters are obtained through corresponding sensors, such as temperature sensors, humidity sensors and particle sensors to measure; next, signal processing techniques, such as fast Fourier transform and wavelet transform, are applied to extract key features from the original signal;

[0067] Subsequently, a machine learning model, such as a support vector machine, is designed, which inputs the extracted key features and environmental parameters and outputs a quality score for the signal; the machine learning model is trained in a controlled environment using known tear events and normal operating data to adjust the model parameters to maximize performance; the trained machine learning model is then deployed in an actual mining environment and its output is monitored in real time, including false positives and false negatives under different environmental conditions; a confusion matrix is ​​used to analyze the model's performance in the field to determine the model's accuracy and reliability.

[0068] To further improve the accuracy and reliability of detection, grid search and cross-validation techniques were used to fine-tune the machine learning model parameters based on the field test results. The adjusted model was tested multiple times, and performance changes were recorded after each iteration. Based on the results of the iterative tests, final optimization was performed to ensure that the model could maintain high accuracy under the widest range of environmental conditions.

[0069] Finally, an optimized environment adaptation algorithm based on machine learning was obtained, which can accurately distinguish between tearing signals and environmental noise and can operate stably in different mining environments.

[0070] S3. Use micro-tear recognition technology to identify the ultrasonic signal processed by the machine learning-based environment adaptation algorithm, introduce the mutation point detection algorithm and pattern recognition and classification algorithm to perform tear feature recognition and tear pattern recognition;

[0071] The specific implementation of micro-tear identification technology is as follows:

[0072] In order to ensure the accuracy and reliability of the belt tear detection system, the ultrasonic signal processed by the environmental adaptation algorithm is first pre-processed by an adaptive filtering algorithm to eliminate noise and interference; the adaptive filtering algorithm optimizes the clarity of the signal by dynamically adjusting the coefficient of the filter according to the real-time statistical characteristics of the signal; the mutation point detection algorithm is then used to identify the tear feature. The mutation point detection algorithm identifies the tear feature by identifying the mutation point in the pre-processed signal. The mutation point indicates the potential tear of the conveyor belt; the mutation point is determined by combining the influence of the first-order derivative and the second-order derivative of the signal through the concept of vector modulus length; the specific implementation mathematical formula is expressed as:

[0073]

[0074] Among them, CPDA(S) represents the time set of all possible tearing points in the ultrasonic signal; θ and are the thresholds of the first-order derivative and the second-order derivative, respectively, which are set based on experience; It is the first-order derivative of the ultrasonic signal S′(t) after being processed by the environmental adaptation algorithm with respect to time, which indicates the rate of change of the signal over time and is used to capture the instantaneous change of the signal; It is the second-order derivative of the ultrasonic signal S′(t) after being processed by the environmental adaptation algorithm with respect to time, which indicates the rate of change of the signal change rate and is used to capture the acceleration of the signal; t is a time variable, which indicates a specific moment in the ultrasonic detection process; It is the vector modulus of the first-order derivative and the second-order derivative of the signal, which combines the instantaneous rate of change of the signal and the change in the rate of change, and is used to detect significant changes in the signal; It is the vector modulus of the threshold, which is used to compare with the modulus of the signal to determine whether the sensitivity threshold of tear detection is exceeded.

[0075] The extreme point analysis method is also used to extract the peak and valley values ​​of the ultrasonic signal. The extreme point is the key indicator of the tearing feature. The extreme point analysis method locates the extreme point by calculating the derivative of the signal and finding its zero point. The specific mathematical formula is expressed as:

[0076]

[0077] EPAM(S) represents the set of all time points that satisfy certain conditions;

[0078] Then, according to the features extracted by the parameter extraction algorithm, a parameter adaptive adjustment strategy is executed. The parameter adaptive adjustment strategy dynamically adjusts the parameters of the ultrasonic sensor, such as sensitivity and sampling frequency, according to the detected features to improve the detection accuracy.

[0079] Next, the tearing pattern is identified using a pattern recognition and classification algorithm. The pattern recognition and classification algorithm is based on a deep learning convolutional neural network (CNN) to identify different tearing patterns, and the following mathematical formula is introduced to implement it:

[0080] PRCA(S features )

[0081] =σ(W k *σ(W k-1 *...σ(W1*S features +b1)...+b k-1 )+b k )

[0082] Among them, S features is a feature vector extracted from the signal, representing the features extracted from the ultrasonic signal, including the signal amplitude, frequency, time interval, and other statistical characteristics that may indicate conveyor belt tearing; PRCA (S features ) represents the output of the convolutional neural network, identifying and classifying tear patterns on the conveyor belt; Wi and b i are the weights and biases of the i-th network layer, i∈[1,k]; k is the number of network layers; σ is the activation function used to introduce nonlinearity; * represents the convolution operation; the above formula starts from the input layer and applies the weight W layer by layer i and bias b i And through the activation function, the output of each layer is used as the input of the next layer. The last layer is usually a fully connected layer that outputs the recognition result;

[0083] S4. Based on the results obtained by the mutation point detection algorithm and the pattern recognition and classification algorithm, a tear event confirmation algorithm is introduced, logical rules are set, a tear position positioning technology is introduced to locate the tear position, and a tear degree assessment algorithm is introduced to assess the tear degree.

[0084] Firstly, considering the results obtained by the mutation point detection algorithm and the pattern recognition and classification algorithm, a tear event confirmation algorithm is introduced, and logical rules are set. For example, if multiple tear features are detected continuously within the set time window, it is confirmed as a tear event. After the tear event is confirmed, the tear position positioning technology is introduced to locate the tear position. The tear position positioning technology uses the spatial resolution of the sensor array to determine the precise position of the tear, and calculates the position of the tear by comparing the signal differences of different sensors. Then, the tear degree assessment algorithm is introduced to assess the tear degree. The tear degree assessment algorithm assesses the severity of the tear according to the depth and length of the tear feature, and provides a quantitative assessment of the tear degree by analyzing the geometric shape and change trend of the tear feature to obtain the tear result. Finally, the tear response decision protocol is implemented according to the tear result. The tear response decision protocol decides whether to shut down immediately or continue to run according to the location and severity of the tear, and monitors the development of the tear to form a report to provide a basis for subsequent maintenance.

[0085] In summary, a belt tear detection method and system for a mining conveyor are completed.

[0086] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for detecting belt tearing of a mining conveyor, characterized in that: The following steps are involved: S1. Adopt adaptive echo sampling method to collect ultrasonic signals, use drift compensation algorithm to correct drift of baseline signal and separate overlapping echoes by overlapping echo separation technology; The specific implementation process of the adaptive echo sampling method is as follows: First, the initial sampling setting is performed. The initial sampling frequency is set based on the design parameters of the conveyor belt. Then, in the initial sampling stage, the collected signal is Fourier transformed to obtain the signal spectrum. Finally, the sampling frequency is dynamically adjusted. The specific adjustment formula is as follows: f sample (t)=f init ·e -α·Δs(t) +f min Among them, f sample (t) is the adjusted sampling frequency; f init is the initial sampling frequency; f min is the minimum frequency to ensure signal sampling; α is the adjustment coefficient used to control the attenuation speed of the sampling frequency; Δs(t) is the stability index of the signal; t is the time variable; S2. Introduce an environmental adaptation algorithm, collect ultrasonic signal data under different environmental conditions, record environmental parameters under different environmental conditions, apply signal processing technology, including fast Fourier transform and wavelet transform, to extract features from ultrasonic signals; build a machine learning model, take the extracted features and environmental parameters as input, output the signal quality score, and adjust the model parameters, and integrate the machine learning-based environmental adaptation algorithm into the belt longitudinal tear detection implementation; S3. Use micro-tear recognition technology to identify the ultrasonic signal processed by the machine learning-based environment adaptation algorithm; the specific implementation of the micro-tear recognition technology is as follows: Firstly, the ultrasonic signal processed by the machine learning-based environment adaptation algorithm is preprocessed using an adaptive filtering algorithm; The mutation point detection algorithm is then used to identify tearing features; the mutation point indicates the potential tearing of the conveyor belt. The specific formula of the mutation point detection algorithm is as follows: Among them, CPDA(S) represents the time set of all possible tearing points in the ultrasonic signal; θ and are the thresholds of the first and second derivatives, respectively; It is the first-order derivative of the ultrasonic signal S′(t) with respect to time after being processed by the environment adaptation algorithm; It is the second-order derivative of the ultrasonic signal S′(t) with respect to time after being processed by the environment adaptation algorithm; is the vector modulus of the first and second derivatives of the signal; is the vector modulus of the threshold; An extreme point analysis method is also used to extract the peak and valley values ​​of the signal. The extreme point is an indicator of the tearing feature. The extreme point analysis method locates the extreme point by calculating the derivative of the signal and finding its zero point. Then, a parameter adaptive adjustment strategy is executed to dynamically adjust the parameters of the ultrasonic sensor, including sensitivity and sampling frequency, according to the detected features. Pattern recognition and classification algorithm is introduced to perform tearing pattern recognition; the specific implementation formula is as follows: PRCA(S features ) =σ(W k *σ(W k-1 *…σ(W1*S features +b1)…+b k-1 )+b k ) Among them, S features is the feature vector extracted from the ultrasonic signal; PRCA(S features ) represents the output of the convolutional neural network, identifying and classifying tear patterns on the conveyor belt; W i and b i are the weight and bias of the i-th network layer, i∈[1,k]; k is the number of network layers; σ is the activation function; * represents the convolution operation; S4. Based on the results obtained by the mutation point detection algorithm and the pattern recognition and classification algorithm, a tear event confirmation algorithm is introduced, logical rules are set, a tear position positioning technology is introduced to locate the tear position, and a tear degree assessment algorithm is introduced to assess the tear degree.

2. The belt tear detection method for a mining conveyor according to claim 1, characterized in that: The S1 specifically includes: Select the ultrasonic sensor according to the width and thickness of the conveyor belt; set the baseline signal level of the ultrasonic sensor; adjust the sensitivity of the ultrasonic sensor to collect ultrasonic signals.

3. The belt tear detection method for a mining conveyor according to claim 1, characterized in that: In the S1, it also includes: During data acquisition, a drift compensation algorithm is applied in real time to correct for drift of the baseline signal.

4. The belt tear detection method for a mining conveyor according to claim 3 is characterized in that: In the S1, it also includes: When the data acquisition process is running, the overlapping echo separation technology is introduced to separate the overlapping echoes.

5. The method for detecting belt tearing of a mining conveyor according to claim 1, characterized in that: The S4 specifically includes: Based on the results obtained by the comprehensive mutation point detection algorithm and the pattern recognition and classification algorithm, a tear event confirmation algorithm is introduced, and logical rules are set to confirm the tear event; after confirming the tear event, the tear position positioning technology is introduced to locate the tear position; then the tear degree assessment algorithm is introduced, which assesses the severity of the tear according to the depth and length of the tear feature, and provides a quantitative assessment of the tear degree by analyzing the geometric shape and change trend of the tear feature to obtain the tear result; finally, the tear response decision protocol is implemented according to the tear result, and the tear response decision protocol decides whether to shut down immediately or continue running according to the location and severity of the tear, monitors the development of the tear, and forms a report.

6. A belt tear detection system for a mining conveyor, applied to the belt tear detection method for a mining conveyor as claimed in claim 1, characterized in that: Includes the following parts: Ultrasonic detection module, environmental adaptation module, micro-tear identification module, modular integration module, self-diagnosis and maintenance module; The ultrasonic detection module selects and deploys ultrasonic sensors, and uses an adaptive echo sampling method to collect ultrasonic signals, and uses a drift compensation algorithm to correct the drift of the baseline signal and separates the overlapping echoes through an overlapping echo separation technique to obtain ultrasonic signal data, and transmits the collected ultrasonic signal data to the environmental adaptation module; The environmental adaptation module collects ultrasonic signal data under different environmental conditions, records environmental parameters under different environmental conditions, builds a machine learning model, and applies the machine learning model to the ultrasonic detection module; The micro-tear recognition module recognizes the ultrasonic signal data processed by the environmental adaptation module using micro-tear recognition technology, further determines the tearing result according to the recognition result, and forms a report; The modular integration module forms an integration plan and implementation guide based on the report of the micro-tear identification module and the design parameters of the conveyor system; The self-diagnosis and maintenance module detects the health status of the system, performs self-diagnosis, triggers the maintenance process, obtains maintenance notifications, system status reports and preventive maintenance suggestions, and feeds back the maintenance notifications and system status reports to the ultrasonic detection module and the environmental adaptation module to optimize the detection parameters; the detection data is the data collected by modules other than the self-diagnosis and maintenance module and the modular integration module, including data provided by the ultrasonic detection module, data processed by the environmental adaptation module, and data provided by the micro-tear identification module.

Images