Intelligent monitoring system for broken belt capturing device

Abstract

The invention relates to the technical field of intelligent monitoring, in particular to an intelligent monitoring system for a broken belt capturing device, which comprises a sensor monitoring module used for monitoring related data of a conveyor belt in real time and identifying potential broken belt risks; and the intelligent processing module is used for carrying out fusion analysis on the collected multi-dimensional data, accurately identifying the belt breakage risk and the wear position and realizing intelligent early warning. The response control module is used for quickly responding to the belt breakage signal and automatically executing the capturing action, so that the capturing accuracy and reliability are ensured; according to the method, the sensor data and the processed image data are combined through the multi-mode deep learning model, so that the belt breakage risk is accurately identified and positioned, and the false alarm rate and the missing report rate are reduced; through real-time monitoring and feedback of the capturing process, the accuracy of the capturing device is improved, it is ensured that the broken belt can be effectively captured, and equipment damage or personnel injury caused by capturing failure is avoided.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent monitoring, and in particular to an intelligent monitoring system for a broken belt capturing device. Background Art

[0002] The broken belt capture device is a device used to quickly respond and automatically capture the broken belt end when the conveyor belt or similar belt-like transmission equipment breaks. Such a device can effectively avoid equipment damage and personal injury caused by splashing or falling of the broken belt. Its working principle is mainly to monitor the tension or position of the conveyor belt. When the conveyor belt breaks, the capture device quickly starts, grabs the broken belt end and fixes it. Then, this capture device faces the following difficulties when working:

[0003] When monitoring the real-time fluctuation of conveyor belt tension, due to the industrial environment with strong vibration, friction and load fluctuation, the signal of the tension sensor is easily interfered by noise, resulting in monitoring errors;

[0004] When using optical or infrared sensors to identify wear, cracks or tears on the belt surface and identify potential belt breaks in advance, the cost of comprehensive detection is high due to the long length of the conveyor belt, and it is also technically difficult to accurately capture images in motion;

[0005] When the capture device is performing the capture action, the capture process must be completed at the moment the belt breaks. It is difficult for existing capture devices to achieve a balance between response speed and capture force. Summary of the invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: an intelligent monitoring system for a broken belt catching device, comprising:

[0008] The sensor monitoring module collects relevant data of the conveyor belt and processes the data, completes data detection according to the data processing results, thereby dynamically adjusting the data collection frequency, and identifying potential belt break risks based on historical belt break data, characterized in that it also includes:

[0009] The intelligent processing module is used to perform secondary screening on data with belt break risk by constructing abnormal data feature vectors, and adaptively adjust the weight of abnormal data feature vectors according to the combination of belt break prediction position and intelligent early warning mechanism, wherein the belt break prediction position also includes:

[0010] The surface data of the conveyor belt is enhanced, and the feature extraction of the image data is completed using a convolutional neural network. The extracted image features are combined with the sensor data through the fully connected layer processing technology, and the feature fusion is completed and output through the output layer. At the same time, a loss function is constructed to accurately locate the broken belt position and intelligently warn.

[0011] It also includes a response control module, which calculates the distance between the broken belt position and the capture device based on the predicted broken belt position, and calculates the time consumed by the broken belt position to move to the capture device, and at the same time collects the basic time to complete the capture action, so as to complete the rapid response of the broken belt capture by comparing the time.

[0012] As a preferred solution of the intelligent monitoring system for the broken belt catching device of the present invention, the data detection is completed according to the data processing result as follows:

[0013] According to the constructed eigenvector, the rate of change of the eigenvector ΔF(t) is calculated, then,

[0014] ΔF(t)=F(t)-F(t-1)

[0015] Wherein, F(t) represents the data feature vector at time t, F(t-1) represents the data feature vector at time t-1, and ΔF(t) represents the rate of change of the data feature vector;

[0016] According to the calculated rate of change of the data feature vector, the acquisition rate is dynamically adjusted. The specific adjustment is as follows:

[0017] If the calculated rate of change of the data feature vector satisfies the formula ΔF(t)≥2(F(t+1)-F(t)), it means that the currently collected data is abnormal, and the current data collection rate is reduced;

[0018] If the calculated rate of change of the data feature vector satisfies the formula ΔF(t)<2(F(t+1)-F(t)), it means that the currently collected data is normal, and the current data collection rate is maintained for data collection.

[0019] As a preferred solution of the intelligent monitoring system for the broken belt catching device of the present invention, the identification of potential broken belt risks based on historical broken belt data is as follows:

[0020] A data feature vector F′ is constructed based on historical belt-breaking events, and the feature values ​​in the constructed data feature vector are consistent with the feature vector constructed using the data collected by the sensor;

[0021] And set the abnormal data threshold S according to the historical belt break time e ;

[0022] Calculate the similarity between the feature vector constructed by historical fault belt data and the feature vector constructed by current data, then we have:

[0023]

[0024] Among them, F represents the feature vector constructed by the current data, F′ represents the feature vector constructed by the historical fault belt data, and S(F,F′) represents the similarity score between the feature vector constructed by the current data and the feature vector constructed by the historical fault belt data, which is used to judge whether the currently collected data has the potential fault belt risk. The specific judgment is as follows:

[0025] If the calculated similarity score is compared with the set abnormal data threshold and satisfies the formula S(F,F′)+σ≤S e When , σ represents the mean value of the feature vector constructed by historical data, indicating that the conveyor belt corresponding to the currently collected data has no risk of belt breaking;

[0026] If the calculated similarity score is compared with the set abnormal data threshold and satisfies the formula S(F,F′)+σ>S e When , σ represents the mean of the feature vector constructed by historical data, indicating that the conveyor belt corresponding to the currently collected data has the risk of belt breakage.

[0027] As a preferred solution of the intelligent monitoring system for the broken belt catching device of the present invention, the secondary screening by constructing the abnormal data feature vector is specifically as follows:

[0028] The data with belt break risk is constructed as the abnormal data feature vector F risk (t);

[0029] And based on the dynamic adaptive weights, the feature fusion of abnormal data is performed, then,

[0030] F risk (t)=w1(t)·T(t)+w2(t)·Vib(t)+w3(t)·Img(t)

[0031] Among them, w1(t), w2(t), w3(t) represent the weights of belt tension value data, conveyor belt vibration intensity data and conveyor belt surface image data respectively, T(t) represents the collected conveyor belt tension value data, Vib(t) represents the collected conveyor belt vibration intensity data, Img(t) represents the collected conveyor belt surface image data, F risk (t) represents abnormal data characteristics.

[0032] As a preferred solution of the intelligent monitoring system for a broken belt catching device of the present invention, the adaptive adjustment of the weight of the abnormal data feature vector is as follows:

[0033] The weights of abnormal data features are adaptively adjusted according to the broken belt position. The specific adjustments are as follows:

[0034] w i (t) = w i (t-1)+α·f(P est (t-1),R(t-1))

[0035] Among them, w i (t) represents the weight of the i-th eigenvalue in the abnormal feature vector at time t, w i (t-1) represents the weight of the i-th eigenvalue in the abnormal feature vector at time t-1, α represents the adjustment coefficient, P est (t-1) represents the broken belt position predicted based on the abnormal data at time t-1, R(t-1) represents the warning status at time t-1, indicating whether the intelligent warning is triggered. If the intelligent warning is triggered, the value is 1, otherwise the value is 0. f(P(t-1), R(t-1)) represents the feedback function, which is set according to the estimated value of the broken belt position and the warning status at time t-1. The specific settings are as follows:

[0036] If P est The position prediction of (t-1) is accurate and triggers intelligent warning, and the feedback function takes the value of +1;

[0037] If P est The position prediction at (t-1) is inaccurate and no intelligent warning is triggered, so the feedback function takes the value of -1;

[0038] When the variable value in the feedback function is other results, the feedback function value is 0.

[0039] As a preferred solution of the intelligent monitoring system for a broken belt catching device of the present invention, the feature extraction of image data using a convolutional neural network is specifically as follows:

[0040] By sliding the filter on the input image to extract the image features, we have:

[0041]

[0042] Among them, I(x,y) represents the input image, which is the fused feature map after multi-scale edge enhancement, K(m,n) represents the convolution kernel of the convolutional neural network, (x,y) represents the position coordinates of the output feature map, and I conv (x, y) represents the image features after being processed by the image processing module.

[0043] As a preferred solution of the intelligent monitoring system for a broken belt catching device described in the present invention, the feature fusion is specifically as follows:

[0044] The features I extracted by the image processing module conv (x,y) and the new eigenvector F fused,risk (t) is used as input and processed by the fully connected layer to complete feature fusion, then we have:

[0045] h=ReLU(W·x+b)

[0046] Among them, W represents the weight matrix of the fully connected layer, x represents the input feature, b represents the bias term, ReLU represents the activation function, and h represents the fused feature;

[0047] The fused features are output through the output layer to complete the prediction of the fault position, then we have:

[0048] P est (t) = W out ·h+b out

[0049] Among them, W out represents the weight matrix of the output layer, b out represents the bias term of the output layer, h represents the fused feature, P est (t) represents the prediction result of the broken belt position at time t.

[0050] As a preferred solution of the intelligent monitoring system for a broken belt catching device of the present invention, the loss function is constructed to complete the precise positioning of the broken belt position and the intelligent early warning as follows:

[0051] The accuracy of the model is evaluated by defining the loss function, specifically:

[0052]

[0053] Among them, P est (t) represents the prediction result of the fault position at time t, N represents the total amount of data involved in the evaluation, i represents the historical fault position index, P(i) represents the i-th historical fault position, which is the actual fault position under the historical fault data. The historical data is the data corresponding to the current fusion feature, and L represents the output result of the loss function, which is used to evaluate the accuracy of the model. The specific evaluation is as follows:

[0054] If the output result of the loss function satisfies the formula L=0, it means that the current model performance is accurate, and the model's prediction position for the broken belt is accurate;

[0055] If the output result of the loss function satisfies the formula L≠0, the intelligent warning threshold L′ is set, and whether to trigger the intelligent warning is determined according to the threshold comparison result, specifically:

[0056] If the output result of the loss function is compared with the intelligent warning threshold and satisfies the formula L≥L′, it means that there is a risk of belt breakage in the current transmission, and the intelligent warning is triggered;

[0057] If the output result of the loss function is compared with the intelligent warning threshold and satisfies the formula L<L′, it means that there is no risk of belt breakage in the current transmission and the intelligent warning is not triggered.

[0058] Beneficial effects of the present invention:

[0059] By combining sensor data and processed image data with a multimodal deep learning model, the risk of belt breakage can be accurately identified and located, reducing false alarm and missed alarm rates.

[0060] Through real-time monitoring and feedback of the capture process, the accuracy of the capture device is improved to ensure that the broken belt can be effectively captured, avoiding equipment damage or personal injury caused by capture failure;

[0061] The system can dynamically adjust monitoring strategies and capture parameters based on historical data and real-time monitoring results, thereby improving the adaptability and intelligence of the overall monitoring system. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0063] Figure 1 The present invention is a schematic diagram of the overall system step structure of an intelligent monitoring system for a broken belt catching device. DETAILED DESCRIPTION

[0064] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0065] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0066] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0067] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0068] Meanwhile, in the description of the present invention, it should be noted that the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0069] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0070] Example 1

[0071] Reference Figure 1 , which is an embodiment of the present invention, provides an intelligent monitoring system for a broken belt capturing device, including a sensor monitoring module, an intelligent processing module and a response control module.

[0072] Specifically, the sensor monitoring module is used to monitor the tension, position, temperature, vibration and belt surface status of the conveyor belt in real time, and identify potential risks of belt breakage. The intelligent processing module is used to perform fusion analysis on the collected multi-dimensional data, accurately identify the risk of belt breakage and wear position, and implement intelligent early warning. The response control module is used to quickly respond to the belt breakage signal and automatically execute the capture action to ensure the accuracy and reliability of the capture.

[0073] Furthermore, the sensor monitoring module is used to monitor relevant data of the conveyor belt in real time and identify potential belt breakage risks, including a data acquisition and processing unit and a data monitoring unit;

[0074] The data acquisition and processing unit collects the relevant data of the conveyor belt by installing data sensors, and performs data preprocessing on the collected data for subsequent data detection, which is specifically implemented as follows:

[0075] A tension sensor, a vibration sensor and an optical sensor are used to collect conveyor belt related data, including conveyor belt tension value data, conveyor belt vibration intensity data and conveyor belt surface image data, and the collected data are constructed into a data feature vector F, and the formula F(t)=[T(t), Vib(t), Img(t)] is satisfied, wherein T(t) represents the collected conveyor belt tension value data, Vib(t) represents the collected conveyor belt vibration intensity data, Img(t) represents the collected conveyor belt surface image data, and F(t) represents the constructed data feature vector;

[0076] At the same time, the collected data will be preprocessed to eliminate abnormal data values, sort the data according to the collected time series, fill in the missing data, and finally normalize the data so that the data of each sensor is maintained within a uniform scale. Data preprocessing and data normalization are both existing public technologies that can be easily associated with those skilled in the art and will not be elaborated here.

[0077] Furthermore, the data detection unit performs data processing on the collected data and completes data detection according to the data processing result, which is specifically implemented as follows:

[0078] According to the constructed eigenvector, the rate of change of the eigenvector ΔF(t) is calculated, then,

[0079] ΔF(t)=F(t)-F(t-1)

[0080] Wherein, F(t) represents the data feature vector at time t, F(t-1) represents the data feature vector at time t-1, and ΔF(t) represents the rate of change of the data feature vector;

[0081] According to the calculated rate of change of the data feature vector, the acquisition rate is dynamically adjusted. The specific adjustment is as follows:

[0082] If the calculated rate of change of the data feature vector satisfies the formula ΔF(t)≥2(F(t+1)-F(t)), it means that the currently collected data is abnormal, and the current data collection rate is reduced;

[0083] If the calculated rate of change of the data feature vector satisfies the formula ΔF(t)<2(F(t+1)-F(t)), it means that the currently collected data is normal, and the current data collection rate is maintained for data collection.

[0084] It should be noted that, for the data collected after the collection rate is adjusted, the corresponding change rate is recalculated, and a secondary determination of the dynamic adjustment of the collection rate is performed until the calculated change rate satisfies the determination formula.

[0085] Furthermore, the identification of potential belt break risks is achieved by comparing the constructed data feature vector with the data in the historical data comparison library, and the specific identification is as follows:

[0086] A data feature vector F′ is constructed based on historical belt-breaking events, and the feature values ​​in the constructed data feature vector are consistent with the feature vector constructed using the data collected by the sensor;

[0087] And set the abnormal data threshold S according to the historical belt break time e ;

[0088] Calculate the similarity between the feature vector constructed from historical fault belt data and the feature vector constructed from current data, then we have:

[0089]

[0090] Among them, F represents the feature vector constructed by the current data, F′ represents the feature vector constructed by the historical fault belt data, and S(F,F′) represents the similarity score between the feature vector constructed by the current data and the feature vector constructed by the historical fault belt data, which is used to judge whether the currently collected data has the potential fault belt risk. The specific judgment is as follows:

[0091] If the calculated similarity score is compared with the set abnormal data threshold and satisfies the formula S(F,F′)+σ≤S e When , σ represents the mean value of the feature vector constructed by historical data, indicating that the conveyor belt corresponding to the currently collected data has no risk of belt breaking;

[0092] If the calculated similarity score is compared with the set abnormal data threshold and satisfies the formula S(F,F′)+σ>S e When , σ represents the mean of the feature vector constructed by historical data, indicating that the conveyor belt corresponding to the currently collected data has the risk of belt breakage.

[0093] Furthermore, the intelligent processing module is used to perform fusion analysis on the collected multi-dimensional data, accurately identify the risk of belt breakage and wear position, and realize intelligent early warning. It performs secondary fusion analysis on the data with belt breakage risk, and accurately locates the belt breakage position based on the analysis results. According to the positioning results, intelligent early warning is realized, including a secondary fusion analysis unit, a belt breakage positioning unit and an intelligent early warning unit.

[0094] Specifically, the secondary fusion analysis unit performs secondary screening analysis on data with a risk of belt breakage, as follows:

[0095] The data with belt break risk is constructed as the abnormal data feature vector F risk (t);

[0096] And based on the dynamic adaptive weights, the feature fusion of abnormal data is performed, then,

[0097] F risk (t)=w1(t)·T(t)+w2(t)·Vib(t)+w3(t)·Img(t)

[0098] Among them, w1(t), w2(t), w3(t) represent the weights of belt tension value data, conveyor belt vibration intensity data and conveyor belt surface image data respectively, T(t) represents the collected conveyor belt tension value data, Vib(t) represents the collected conveyor belt vibration intensity data, Img(t) represents the collected conveyor belt surface image data, F risk (t) represents abnormal data characteristics;

[0099] The weights of abnormal data features are adaptively adjusted according to the broken belt position. The specific adjustments are as follows:

[0100] w i (t) = w i (t-1)+α·f(P est (t-1),R(t-1))

[0101] Among them, w i (t) represents the weight of the i-th eigenvalue in the abnormal feature vector at time t, w i (t-1) represents the weight of the i-th eigenvalue in the abnormal feature vector at time t-1, α represents the adjustment coefficient, which is set by the implementer according to the actual application scenario, and P est (t-1) represents the broken belt position predicted based on the abnormal data at time t-1, R(t-1) represents the warning status at time t-1, indicating whether the intelligent warning is triggered. If the intelligent warning is triggered, the value is 1, otherwise the value is 0. f(P(t-1), R(t-1)) represents the feedback function, which is set according to the estimated value of the broken belt position and the warning status at time t-1. The specific settings are as follows:

[0102] If P est The position prediction at (t-1) is accurate and triggers the intelligent warning. The feedback function value at this time is +1;

[0103] If P est The position prediction at (t-1) is inaccurate and no intelligent warning is triggered. The feedback function value at this time is -1;

[0104] When the variable value in the feedback function is other results, the feedback function takes the value of 0. The variable value in the feedback function is other results, which are not described one by one in this embodiment, and the implementer can set it according to the actual application scenario.

[0105] Furthermore, the broken belt positioning unit performs feature enhancement and extraction on the conveyor belt surface image data, dynamically combines the processed image data with other sensor data, and comprehensively analyzes the conveyor belt state through a deep learning algorithm to achieve accurate positioning of the broken belt position;

[0106] The feature enhancement and extraction of the conveyor belt surface image data is specifically implemented as follows:

[0107] Generate multiple smoothed images using Gaussian filters of different sizes Where k represents the scale index of edge detection, represents the smoothed image at the kth index;

[0108] By fusing the smooth images at all scales and enhancing the features of the image data, we have:

[0109]

[0110] Among them, K represents the total number of all scales, represents the smoothed image at the kth index, I fused (t) represents the fused feature map after multi-scale edge enhancement.

[0111] Furthermore, the comprehensive analysis of the conveyor belt status by deep learning algorithm is to combine the sensor data with the processed image data, use the deep learning model for joint training, complete the analysis of abnormal data features, and locate the possible position of the broken belt, which is specifically implemented as follows:

[0112] The sensor abnormal data is fused with the processed image data to form a new feature vector F fused,risk (t);

[0113] A multimodal deep learning model is constructed based on the deep learning algorithm, including an image processing module and a feature processing module, as follows:

[0114] The image processing module uses a convolutional neural network to extract the features of image data, and uses a filter to slide on the input image to extract the image features. Then,

[0115]

[0116] Among them, I(x,y) represents the input image, which is the fused feature map after multi-scale edge enhancement, K(m,n) represents the convolution kernel of the convolutional neural network, (x,y) represents the position coordinates of the output feature map, and I conv (x, y) represents the image features after being processed by the image processing module;

[0117] The feature processing module is used to process the features extracted by the image processing module. conv (x,y) and the new eigenvector F fused,risk (t) is used as input and processed by the fully connected layer to complete feature fusion, then we have:

[0118] h=ReLU(W·x+b)

[0119] Among them, W represents the weight matrix of the fully connected layer, x represents the input feature, b represents the bias term, ReLU represents the activation function, and h represents the fused feature;

[0120] The fused features are output through the output layer to complete the prediction of the fault position, then we have:

[0121] P est (t) = W out ·h+b out

[0122] Among them, W out represents the weight matrix of the output layer, b out represents the bias term of the output layer, h represents the fused feature, P est (t) represents the prediction result of the broken belt position at time t;

[0123] The accuracy of the model is evaluated by defining the loss function, specifically:

[0124]

[0125] Among them, P est (t) represents the prediction result of the fault position at time t, N represents the total amount of data involved in the evaluation, i represents the historical fault position index, P(i) represents the i-th historical fault position, which is the actual fault position under the historical fault data. The historical data is the data corresponding to the current fusion feature, and L represents the output result of the loss function, which is used to evaluate the accuracy of the model. The specific evaluation is as follows:

[0126] If the output result of the loss function satisfies the formula L=0, it means that the current model performance is accurate, and the model's prediction position for the broken belt is accurate;

[0127] If the output result of the loss function satisfies the formula L≠0, the intelligent warning threshold L′ is set, and whether to trigger the intelligent warning is determined according to the threshold comparison result, specifically:

[0128] If the output result of the loss function is compared with the intelligent warning threshold and satisfies the formula L≥L′, it means that there is a risk of belt breakage in the current transmission, and the intelligent warning is triggered;

[0129] If the output result of the loss function is compared with the intelligent warning threshold and satisfies the formula L<L′, it means that there is no risk of belt breakage in the current transmission and the intelligent warning is not triggered.

[0130] Furthermore, the response control module is used to quickly respond to the broken belt signal and automatically perform the capture action to ensure the accuracy and reliability of the capture. It is a quick response to the capture action based on the prediction result of the broken belt position. The specific implementation is as follows:

[0131] For the predicted fault position P est (t), calculate the distance l1(t) between the broken belt position and the capture device, and collect the basic time t2 for the capture device to complete the capture action;

[0132] Collect the moving speed v(t) of the conveyor belt and calculate the time t1 consumed by the broken belt position to move to the capture device, then we have:

[0133]

[0134] Among them, l1(t) represents the distance between the predicted position of the broken belt and the capture device, v(t) represents the moving speed of the conveyor belt, and t1 represents the time consumed by the broken belt position to move to the capture device, which is used to determine the rapid response level, as follows:

[0135] If the time taken for the broken belt to move to the capture device and the basic time for the capture device to complete the capture action satisfy the formula t1≥t2, the rapid response level is level 1, and the current capture action signal is maintained to complete the capture of the broken belt;

[0136] If the time consumed by the broken belt position to move to the capture device and the basic time for the capture device to complete the capture action meet the formula t1<t2<2t1, the rapid response level is level 2, the capture response signal is increased, and the capture action of the capture device is executed in advance to ensure that the capture action is completed when the broken belt position reaches the capture device;

[0137] If the time consumed for the broken belt position to move to the capturing device is compared with the basic time for the capturing device to complete the capturing action and satisfies the formula 2t1<t2, the rapid response level is level three. At this time, the capturing device cannot capture the current broken belt. The machine is stopped immediately and relevant staff are notified to handle it.

[0138] It should be noted that the capture action of the capture device when the rapid response level is level 2 is achieved by increasing the force applied to the capture device to shorten the overall time from the beginning to the end of the capture action, which is specifically implemented as follows:

[0139] For the basic time t2 to complete the capture action, by increasing the force applied to the capture device, the overall time from the beginning to the end of the capture action can be shortened, then,

[0140] The enhancement coefficient of the external force of the capture device is determined according to the time t1 consumed by the broken belt position moving to the capture device and the basic time t2 for the capture device to complete the capture action, specifically:

[0141]

[0142] And according to the determined enhancement coefficient, the external force of the capture device is adjusted, then,

[0143] F new =F out ·k f

[0144] Among them, t1 represents the time consumed by the broken belt position to move to the capture device, t2 represents the basic time for the capture device to complete the capture action, and k f Indicates the external force enhancement coefficient of the capture device, F out represents the initial external force of the capture device, F new Indicates the external force of the adjusted capture device.

[0145] For the capture device after the external force is adjusted, the capture speed is detected and fed back. If the adjusted capture device cannot achieve accurate capture of the broken belt, the external force is continuously adjusted until accurate capture of the broken belt is achieved.

[0146] Furthermore, if the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the system described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0147] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0148] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0149] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0150] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0151] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent monitoring system for a broken belt catching device, comprising: The sensor monitoring module collects relevant data of the conveyor belt and processes the data, completes data detection according to the data processing results, thereby dynamically adjusting the data collection frequency, and identifying potential belt break risks based on historical belt break data, characterized in that it also includes: The intelligent processing module is used to perform secondary screening on data with belt break risk by constructing abnormal data feature vectors, and adaptively adjust the weight of abnormal data feature vectors according to the combination of belt break prediction position and intelligent early warning mechanism, wherein the belt break prediction position also includes: The surface data of the conveyor belt is enhanced, and the feature extraction of the image data is completed using a convolutional neural network. The extracted image features are combined with the sensor data through the fully connected layer processing technology, and the feature fusion is completed and output through the output layer. At the same time, a loss function is constructed to accurately locate the broken belt position and intelligently warn. It also includes a response control module, which calculates the distance between the broken belt position and the capture device based on the predicted broken belt position, and calculates the time consumed by the broken belt position to move to the capture device, and at the same time collects the basic time to complete the capture action, so as to complete the rapid response of the broken belt capture by comparing the time.

2. The intelligent monitoring system for a broken belt catching device according to claim 1, characterized in that: The data detection is completed according to the data processing result as follows: According to the constructed eigenvector, the rate of change of the eigenvector ΔF(t) is calculated, then, ΔF(t)=F(t)-F(t-1) Wherein, F(t) represents the data feature vector at time t, F(t-1) represents the data feature vector at time t-1, and ΔF(t) represents the rate of change of the data feature vector; According to the calculated rate of change of the data feature vector, the acquisition rate is dynamically adjusted. The specific adjustment is as follows: If the calculated rate of change of the data feature vector satisfies the formula ΔF(t)≥2(F(t+1)-F(t)), it means that the currently collected data is abnormal, and the current data collection rate is reduced; If the calculated rate of change of the data feature vector satisfies the formula ΔF(t)<2(F(t+1)-F(t)), it means that the currently collected data is normal, and the current data collection rate is maintained for data collection.

3. The intelligent monitoring system for a broken belt catching device according to claim 2, characterized in that: The identification of potential belt break risks based on historical belt break data is as follows: A data feature vector F′ is constructed based on historical belt-breaking events, and the feature values ​​in the constructed data feature vector are consistent with the feature vector constructed using the data collected by the sensor; And set the abnormal data threshold S according to the historical belt break time e ; Calculate the similarity between the feature vector constructed by historical fault belt data and the feature vector constructed by current data, then we have: Among them, F represents the feature vector constructed by the current data, F′ represents the feature vector constructed by the historical fault belt data, and S(F,F′) represents the similarity score between the feature vector constructed by the current data and the feature vector constructed by the historical fault belt data, which is used to judge whether the currently collected data has the potential fault belt risk. The specific judgment is as follows: If the calculated similarity score is compared with the set abnormal data threshold and satisfies the formula S(F,F′)+σ≤S e When , σ represents the mean value of the feature vector constructed by historical data, indicating that the conveyor belt corresponding to the currently collected data has no risk of belt breaking; If the calculated similarity score is compared with the set abnormal data threshold and satisfies the formula S(F,F′)+σ>S e When , σ represents the mean of the feature vector constructed by historical data, indicating that the conveyor belt corresponding to the currently collected data has the risk of breaking.

4. The intelligent monitoring system for a broken belt catching device according to claim 3, characterized in that: The secondary screening by constructing the abnormal data feature vector is specifically as follows: The data with belt break risk is constructed as the abnormal data feature vector F risk (t); And based on the dynamic adaptive weights, feature fusion of abnormal data is performed, then, F risk (t)=w1(t)·T(t)+w2(t)·Vib(t)+w3(t)·Img(t) Among them, w1(t), w2(t), w3(t) represent the weights of belt tension value data, conveyor belt vibration intensity data and conveyor belt surface image data respectively, T(t) represents the collected conveyor belt tension value data, Vib(t) represents the collected conveyor belt vibration intensity data, Img(t) represents the collected conveyor belt surface image data, F risk (t) represents abnormal data characteristics.

5. The intelligent monitoring system for a broken belt catching device according to claim 4, characterized in that: The adaptive adjustment of the abnormal data feature vector weight is specifically as follows: The weights of abnormal data features are adaptively adjusted according to the broken belt position. The specific adjustments are as follows: w i (t)=w i (t-1)+α·f(P est (t-1),R(t-1)) Among them, w i (t) represents the weight of the i-th eigenvalue in the abnormal feature vector at time t, w i (t-1) represents the weight of the i-th eigenvalue in the abnormal feature vector at time t-1, α represents the adjustment coefficient, P est (t-1) represents the broken belt position predicted based on the abnormal data at time t-1, R(t-1) represents the warning status at time t-1, indicating whether the intelligent warning is triggered. If the intelligent warning is triggered, the value is 1, otherwise the value is 0. f(P(t-1), R(t-1)) represents the feedback function, which is set according to the estimated value of the broken belt position and the warning status at time t-1. The specific settings are as follows: If P est The position prediction of (t-1) is accurate and triggers intelligent warning, and the feedback function takes the value of +1; If P est The position prediction at (t-1) is inaccurate and no intelligent warning is triggered, so the feedback function takes the value of -1; When the variable value in the feedback function is other results, the feedback function value is 0.

6. The intelligent monitoring system for a broken belt catching device according to claim 5, characterized in that: The feature extraction of image data using convolutional neural network is specifically as follows: By sliding the filter on the input image to extract the image features, we have: Among them, I(x,y) represents the input image, which is the fused feature map after multi-scale edge enhancement, K(m,n) represents the convolution kernel of the convolutional neural network, (x,y) represents the position coordinates of the output feature map, and I conv (x, y) represents the image features after being processed by the image processing module.

7. The intelligent monitoring system for a broken belt catching device according to claim 6, characterized in that: The feature fusion is specifically as follows: The features I extracted by the image processing module conv (x,y) and the new eigenvector F fused,risk (t) is used as input and processed by the fully connected layer to complete feature fusion, then we have: h=ReLU(W·x+b) Among them, W represents the weight matrix of the fully connected layer, x represents the input feature, b represents the bias term, ReLU represents the activation function, and h represents the fused feature; The fused features are output through the output layer to complete the prediction of the fault position, then we have: P est (t)=W out ·h+b out Among them, W out represents the weight matrix of the output layer, b out represents the bias term of the output layer, h represents the fused feature, P est (t) represents the prediction result of the broken belt position at time t.

8. The intelligent monitoring system for a broken belt catching device according to claim 7, characterized in that: The loss function is constructed to achieve accurate positioning of the broken belt position and intelligent early warning as follows: The accuracy of the model is evaluated by defining the loss function, specifically: Among them, P est (t) represents the prediction result of the fault position at time t, N represents the total amount of data involved in the evaluation, i represents the historical fault position index, P(i) represents the i-th historical fault position, which is the actual fault position under the historical fault data. The historical data is the data corresponding to the current fusion feature, and L represents the output result of the loss function, which is used to evaluate the accuracy of the model. The specific evaluation is as follows: If the output result of the loss function satisfies the formula L=0, it means that the current model performance is accurate, and the model's prediction position for the broken belt is accurate; If the output result of the loss function satisfies the formula L≠0=, the intelligent warning threshold L′ is set, and whether to trigger the intelligent warning is determined according to the threshold comparison result, specifically: If the output result of the loss function is compared with the intelligent warning threshold and satisfies the formula L≥L′, it means that there is a risk of belt breakage in the current transmission, and the intelligent warning is triggered; If the output result of the loss function is compared with the intelligent warning threshold and satisfies the formula L<L′, it means that there is no risk of belt breakage in the current transmission and the intelligent warning is not triggered.

Images