Data-driven material conveyor fault early warning method and system

Abstract

The invention discloses a data-driven material conveyor fault early warning method and system, and relates to the field related to data early warning, and the method comprises the steps: arranging a plurality of sensors for real-time collection, obtaining a real-time operation data set of a material conveyor for multi-dimensional analysis, determining multi-dimensional operation data for time series decomposition, extracting a plurality of operation features, and carrying out the early warning of the faults of the material conveyor; learning training is carried out according to the historical operation record log to obtain a fault prediction model, the real-time operation data set is synchronized to the fault prediction model for prediction calculation, and the operation health score of the material conveyor is determined to carry out fault judgment on the material conveyor. And determining a fault risk level, generating an early warning signal, and sending the early warning signal to a remote terminal for intelligent early warning feedback of the fault. The technical problem that intelligent maintenance is difficult to achieve due to the fact that fault monitoring of the material conveyor is not timely and the early warning precision is low is solved, real-time monitoring and intelligent feedback of the equipment state are achieved through data-driven fault prediction and real-time early warning, and the technical effect of improving the operation and maintenance efficiency is achieved.

Description

Technical Field

[0001] The present application relates to data warning related fields, and in particular to a data-driven material conveyor fault warning method and system. Background Art

[0002] With the continuous development of industrial automation and intelligence, material conveyors, as important conveying equipment in production lines, are widely used in mining, logistics, manufacturing and other fields. Their stable operation is directly related to the efficiency and safety of production lines. However, the fault monitoring method of traditional material conveyors mainly relies on regular manual inspections and equipment shutdown maintenance, which is not only inefficient, but also unable to detect potential faults in time, which can easily lead to unexpected equipment shutdowns or even major production accidents. Especially in scenarios with complex operating environments and frequent load changes, traditional monitoring methods are difficult to meet the needs of efficient production and safe operation.

[0003] In the existing technology, some material conveyors collect data by installing simple sensors on key components, such as monitoring the tension, vibration, motor temperature and other parameters of the conveyor belt. However, these data are usually only used for monitoring a single indicator and simple threshold alarms, lacking the ability to comprehensively analyze multi-dimensional data and difficult to capture potential fault characteristics. In addition, the warning of traditional methods is mainly based on rule algorithms, and the definition of abnormalities relies on fixed empirical thresholds. It lacks dynamic adjustment capabilities and is prone to false alarms or missed alarms when dealing with complex operating conditions. Summary of the invention

[0004] This application solves the technical problem of difficulty in realizing intelligent maintenance due to untimely fault monitoring and low warning accuracy of material conveyor by providing a data-driven material conveyor fault warning method and system. Through data-driven fault prediction and real-time warning, real-time monitoring and intelligent feedback of equipment status are realized, achieving the technical effect of improving operation and maintenance efficiency.

[0005] The present application provides a data-driven material conveyor fault warning method, which is applied to a data-driven material conveyor fault warning system, and includes: traversing the material conveyor and deploying a plurality of sensors for real-time collection to obtain a real-time operation data set of the material conveyor; Perform multidimensional analysis based on the real-time operation data set to determine multidimensional operation data, perform time series decomposition based on the multidimensional operation data, and extract multiple operation features; Retrieving the historical operation record log of the material conveyor, performing learning and training according to the historical operation record log, and obtaining a fault prediction model; Synchronizing the real-time operation data set to the fault prediction model for prediction calculation to determine the operation health score of the material conveyor; According to the operation health score, the material conveyor is fault-judged to determine the fault risk level, an early warning signal is generated according to the fault risk level, and the early warning signal is sent to a remote terminal for intelligent early warning feedback of the fault.

[0006] In a possible implementation, a multidimensional analysis is performed based on the real-time operation data set to determine the multidimensional operation data, and the following processing is performed: data cleaning is performed based on the real-time operation data set to obtain a real-time operation standard data set; similarity calculation is performed on the real-time operation standard data set to determine a plurality of similarity coefficients; a plurality of cluster centers are determined according to the plurality of similarity coefficients, and the real-time operation standard data set is divided into a plurality of data clusters according to the plurality of cluster centers; and the plurality of data clusters are added to the multidimensional operation data.

[0007] In a possible implementation, time series decomposition is performed according to the multidimensional operation data, multiple operation features are extracted, and the following processing is performed: multiple data timestamps of the multiple cluster centers are extracted, the multiple data timestamps are arranged in sequence, and multiple time series are constructed, and the multiple time series have a corresponding relationship with the multidimensional operation data; the multidimensional operation data group is sorted according to the multiple time series in the time dimension to construct a time series model; the time series model is used to decompose to obtain multiple decomposition components, and the multiple decomposition components are integrated to generate the multiple operation features.

[0008] In a possible implementation, learning and training are performed according to the historical operation record logs to obtain a fault prediction model, and the following processing is performed: traverse the historical operation record logs to perform operation fault analysis and determine multiple fault operation characteristics; feature annotate the historical operation record logs based on the multiple fault operation characteristics and determine multiple fault annotation information; perform data segmentation on the historical operation record logs according to the multiple fault annotation information to determine a training data set and a verification data set; perform supervised training iterations based on the training data set, and when the training data set tends to converge, use the verification data set for cross-validation, and obtain the fault prediction model based on the verification result.

[0009] In a possible implementation, the real-time operation data set is synchronized to the fault prediction model for predictive calculation, the operation health score of the material conveyor is determined, and the following processing is performed: the fault prediction model is loaded, and the real-time operation data set is synchronized to the fault prediction model to perform operation prediction on the material conveyor, and an operation fault prediction result is generated; the historical operation record log is mapped to the operation fault prediction result for association, and an operation fault association list is determined; fault calculation is performed according to the operation fault association list to generate multiple fault probabilities, and a health assessment is performed based on the multiple fault probabilities to generate the operation health score.

[0010] In a possible implementation, the material conveyor is fault-judged according to the operation health score, the fault risk level is determined, a warning signal is generated according to the fault risk level, and the warning signal is sent to a remote terminal for intelligent warning feedback of the fault, and the following processing is performed: an expected operation threshold is set based on the historical operation record log, and the material conveyor is fault-judged according to the expected operation threshold and the operation health score to generate a fault judgment result; a critical analysis is performed according to the fault judgment result to determine a fault risk data matrix, and the fault risk data matrix is ​​corrected based on the multiple operation characteristics to generate a fault risk level; the fault risk level is matched to the operation health score to generate a state signal set, and the state signal set includes a normal state signal, a warning state signal, and a fault state signal; the normal state signal, the warning state signal, and the fault state signal are screened according to the fault risk level, and the warning signal is determined and sent to a remote terminal for warning feedback.

[0011] In a possible implementation, a fault determination is performed on the material conveyor according to the expected operating threshold and the operating health score, a fault determination result is generated, and the following processing is performed: the operating health score is compared with the expected operating threshold to determine whether the operating health score is greater than or equal to the expected operating threshold; if the operating health score is greater than or equal to the expected operating threshold, normal operating data is obtained, continuous operation monitoring is performed based on the normal operating data, an operation monitoring instruction is generated, and the operation monitoring instruction is added to the fault determination result; if the operating health score is less than the expected operating threshold, an abnormal operating data set is generated, a fault impact analysis is performed based on the abnormal operating data set, and an operating impact factor is generated; weights are assigned to the abnormal operating data set based on the operating impact factor to determine multiple weight coefficients; the multiple weight coefficients are associated with the abnormal operating data set to generate the fault determination result.

[0012] The present application also provides a data-driven material conveyor fault warning system, including: a data acquisition module, which is used to traverse the material conveyor and deploy multiple sensors for real-time data collection to obtain a real-time operation data set of the material conveyor; a feature extraction module, which is used to perform multi-dimensional analysis based on the real-time operation data set, determine the multi-dimensional operation data, perform time series decomposition based on the multi-dimensional operation data, and extract multiple operation features; a learning and training module, which is used to retrieve the historical operation record log of the material conveyor, perform learning and training based on the historical operation record log, and obtain a fault prediction model; a prediction calculation module, which is used to synchronize the real-time operation data set to the fault prediction model for prediction calculation, and determine the operation health score of the material conveyor; a fault judgment module, which is used to perform fault judgment on the material conveyor according to the operation health score, determine the fault risk level, generate a warning signal based on the fault risk level, and send the warning signal to a remote terminal for intelligent warning feedback of the fault.

[0013] One or more technical solutions provided in this application have at least the following technical effects or advantages: The data-driven material conveyor fault warning method and system provided in this application relate to the field of data warning technology, and solve the technical problem that material conveyor fault monitoring is not timely and the warning accuracy is low, which makes it difficult to achieve intelligent maintenance. Through data-driven fault prediction and real-time warning, real-time monitoring and intelligent feedback of equipment status are realized, achieving the technical effect of improving operation and maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical scheme of the embodiment of the present application, the accompanying drawings of the embodiment of the present application will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or a certain step or several steps of operations can be removed from these processes.

[0015] Figure 1 A schematic flow chart of a data-driven material conveyor fault warning method provided in an embodiment of the present application; Figure 2 A schematic structural diagram of a data-driven material conveyor fault warning system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0017] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0018] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0019] The embodiment of the present application provides a data-driven material conveyor fault warning method, which is applied to a data-driven material conveyor fault warning system, such as Figure 1 As shown, the method includes: Step A100, traversing the material conveyor and deploying multiple sensors to collect data in real time, so as to obtain a real-time operation data set of the material conveyor; First, according to the characteristics and operation requirements of the material conveyor, the parameters that need to be monitored, such as temperature, pressure, vibration, flow, speed, etc., are clearly defined, and then the appropriate sensor type is selected according to the monitoring target. For example, temperature sensors are used to monitor equipment temperature changes, vibration sensors are used to monitor equipment vibration, and photoelectric sensors are used for object detection, distance measurement, and speed monitoring. At the same time, according to the structure and layout of the material conveyor, a sensor layout plan is formulated. Determine the installation location, quantity, and connection method of the sensor to ensure that the required data can be collected comprehensively and accurately, and install sensors at key locations of the material conveyor according to the layout plan.

[0020] Furthermore, the sensor is connected to a data acquisition system (such as a PLC). The connection method can be wired or wireless, depending on the characteristics of the equipment and the on-site environment. The data acquisition system collects the raw data transmitted by the sensor. To ensure the real-time, accuracy and completeness of data acquisition, the collected raw data can be processed, including data filtering, data sampling, data encoding, etc., and then the processed data is stored in the memory or external storage device of the data acquisition system for subsequent analysis and processing, so as to realize real-time acquisition of multiple sensors on the material conveyor and obtain the real-time operation data set of the material conveyor. This helps to improve the management and maintenance level of equipment, reduce unexpected downtime, and improve production efficiency and equipment life.

[0021] Execute step A200, perform multidimensional analysis based on the real-time operation data set, determine multidimensional operation data, perform time series decomposition based on the multidimensional operation data, and extract multiple operation features; in a possible implementation, step A200 further includes step A210, perform data cleaning based on the real-time operation data set to obtain a real-time operation standard data set; execute step A220, traverse the real-time operation standard data set to perform similarity calculation, and determine multiple similarity coefficients; execute step A230, determine multiple clustering centers according to the multiple similarity coefficients, and divide the real-time operation standard data set into multiple data clusters according to the multiple clustering centers; execute step A240, add the multiple data clusters to the multidimensional operation data.

[0022] By checking the outliers in the real-time operation data set obtained above, the outliers may be caused by sensor failure, data transmission error or abnormal operation of the equipment. At the same time, statistical methods or machine learning algorithms (such as isolation forests) can be used to identify and process these outliers, and data of different dimensions and ranges are standardized to ensure that they are comparable in similarity calculation and cluster analysis, thereby obtaining a trial operation standard data set. Further, according to the characteristics of the data and business needs, similarity measurement methods such as Euclidean distance, Manhattan distance, cosine similarity, Pearson correlation coefficient, etc. can be used to traverse the real-time operation standard data set, calculate the similarity coefficient between each pair of data points, and then use clustering algorithms such as K-means, DBSCAN, hierarchical clustering, etc. to perform cluster analysis on the trial operation standard data set according to multiple similarity coefficients to determine multiple cluster centers, and the multiple cluster centers contain center points of different categories in the real-time operation data set. Further, according to the cluster centers and similarity coefficients, the real-time operation standard data set is divided into multiple data clusters, which means that the data points in each data cluster have a higher similarity in similarity, while the data points between different data clusters have a lower similarity.

[0023] Finally, a label of the data cluster to which it belongs is added to each data point. The label can be used for subsequent data analysis and visualization, providing strong support for subsequent data analysis of material conveyors.

[0024] In one possible implementation, step A200 further includes step A250, extracting multiple data timestamps of the multiple cluster centers, arranging the multiple data timestamps in sequence, constructing multiple time series, and the multiple time series correspond to the multidimensional operation data; executing step A260, sorting the multidimensional operation data group according to the multiple time series in time dimension, and constructing a time series model; executing step A270, decomposing using the time series model to obtain multiple decomposition components, integrating the multiple decomposition components, and generating the multiple operation characteristics.

[0025] First, the data timestamp corresponding to each cluster center is extracted from the multiple cluster centers obtained by the cluster analysis. The extracted timestamp is used to represent the time position of the cluster center in the data set. At the same time, the multiple extracted data timestamps are arranged in sequence to construct a time series framework. The arrangement order can be based on the sequence of timestamps or specific needs. According to the arranged data timestamps, multiple time series are constructed, and each time series corresponds to the time evolution of a cluster center or a group of related data. At the same time, it is necessary to ensure that there is a one-to-one correspondence between the constructed time series and the multidimensional operating data.

[0026] Furthermore, the multidimensional operation data is organized according to its corresponding characteristic dimensions to form a multidimensional operation data group, each data group in the multidimensional operation data group contains multiple operation characteristic values ​​at the same time point, and at the same time, the multidimensional operation data group is sorted in the time dimension according to the constructed time series to ensure that the data groups are arranged in the order of the time series to reflect the time evolution law of the data, and the sorted multidimensional operation data group is used to construct a time series model, and the time series model can capture the trend, periodicity and seasonality of the data over time. The constructed time series model is used to decompose the data to obtain multiple decomposition components, and the multiple decomposition components can include trend decomposition, periodicity decomposition, and random term decomposition. The trend decomposition is to use moving averages to generate a series of decomposition components. The averaging method or Hodrick-Prescott filter extracts the long-term change trend of the data (such as the gradual increase in temperature or the long-term increase in vibration amplitude). The periodic decomposition is to extract periodic features (such as the change in vibration frequency of equipment operation and the daily change law of temperature) through Fourier transform or wavelet decomposition. The random term decomposition is to separate the random fluctuations outside the trend and periodic parts, extract unpredictable disturbance characteristics, ensure that the impact of noise on the extraction of main features is minimized, and extract features from the decomposed time series. For example, the change rate and inflection point features can be extracted from the trend component, the main frequency, amplitude and phase can be extracted from the periodic component, and the variance, peak and other indicators can be extracted from the random component to evaluate the data volatility. Finally, the multiple decomposed components are integrated according to the time dimension to generate the multiple operating features, which provide important data support for subsequent fault warning, health assessment and maintenance decisions.

[0027] Execute step A300, retrieve the historical operation record log of the material conveyor, perform learning and training according to the historical operation record log, and obtain a fault prediction model; in a possible implementation, step A300 further includes step A310, traverse the historical operation record log to perform operation fault analysis, and determine multiple fault operation characteristics; execute step A320, feature annotate the historical operation record log based on the multiple fault operation characteristics, and determine multiple fault annotation information; execute step A330, perform data segmentation on the historical operation record log according to the multiple fault annotation information, and determine a training data set and a verification data set; execute step A340, perform supervised training iterations based on the training data set, and when the training data set tends to converge, use the verification data set for cross-validation, and obtain the fault prediction model according to the verification result.

[0028] First, the operation data in the historical operation record log is analyzed to extract the features related to the fault. The features related to the fault may include the operation parameters of the equipment (such as temperature, pressure, vibration, etc.), time series data (such as change trends, periodicity, etc.) and environmental factors (such as temperature, humidity, etc.), and the time point when the fault occurred and the corresponding operation features are identified based on the known equipment fault records and log information, so as to determine multiple fault operation features. The multiple fault operation features are parameters or patterns that change significantly when the fault occurs. Based on the extracted fault operation features, the historical operation record log is feature labeled, which means adding labeling information such as fault type and fault level to each time point or data point to form multiple fault labeling information, and the labeled historical operation record log is sorted into multiple fault labeling information. The multiple fault labeling information may include the original operation data and the corresponding fault labeling information.

[0029] Further, according to the size and characteristics of the labeled data set, it is divided into a training data set and a validation data set. The training data set is used to train the fault prediction model, and the validation data set is used to evaluate the performance of the model and perform cross-validation. If there is an imbalance in the number of samples of different fault types in the labeled data set, it is also necessary to perform data balancing through methods such as oversampling, undersampling, or synthetic minority class oversampling technology. The model is trained iteratively with the training data set in a supervised manner. At the same time, the parameters and structure of the model are continuously adjusted during the training process to improve the performance of the model. By observing the changes in indicators such as the loss function value and accuracy during the training process, it is determined whether the model tends to converge. When the performance of the model on the training data set is no longer significantly improved, it can be considered that the model has converged. When the training data set tends to converge, the validation data set is used for cross-validation. The generalization ability of the model is evaluated through cross-validation to avoid overfitting or underfitting. According to the results of cross-validation, the model is further optimized and adjusted. This means that the model's performance can be improved by adjusting the model's parameters, adding regularization terms, using ensemble learning methods, etc. When the model's performance on the validation data set reaches a satisfactory level, it is used as the final fault prediction model, providing strong support for fault prediction and preventive maintenance of material conveyors.

[0030] Execute step A400, synchronize the real-time operation data set to the fault prediction model for predictive calculation, and determine the operation health score of the material conveyor; in one possible implementation, step A400 further includes step A410, load the fault prediction model, synchronize the real-time operation data set to the fault prediction model to perform operation prediction on the material conveyor, and generate an operation fault prediction result; execute step A420, map the historical operation record log to the operation fault prediction result for association, and determine an operation fault association list; execute step A430, perform fault calculation according to the operation fault association list, generate multiple fault probabilities, perform health assessment based on the multiple fault probabilities, and generate the operation health score.

[0031] Load the trained fault prediction model from the storage location, and then synchronize the real-time operation data set to the fault prediction model, while ensuring the real-time, accuracy and completeness of the data. Use the fault prediction model to process the real-time operation data set, including the time series trend, periodic characteristics, and random fluctuation characteristics of the real-time operation data set, to generate an operation fault prediction result, which may include fault type, fault probability, fault occurrence time, etc. At the same time, map the historical operation record log to the operation fault prediction result for association, which means matching through key fields such as timestamps and device IDs to ensure the correspondence of the data. Further, based on the mapping results, generate an operation fault association list, which may include the correspondence between fault information in historical records and real-time prediction results.

[0032] Furthermore, according to the operation fault association list, the faults in the real-time prediction results are counted and analyzed, the probability of each fault type is calculated, and multiple fault probabilities are generated for equipment health assessment, which means health score = 100-(fault probability × weight 1 + slight abnormality probability × weight 2), and the operation health score of the material conveyor is calculated (for example, the score range is 0 to 100 points, and the higher the score, the better the health status of the equipment). The score can also be divided into different levels, such as excellent, good, general, poor, bad, etc., which provides strong support for preventive maintenance, optimized operation and extended service life of material conveyors.

[0033] Finally, step A500 is executed to perform fault determination on the material conveyor according to the operation health score, determine the fault risk level, generate a warning signal according to the fault risk level, and send the warning signal to the remote terminal for intelligent warning feedback of the fault.

[0034] In one possible implementation, step A500 further includes step A510, setting an expected operation threshold based on the historical operation record log, performing fault determination on the material conveyor according to the expected operation threshold and the operation health score, and generating a fault determination result; in one possible implementation, step A510 further includes step A511, comparing the operation health score with the expected operation threshold to determine whether the operation health score is greater than or equal to the expected operation threshold; executing step A512, if the operation health score is greater than or equal to the expected operation threshold, obtaining normal operation data, performing continuous operation monitoring according to the normal operation data, generating operation monitoring instructions, and adding the operation monitoring instructions to the fault determination result; executing step A513, if the operation health score is less than the expected operation threshold, generating an abnormal operation data set, performing fault impact analysis according to the abnormal operation data set, and generating an operation impact factor; executing step A514, performing weight assignment on the abnormal operation data set according to the operation impact factor, and determining multiple weight coefficients; executing step A515, associating the multiple weight coefficients with the abnormal operation data set to generate the fault determination result.

[0035] Based on the historical operation record log, the historical operation analysis of the material conveyor is carried out, the data under normal operation is extracted and integrated, and the expected operation threshold of the boring conveyor is set according to the preset normal operation data of each dimension. Furthermore, the operation health score is compared with the expected operation threshold to determine whether the operation health score is greater than or equal to the expected operation threshold. If the operation health score is greater than or equal to the expected operation threshold, it is considered that the material conveyor is in a normal operation state, and normal operation data is obtained, which may include key operation parameters, time series data, etc. At the same time, the normal operation data is continuously monitored to monitor whether the operation state of the material conveyor is stable and whether there are abnormal changes. Based on the continuous monitoring results, an operation monitoring instruction is generated, which may include continuing monitoring, adjusting operation parameters, performing preventive maintenance, etc. The operation monitoring instruction is then added to the fault judgment result, indicating that the material conveyor is currently in a normal operation state, and the corresponding monitoring instruction is given.

[0036] Furthermore, if the operation health score is less than the expected operation threshold, it is considered that the material conveyor is in an abnormal operation state, and an abnormal operation data set is generated, including abnormal operation parameters, abnormal occurrence time, etc., and a fault impact analysis is performed on the abnormal operation data set to evaluate the impact of abnormal operation on the performance, safety, reliability, etc. of the material conveyor. According to the fault impact analysis results, an operation impact factor is generated, and the operation impact factor may include fault type, fault probability, fault severity, etc., and then the abnormal operation data set is weighted according to the operation impact factor, and multiple weight coefficients are determined to reflect the importance of different abnormal operation data to fault judgment, and multiple weight coefficients are associated with the abnormal operation data set, which means that the degree of influence of an operation data set on the material conveyor is matched with multiple weight coefficients. The greater the degree of influence, the higher the weight coefficient. Then, according to the association result and the weight coefficient, a fault judgment result is generated, and the fault judgment result may include fault type, fault level, recommended maintenance measures, etc., to ensure the safe, stable and efficient operation of the material conveyor.

[0037] Execute step A520, perform critical analysis according to the fault judgment result, determine the fault risk data matrix, modify the fault risk data matrix based on the multiple operating characteristics, and generate a fault risk level; execute step A530, match the fault risk level to the operating health score, and generate a status signal set, wherein the status signal set includes a normal status signal, a warning status signal, and a fault status signal; execute step A540, perform signal screening on the normal status signal, the warning status signal, and the fault status signal according to the fault risk level, determine the warning signal, and send it to the remote terminal for warning feedback.

[0038] First, a critical analysis is performed on the fault judgment results. The possibility and urgency of the fault occurrence are determined based on factors such as the fault type, fault probability, and fault impact, and a critical analysis result is generated. According to the critical analysis result, the risk levels of different fault types under different conditions are analyzed, and a fault risk data matrix is ​​constructed. The fault risk data matrix is ​​corrected based on multiple operating characteristics, which means that an operating trend analysis can be performed based on multiple operating characteristics, and the fault risk data matrix is ​​synchronously updated according to trend changes to generate a more accurate fault risk level. The fault risk level is further matched with the operating health score, and it is ensured that the status signal set can accurately reflect the current state of the material conveyor. The corresponding status signal is constructed based on the status information, and the status signal set is determined. The status signal set includes a normal status signal, a warning status signal, Fault status signal: the normal status signal can be the fault risk level corresponding to the operation health score greater than or equal to 80 points, the warning status signal can be the fault risk level corresponding to the operation health score less than 80 points and greater than or equal to 50 points, and the fault status signal can be the fault risk level corresponding to less than 50 points. Each signal should have a clear definition and trigger condition. Finally, the status signal set is screened according to the fault risk level, which means that the warning status signal and the fault status signal are processed first, and the screened warning signal is sent to the remote terminal to ensure that the signal can be conveyed to the relevant personnel in a timely and accurate manner. After receiving the warning signal, the remote terminal performs warning feedback processing, including notifying relevant personnel, launching emergency plans, adjusting equipment operating parameters and other measures, which provides strong support for preventive maintenance, timely response and reduction of failure losses of material conveyors.

[0039] The embodiments of the present application solve the technical problem that material conveyor fault monitoring is not timely and the early warning accuracy is low, making it difficult to achieve intelligent maintenance. Through data-driven fault prediction and real-time early warning, real-time monitoring and intelligent feedback of equipment status are achieved, achieving the technical effect of improving operation and maintenance efficiency.

[0040] In the above, refer to Figure 1 The data-driven material conveyor fault warning method according to the embodiment of the present application is described in detail. Figure 2 A data-driven material conveyor fault warning system according to an embodiment of the present application is described.

[0041] The data-driven material conveyor fault warning system according to the embodiment of the present application is used to solve the technical problem that the material conveyor fault monitoring is not timely and the warning accuracy is low, which makes it difficult to achieve intelligent maintenance. Through data-driven fault prediction and real-time warning, real-time monitoring and intelligent feedback of equipment status are achieved, achieving the technical effect of improving operation and maintenance efficiency. The data-driven material conveyor fault warning system includes: a data acquisition module 10, a feature extraction module 20, a learning and training module 30, a prediction and calculation module 40, and a fault determination module 50.

[0042] The data acquisition module 10 is used to traverse the material conveyor and deploy multiple sensors to perform real-time data acquisition to obtain a real-time operation data set of the material conveyor; A feature extraction module 20, configured to perform multidimensional analysis based on the real-time operation data set, determine multidimensional operation data, perform time series decomposition based on the multidimensional operation data, and extract multiple operation features; A learning and training module 30 is used to retrieve the historical operation record log of the material conveyor, perform learning and training according to the historical operation record log, and obtain a fault prediction model; A prediction calculation module 40, used to synchronize the real-time operation data set to the fault prediction model for prediction calculation to determine the operation health score of the material conveyor; The fault determination module 50 is used to determine the fault of the material conveyor according to the operation health score, determine the fault risk level, generate an early warning signal according to the fault risk level, and send the early warning signal to the remote terminal for intelligent early warning feedback of the fault.

[0043] The specific configuration of the feature extraction module 20 will be described in detail below. As described above, based on the real-time operation data set, multi-dimensional analysis is performed to determine the multi-dimensional operation data, and the feature extraction module 20 may further include: performing data cleaning based on the real-time operation data set to obtain a real-time operation standard data set; traversing the real-time operation standard data set to perform similarity calculations to determine multiple similarity coefficients; determining multiple cluster centers according to the multiple similarity coefficients, and dividing the real-time operation standard data set into multiple data clusters according to the multiple cluster centers; and adding the multiple data clusters to the multi-dimensional operation data.

[0044] The specific configuration of the feature extraction module 20 will be described in detail below. As described above, the time series decomposition is performed according to the multidimensional operation data to extract multiple operation features. The feature extraction module 20 may further include: extracting multiple data timestamps of the multiple cluster centers, arranging the multiple data timestamps in sequence, constructing multiple time series, and the multiple time series have a corresponding relationship with the multidimensional operation data; sorting the multidimensional operation data group according to the multiple time series by time dimension to construct a time series model; decomposing using the time series model to obtain multiple decomposition components, integrating the multiple decomposition components, and generating the multiple operation features.

[0045] The specific configuration of the learning and training module 30 will be described in detail below. As described above, learning and training are performed according to the historical operation record log to obtain a fault prediction model. The learning and training module 30 may further include: traversing the historical operation record log to perform operation fault analysis and determine multiple fault operation features; feature labeling the historical operation record log based on the multiple fault operation features to determine multiple fault labeling information; data segmentation of the historical operation record log according to the multiple fault labeling information to determine a training data set and a verification data set; supervised training iterations are performed based on the training data set, and when the training data set tends to converge, cross-validation is performed using the verification data set, and the fault prediction model is obtained according to the verification result.

[0046] The specific configuration of the prediction and calculation module 40 will be described in detail below. As described above, the real-time operation data set is synchronized to the fault prediction model for prediction and calculation to determine the operation health score of the material conveyor. The prediction and calculation module 40 may further include: loading the fault prediction model, synchronizing the real-time operation data set to the fault prediction model to perform operation prediction on the material conveyor, and generating an operation fault prediction result; mapping the historical operation record log to the operation fault prediction result for association, and determining an operation fault association list; performing fault calculation according to the operation fault association list, generating multiple fault probabilities, performing health assessment based on the multiple fault probabilities, and generating the operation health score.

[0047] The specific configuration of the fault determination module 50 will be described in detail below. As described above, the material conveyor is fault-determined according to the operation health score, the fault risk level is determined, an early warning signal is generated according to the fault risk level, and the early warning signal is sent to the remote terminal for intelligent early warning feedback of the fault. The fault determination module 50 may further include: setting an expected operation threshold based on the historical operation record log, performing fault determination on the material conveyor according to the expected operation threshold and the operation health score, and generating a fault determination result; performing critical analysis according to the fault determination result, determining a fault risk data matrix, and modifying the fault risk data matrix based on the multiple operation characteristics to generate a fault risk level; matching the fault risk level to the operation health score, generating a state signal set, and the state signal set includes a normal state signal, an early warning state signal, and a fault state signal; performing signal screening on the normal state signal, the early warning state signal, and the fault state signal according to the fault risk level, determining the early warning signal and sending it to the remote terminal for early warning feedback.

[0048] The specific configuration of the fault determination module 50 will be described in detail below. As described above, the material conveyor is subjected to fault determination according to the expected operation threshold and the operation health score to generate a fault determination result. The fault determination module 50 may further include: comparing the operation health score with the expected operation threshold to determine whether the operation health score is greater than or equal to the expected operation threshold; if the operation health score is greater than or equal to the expected operation threshold, normal operation data is obtained, continuous operation monitoring is performed according to the normal operation data, operation monitoring instructions are generated, and the operation monitoring instructions are added to the fault determination result; if the operation health score is less than the expected operation threshold, an abnormal operation data set is generated, and fault impact analysis is performed according to the abnormal operation data set to generate an operation impact factor; weights are assigned to the abnormal operation data set according to the operation impact factor to determine a plurality of weight coefficients; the plurality of weight coefficients are associated with the abnormal operation data set to generate the fault determination result.

[0049] The data-driven material conveyor fault warning system provided in the embodiments of the present application can execute the data-driven material conveyor fault warning method provided in any embodiment of the present application, and has the corresponding functional modules and beneficial effects of the execution method.

[0050] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present application.

[0051] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A data-driven material conveyor fault early warning method, characterized in that: The method comprises: Multiple sensors are deployed along the material conveyor to collect data in real time, and obtain the real-time operation data set of the material conveyor; Perform multidimensional analysis based on the real-time operation data set to determine multidimensional operation data, perform time series decomposition based on the multidimensional operation data, and extract multiple operation features; Retrieving the historical operation record log of the material conveyor, performing learning and training according to the historical operation record log, and obtaining a fault prediction model; Synchronizing the real-time operation data set to the fault prediction model for prediction calculation to determine the operation health score of the material conveyor; According to the operation health score, the material conveyor is fault-judged to determine the fault risk level, an early warning signal is generated according to the fault risk level, and the early warning signal is sent to a remote terminal for intelligent early warning feedback of the fault.

2. The data-driven material conveyor fault early warning method according to claim 1, characterized in that: Performing multidimensional analysis based on the real-time operation data set to determine multidimensional operation data, the method comprising: Performing data cleaning based on the real-time operation data set to obtain a real-time operation standard data set; Traversing the real-time running standard data set to perform similarity calculation and determine multiple similarity coefficients; Determine a plurality of cluster centers according to the plurality of similarity coefficients, and divide the real-time operation standard data set into a plurality of data clusters according to the plurality of cluster centers; The plurality of data clusters are added to the multi-dimensional operational data.

3. The data-driven material conveyor fault early warning method according to claim 2, characterized in that: Performing time series decomposition according to the multi-dimensional operation data to extract multiple operation features, the method includes: Extracting multiple data timestamps of the multiple cluster centers, arranging the multiple data timestamps in sequence, and constructing multiple time series, wherein the multiple time series have a corresponding relationship with the multidimensional operation data; Sorting the multidimensional operation data group by time dimension according to the multiple time series to construct a time series model; The time series model is used for decomposition to obtain a plurality of decomposition components, and the plurality of decomposition components are integrated to generate the plurality of operation characteristics.

4. The data-driven material conveyor fault early warning method according to claim 1, characterized in that: Performing learning and training according to the historical operation record log to obtain a fault prediction model, the method includes: Traversing the historical operation record log to perform operation fault analysis and determine multiple fault operation characteristics; Performing feature annotation on the historical operation record log based on the multiple fault operation features to determine multiple fault annotation information; Segment the historical operation record log according to the multiple fault annotation information to determine a training data set and a verification data set; Supervised training iterations are performed based on the training data set. When the training data set tends to converge, cross-validation is performed using the validation data set, and the fault prediction model is obtained according to the validation result.

5. The data-driven material conveyor fault early warning method according to claim 1, characterized in that: The real-time operation data set is synchronized to the fault prediction model for prediction calculation to determine the operation health score of the material conveyor, the method comprising: Loading the fault prediction model, synchronizing the real-time operation data set to the fault prediction model to perform operation prediction on the material conveyor, and generating an operation fault prediction result; Mapping the historical operation record log to the operation fault prediction result for association, and determining an operation fault association list; Fault calculation is performed according to the operation fault association list to generate multiple fault probabilities, and health assessment is performed based on the multiple fault probabilities to generate the operation health score.

6. The data-driven material conveyor fault early warning method according to claim 1, characterized in that: According to the operation health score, the material conveyor is fault-determined, the fault risk level is determined, an early warning signal is generated according to the fault risk level, and the early warning signal is sent to a remote terminal for intelligent early warning feedback of the fault, the method comprising: Setting an expected operation threshold based on the historical operation record log, performing a fault determination on the material conveyor according to the expected operation threshold and the operation health score, and generating a fault determination result; Performing critical analysis according to the fault determination result to determine a fault risk data matrix, and modifying the fault risk data matrix based on the multiple operating characteristics to generate a fault risk level; Matching the fault risk level to the operation health score to generate a state signal set, wherein the state signal set includes a normal state signal, a warning state signal, and a fault state signal; The normal state signal, the warning state signal, and the fault state signal are screened according to the fault risk level, and the warning signal is determined and sent to a remote terminal for warning feedback.

7. The data-driven material conveyor fault early warning method according to claim 6, characterized in that: Performing a fault determination on the material conveyor according to the expected operation threshold and the operation health score, and generating a fault determination result, the method comprising: Comparing the operation health score with the expected operation threshold to determine whether the operation health score is greater than or equal to the expected operation threshold; If the operation health score is greater than or equal to the expected operation threshold, normal operation data is obtained, continuous operation monitoring is performed according to the normal operation data, an operation monitoring instruction is generated, and the operation monitoring instruction is added to the fault determination result; If the operation health score is less than the expected operation threshold, an abnormal operation data set is generated, and a fault impact analysis is performed based on the abnormal operation data set to generate an operation impact factor; According to the operation influencing factors, weights are assigned to the abnormal operation data set to determine a plurality of weight coefficients; The plurality of weight coefficients are associated with the abnormal operation data set to generate the fault determination result.

8. Data-driven material conveyor fault warning system, characterized in that, The system is used to implement the data-driven material conveyor fault early warning method according to any one of claims 1 to 7, comprising: A data acquisition module is used to traverse the material conveyor and deploy multiple sensors to perform real-time data collection to obtain a real-time operating data set of the material conveyor; A feature extraction module, configured to perform multidimensional analysis based on the real-time operation data set, determine multidimensional operation data, perform time series decomposition based on the multidimensional operation data, and extract multiple operation features; A learning and training module, used to retrieve the historical operation record log of the material conveyor, perform learning and training according to the historical operation record log, and obtain a fault prediction model; A prediction calculation module, used for synchronizing the real-time operation data set to the fault prediction model for prediction calculation to determine the operation health score of the material conveyor; The fault determination module is used to determine the fault of the material conveyor according to the operation health score, determine the fault risk level, generate an early warning signal according to the fault risk level, and send the early warning signal to the remote terminal for intelligent early warning feedback of the fault.

Images