Multi-device collaborative analysis system and method for coal quantity and quality fusion and anomaly recognition

Abstract

The invention relates to a multi-device collaborative analysis system for coal quantity and quality fusion and anomaly recognition, and the system comprises a multi-source data collection module; the edge computing node module is used for acquiring edge node data acquired by various metering and detecting devices; the central data center is used for processing the complete coal quantity quality data by using a preset error prediction model and a preset regression prediction model; the centralized monitoring and alarm module is used for displaying the processing data on a visual interface and checking whole-network real-time data quality data, historical trends, abnormal events and alarm information of the coal; and the collaborative analysis module is used for carrying out multi-device collaboration oriented to coal quantity and quality fusion and anomaly recognition according to the model and the control strategy. Therefore, the problems that the requirements of real-time fusion monitoring and intelligent anomaly identification of coal full-flow-number quality data cannot be met, and digital management and efficient decision making of coal production and supply chains are restricted in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-device collaborative analysis, and particularly relates to a multi-device collaborative analysis system and method for coal quantity and quality fusion and abnormality identification. BACKGROUND

[0002] In the links of coal mine production, coal washing, transportation and utilization, various types of metering devices (such as belt scales, rail scales, truck scales, etc.) and coal quality detection devices (such as manual sampling and testing) have traditionally operated independently, lacking a unified data fusion and monitoring platform. Existing coal quantity and quality management usually records data by each device separately, and accounts manually or after the fact, which is not only inefficient but also lacks real-time performance. For example, coal needs to be weighed and sampled separately at the loading station and unloading acceptance, and the traditional testing cycle is long, with data lag, leading to possible inconsistencies in upstream and downstream test results. The measurement data of each device often has deviations due to environmental factors and calibration differences (such as belt running speed, sensor drift, human operation error, etc.), and even different devices give contradictory results for the quantity and quality of the same batch of coal, making it difficult to discover and correct in a timely manner. In addition, the existing system uses manual experience or fixed thresholds for data verification and alarm, lacks intelligent analysis means, and the abnormality identification is not sensitive enough. The traditional coal quality detection process is complex and time-consuming (at least several hours), the settlement cycle is long, the data is lagging and there are human operation differences, making it difficult to support production and trade decisions in a timely manner.

[0003] Overall, the prior art has the following prominent defects: (1) Data islands, lack of centralized monitoring: The use of measurement and coal quality equipment in coal mines, washing plants, power plants, ports and other links is scattered, and there is a lack of a centralized monitoring platform in the total dispatching room, which cannot grasp the real-time data of each place globally. (2) Lack of fusion and verification of multi-source data: Each device (belt scale, track scale, truck scale, etc.) is independent, and the quantity data measured is not cross-verified. A batch of coal often has only single-point data support, and lacks consistency verification between multiple devices, resulting in measurement deviation not being discovered in time. In terms of coal quality, the results of on-site rapid detection and laboratory testing are often inconsistent due to sampling errors, and there is a lack of effective fusion and comparison means. (3) Lack of intelligent correction and abnormality detection: Traditional systems mainly rely on manual experience to handle measurement deviation, and lack prediction and automatic correction mechanisms for device drift or abnormalities. At the same time, abnormality monitoring is mostly limited to simple threshold alarms, and cannot identify complex pattern abnormalities (such as systematic deviation, sensor failure, etc.), which can easily miss or misreport. (4) Lack of hierarchical optimization in architecture: Most existing solutions directly upload data to the background or cloud for processing, and lack edge computing support on site, resulting in poor real-time response and bandwidth utilization. While the pure centralized processing model has a long training period and cannot fully utilize on-site computing resources, the overall architecture is difficult to balance real-time performance and deep analysis capabilities. (5) Insufficient data traceability: At present, there is a lack of unified data tags and complete archives for each batch of coal, and the data at each link cannot be effectively linked. In the event of quality disputes or measurement abnormalities, it is difficult to trace the source and reconstruct the whole process data in a timely manner, which poses risks to production management and trade settlement. SUMMARY

[0004] The present application provides a multi-device collaborative analysis system for coal number and quality fusion and abnormality identification, to solve the problem that related technologies cannot meet the real-time fusion monitoring and intelligent abnormality identification needs of coal whole-process number and quality data, which restricts digital management and efficient decision-making of coal production and supply chain, etc.

[0005] The first aspect embodiment of the present application provides a multi-device collaborative analysis system for coal quantity and quality fusion and abnormality identification, comprising: a multi-source data acquisition module, configured to acquire quantity data and quality data of coal; an edge computing node module, connected with a plurality of metering and detection devices, configured to acquire edge node data collected by the plurality of metering and detection devices, and preprocess the edge node to obtain target data; a central data platform, configured to analyze the quantity data, the quality data and the target data to generate complete coal quantity and quality data, and use a preset error prediction model and a preset regression prediction model to process the complete coal quantity and quality data respectively to generate processing data of the coal; a centralized monitoring and alarming module, configured to display the processing data on a visual interface and view real-time quantity and quality data, historical trends, abnormal events and alarm information of the coal; and a collaborative analysis module, configured to record metering values, quality indexes, model correction values, abnormal markers and flow time and location of the coal to a full-process data archive of the coal based on the real-time quantity and quality data, the historical trends, the abnormal events and the alarm information, and to issue a model and a control strategy to the edge computing node module based on the full-process data archive, and to perform multi-device collaboration for coal quantity and quality fusion and abnormality identification according to the model and the control strategy.

[0006] Optionally, in an embodiment of the present application, the multi-source data acquisition module is further configured to access at least one heterogeneous device of a belt scale, a quantitative loading system, a rail scale, a truck scale, a water gauge, a sampling machine, a coal quality rapid detection device and a laboratory manual testing device to an edge computing node to acquire quantity and quality data of the coal.

[0007] Optionally, in an embodiment of the present application, the edge computing node module comprises: a filtering unit, configured to convert data of different communication protocols into unified format data, and filter data satisfying a preset abnormal condition in the unified format data to generate data satisfying a preset standard condition; and a fusion unit, configured to fuse contemporaneous data of a plurality of devices according to the data satisfying the preset standard condition to generate the target data.

[0008] Optionally, in an embodiment of the present application, the central data platform comprises: a data storage unit configured to store the quantity data and quality data of the coal and the target data into a batch archive, to generate storage data; a data processing unit configured to perform large-scale data cleaning on the storage data, to generate cleaning data; a model training unit configured to establish the preset error prediction model according to the environmental parameters, the equipment state parameters and the operation conditions, to train the preset error prediction model to correct the measurement deviation, to generate a correction result, and to establish the preset regression prediction model according to the coal quality index, to correct the coal quality data by using the preset regression prediction model, and to generate the processing data.

[0009] Optionally, in an embodiment of the present application, the centralized monitoring and alarming module comprises: an anomaly identification unit configured to train a clustering model by using historical data, to identify data distribution characteristics, to mark batches deviating from a main clustering center according to the data distribution characteristics, to generate marked data, and to determine samples satisfying a preset high isolation degree condition in the marked data as anomaly points; an anomaly calculation unit configured to fuse a plurality of detection results by using a Bayes probability model, to calculate a comprehensive anomaly probability, and to generate the anomaly event according to the anomaly points and the anomaly probability; and a display unit configured to display the real-time quantity data of the whole network, the historical trend, the anomaly event and the alarm information by using the visual interface.

[0010] Optionally, in an embodiment of the present application, the collaborative analysis module comprises: an establishing unit configured to assign a unique tag to each batch of coal, and to establish a whole-process data archive corresponding to the unique tag, wherein the whole-process data archive comprises at least one of a measurement value, a quality index, a model correction value, an anomaly mark and a flow time and location; and a collaborative analysis unit configured to collaboratively analyze, by using the whole-process data archive, a starting location and time, a final receiving location and time, measurement data of each link, coal quality data of each link, a model correction value, an anomaly alarm log, a person handling and a time stamp of a plurality of devices in a process from production to utilization of the each batch of coal, to generate a collaborative analysis result.

[0011] The second aspect embodiment of the application provides a multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification, comprising the following steps: collecting quantity data and quality data of coal; obtaining edge node data collected by the plurality of metering and detection devices, and preprocessing the edge nodes to obtain target data; analyzing the quantity data, the quality data and the target data to generate complete coal quantity and quality data, and processing the complete coal quantity and quality data using a preset error prediction model and a preset regression prediction model to generate processing data of the coal; displaying the processing data on a visual interface and viewing the full-network real-time quantity and quality data, historical trends, abnormal events and alarm information of the coal; based on the full-network real-time quantity and quality data, the historical trends, the abnormal events and the alarm information, recording the metering value, the quality index, the model correction value, the abnormal marker and the flow time and location of the coal to a full-process data archive of the coal, so as to issue a model and a control strategy to the edge computing node module based on the full-process data archive, and perform multi-device collaboration for coal quantity and quality fusion and abnormality identification according to the model and the control strategy.

[0012] Optionally, in an embodiment of the application, the obtaining edge node data collected by the plurality of metering and detection devices and preprocessing the edge nodes to obtain target data comprises: converting data of different communication protocols into unified format data, and filtering data in the unified format data that meets a preset abnormal condition to generate data that meets a preset standard condition; fusing contemporaneous data of multiple devices according to the data that meets the preset standard condition to generate the target data.

[0013] Optionally, in an embodiment of the application, the processing the complete coal quantity and quality data using a preset error prediction model and a preset regression prediction model to generate processing data of the coal comprises: a data storage unit configured to store the quantity data and the quality data of the coal and the target data to a batch archive to generate storage data; a data processing unit configured to perform large-scale data cleaning on the storage data to generate cleaned data; a model training unit configured to establish the preset error prediction model according to environmental parameters, device state parameters and operation conditions, train the preset error prediction model to correct measurement deviation to generate a correction result, and establish the preset regression prediction model according to coal quality indexes to correct coal quality data using the preset regression prediction model to generate the processing data.

[0014] Optionally, in an embodiment of the present application, the displaying the processing data on the visualization interface and viewing the full-network real-time quantity and quality data, historical trend, abnormal event and alarm information of the coal includes: training a clustering model using historical data to identify data distribution characteristics, and marking batches deviating from the main clustering center according to the data distribution characteristics to generate marked data, and determining samples in the marked data satisfying a preset high-isolation condition as abnormal points; fusing multiple detection results using a Bayesian probability model to calculate a comprehensive abnormal probability, and generating the abnormal event according to the abnormal points and the abnormal probability; and displaying the full-network real-time quantity and quality data, the historical trend, the abnormal event and the alarm information on the visualization interface.

[0015] Optionally, in an embodiment of the present application, the model and control strategy are issued to the edge computing node module based on the full-process data archive to perform multi-device collaboration for coal quantity and quality fusion and abnormality identification according to the model and the control strategy, including: giving each batch of coal a unique label, and establishing a full-process data archive corresponding to the unique label, wherein the full-process data archive includes at least one of a measurement value, a quality index, a model correction value, an abnormality label and a flow time and location; and using the full-process data archive to collaboratively analyze the starting location and time, the final receiving location and time, the measurement data of each link, the coal quality data of each link, the model correction value, the abnormality alarm log, the person handling and the time stamp of the multi-device of each batch of coal from production to utilization to generate a collaborative analysis result.

[0016] The third aspect embodiment of the present application provides an electronic device, including: a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the program to implement the multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification as described in the above embodiments.

[0017] The fourth aspect embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification as described above.

[0018] The embodiments of the present application can perform centralized monitoring of quantity and quality data in multiple scenarios, so that the total dispatching room can master the coal production, transportation and sales situation of each site in real time; the quantity and quality data of coal can be collected by multiple devices and managed uniformly; the quantity and quality data generated by different devices are aligned, compared, fused and calculated, and the consistency of multi-source data is automatically verified to ensure that the measurement results and quality indexes of the same batch of coal are consistent in each link; a machine learning model is introduced to predict the systematic error and random error in the coal measurement process, and a regression model is established for key indicators of coal quality to realize the correlation prediction and correction of online rapid detection results and laboratory test results, thereby improving the accuracy of quantity and quality data; measurement abnormalities or coal quality abnormalities can be automatically found and real-time alarms can be given; the combination of rapid response on the edge side and global optimization on the center side is realized; a unique data tag is generated for each batch of coal, a holographic circulation file is established, and the quantity, quality and related processing information of the batch of coal from the production site to the user are recorded to enhance the transparency and traceability of data. Thus, the problems that the related art cannot meet the real-time fusion monitoring and intelligent abnormality identification of quantity and quality data in the whole process of coal, and restricts the digital management and efficient decision-making of coal production and supply chain are solved.

[0019] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0020] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the following description, taken in conjunction with the accompanying drawings, in which: Figure 1 A structural schematic diagram of a multi-device collaborative analysis system for coal quantity and quality fusion and abnormality identification according to an embodiment of the present application is provided. Figure 2 A schematic diagram of a coal quantity and quality multi-device collaborative analysis system architecture according to an embodiment of the present application is provided. Figure 3 A flowchart of a multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification according to an embodiment of the present application is provided. Figure 4 A structural schematic diagram of an electronic device according to an embodiment of the present application is provided. DETAILED DESCRIPTION

[0021] The embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings, in which the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by reference to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as limiting the present application.

[0022] A multi-device collaborative analysis system for coal number-quality fusion and abnormality identification is described below with reference to the accompanying drawings. In view of the fact that the related technologies mentioned in the background art cannot meet the real-time fusion monitoring and intelligent abnormality identification requirements of the number-quality data of the whole process of coal, which restricts the digital management and efficient decision-making of the coal production and supply chain, the present application provides a multi-device collaborative analysis system for coal number-quality fusion and abnormality identification. In this system, the number-quality data of multiple scenarios can be centrally monitored, enabling the total dispatching room to keep abreast of the coal production, transportation and marketing at each location; the multi-device collaborative collection and unified management of the number and quality data of coal can be realized; the number-quality data generated by different devices can be aligned, compared, fused and calculated, and the consistency of multi-source data can be automatically verified to ensure that the measurement results and quality indicators of the same batch of coal remain consistent at each link; machine learning models are introduced to predict systematic errors and random errors in the coal measurement process, and regression models are established for key indicators of coal quality to realize the correlation prediction and correction of online rapid detection results and laboratory test results, thereby improving the accuracy of number-quality data; measurement abnormalities or coal quality abnormalities can be automatically discovered and real-time alarms can be given; the combination of rapid response on the edge side and global optimization on the center side can be realized; a unique data tag is generated for each batch of coal to establish a holographic transfer file and record the quantity, quality and related processing information of the batch of coal from the production site to the user, thereby enhancing the transparency and traceability of data. Thus, the problems of related technologies that cannot meet the real-time fusion monitoring and intelligent abnormality identification requirements of the number-quality data of the whole process of coal, which restrict the digital management and efficient decision-making of the coal production and supply chain, are solved.

[0023] Specifically, Figure 1 A structure diagram of a multi-device collaborative analysis system for coal number-quality fusion and abnormality identification provided by the embodiments of the present application is shown.

[0024] As Figure 1 shown, the multi-device collaborative analysis system for coal number-quality fusion and abnormality identification 10 includes a multi-source data acquisition module 100, an edge computing node module 200, a central data platform 300, a centralized monitoring and alarm module 400, and a collaborative analysis module 500.

[0025] Specifically, the multi-source data acquisition module 100 is configured to acquire the number data and quality data of coal.

[0026] It can be understood that, as Figure 2 shown, the multi-source data acquisition module 100 in the embodiments of the present application can be a multi-source data acquisition device layer.

[0027] In actual implementation, the quantity data and quality data of the coal can be collected by the multi-source data collection module 100, the multi-scene centralized monitoring problem is solved, the quantity and quality data of various scenes such as coal mines, coal preparation plants, power plants, ports, stations and chemical plants are centrally monitored, and the total dispatching room can master the coal production, transportation and sales situation of each place in real time.

[0028] Optionally, in an embodiment of the present application, the multi-source data collection module 100 is further configured to connect at least one of a belt scale, a quantitative loading system, a track scale, a truck scale, a water gauge, a sampling machine, a coal quality rapid detection device and a laboratory manual testing device to the edge computing node to obtain the quantity and quality data of the coal.

[0029] It can be understood that the embodiment of the present application supports the access and integration of various data sources such as belt scales, quantitative loading systems, track scales, truck scales, water gauges, sampling machines, coal quality rapid detection devices and manual testing devices, realizes the multi-device collaborative collection and unified management of the quantity and quality data of the coal, and solves the multi-source heterogeneous data integration problem.

[0030] In the embodiment of the present application, the multi-source data collection module 100 is located at the leftmost side of the edge computing node, and includes a belt scale, a quantitative loading system, a track scale, a truck scale, a water gauge (ship draft measurement), a sampling machine, a coal quality rapid detection device and a laboratory manual testing device, and the like. These heterogeneous devices access the edge computing node through wired or wireless networks, and continuously generate the quantity and quality raw data of the coal. Figure 2

[0031] ​Further, the application can perform multi-source data alignment and fusion: for heterogeneous data output by different devices, the application designs a unified data representation and alignment mechanism. First, a unique batch ID and timestamp label is generated for each batch of coal, and the quantity and quality data generated by each device for the same batch are associated through the label (for example, the track scale measurement record of a train and the corresponding quantitative loading system record are matched, and the water gauge measurement of a ship berth and the corresponding track scale record are matched). Based on data alignment, a self-developed data fusion algorithm is used to comprehensively calculate the multi-source measurement results: for quantity (weight) data, weighted average method or least squares calibration method can be used to fuse the readings of belt scale, track scale, truck scale, etc. to reduce the error of a single device; for quality index data, deviation compensation and consistency verification are performed based on on-site rapid detection and laboratory test results. For example, the rapid ash content detection value can be calibrated to the standard laboratory benchmark through linear regression. In the data fusion process, a cross-validation mechanism is introduced: the data of each device is compared with each other, and if it is found that the reading of a certain device deviates from other sources by more than a set threshold, it is marked as abnormal and triggers the calibration or maintenance process. This multi-device data fusion and consistency verification method can effectively eliminate the influence of single-point failure or reading error and ensure the authenticity and reliability of the data.

[0032] The embodiment of the application has comprehensive data fusion: it breaks the pattern of scattered and isolated quantity data and quality data in traditional systems and builds a complete data monitoring chain covering all links and processes. It integrates the data generated by various devices such as belt scales, track scales, rapid detection equipment, etc. to achieve panoramic control of coal quantity and quality. This not only improves data utilization, but also provides rich and reliable first-hand information for subsequent analysis.

[0033] The edge computing node module 200 is connected with a plurality of measurement and detection devices, and is used to obtain edge node data collected by the plurality of measurement and detection devices, and to preprocess the edge node to obtain target data.

[0034] It can be understood that the edge computing node module 200 in the embodiment of the application can be an edge computing node (field end).

[0035] In actual execution process, the edge computing node module 200 in the embodiment of the application is the node in the middle left of the figure, and the edge computing node module is deployed in each production or transportation site, and is connected with a plurality of measurement and detection devices, to realize real-time data collection, preprocessing and local fusion, to obtain target data.

[0036] The embodiments of the present application can realize the improvement of measurement accuracy and consistency: through cross-device data alignment and cross-validation, the present application can automatically find and correct the deviation between different measurement devices. For example, the edge node can correct the reading difference of the belt scale and the track scale in time, and the center model can predict the influence of environmental factors on measurement to compensate, so that the finally aggregated weight data is closer to the true value. In practice, the fusion algorithm of the present application significantly reduces the difference between the measurement results of multiple devices, greatly improves the measurement consistency, and reduces disputes and losses caused by inconsistent data.

[0037] Optionally, in an embodiment of the present application, the edge computing node module 200 comprises: a filtering unit, configured to convert data of different communication protocols into unified format data, and filter data in the unified format data that meets a preset abnormal condition, to generate data that meets a preset standard condition; and a fusion unit, configured to fuse contemporaneous data of multiple devices according to the data that meets the preset standard condition, to generate target data.

[0038] It can be understood that the data that meets the preset standard condition in the embodiments of the present application can be standard data.

[0039] Specifically, the edge computing node 200 in the embodiments of the present application is deployed at each data acquisition site (such as a mine dispatching room, a port site machine room, etc.). The edge node is directly connected with the on-site devices, and performs real-time data acquisition, caching and preliminary processing functions, including: converting data of different protocols into a unified format, filtering missing or abnormal data, generating data that meets a preset standard condition, and simply fusing and comparing contemporaneous data from multiple devices to generate target data. The edge node also undertakes local alarm (such as on-site device failure reminder) and edge reasoning (deploying the model trained by the center for real-time error correction), etc.

[0040] The embodiments of the present application can reduce the load of the center through edge computing, and guarantee the basic availability of on-site data in the case of network interruption. The present application proposes a self-developed data fusion and cross-validation algorithm, which aligns, compares, fuses and calculates the quantity data generated by different devices, and automatically verifies the consistency of multi-source data, ensures that the measurement results and quality indicators of the same batch of coal are consistent at each link, and solves the problem of cross-device data fusion and consistency verification.

[0041] The center data platform 300 is configured to analyze the quantity data, the quality data and the target data to generate complete coal quantity and quality data, and process the complete coal quantity and quality data by using a preset error prediction model and a preset regression prediction model, to generate processing data of the coal.

[0042] It can be understood that the center data platform 300 in the embodiments of the present application can be Figure 2Large nodes in the middle right, usually deployed in the data center of the enterprise headquarters or cloud platform.

[0043] In actual execution process, the center data platform 300 in the embodiment of the application is connected with each edge node through a secure network, is used for analyzing quantity data, quality data and target data, and receives complete coal quantity and quality data of the whole company. The center data platform 300 processes the complete coal quantity and quality data by using a preset error prediction model and a preset regression prediction model respectively, to generate processing data of coal.

[0044] The embodiment of the application can improve the efficiency and accuracy of coal quality detection: by means of machine learning model, real-time regression prediction is performed on coal quality indicators, and laboratory test results that originally take several hours to obtain can be estimated within a few minutes, realizing rapid quality evaluation. At the same time, the correction of the model to the rapid detection data makes the online detection accuracy close to the traditional test level. This means that users can obtain more accurate coal quality information in real time, speed up trade settlement and production adjustment, and avoid waiting and uncertainty caused by unknown quality. The application constructs a hierarchical architecture of “edge computing node + center data platform”, deploys edge computing nodes near the equipment end to perform real-time data acquisition and preliminary fusion processing, and constructs a data platform at the center end to be responsible for large-scale data storage, model training and deep analysis, realizes the combination of rapid response on the edge side and global optimization on the center side, and solves the cloud-edge collaborative architecture optimization problem.

[0045] Optionally, in an embodiment of the application, the center data platform 300 comprises: a data storage unit, configured to store the quantity data and quality data and target data of coal into a batch archive, to generate storage data; a data processing unit, configured to perform large-scale data cleaning on the storage data, to generate cleaning data; a model training unit, configured to establish a preset error prediction model according to environmental parameters, equipment state parameters and operation conditions, and train the preset error prediction model to correct measurement deviation, to generate correction results, and establish a preset regression prediction model according to coal quality indicators, to correct coal quality data by using the preset regression prediction model, to generate processing data.

[0046] It can be understood that the embodiment of the application introduces a machine learning model to predict system error and random error in the coal metering process, and models and compensates for the deviation that may be generated by belt scales, track scales and other equipment; at the same time, a regression model is established for key indicators of coal quality, to realize the associated prediction and correction of online rapid detection results and laboratory test results, improve the accuracy of quantity and quality data, and solve the intelligent correction problem of metering error.

[0047] In actual execution process, the center data platform 300 in the embodiment of the application can utilize the data storage unit for big data storage, batch archive library, etc., store the quantity data and quality data and target data of coal into the batch archive library, generate storage data; the center data platform 300 can utilize the data processing unit for large-scale data cleaning, complex fusion calculation, etc., clean the storage data in large scale to generate cleaning data; the model training unit in the center data platform 300 includes global historical data training / updates measurement error prediction model, coal quality regression model, anomaly detection model, etc., establishes a preset error prediction model according to environmental parameters, equipment state parameters and operation conditions, and trains the preset error prediction model to correct measurement deviation to generate correction results, and establishes a preset regression prediction model according to coal quality indexes to correct coal quality data by using the preset regression prediction model to generate processing data, and the center data platform 300 further includes a business application unit. The center data platform 300 performs deep fusion analysis on the data uploaded from each edge, and stores and archives the results. At the same time, through a model issuing mechanism, new correction parameters or algorithms are issued to the edge nodes for execution, so that the system is continuously self-optimized.

[0048] In the system, the overall architecture adopts a collaborative architecture of “edge computing node + center data platform”. The center data platform 300 is deployed at the enterprise headquarters or cloud end, gathers the data uploaded by each edge node, undertakes machine learning model training, big data storage and deep analysis functions, and provides a unified monitoring and management interface. The overall dispatching room realizes centralized monitoring and dispatching command of global data through the center platform. The architecture deploys data processing in layers according to real-time performance and complexity: the edge side processes locally quickly to reduce bandwidth occupation, the center side performs global optimization analysis and model iteration, and the effective balance between low latency and high computing power demand is realized.

[0049] The application builds a machine learning model library on the central data platform 300. Among them, the prediction model of quantity measurement error is obtained by supervised learning training: taking the difference value of the readings of different measurement devices in the historical batch as the target, selecting environmental parameters (temperature, humidity), device state parameters (sensor zero drift, belt speed), operation conditions, etc. as features, training a random forest regression or neural network model, which can predict the possible measurement deviation of a device when a new batch arrives, and accordingly intelligently correct its real-time readings. For coal quality indicators such as calorific value, ash content, sulfur content, etc., a regression prediction model (such as based on XGBoost or multiple linear regression) is established: input the spectral or composition data provided by the on-site online rapid detection device, and output the predicted laboratory standard analysis value, thereby realizing intelligent conversion and correction of coal quality data. These models are ensured to have generalization performance through cross-validation, and after training on the central platform, they can be distributed to the edge nodes to perform real-time inference. Through the introduction of machine learning, the application can dynamically calibrate the measurement device readings and quickly estimate the coal quality indicators, significantly improving the data accuracy and reducing the frequency of manual testing.

[0050] Further, the data platform and monitoring display of the embodiments of the application can be performed: in the central data platform 300, a unified data storage and service architecture is designed, and a coal quantity and quality database and a model library are constructed. All data from the edge nodes are aggregated into the database through high-speed network, and are organized according to the dimensions of coal batch, device category, time sequence, etc. The central data platform 300 provides a visual monitoring interface for the dispatching management personnel, including key indicator boards, real-time curves, geographic distribution maps, etc., realizing global transparent supervision of coal quantity and quality in the whole company. The total dispatching room can view the real-time data and historical trends of each mine, coal preparation plant and transportation route through the platform, and obvious warnings will be given once there is an abnormality (such as the data of the belt scale and the track scale of a mine do not match, the calorific value of a batch of coal abnormally decreases, etc.). The platform also integrates report and analysis tools, supports daily, monthly statistics of warehouse total quantity, calorific value distribution, device running state, and correction effect of the model on error, etc., providing a basis for management decision. Through edge-center cooperation, the application constructs a data monitoring chain throughout the whole process of production, transportation and sales, realizing "panoramic, real, accurate, fast and transparent" management and control of coal quantity and quality.

[0051] The embodiments of the present application can perform architecture optimization and resource efficient utilization: after adopting the cloud-edge collaborative architecture, the system fully utilizes the edge computing capability while ensuring centralized management, realizing the organic combination of low delay, localized processing and global deep learning. On the one hand, the edge node reduces the pressure of the network and the center, and can still guarantee the operation of the key function in the unstable network scene; on the other hand, the center platform coordinates the training and analysis of the whole network data, the model accuracy is higher, and the system can be updated as a whole to keep the overall intelligence level evolving. This architecture improves the robustness and scalability of the system, and can still run smoothly when the device and data scale further grow. The present application designs an intelligent abnormality identification mechanism combining expert rules and machine learning, and comprehensively uses clustering analysis, isolation forest, Bayesian probability model and other algorithms to detect abnormalities in multi-dimensional data, which can automatically find metering abnormalities (such as sensor misalignment, data forgery) or coal quality abnormalities (such as quality mutation, abnormal mixed coal), and real-time alarm, solving the problem of automatic identification of abnormalities.

[0052] The centralized monitoring and alarm module 400 is used to display the processing data on the visual interface and view the real-time quantity data of the coal, historical trends, abnormal events and alarm information of the whole network.

[0053] It can be understood that the centralized monitoring and alarm module 400 in the embodiments of the present application can be a centralized monitoring and alarm platform, which is located Figure 2 in the rightmost part of the center data platform 300, and is used for the monitoring interface used by the dispatching and management personnel.

[0054] In the actual execution process, the embodiments of the present application can obtain the processing data and analysis results from the center data platform 300, and provide visual display and interactive operation. The user can view the real-time quantity data of the whole network, historical trends, and abnormal events and alarm information detected by the model. The platform supports multi-dimensional filtering and drilling, such as viewing specific data according to coal mine, transportation line and customer. When an abnormality occurs, the platform alarms in real time in the form of pop-up window, sound, etc., prompts the relevant batch and reason, and assists the manager to respond and handle quickly.

[0055] The abnormal response of the embodiments of the present application is more intelligent and timely: the centralized monitoring and alarm module 400 of the present application can learn the historical normal mode and automatically adapt to the new situation, so the false negative rate and the false positive rate are low. For example, when a device gradually deviates from the calibration, the model can detect the trend abnormality in advance and issue a reminder to prevent a problem from happening; for example, when abnormal coal quality (mixed with heterogeneous coal) occurs, the clustering algorithm can distinguish it and notify the manager. Real-time intelligent alarm enables the staff to take measures (such as repairing the device and rejecting abnormal coal) in time, and reduces the risk to the minimum.

[0056] Optionally, in an embodiment of the present application, the centralized monitoring and alarming module 400 comprises: an anomaly identification unit, configured to train a clustering model using historical data to identify data distribution characteristics, and mark batches deviating from the main clustering center according to the data distribution characteristics to generate marked data, and determine samples in the marked data satisfying a preset high isolation degree condition as anomaly points; an anomaly calculation unit, configured to fuse multiple detection results using a Bayesian probability model to calculate a comprehensive anomaly probability, and generate an anomaly event according to the anomaly points and the anomaly probability; and a display unit, configured to display network-wide real-time quantity and quality data, historical trends, anomaly events and alarm information using a visual interface.

[0057] In actual implementation, the embodiment of the present application can implement an intelligent anomaly identification mechanism: to achieve automatic discovery of metering and coal quality anomalies, the present application fuses a rule method and a statistical learning algorithm to construct an anomaly detection module. First, a rule library is set according to industry experience: for example, a difference between metering results of different batches of equipment exceeding a certain percentage is determined as an anomaly, and a coal quality index exceeding a reasonable range triggers a warning, etc. These rules can quickly capture obvious anomalies. Second, an unsupervised learning algorithm is introduced to detect subtle or complex anomaly patterns: a clustering model (such as K-Means) is trained using historical normal data to identify data distribution, and batches deviating from the main clustering center are marked; an isolation forest algorithm is used to model multi-dimensional data (including quantity, quality, equipment state, etc. feature vectors), and samples with high isolation degree are determined as anomaly points. In addition, a Bayesian probability model is used to fuse multiple detection results to calculate a comprehensive anomaly probability: when both the rule determination and the isolation forest indicate an anomaly, the confidence of the anomaly event is increased; if there is only a slight deviation, the alarm level is reduced. This multi-level anomaly identification mechanism can monitor anomalies in data flow in real time, and immediately issue an alarm through the central monitoring platform once a suspicious situation is found, and record relevant information for operation and maintenance personnel to check and handle.

[0058] The anomaly identification in the embodiment of the present application has higher sensitivity and accuracy, and can timely discover metering equipment hidden dangers and coal quality fluctuations to ensure production and transaction fairness.

[0059] The collaborative analysis module 500 is configured to record metering values, quality indexes, model correction values, anomaly labels and flow time and location of the coal to a full-process data archive of the coal based on network-wide real-time quantity and quality data, historical trends, anomaly events and alarm information, to issue a model and a control strategy to the edge computing node module 200 based on the full-process data archive, and to perform multi-device collaboration for coal quantity and quality fusion and anomaly identification according to the model and the control strategy.

[0060] It can be understood that the collaborative analysis module 500 in the embodiment of the present application can correspond to a coal batch holographic archive part, Figure 2The data storage unit at the lower right side is used for saving the whole-process data file of each batch of coal.

[0061] As a possible implementation, the embodiment of the application is based on real-time network data, historical trends, abnormal events and alarm information. The center data platform 300 writes all information related to the batch into the archive, including measurement values, quality indicators, model correction values, abnormal markers, flow time and location, etc., to realize whole-process trace management of one batch one file. The monitoring platform supports the query and review of the file, facilitating traceability verification. Based on the whole-process data file, the embodiment of the application issues models and control strategies to the edge computing node module 200 to perform multi-device collaboration for coal number and quality fusion and abnormality identification according to the models and control strategies.

[0062] Figure 2 In the architecture shown, the data flow direction of each part is indicated by the arrow connection line: the data of the field multi-source device is first aggregated to the edge node, and after preprocessing, it is uploaded to the center data platform 300; the center data platform 300 stores and analyzes the data, on the one hand, sends the results to the monitoring platform for personnel to view, and on the other hand, writes the key data into the batch archive for saving. In addition, the center platform issues models and control strategies (such as periodically issuing new calibration coefficients, abnormal threshold values, etc.) to the edge node through the dashed arrow to optimize the field data processing. The bidirectional flow of such information ensures the close collaboration between the edge and the center. The system built by the application realizes the full-link data connection from the field device to the central platform in the vertical direction, and covers the fusion analysis of the quantity measurement and quality detection two data domains in the horizontal direction. The system structure closely surrounds the actual needs of the coal production and transportation process, considering both real-time and overall optimization and data sedimentation, providing a solid technical support for realizing intelligent management and control of coal number and quality.

[0063] Through multi-aspect innovation, the application realizes more intelligent and reliable management of coal number and quality data, and provides a complete technical solution from device access, data fusion, intelligent analysis to system architecture. The collaborative action of each innovative module enables the coal production and supply chain to realize multi-source data fusion and intelligent management and control: the edge computing node module 200 (field edge node) ensures timely and accurate data acquisition and preliminary consistency verification, the center data platform 300 realizes fine calibration and abnormality deep mining through advanced machine learning and big data analysis, and finally the results are fed back to the user through the unified platform and archived. The application has a significant improvement in data fusion depth and intelligent level, and can effectively solve the pain points in the prior art.

[0064] Optionally, in an embodiment of the present application, the collaborative analysis module 500 comprises: a establishing unit for assigning a unique tag to each batch of coal and establishing a full-process data file corresponding to the unique tag, wherein the full-process data file comprises at least one of the metering value, the quality index, the model correction value, the abnormality mark and the flow time and location; and a collaborative analysis unit for using the full-process data file to collaboratively analyze the starting location and time, the final receiving location and time, the metering data of each link, the coal quality data of each link, the model correction value, the abnormality alarm log, the person handling and the time stamp of each batch of coal from production to utilization, so as to generate a collaborative analysis result.

[0065] It can be understood that the unique tag in the embodiment of the present application can adopt the form of batch number, RFID or blockchain ID, etc. The present application can generate a unique data tag for each batch of coal, establish a holographic flow file, record the quantity, quality and related processing information of the batch of coal from the production place to the user, enhance the transparency and traceability of the data, and solve the problem of batch data full-process traceability in the traditional mode, in which the data chain is broken and the responsibility is difficult to define.

[0066] In actual execution process, the embodiment of the present application comprises batch data tag and holographic file: in order to enhance data traceability, the present application assigns a unique tag (which can adopt the form of batch number, RFID or blockchain ID, etc.) to each batch of coal, and establishes a corresponding holographic flow file in the data platform. The file records all the key data of the batch of coal from production to utilization, including: starting location and time, final receiving location and time, metering data of each link (loading weight, weighing during transportation, unloading weight, etc.), coal quality data of each link (mine-side rapid inspection, third-party inspection, power plant acceptance test results, etc.), model correction value, abnormality alarm log, person handling and time stamp, etc. The file is also associated with related electronic bills and quality inspection reports, forming a complete "digital fingerprint" of the batch of coal. When a quality dispute or supervision and inspection occurs, the file can be called out to realize data traceability; managers can also analyze cumulative deviation and transportation loss law through the file. In order to prevent data tampering, blockchain technology can be selected to chain the key data and file summary for storage, further improving the traceability credibility. Through the batch holographic file mechanism, the present application greatly improves the transparency and traceability of coal quality management, and provides a technical means for solving the long-standing quality trust problem between buyers and sellers.

[0067] This application's embodiments enhance data transparency and traceability: a holographic data archive is established for each batch of coal, accompanied by a label, giving it a data "identity card" within the supply chain. Any questions regarding the measurement or quality of that batch of coal can be answered authoritatively by consulting the archive. This not only improves internal management efficiency but also helps build trust between trading parties. For example, if a power plant has doubts about the calorific value of received coal, it can compare the archive with the mine's factory testing results and its own laboratory test results to confirm the source of the discrepancy and avoid disputes. The immutability of the data archive provides a transparent basis for regulatory oversight. In summary, this application significantly improves the transparency and fairness of coal quantity and quality management, promoting the industry towards digital trustworthiness. This application also achieves economic and safety benefits: enterprises can reduce direct economic losses and settlement dispute costs caused by measurement errors, while timely detection of coal quality anomalies can prevent substandard coal from flowing into subsequent stages, causing equipment wear or environmental penalties. Furthermore, anomaly monitoring provides early warnings of equipment malfunctions and safety hazards, helping to prevent accidents. Overall, this application will bring considerable economic benefits (improving the accuracy of measurement and settlement, shortening the settlement cycle, and reducing inventory backlog) and safety benefits (reducing the production accident rate) to coal production and utilization enterprises, and play a positive role in the construction of a green and efficient coal supply chain.

[0068] It should be noted that this application also includes extensions: (1) Alternative data fusion algorithm: In the embodiments, graph models or deep learning methods can be introduced to handle the multi-device data fusion problem. Each metering device can be regarded as a node in a graph network, and the measurement relationship can be regarded as an edge. A GNN model can be introduced to learn the correlation of measurement deviations between devices, thereby achieving a similar cross-device correction effect. Compared with the self-developed algorithm, this scheme has higher computational complexity, but it can be used as a supplement in scenarios with extremely complex device relationships.

[0069] (2) Architectural variations: In specific applications, the division of responsibilities between the cloud and the edge can be adjusted according to actual needs. For scenarios with good network conditions and low real-time requirements, a centralized architecture can be adopted, where all data is directly uploaded to the cloud for unified processing, with only simple caching on the device side. This solution simplifies the functions of edge nodes but relies on a stable network. Conversely, for mines with limited network access or requiring local autonomy, a localized architecture can be adopted, where the main data fusion and analysis are completed by the local server (edge) in the mining area, and the summarized results are only periodically uploaded to the center. Regardless of how the responsibilities between the cloud and the edge are adjusted, the core ideas of data fusion and anomaly detection remain consistent, and the purpose of the invention can be achieved.

[0070] (3) Replacement of anomaly detection algorithm: The application uses clustering, isolation forest and other unsupervised algorithms and Bayesian inference for anomaly recognition. In the alternative, other machine learning algorithms can also be used to achieve similar functions, for example: using an autoencoder neural network to establish a normal data reconstruction model, and determining an anomaly when the reconstruction error exceeds a threshold; or using a support vector machine (One-class SVM) for single-class training to detect anomalies. These algorithms can replace clustering or isolation forest under certain conditions to achieve similar anomaly detection results.

[0071] (4) Security variant of batch archives: Although the application refers to optional blockchain storage technology, more traditional digital signature and permission management methods can also be used to ensure that the archives are not tampered with in other implementations. For example, each archive record is automatically stamped with a digital signature when it is generated, and the archives are regularly exported for backup for verification; At the same time, strict access control policies are set, only authorized users can query or update the archives. Such alternative measures can also ensure the reliability of data traceability and prevent others from avoiding the core of the scheme described in this patent to achieve similar purposes.

[0072] (5) Hardware and communication alternatives: The application does not limit the specific protocol and hardware for device data acquisition and transmission. For example, the edge node and the device can communicate through industrial Ethernet / OPC UA, or can obtain data through a 5G industrial wireless network; The center platform can be deployed in the enterprise's own machine room, or in a private cloud / public cloud environment. These changes do not affect the realization of the application, and are equivalent alternatives made by those skilled in the art according to their needs.

[0073] In summary, the application focuses on achieving the purpose of coal quantity and quality multi-device collaborative analysis, and provides multiple implementation approaches in specific means, realizing intelligent fusion, consistency verification and automatic anomaly recognition of different device quantity and quality data, and having the characteristics of edge and center collaborative architecture and batch traceability archives.

[0074] The multi-device collaborative analysis system for coal quantity and quality fusion and abnormality identification provided in the embodiments of the present application can perform centralized monitoring of quantity and quality data in multiple scenarios, so that the total dispatching room can master the coal production, transportation and sales situation of each site in real time; multi-device collaborative collection and unified management of coal quantity and quality data are realized; quantity and quality data generated by different devices are aligned, compared, fused and calculated, and the consistency of multi-source data is automatically verified to ensure that the measurement results and quality indexes of the same batch of coal are consistent at each link; a machine learning model is introduced to predict systematic errors and random errors in the coal measurement process, and a regression model is established for key coal quality indexes to realize the associated prediction and correction of online rapid detection results and laboratory test results, thereby improving the accuracy of quantity and quality data; measurement abnormalities or coal quality abnormalities can be automatically discovered and real-time alarms can be given; the combination of rapid response on the edge side and global optimization on the center side is realized; a unique data tag is generated for each batch of coal, a holographic circulation file is established, and the quantity, quality and related processing information of the batch of coal from the production site to the user are recorded to enhance the transparency and traceability of data. Thus, the problem that related technologies cannot meet the real-time fusion monitoring and intelligent abnormality identification requirements of quantity and quality data in the whole process of coal is solved, and the digital management and efficient decision-making of coal production and the supply chain are restricted. The present application solves the key technical problems in coal quantity and quality management from the aspects of data fusion, intelligent analysis and system architecture, and greatly improves the data monitoring capability and abnormality management level of the industry.

[0075] Secondly, a multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification provided by the embodiments of the present application is described with reference to the accompanying drawings.

[0076] Figure 3 is a flowchart of the multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification of the embodiments of the present application.

[0077] As shown in Figure 3 , the multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification includes the following steps: In step S301, quantity data and quality data of coal are collected.

[0078] In step S302, edge node data collected by multiple measurement and detection devices is acquired, and the edge nodes are preprocessed to obtain target data.

[0079] In step S303, the quantity data, quality data and target data are analyzed to generate complete coal quantity and quality data, and a preset error prediction model and a preset regression prediction model are used to process the complete coal quantity and quality data respectively to generate processing data of the coal.

[0080] In step S304, the processing data is displayed on the visualization interface, and the full-network real-time quantity and quality data, historical trends, abnormal events and alarm information of the coal are viewed.

[0081] In step S305, based on the full-network real-time quantity and quality data, historical trends, abnormal events and alarm information, the measurement value, quality index, model correction value, abnormal mark and flow time and location of the coal are recorded to the full-process data archive of the coal, so as to issue the model and control strategy to the edge computing node module based on the full-process data archive, and to perform multi-device collaboration for coal quantity and quality fusion and abnormal identification according to the model and control strategy.

[0082] Optionally, in an embodiment of the present application, the edge node data collected by multiple measurement and detection devices is acquired, and the edge node is preprocessed to obtain target data, including: converting data of different communication protocols into unified format data, and filtering data in the unified format data that meets a preset abnormal condition to generate data that meets a preset standard condition; fusing contemporaneous data of multiple devices according to the data that meets the preset standard condition to generate the target data.

[0083] Optionally, in an embodiment of the present application, a complete coal quantity and quality data is processed by using a preset error prediction model and a preset regression prediction model to generate processing data of the coal, including: a data storage unit for storing the quantity data and quality data of the coal and the target data to a batch archive library to generate storage data; a data processing unit for performing large-scale data cleaning on the storage data to generate cleaning data; a model training unit for establishing a preset error prediction model according to environmental parameters, device state parameters and operation conditions, training the preset error prediction model to correct measurement deviation to generate a correction result, and establishing a preset regression prediction model according to coal quality indexes to correct coal quality data by using the preset regression prediction model to generate the processing data.

[0084] Optionally, in an embodiment of the present application, the processing data is displayed on the visualization interface, and the full-network real-time quantity and quality data, historical trends, abnormal events and alarm information of the coal are viewed, including: training a clustering model by using historical data to identify data distribution characteristics, and marking batches deviating from the main clustering center according to the data distribution characteristics to generate marked data, and determining samples in the marked data that meet a preset high isolation condition as abnormal points; fusing multiple detection results by using a Bayes probability model to calculate a comprehensive abnormal probability, and generating abnormal events according to the abnormal points and the abnormal probability; displaying the full-network real-time quantity and quality data, historical trends, abnormal events and alarm information on the visualization interface.

[0085] Optionally, in an embodiment of the present application, based on the whole-process data archive, the model and control strategy are issued to the edge computing node module to perform multi-device collaboration for coal quantity and quality fusion and abnormality identification according to the model and control strategy, including: giving each batch of coal a unique label, and establishing a whole-process data archive corresponding to the unique label, wherein the whole-process data archive includes at least one of the measurement value, the quality index, the model correction value, the abnormality mark and the flow time and place; and using the whole-process data archive to collaboratively analyze the starting place and time, the final receiving place and time, the measurement data of each link, the coal quality data of each link, the model correction value, the abnormality alarm log, the person in charge and the time stamp of each batch of coal from the production to the utilization process, to generate a collaborative analysis result.

[0086] It should be noted that the foregoing explanation and description of the embodiment of the multi-device collaborative analysis system for coal quantity and quality fusion and abnormality identification also applies to the embodiment of the multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification, which will not be repeated here.

[0087] The multi-device collaborative analysis method for coal quantity and quality fusion and abnormality identification according to the embodiment of the present application can perform centralized monitoring of quantity and quality data in multiple scenarios, so that the total dispatching room can master the coal production, transportation and marketing situation of each place in real time; realizes multi-device collaborative collection and unified management of coal quantity and quality data; aligns and compares, fuses and calculates the quantity and quality data generated by different devices, and automatically verifies the consistency of multi-source data, to ensure that the measurement results and quality indexes of the same batch of coal are consistent at each link; introduces a machine learning model to predict the systematic error and random error in the coal measurement process, and establishes a regression model for key coal quality indexes to realize the correlation prediction and correction of online rapid detection results and laboratory test results, and improve the accuracy of quantity and quality data; can automatically find measurement abnormalities or coal quality abnormalities and alarm in real time; realizes the combination of rapid response on the edge side and global optimization on the center side; generates a unique data label for each batch of coal, establishes a holographic flow archive, records the quantity, quality and related processing information of the batch of coal from the production place to the user, and enhances the transparency and traceability of the data. Thus, the problem that the related technology cannot meet the real-time fusion monitoring and intelligent abnormality identification needs of the coal whole-process quantity and quality data, and restricts the digital management and efficient decision-making of the coal production and supply chain, is solved.

[0088] Figure 4 The structure schematic diagram of the electronic device provided by the embodiment of the present application is shown in the figure. The electronic device can include: The memory 401, the processor 402 and the computer program stored in the memory 401 and executable on the processor 402.

[0089] The processor 402 implements the multi-device collaborative analysis method for coal number and mass fusion and abnormality identification provided in the above embodiments when executing a program.

[0090] Further, the electronic device further comprises: The communication interface 403 is configured to communicate between the memory 401 and the processor 402.

[0091] The memory 401 is configured to store a computer program executable on the processor 402.

[0092] The memory 401 can include a high-speed RAM memory, and can further include a non-volatile memory, for example, at least one disk memory.

[0093] If the memory 401, the processor 402 and the communication interface 403 are independently implemented, the communication interface 403, the memory 401 and the processor 402 can be connected to each other through a bus and complete communication between each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 4 Only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus.

[0094] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can complete communication between each other through an internal interface.

[0095] The processor 402 can be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0096] The embodiments also provide a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the multi-device collaborative analysis method for coal number and mass fusion and abnormality identification as above.

[0097] In the description of the application, reference to "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that a particular feature, structure, material, or characteristic being described is included in at least one embodiment or example of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment or example. Furthermore, the described specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples. In addition, the usage of "N" means at least two, for example, two, three or the like, unless explicitly stated otherwise.

[0098] Furthermore, the terms "first", "second", or the like, are used merely as a designation of certain elements or features, and do not imply or connote relative importance or a specific order of categorization thereof. Accordingly, features described as "first" or "second" can implicitly or explicitly include at least one of the features. In the description of the application, the term "N" means at least two, for example, two, three, etc., unless otherwise specifically stated.

[0099] Any process or method descriptions or blocks in flow charts or otherwise described herein represent embodiments which can be managed as one or more modules, segments, or portions of code which include one or more executable instructions for implementing specific logic functions or steps, and alternate implementations are possible. In some embodiments, the processes and methods described can be executably encoded on a machine- readable medium in a data signal embodied in an electromagnetic signal, a wireless signal, or a propagated signal.

[0100] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0101] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0102] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0103] In addition, each of the functional units in the various embodiments of the present application can be integrated in one processing module, or each of the units can be physically present separately, or two or more units can be integrated in one module. The integrated module can be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0104] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application.

Claims

1. A multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification, characterized in that, include: Multi-source data acquisition module, used to collect quantity and quality data of coal; An edge computing node module is connected to various metering and testing devices to acquire edge node data collected by the various metering and testing devices and preprocess the edge nodes to obtain target data. The central data platform is used to analyze the quantity data, the quality data, and the target data to generate complete coal quantity and quality data. It also uses a preset error prediction model and a preset regression prediction model to process the complete coal quantity and quality data to generate coal processing data. The centralized monitoring and alarm module is used to display the processing data on a visual interface and view the real-time quantity and quality data, historical trends, abnormal events and alarm information of the coal across the entire network. The collaborative analysis module is used to record the coal's measurement value, quality indicators, model correction value, anomaly markers, and transfer time and location into the coal's full-process data archive based on the real-time data and quality data of the entire network, the historical trends, the abnormal events, and the alarm information. Based on the full-process data archive, the module distributes models and control strategies to the edge computing node module to perform multi-device collaboration for coal data and quality fusion and anomaly identification according to the models and control strategies.

2. The multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification according to claim 1, characterized in that, The multi-source data acquisition module is further used to connect at least one heterogeneous device among belt scales, quantitative loading systems, rail scales, truck scales, water level gauges, samplers, rapid coal quality testing equipment, and laboratory manual testing equipment to the edge computing node in order to obtain the quantity and quality data of the coal.

3. The multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification according to claim 1, characterized in that, The edge computing node module includes: The filtering unit is used to convert data from different communication protocols into data in a unified format, and filter the data in the unified format data that meets preset abnormal conditions, so as to generate data that meets preset standard conditions. The fusion unit is used to fuse synchronous data from multiple devices based on the data that meets the preset standard conditions, in order to generate the target data.

4. The multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification according to claim 1, characterized in that, The central data platform includes: A data storage unit is used to store the quantity and quality data of the coal and the target data into a batch archive to generate stored data; The data processing unit is used to perform large-scale data cleaning on the stored data to generate cleaned data; The model training unit is used to establish the preset error prediction model based on environmental parameters, equipment status parameters and operating conditions, train the preset error prediction model to correct measurement deviations, generate correction results, and establish the preset regression prediction model based on coal quality indicators to correct coal quality data and generate processing data.

5. The multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification according to claim 1, characterized in that, The centralized monitoring and alarm module includes: An anomaly identification unit is used to train a clustering model using historical data to identify data distribution characteristics, and to mark batches that deviate from the main cluster center according to the data distribution characteristics, generating labeled data, and identifying samples in the labeled data that meet the preset high isolation condition as anomalies. An anomaly calculation unit is used to fuse multiple detection results using a Yeats probability model to calculate a comprehensive anomaly probability, and to generate the anomaly event based on the anomaly point and the anomaly probability. The display unit is used to display the real-time data quality of the entire network, the historical trends, the abnormal events, and the alarm information through the visualization interface.

6. The multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification according to claim 1, characterized in that, The collaborative analysis module includes: A unit is established to assign a unique label to each batch of coal and to establish a full-process data archive corresponding to the unique label. The full-process data archive includes at least one of the following: measurement value, quality index, model correction value, anomaly marker, and transfer time and location. The collaborative analysis unit is used to collaboratively analyze the starting location and time of multiple devices, the final receiving location and time, the measurement data of each link, the coal quality data of each link, the model correction value, the abnormal alarm log, the handler and the timestamp of each link in the process of each batch of coal from production to utilization, in order to generate collaborative analysis results.

7. A multi-device collaborative analysis method for coal quantity and quality fusion and anomaly identification, employing the multi-device collaborative analysis system for coal quantity and quality fusion and anomaly identification as described in any one of claims 1-6, characterized in that, Includes the following steps: Collect quantity and quality data of coal; The edge node data collected by the various metering and testing devices is acquired, and the edge nodes are preprocessed to obtain the target data; The quantitative data, the quality data, and the target data are analyzed to generate complete coal quantity and quality data. The complete coal quantity and quality data are then processed using a preset error prediction model and a preset regression prediction model to generate coal processing data. The processing data is displayed on a visual interface, and the real-time quantity and quality data, historical trends, abnormal events, and alarm information of the coal across the entire network are viewed. Based on the real-time data and quality information of the entire network, the historical trends, the abnormal events, and the alarm information, the measurement value, quality indicators, model correction values, anomaly markers, and transfer time and location of the coal are recorded in the full-process data archive of the coal. Based on the full-process data archive, models and control strategies are distributed to the edge computing node modules to enable multi-device collaboration for coal data and quality fusion and anomaly identification according to the models and control strategies.

8. The multi-device collaborative analysis method for coal quantity and quality fusion and anomaly identification according to claim 7, characterized in that, The process of acquiring edge node data collected by the various metering and testing devices and preprocessing the edge nodes to obtain target data includes: Data from different communication protocols is converted into a unified format, and data that meets preset abnormal conditions is filtered out from the unified format data to generate data that meets preset standard conditions. Based on the data that meets the preset standard conditions, synchronous data from multiple devices are fused to generate the target data.

9. An electronic device, characterized in that, include: The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the multi-device collaborative analysis method for coal quantity and quality fusion and anomaly identification as described in any one of claims 7-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the multi-device collaborative analysis method for coal quantity and quality fusion and anomaly identification as described in any one of claims 7-8.

Images