A blockchain-based intelligent spray management method and system for ore terminals

Abstract

The present invention relates to the field of blockchain technology, and in particular to a blockchain-based intelligent spray management method and system for ore terminals. The method comprises the following steps: obtaining ore stockpile data; verifying the ore stockpile data through a blockchain consensus mechanism to obtain ore stockpile block data; generating a spray strategy based on the ore stockpile block data to obtain spray control strategy data; controlling the spray equipment based on the spray control strategy data to obtain spray implementation data to record and trace the spray process. The present invention records the entire process from collection, storage to strategy execution in the blockchain, forming a transparent management process, supporting multi-party auditing and compliance verification, and all decision-making and implementation records of the spray process can be traced, which is convenient for problem troubleshooting and optimization and improvement.

Description

Technical Field

[0001] The present invention relates to the field of blockchain technology, and in particular to a blockchain-based intelligent spray management method and system for an ore terminal. Background Art

[0002] In the modern industrial field, ore terminals are an important link in the transportation and storage of minerals. Due to the large amount of dust generated during the stacking, loading and unloading and transportation of ore, it not only causes serious pollution to the surrounding environment, but also endangers the health of on-site workers. Therefore, how to efficiently control the spread of dust and optimize the resource utilization of the spraying system has become an important challenge in the management of ore terminals. In the traditional management model, environmental data recording and spraying process lack transparency, data can be easily tampered with, and it is difficult to meet the supervision and traceability needs of multiple parties. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a blockchain-based intelligent spraying management method and system for ore terminals to solve at least one of the above technical problems.

[0004] This application provides a blockchain-based intelligent spray management method for ore terminals, comprising the following steps:

[0005] Step S1: Acquire ore stockpile data;

[0006] Step S2: Verify the ore stockpile data through the blockchain consensus mechanism to obtain the ore stockpile block data;

[0007] Step S3: generating a spraying strategy according to the block data of the ore stockpile yard to obtain spraying control strategy data;

[0008] Step S4: Control the spraying equipment according to the spraying control strategy data to obtain the spraying implementation data for spraying process recording and tracing operations.

[0009] In the present invention, data of multiple modes (such as temperature, humidity, dust concentration, etc.) can be collected simultaneously to provide a basis for multi-dimensional spray optimization and improve the environmental governance effect. The blockchain consensus mechanism is used to verify the authenticity of the data, prevent the data from being tampered with or forged, and improve the credibility of the system. The spray data of the ore terminal is stored on the blockchain to ensure the transparency and traceability of the data. Based on blockchain data and intelligent algorithms (such as deep learning or optimization models), the spray strategy is dynamically generated to ensure the optimal configuration of the spray range, time and intensity. The environmental change trend can be predicted based on real-time data and historical data, and adaptive strategies can be formulated in advance. The spray strategy can be dynamically adjusted according to environmental changes, for example, the spray intensity can be reduced when the wind speed is too high to prevent the waste of spray water or the reduction of the effect. The operation data of the spray equipment, the spray effect and the environmental changes are all recorded on the blockchain to achieve traceability of the whole process. When a problem occurs (such as insufficient spray effect or equipment failure), the source of the problem can be quickly traced back and a solution can be formulated.

[0010] Preferably, step S1 specifically includes:

[0011] Step S11: deploying an ore stockpile monitoring sensor network to cover key areas of the ore stockpile to monitor temperature, humidity, and dust concentration, thereby obtaining real-time collected data;

[0012] Step S12: performing edge multimodal fusion on the real-time collected data to obtain multimodal data of the ore stockpile yard;

[0013] Step S13: constructing sequence features for the multimodal data of the ore stockpile to obtain sequence feature data of the ore stockpile;

[0014] Step S14: encrypting the ore stockpile sequence characteristic data and uploading the data to obtain the ore stockpile data.

[0015] In the present invention, a sensor network is deployed in key areas of the ore stockpile to monitor key environmental parameters such as temperature, humidity, and dust concentration. The sensor coverage and density are optimized according to the site terrain and stockpile characteristics to ensure the comprehensiveness and accuracy of the monitoring. The multimodal data (temperature, humidity, dust concentration, etc.) collected in real time are fused through the edge computing node. A multimodal fusion algorithm (such as a fusion model based on weighted mutual information) is used to integrate data of different modes into a consistent data format. Sequence feature extraction is performed on the fused multimodal data to construct a key feature sequence (such as dust concentration change trend, humidity fluctuation characteristics, etc.). The generated sequence feature data is encrypted (such as AES or RSA encryption) to ensure the security of the data during transmission. The encrypted data is uploaded to the blockchain node to ensure the transparency and non-tamperability of the data.

[0016] Preferably, step S13 is specifically:

[0017] Constructing differential sequence features of the multimodal data of the ore stockpile to obtain sequence feature data of the first ore stockpile;

[0018] Constructing pattern sequence features of the multimodal data of the ore stockpile to obtain sequence feature data of the second ore stockpile;

[0019] A multi-sequence weight graph is constructed by performing a multi-sequence weight graph on the first ore stockpile sequence characteristic data and the second ore stockpile sequence characteristic data to obtain the ore stockpile sequence characteristic data.

[0020] The differential sequence feature construction in the present invention extracts the dynamic change characteristics in the time dimension by performing differential analysis on the time series of multimodal data. The key features of the environmental changes in the ore stockpile are captured by differential calculation (such as first-order difference, second-order difference) and change rate analysis methods. The pattern sequence feature construction extracts the long-term characteristics of environmental data by mining the potential laws and repeated patterns in multimodal data (such as periodic fluctuations in dust concentration). Pattern mining helps to predict future environmental changes (such as the possibility of increased dust concentration in a certain period of time) and realize forward-looking management. The differential sequence feature data and the pattern sequence feature data are fused through a multi-sequence weight graph. The nodes in the graph represent features, and the edge weights reflect the strength of association between features. A graph construction algorithm (such as minimum spanning tree MST and graph embedding technology) is used to integrate multidimensional features into unified feature data.

[0021] Preferably, the differential sequence feature is constructed as follows:

[0022] Dynamically divide the span of multimodal data of ore stockpile to obtain divided sequence data;

[0023] Perform multimodal cross-difference calculation on the divided sequence data to obtain cross-difference matrix data;

[0024] Perform differential trend cluster analysis on the cross-difference matrix data to obtain differential sequence cluster data;

[0025] Perform differential feature extraction on differential sequence clustering data to obtain differential sequence feature data;

[0026] The differential sequence feature data is embedded into a differential feature sequence to obtain the first ore stockpile sequence feature data.

[0027] In the present invention, the multimodal data is span-divided using an adaptive window mechanism according to the time dimension and data change characteristics. The partitioning strategy dynamically adjusts the window size to adapt to the change rate of different data modes. The dynamic partitioning strategy of different modal data avoids the information loss problem caused by fixed windows and improves the applicability of the analysis. On the basis of time span division, differential calculations are performed on the data within and between modes. The differential calculation includes first-order differences, second-order differences, and change rate analysis to generate a cross-difference matrix that reflects the correlation between modes and the time series dynamics. Based on the cross-difference matrix, dynamic time warping or K-Shape algorithm is used to cluster the differential trends. The clustering results group similar time series to form differential sequence clustering data. Key features are extracted from the clustered differential sequences, including change rate, peak position, change direction, etc. The most representative differential features are extracted using feature selection algorithms (such as Lasso regression or random forest feature importance evaluation). Differential feature extraction can filter redundant information and extract the features that are most sensitive to environmental changes. The differential features are embedded into a low-dimensional feature space using a graph neural network (GNN) or time series embedding technology (such as a time convolutional network (TCN)). The embedded sequence feature vector is used to represent the overall environmental dynamic characteristics of the ore stockpile. The embedding technology captures the potential correlation between features and provides a more refined expression of complex environmental changes.

[0028] Preferably, the pattern sequence feature construction is specifically as follows:

[0029] Perform mutation point detection on multimodal data of ore stockpile to obtain multimodal mutation point data;

[0030] Performing event modal mapping on multimodal mutation point data to obtain event modal mapping data;

[0031] Perform inter-modal pattern association mining on event modal mapping data to obtain modal association data;

[0032] Reconstructing the mode spectrum of the modal association data to obtain the modal spectrum data;

[0033] The modal spectrum data is subjected to a pattern dynamic sequence generation to obtain the second ore stockpile sequence characteristic data.

[0034] The present invention uses a mutation point detection algorithm (such as the Pelt algorithm or an online change point detection algorithm based on Bayesian) to identify significant change points in the multimodal data of the ore stockpile. Mutation points reflect the occurrence of key events in the environment (such as a sudden increase in dust concentration and a sharp drop in humidity). Mutation point detection can accurately identify abnormal changes in multimodal data and quickly locate key events in the environment. A mapping relationship is established between mutation points and related multimodal data to generate a bidirectional link between events and modalities. The mapping rules are dynamically adjusted based on the importance weight of the modality and the intensity of the impact of the mutation point on the environment. Modal mapping enables the system to focus on modal features that are strongly related to events and improve the accuracy of spray strategy generation. Association mining makes pattern features no longer limited to a single mode, but integrates multiple modal characteristics to enhance the expression ability of the mode. The modal association data is converted into a graph structure, the node represents the modal event, the edge represents the association relationship between the modes, the weight reflects the association strength, and the mode spectrum expresses the modal association and its strength in the form of a graph structure, which is convenient for capturing the characteristics of global and local modes. Based on the reconstructed modal spectrum data, dynamic pattern sequences are generated to capture the pattern evolution in the time dimension. The dynamic pattern sequences strengthen the time series characteristics and improve the temporal consistency of pattern features.

[0035] Preferably, the multi-sequence weight graph is constructed as follows:

[0036] Performing multi-sequence feature fusion initialization on the first ore stockpile sequence feature data and the second ore stockpile sequence feature data to obtain initial feature fusion matrix data;

[0037] Perform multi-dimensional feature association mapping on the initial feature fusion matrix data to obtain feature association matrix data;

[0038] Perform sequence difference weight assignment on feature association matrix data to obtain difference weight matrix data;

[0039] The modal weight map is constructed for the difference weight matrix data to obtain the preliminary ore stockpile sequence characteristic data;

[0040] Extracting characteristic subgraphs from the preliminary ore stockpile sequence characteristic data to obtain ore stockpile sequence characteristic data;

[0041] The feature subgraph extraction is as follows:

[0042] Extract characteristic nodes from the preliminary ore stockpile sequence characteristic data and construct time series characteristics to obtain characteristic node graph data and time series characteristic edge data respectively;

[0043] Perform cross-modal weight balancing based on feature node graph data and time series feature edge data to obtain cross-modal fusion matrix data;

[0044] Perform sub-graph aggregation on the cross-modal fusion matrix data to obtain cross-modal sub-graph data;

[0045] Extracting sub-graph patterns from cross-modal sub-graph data to obtain sub-graph pattern data;

[0046] The graph features of the cross-modal sub-graph data are embedded according to the sub-graph mode data to obtain the sequence feature data of the ore stockpile.

[0047] In the present invention, key feature nodes (such as dust concentration, humidity, temperature, etc.) are extracted from the preliminary ore stockpile sequence feature data. Time correlation or dynamic time warping algorithm is used to construct the time series feature edges between feature nodes. The constructed time series feature edges capture the dynamic association between features in the time dimension and provide a deep understanding of environmental changes. Dynamically adjust the weight balance to make multimodal data more coordinated during fusion to avoid excessive influence of a certain modal feature on the analysis results. Based on the cross-modal fusion matrix data, the nodes and edges are aggregated into cross-modal subgraphs using a graph clustering algorithm (such as Louvain algorithm or K-Means). The subgraph aggregation divides the original graph structure into multiple small-scale subgraphs to reduce the complexity of the analysis. The subgraph pattern reflects the high-frequency patterns and potential laws in the ore stockpile environment (such as the interaction pattern between specific humidity and dust concentration). The embedding result represents the overall characteristics and dynamic changes of the ore stockpile. The embedded feature vector retains the global and local characteristics of the original graph, providing efficient input for spray strategy prediction.

[0048] Preferably, step S2 specifically includes:

[0049] Step S21: starting a consensus voting mechanism for the ore stockpile data to obtain voting status record data;

[0050] Step S22: Performing data credibility scoring according to the voting status record data to obtain credibility scoring data;

[0051] Step S23: isolating abnormal data of the ore stockpile data according to the trustworthy scoring data to obtain abnormal isolation record data;

[0052] Step S24: Generate blockchain blocks based on the trustworthy score data and the abnormal isolation record data to obtain preliminary ore stockpile block data;

[0053] Step S25: Perform cross-chain verification on the preliminary ore stockpile block data to obtain the ore stockpile block data.

[0054] The present invention uses the consensus algorithm of blockchain (such as PBFT, PoS or Raft) to perform node voting verification on the ore stockpile data. Each node votes according to the integrity and legitimacy of the data and records the voting status. The consensus voting mechanism ensures the authenticity of the uploaded data through independent verification of multiple nodes to avoid cheating or tampering of data by a single node. The trust score of the data is calculated based on the voting results, and the score is based on data consistency, node historical trust and voting status. The data trust score is used to measure the trustworthiness of the data in the network, and the trust score is used to filter out data that may be untrustworthy or controversial. According to the trust score, abnormal data (such as sensor failure, noise or human tampering data) is identified and isolated. Abnormal data is marked and stored in an independent isolation record to prevent it from affecting the main data chain. The data that passes the trust score and the abnormal isolation record data are organized into a new block according to the blockchain rules. The block contains a timestamp, a data hash value and a hash value of the previous block to ensure the chain structure and immutability of the data. Through cross-chain technology (such as Polkadot or Cosmos), the preliminary block data is secondary verified with other related chains to ensure data consistency and sharing. The verification results are stored on the main chain to form the final ore stockpile block data.

[0055] Preferably, step S3 specifically includes:

[0056] Step S31: extracting data according to the block data of the ore stockpile to obtain basic data of the ore stockpile;

[0057] Step S32: constructing an ore stockpile model based on the ore stockpile basic data to obtain an ore stockpile model;

[0058] Step S33: performing feature correlation analysis on the ore stockpile model to obtain feature matrix data of the ore stockpile;

[0059] Step S34: Calculate the spraying demand for the characteristic matrix data of the ore stockpile to obtain the spraying demand data;

[0060] Step S35: performing spray control simulation on the ore stockpile model according to the spray demand data to obtain spray control simulation data;

[0061] Step S36: Generate a spray strategy based on the spray control simulation data and the spray demand data to obtain spray control strategy data.

[0062] In the present invention, key basic data (such as temperature and humidity, dust concentration, wind speed, etc.) and historical operation data are extracted from the block data of the ore stockpile. Data extraction is screened by specific rules (such as time range, geographical location, modal importance, etc.) to ensure that the extracted content is related to the current spraying demand. A mathematical model of the ore stockpile is constructed, and the environmental parameters, historical data and equipment characteristics are integrated to simulate the environmental dynamics of the stockpile. The model is flexibly adjusted according to the characteristics of different stockpile yards (such as stockpile shape, sensor layout) to improve applicability. Feature association analysis techniques (such as canonical correlation analysis and Pearson correlation analysis) are used to reveal the correlation between the environmental parameters of the ore stockpile. A feature matrix is ​​generated to reflect the interaction between each environmental variable (such as humidity, dust concentration) and the spraying demand. The feature association analysis reveals the potential relationship between environmental variables (such as the relationship between decreased humidity and increased dust concentration), thereby improving data interpretation. Based on the characteristic matrix, the spray demand calculation model (such as linear programming, random forest regression) is used to calculate the spray demand of the current environment. The calculation content includes spray volume, spray time, spray area, etc. According to the calculation results, the key areas are accurately covered to minimize dust and other environmental risks. The optimal strategy is used to effectively reduce dust concentration and environmental pollution, and improve the overall environmental quality of the ore stockpile.

[0063] Preferably, step S4 is specifically:

[0064] Step S41: parsing the equipment control instruction according to the spray control strategy data to obtain the spray equipment control instruction data;

[0065] Step S42: activating the spray equipment and monitoring the spraying in real time according to the spray equipment control instruction data to obtain real-time monitoring data of the spraying;

[0066] Step S43: Analyze the spraying effect of the real-time monitoring data of the spraying to obtain the spraying implementation data for recording and tracing the spraying process.

[0067] In the present invention, the spray control strategy data is parsed into specific instructions that can be executed by the device, including nozzle position, spray time, spray intensity, and equipment start and stop status. The parsing process is connected to the device communication protocol through an instruction mapping algorithm (such as JSON parsing or binary protocol conversion). According to the parsed control instructions, the spray equipment is started and the spray operation is performed. The key parameters such as the equipment status, spray range, and water flow pressure during the spraying process are tracked and recorded in real time through a real-time monitoring system (such as IoT sensors and cameras). Using the real-time monitoring data of the spray, the spray effect is evaluated through an analysis algorithm (such as dust reduction rate calculation and humidity change analysis). Spray implementation data is generated, including key information such as spray coverage, dust reduction effect, and water resource consumption. The spray implementation data is uploaded to the blockchain to complete the recording and traceability of the spray process. The recorded content includes spray time, equipment status, dust reduction effect, environmental changes, etc., forming an unalterable data archive.

[0068] Preferably, the present application also provides a blockchain-based ore terminal intelligent spray management system, which is used to execute the blockchain-based ore terminal intelligent spray management method as described above, and the blockchain-based ore terminal intelligent spray management system includes:

[0069] Ore stockpile data acquisition module, used to obtain ore stockpile data;

[0070] The ore stockpile block generation module is used to verify the ore stockpile data through the blockchain consensus mechanism to obtain the ore stockpile block data;

[0071] A spray control strategy generation module is used to generate a spray strategy according to the block data of the ore stockpile yard to obtain spray control strategy data;

[0072] The spray equipment control module is used to control the spray equipment according to the spray control strategy data and obtain the spray implementation data to record and trace the spray process.

[0073] The beneficial effect of the present invention is that through edge computing and dynamic sequence feature analysis, the dynamic changes of multimodal data such as dust concentration, humidity, and temperature can be captured in real time. The ore stockpile data is verified through the blockchain consensus mechanism, and distributed node voting is used to ensure that the data source is authentic and the process is transparent. The immutability of blockchain storage ensures the credibility of the environmental data of the ore terminal and eliminates the risk of human tampering or data loss. Through the trusted scoring and abnormal data isolation steps, invalid data caused by sensor failure or network fluctuations can be effectively eliminated, improving data quality and reliability. Spray strategy generation integrates dynamic demand prediction, resource utilization optimization and environmental response speed through environmental modeling and feature association analysis, and can generate the optimal spray strategy in real time. Through real-time spray monitoring and feedback mechanism, the spray equipment can adaptively adjust the working mode (such as adjusting the spray intensity, coverage and time) according to the dynamic changes of the environment to achieve precise dust reduction and resource conservation. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings:

[0075] Figure 1 A flowchart of a blockchain-based intelligent spray management method for an ore terminal according to an embodiment is shown;

[0076] Figure 2 A flowchart showing the steps of a method for collecting data from an ore stockpile yard according to an embodiment is shown;

[0077] Figure 3 A flowchart showing the steps of a method for generating an ore stockpile block according to an embodiment is shown;

[0078] Figure 4 A flow chart showing the steps of a method for generating a spray control strategy according to an embodiment is shown;

[0079] Figure 5 A flow chart showing the steps of a method for controlling a spraying device according to an embodiment is shown. DETAILED DESCRIPTION

[0080] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0081] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0082] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0083] See also Figures 1 to 5 , this application provides a blockchain-based intelligent spray management method for ore terminals, comprising the following steps:

[0084] Step S1: Acquire ore stockpile data;

[0085] Specifically, a multi-type sensor network is deployed in the ore terminal stockpile yard, including temperature, humidity, dust concentration and stockpile height sensors. The sensors collect data at fixed time intervals (for example, every minute) and upload the real-time collected data to the edge computing device through a wireless network. The edge computing device performs data preprocessing to form a basic data table containing multimodal information, such as the temperature of a certain area is 25.4°C, the humidity is 65%, the dust concentration is 85µg / m³, and the stockpile height is 7.5 meters. After preprocessing, the data is batch sorted and uploaded to the blockchain network for subsequent verification and analysis.

[0086] Step S2: Verify the ore stockpile data through the blockchain consensus mechanism to obtain the ore stockpile block data;

[0087] Specifically, the data of the ore stockpile is received by multiple nodes and participates in the blockchain consensus verification. Each node independently audits the integrity and credibility of the data, such as detecting whether the data is within a reasonable range (such as the temperature is between -10°C and 50°C). After the audit is completed, the nodes vote and record the number of accepted votes and rejected votes. Trusted data is marked as "valid" and abnormal data is isolated. The valid data that passes the consensus is packaged to generate blocks, and a unique hash value is attached to each block. Each block contains a hash link to the previous block to ensure that the data cannot be tampered with. The ore stockpile block data is stored in the blockchain as a trusted input data source.

[0088] Specifically, the ore stockyard data is hashed to generate a unique identifier. The hash value is broadcast to each node in the ore terminal blockchain network, and the nodes compare the data integrity and authenticity. If more than 2 / 3 of the nodes feedback successful verification, it is considered passed.

[0089] Step S3: generating a spraying strategy according to the block data of the ore stockpile yard to obtain spraying control strategy data;

[0090] Specifically, by analyzing the ore stockpile data stored in the blockchain, combined with the stockpile area division, environmental conditions and spraying requirements, a digital model of the stockpile is constructed. Based on the model analysis, for example, if the humidity in the current area is insufficient and the dust concentration is high, the system calculates the amount of humidity that needs to be increased and the target value for reducing the dust concentration. Based on these requirements, a spraying strategy is generated, including spraying intensity, spraying time and spraying location. For example, for area A, the recommended spraying time is 9 to 10 in the morning, 50 liters of water are sprayed per minute, and the spraying coverage area is the central area of ​​the stockpile. The generated spraying strategy is recorded and sent to the equipment control module.

[0091] Step S4: Control the spraying equipment according to the spraying control strategy data to obtain the spraying implementation data for spraying process recording and tracing operations.

[0092] Specifically, according to the spraying strategy, the control module starts the spraying equipment to perform the spraying operation. The control signal includes specific parameters of spraying intensity, spraying time and spraying range. For example, after the spraying equipment is started, it starts spraying at an intensity of 50 liters per minute in area A for 1 hour. During the spraying process, the sensor monitors the environmental data in real time and records the spraying effect, including the increase in humidity and the decrease in dust concentration. These data are stored in the blockchain together with the operation log of the equipment to form a complete record of the spraying process. Through the traceability of the blockchain, it is possible to verify whether each spraying operation meets the expected goals and provide data support for subsequent optimization.

[0093] Preferably, step S1 specifically includes:

[0094] Step S11: deploying an ore stockpile monitoring sensor network to cover key areas of the ore stockpile to monitor temperature, humidity, and dust concentration, thereby obtaining real-time collected data;

[0095] Specifically, a series of environmental monitoring sensors, including temperature, humidity sensors, dust concentration sensors, and stockpile height measuring devices, are deployed in key areas of the ore stockpile. The sensors cover the entire stockpile in a uniformly distributed or targeted manner. For example, sensors are installed in the center, edge, and key pollution areas of the stockpile to ensure that environmental data can be fully collected. The sensors are connected to edge computing devices via wireless communication networks (such as LoRa or Wi-Fi) and collect and transmit data in real time at a fixed sampling period (for example, every minute). These data include parameters such as ambient temperature, air humidity, PM2.5 dust concentration, and the average height of the ore stockpile.

[0096] Step S12: performing edge multimodal fusion on the real-time collected data to obtain multimodal data of the ore stockpile yard;

[0097] Specifically, after receiving real-time data from multiple sensors, the edge computing device first performs data preprocessing, including denoising and data formatting, to unify different types of data into a standard format. For example, temperature data is expressed in degrees Celsius and dust concentration is expressed in micrograms per cubic meter. Subsequently, the edge computing device performs multimodal fusion of the data through spatial weighting and time synchronization technology. For example, the dust concentration is combined with the temperature and humidity data, and its distribution characteristics in different areas are analyzed to obtain a complete set of multimodal data of the ore stockpile. The fused data not only reflects the environmental conditions, but also provides comprehensive status information of the stockpile.

[0098] Step S13: constructing sequence features for the multimodal data of the ore stockpile to obtain sequence feature data of the ore stockpile;

[0099] Specifically, on the edge computing device, the multimodal data of the ore stockpile is analyzed in time series using data within a certain time range (e.g., the past 24 hours). By comparing the data changes at consecutive moments, the sequence characteristics of the ore stockpile are constructed. For example, the rising trend of humidity corresponds to the falling trend of dust concentration, and the change in stockpile height affects the local dust concentration. Features are stored in the form of sequences and contain statistical values ​​(such as average values, fluctuation ranges) and dynamic change patterns (such as rising or falling trends). These sequence features can provide a reference for subsequent spraying strategies.

[0100] Step S14: encrypting the ore stockpile sequence characteristic data and uploading the data to obtain the ore stockpile data.

[0101] Specifically, after completing the construction of sequence features, the system encrypts the data to protect its privacy and security. For example, an encryption algorithm is used to generate irreversible ciphertext for characteristic data such as humidity and dust concentration to ensure that the data will not be tampered with during transmission. The encrypted data is uploaded to the blockchain network via a wireless network and stored and managed by multiple nodes. Each piece of data is attached with a timestamp and source node information to achieve data traceability and verification. The uploaded data constitutes the basic data set of the ore stockpile.

[0102] Preferably, step S13 is specifically:

[0103] Constructing differential sequence features of the multimodal data of the ore stockpile to obtain sequence feature data of the first ore stockpile;

[0104] Specifically, the multimodal data of the ore stockpile are analyzed in time series to extract the data change characteristics between consecutive time points. For example, the humidity, temperature and dust concentration data are differentiated hour by hour, and their changes are recorded. Through these differential data, the rate and direction of environmental changes in the stockpile can be identified, such as whether the humidity is gradually decreasing, whether the dust concentration is rapidly increasing, etc. The system performs statistical processing on these differential changes, extracts characteristic values ​​(such as average change and fluctuation range), and forms the first ore stockpile sequence characteristic data. These characteristics reflect the dynamic changes in the environment in the stockpile.

[0105] Constructing pattern sequence features of the multimodal data of the ore stockpile to obtain sequence feature data of the second ore stockpile;

[0106] Specifically, according to the changing patterns of multimodal data in different time periods, repetitive or regular changing patterns are identified and classified. For example, humidity may rise in the early morning, while dust concentration will increase significantly during the peak period of stockpiling operations. The system sorts out these changing patterns, divides them into several pattern categories, and encodes and counts the patterns. These pattern data describe the evolution trend of the ore stockpile environment under typical conditions, and constitute the second ore stockpile sequence characteristic data.

[0107] A multi-sequence weight graph is constructed by performing a multi-sequence weight graph on the first ore stockpile sequence characteristic data and the second ore stockpile sequence characteristic data to obtain the ore stockpile sequence characteristic data.

[0108] Specifically, the first ore stockpile sequence feature data is integrated with the second ore stockpile sequence feature data to construct a weight graph describing the correlation between them. For example, if a rapid drop in humidity is usually accompanied by a significant increase in dust concentration, the system will assign a higher weight to the relationship between the two features. At the same time, the dynamic changes of differential features and the repeatability of pattern features are comprehensively evaluated to further optimize the correlation strength between features. The generated ore stockpile sequence feature data is a structured data set that comprehensively describes the environmental state, dynamic changes, and pattern associations, which can provide all-round information support for the formulation of spraying strategies.

[0109] Preferably, the differential sequence feature is constructed as follows:

[0110] Dynamically divide the span of multimodal data of ore stockpile to obtain divided sequence data;

[0111] Specifically, the multimodal data of the ore stockpile are dynamically divided into time spans. For example, according to the rate of change of the environmental data of the stockpile, the period of stable change is divided into a larger time span, while the period of drastic change is divided into a smaller time span. Taking one day as a unit, if the humidity and temperature change amplitudes are small in some periods, they are merged into one period; and when the dust concentration fluctuates significantly during the peak period of stockpile operation, it is divided into shorter time periods. The result of dynamic division is a number of sequences with different time spans, each of which corresponds to a relatively stable or significantly changing environmental state.

[0112] Specifically, based on the timestamp of the calibration data frame, an online change point detection algorithm is used to dynamically divide the time series window. The amount of data and span in each time window are dynamically adjusted to ensure that sufficient change information is captured. After the window is divided, the data is indexed to support differential calculation.

[0113] Perform multimodal cross-difference calculation on the divided sequence data to obtain cross-difference matrix data;

[0114] Specifically, for the data in each divided time period, the cross-differences between different modes are calculated. For example, in the same time span, the difference changes between temperature and humidity, and the difference changes between humidity and dust concentration are calculated. By analyzing these cross-difference data, the coordinated change relationship between modes can be identified. For example, when the temperature rises rapidly, the humidity decreases, and at the same time, the dust concentration shows an upward trend. The cross-difference results are organized into cross-difference matrix data, which describes the dynamic interaction between multi-modal variables.

[0115] Specifically, for the data in each time window, a difference-based method is used to calculate the intra-modal and inter-modal changes. Intra-modal difference: calculates the difference between adjacent time points of the same mode (such as dust concentration). Inter-modal difference: calculates the correlation difference between different modes (such as the correlation difference between humidity change and dust concentration). The difference calculation uses a weight adjustment model optimized by reinforcement learning to dynamically assign different importance to the inter-modal differences.

[0116] Perform differential trend cluster analysis on the cross-difference matrix data to obtain differential sequence cluster data;

[0117] Specifically, based on the cross-difference matrix data, cluster analysis is performed on the differential change trend within each time span. For example, the time periods with rapid changes in humidity and dust concentration are classified into one category, and the time periods with slow changes in temperature and humidity are classified into another category. This trend clustering can classify the dynamic characteristics of the environment within the time span to form differentiated environmental state groups. The result of the cluster analysis is a number of cluster centers, each of which represents a typical differential change pattern.

[0118] Perform differential feature extraction on differential sequence clustering data to obtain differential sequence feature data;

[0119] Specifically, the results of cluster analysis are further extracted with features, such as the average change value, change rate, and proportion of total time of each differential trend. Features can describe the core characteristics of each differential pattern. For example, the features of a certain cluster center show a rapid decrease in humidity and a significant increase in dust concentration, corresponding to the peak period of stockpiling operations. The extracted features are organized into differential sequence feature data, which contains the key statistical information of the differential pattern.

[0120] The differential sequence feature data is embedded into a differential feature sequence to obtain the first ore stockpile sequence feature data.

[0121] Specifically, the extracted differential feature data is converted into an embedded sequence to describe the dynamic changes of the environmental state. For example, the feature vector of the differential pattern is embedded into a multidimensional sequence to generate a continuous time feature representation. Through this embedding method, the change trajectory of the environment in the stockpile over time can be intuitively described. These embedded sequences are organized into the first ore stockpile sequence feature data.

[0122] Preferably, the pattern sequence feature construction is specifically as follows:

[0123] Perform mutation point detection on multimodal data of ore stockpile to obtain multimodal mutation point data;

[0124] Specifically, in the multimodal data of the ore stockpile, the time variation trends of parameters such as temperature, humidity, and dust concentration are analyzed to detect mutation points. For example, when the humidity suddenly drops below the set threshold at a certain point in time, and the dust concentration rises significantly during the same period, the system marks this moment as a mutation point. The detection results of the mutation point include the timestamp and the specific modal information of the mutation, such as "humidity mutation" and "dust concentration mutation". Multimodal mutation point data are recorded.

[0125] Performing event modal mapping on multimodal mutation point data to obtain event modal mapping data;

[0126] Specifically, the detected multimodal mutation points are mapped to specific event types. For example, when the temperature rises and the humidity decreases, it corresponds to a weather change event; when the dust concentration increases significantly, it corresponds to a stockpiling operation event. The system assigns an event tag to each mutation point and generates event modal mapping data, which describes the relationship between the mutation point and the actual event. For example, the event modal mapping data at a certain moment represents "time point t1: dust concentration mutation-stockpiling operation".

[0127] Perform inter-modal pattern association mining on event modal mapping data to obtain modal association data;

[0128] Specifically, based on the event modal mapping data, the association patterns between different modal events are analyzed. For example, when the stacking operation starts, the dust concentration usually rises in a short period of time, and the frequency of humidity decreases also increases. The system mines the association relationship between these modes and records the frequency and intensity of the mode occurrence. For example, it can be concluded that "the association intensity between the stacking operation event and the dust concentration increase event is high", forming modal association data. The structure of the event modal mapping data is: ,in It is The timestamp of the event. is a sequence item, with a value range of 1, 2, ..., n, It is event modes (e.g. "humidity drops", "dust concentration increases", etc.). Two modal events , The association pattern is defined as: temporal proximity: , The time interval between occurrences does not exceed the threshold (i.e. preset time threshold), modal dependency: event The probability of occurrence depends significantly on the event Traverse the event data and filter the modal event pairs that are associated based on the time attribute. For example: , “humidity drop event”; , “dust concentration increase event”, and , then the modal event pair is recorded as: For each candidate modal event pair , calculate the event Conditional probability of occurrence: , Humidity decreases, with 100 occurrences; The dust concentration increased It happened 60 times in the last 10 minutes. , if the conditional probability If the correlation value is greater than the set correlation threshold (such as 0.5), the event is considered and There is a correlation. By screening modal event pairs with high correlation strength, event modal mapping data is generated, for example: humidity decreases → dust concentration increases (correlation strength: 0.6); temperature increases → humidity decreases (correlation strength: 0.7); each rule records the following information: antecedent event (causal modality); post-event (result modality); correlation strength (conditional probability).

[0129] Reconstructing the mode spectrum of the modal association data to obtain the modal spectrum data;

[0130] Specifically, a pattern graph between modalities is constructed based on the modal association data. The nodes of the pattern graph represent specific modal events, such as "humidity decreases" or "dust concentration increases", while the edges represent the associations between them. The weights of the edges reflect the strength of the event associations. For example, a strong association between certain modal events indicates that they often occur at the same time. The reconstructed pattern graph describes the typical pattern relationships in the ore stockpile environment.

[0131] The modal spectrum data is subjected to a pattern dynamic sequence generation to obtain the second ore stockpile sequence characteristic data.

[0132] Specifically, dynamic sequences are extracted from the pattern map to describe the evolution of modal events over time. For example, the system rearranges events according to timestamps to form a complete dynamic sequence, such as "humidity decreases → dust concentration increases → stockpile height increases". The dynamic sequence not only describes the relationship between events, but also provides their time evolution trajectory. The generated second ore stockpile sequence feature data is presented in the form of a dynamic sequence, providing an important reference for subsequent decision-making and strategy formulation.

[0133] Preferably, the multi-sequence weight graph is constructed as follows:

[0134] Performing multi-sequence feature fusion initialization on the first ore stockpile sequence feature data and the second ore stockpile sequence feature data to obtain initial feature fusion matrix data;

[0135] Specifically, the first ore stockpile sequence feature data (such as differential features) and the second ore stockpile sequence feature data (such as pattern features) are integrated into a matrix to form the initial feature fusion matrix data. For example, the differential feature data describes the dynamic changes of humidity and dust concentration, while the pattern feature data records the typical change pattern of environmental events. Through normalization processing, the numerical ranges of the two sets of feature data are adjusted to be consistent. The feature fusion matrix contains the comprehensive information of each modal feature and can fully reflect the dynamic changes of the stockpile.

[0136] Perform multi-dimensional feature association mapping on the initial feature fusion matrix data to obtain feature association matrix data;

[0137] Specifically, based on the initial feature fusion matrix, the correlation between different features is analyzed. For example, changes in humidity are correlated with changes in dust concentration, and certain high-frequency events in pattern features are closely related to specific dynamic change features. The system maps these correlations into a feature correlation matrix, where each element of the matrix represents the strength of correlation between two features. The correlation mapping reveals the interaction patterns between multimodal feature data.

[0138] Perform sequence difference weight assignment on feature association matrix data to obtain difference weight matrix data;

[0139] Specifically, according to the feature association matrix, the difference of each feature in the entire sequence is calculated, such as whether the change in humidity is significant in different time periods, or whether the frequency of pattern events fluctuates. Features with higher differences are given higher weights, reflecting their significant impact on the dynamics of the stockyard environment. The difference weight matrix records the importance of each feature in the multidimensional feature association.

[0140] The modal weight map is constructed for the difference weight matrix data to obtain the preliminary ore stockpile sequence characteristic data;

[0141] Specifically, the feature nodes and the associations between them are represented as a modal weight graph using a difference weight matrix. The nodes of the weight graph represent different features, such as humidity, dust concentration, and mode events, and the weights of the edges represent the strength of the associations between the features. The weight graph can clearly describe the interrelationships of multimodal features in the ore stockpile environment. The modal weight graph not only contains the association information between features, but also reflects the importance of each feature, providing a basis for further analysis.

[0142] Extracting characteristic subgraphs from the preliminary ore stockpile sequence characteristic data to obtain ore stockpile sequence characteristic data;

[0143] Specifically, feature subgraphs are extracted from the modal weight graph, focusing on analyzing features with higher weights and strong correlations between them. For example, some subgraphs focus on the dynamic changes of humidity and dust concentration, while others reflect the high-frequency correlations of pattern events. Through the extraction and division of subgraphs, the weight graph can be decomposed into several substructures, which describe the dynamics of different aspects of the stockpile environment. The ore stockpile sequence feature data is composed of these subgraphs, which can fully reflect the multimodal environmental characteristics of the stockpile.

[0144] The feature subgraph extraction is as follows:

[0145] Extract characteristic nodes from the preliminary ore stockpile sequence characteristic data and construct time series characteristics to obtain characteristic node graph data and time series characteristic edge data respectively;

[0146] Specifically, in the preliminary ore stockpile sequence feature data, each feature (such as humidity, temperature, and dust concentration) is extracted as a node of the graph, and the node set reflects the main modal features in the stockpile environment. For example, "humidity decreases" or "dust concentration increases" are regarded as feature nodes. At the same time, based on the dynamic changes of these features over time, temporal feature edges between nodes are constructed. The weight of the edge reflects the temporal correlation or causal relationship between two nodes. For example, when the humidity drops significantly, the probability of an increase in dust concentration is higher, and the edge weight between the two is larger. The generated feature node graph data and temporal feature edge data describe the spatial distribution and temporal variation relationship of the ore stockpile characteristics.

[0147] Perform cross-modal weight balancing based on feature node graph data and time series feature edge data to obtain cross-modal fusion matrix data;

[0148] Specifically, the importance of different modal features is weighted and balanced based on the feature node graph data and the time series feature edge data. For example, changes in humidity may have a greater impact on changes in dust concentration in some cases, while in other cases, modal features such as the frequency of stockpiling operations may be more important. The system allocates cross-modal balance factors based on node weights and edge weights to generate cross-modal fusion matrix data. The matrix reflects the global weight distribution of feature nodes and their associations, laying the foundation for further analysis. For each node Calculate its weight , combined with the modal eigenvalues ​​of the nodes and the edge connectivity in the graph : ,in is the weight coefficient of the modal eigenvalue of the node, is the weight coefficient of the edge connectivity in the graph. According to the node weight and edge weight, a cross-modal fusion matrix is ​​constructed. , whose elements are , For the The weight corresponding to each node is For the Line The node corresponding to the column, For the The weight corresponding to each node.

[0149] Perform sub-graph aggregation on the cross-modal fusion matrix data to obtain cross-modal sub-graph data;

[0150] Specifically, based on the cross-modal fusion matrix data, nodes and edges with strong correlation are aggregated to form a subgraph. For example, nodes with high correlation between humidity and dust concentration are aggregated into the same subgraph, while nodes describing the stockpile operation mode are classified into another subgraph. Each subgraph contains a set of nodes and a set of edges, which describe different aspects of the stockpile environment. The subgraph aggregation process decomposes the global graph into several small subgraphs, each of which reflects the local dynamics of a certain type of modal feature.

[0151] Extracting sub-graph patterns from cross-modal sub-graph data to obtain sub-graph pattern data;

[0152] Specifically, common patterns are extracted from cross-modal subgraph data. For example, in some subgraphs, the decrease in humidity and the increase in dust concentration form a high-frequency three-node pattern. By counting the structural features and node connection methods of each subgraph, the system identifies representative subgraph pattern data. The pattern reflects the typical associations in the stockyard environment. For example, the pattern "decreased humidity - increased dust - increased temperature" corresponds to hot weather or the peak period of stockyard operations.

[0153] The graph features of the cross-modal sub-graph data are embedded according to the sub-graph mode data to obtain the sequence feature data of the ore stockpile.

[0154] Specifically, the extracted subgraph pattern data is converted into embedded feature representations to generate global sequence feature data of the ore stockpile. For example, the features of each subgraph pattern are mapped to a multidimensional embedding space to describe its contribution to the overall environmental dynamics. The embedded data can represent the multimodal feature dynamics of the ore stockpile in vector form, supporting further analysis and decision-making. The generated ore stockpile sequence feature data integrates the embedded representations of all subgraphs and comprehensively describes the diversity and changing laws of the stockpile environmental characteristics.

[0155] Preferably, step S2 specifically includes:

[0156] Step S21: starting a consensus voting mechanism for the ore stockpile data to obtain voting status record data;

[0157] Specifically, the monitoring data of the ore stockpile is collected through the sensor network and then encrypted and broadcast to multiple nodes of the blockchain network. After receiving the data, each node decrypts and independently verifies the integrity and logical consistency of the data. For example, verify whether the humidity is within a reasonable range (such as 10%-90%) and whether the dust concentration is within the standard value range under normal working conditions (such as 0-300µg / m³). If the data meets the requirements, the node will vote "accept"; if the data exceeds the threshold or is missing, it will vote "reject". Each node records its voting results as voting status data, including voting time, data source and voting results.

[0158] Step S22: Performing data credibility scoring according to the voting status record data to obtain credibility scoring data;

[0159] Specifically, the blockchain network counts the voting results of each piece of data, including the number of accepted votes and the number of rejected votes. A credibility score is calculated for each piece of data based on the voting results. For example, if most nodes vote "accept", the credibility score of the data is high. The credibility score is used to reflect the reliability of the data, and the scoring results are recorded as a set of credibility score data for subsequent analysis and processing.

[0160] Step S23: isolating abnormal data of the ore stockpile data according to the trustworthy scoring data to obtain abnormal isolation record data;

[0161] Specifically, based on the credibility score data, the system marks and isolates the data with low credibility scores. For example, the credibility score of humidity data is too low because it is an outlier caused by sensor failure. The isolated data is recorded separately as abnormal isolation record data, including the identification information of the abnormal data, the type of abnormality detected, and the reason for isolation. The data is removed from the normal data stream.

[0162] Step S24: Generate blockchain blocks based on the trustworthy score data and the abnormal isolation record data to obtain preliminary ore stockpile block data;

[0163] Specifically, for data that has passed the credibility score verification, the system packages it into a new block. Each block contains the following information: the content of the data that has passed the verification; the credibility score of the data; the isolation record of abnormal data; the timestamp of the current block; and the hash value of the previous block to ensure the integrity and immutability of the blockchain.

[0164] The generated preliminary block stores a complete record of the ore stockpile environmental monitoring data, with credibility scores and exception handling information.

[0165] Step S25: Perform cross-chain verification on the preliminary ore stockpile block data to obtain the ore stockpile block data.

[0166] Specifically, the initially generated blocks are verified across multiple blockchain networks through a cross-chain mechanism. Nodes in other blockchain networks verify the integrity and consistency of the blocks, such as checking whether the data content is consistent with the hash value and whether the block meets the network consensus rules. If the verification is passed, the block is marked as "trusted"; if the verification is not passed, the system will record the reason for the verification failure. The cross-chain verified ore stockpile block data is stored in the blockchain network, ensuring the reliability and transparency of the data.

[0167] Preferably, step S3 specifically includes:

[0168] Step S31: extracting data according to the block data of the ore stockpile to obtain basic data of the ore stockpile;

[0169] Specifically, basic information is extracted from the block data of the ore stockpile stored in the blockchain, including key parameters such as ambient temperature and humidity, dust concentration, stockpile height, timestamp, and processed structured data (such as structured data generated by the processing method provided above). The data is screened and sorted to form a set of basic data of the ore stockpile. For example, the temperature in a certain area is 25°C, the humidity is 60%, the dust concentration is 120µg / m³, and the stockpile height is 8 meters. The data extraction results record the current and historical environmental status of the stockpile.

[0170] Step S32: constructing an ore stockpile model based on the ore stockpile basic data to obtain an ore stockpile model;

[0171] Specifically, based on the extracted basic data, a digital model of the ore stockpile was established through multivariate regression calculation. The model structure reflects the relationship between temperature and humidity, dust concentration and stockpile height. For example, changes in temperature and humidity affect the diffusion of dust concentration, and stockpile height may affect the coverage and effect of spraying. Through the statistics and analysis of historical data, the model dynamically describes the evolution of the stockpile environment under different conditions, laying a scientific basis for further analysis.

[0172] Step S33: performing feature correlation analysis on the ore stockpile model to obtain feature matrix data of the ore stockpile;

[0173] Specifically, in the ore stockpile model, the correlation between each feature is analyzed. For example, there is a negative correlation between humidity and dust concentration, and an increase in humidity leads to a decrease in dust concentration. In addition, there is a positive correlation between stockpile height and spray coverage. Through analysis and induction, the characteristic matrix data of the ore stockpile is generated. The matrix records the correlation strength and direction between each feature, reflecting the overall characteristics of the stockpile environment.

[0174] Step S34: Calculate the spraying demand for the characteristic matrix data of the ore stockpile to obtain the spraying demand data;

[0175] Specifically, the spraying demand of each area is calculated by combining the characteristic matrix data and the current environmental status of the stockpile. For example, if the humidity in a certain area is lower than the target value and the dust concentration is high, the system will calculate the amount of humidity that needs to be supplemented and the spraying intensity and duration required to reduce the dust concentration. At the same time, according to the stockpile height and stockpile coverage, the required spraying water volume and nozzle layout parameters for each area are determined, and finally the spraying demand data is generated.

[0176] Step S35: performing spray control simulation on the ore stockpile model according to the spray demand data to obtain spray control simulation data;

[0177] Specifically, the spray demand data is used to simulate the ore stockpile model to predict the environmental changes after the spray operation. For example, the effects of humidity increase and dust concentration reduction during the spray process are simulated to verify whether the spray demand meets the target value. If the simulation results show that some areas do not achieve the expected effect, the system will adjust the spray parameters (such as increasing the spray duration or intensity) until the simulation results meet the requirements.

[0178] Step S36: Generate a spray strategy based on the spray control simulation data and the spray demand data to obtain spray control strategy data.

[0179] Specifically, a specific spray control strategy is generated based on the results of the spray control simulation. For example, the system recommends spraying the central area of ​​the stockpile at a spray intensity of 60 liters per minute from 10 a.m. to 10:30 a.m., covering all stockpile height layers in the area. The spray strategy also includes detailed parameters such as equipment start and stop time, nozzle position and coverage angle to ensure the accuracy and efficiency of the spray operation. The generated spray control strategy data is sent to the control module of the spray equipment to guide the actual spray operation.

[0180] Preferably, step S4 is specifically:

[0181] Step S41: parsing the equipment control instruction according to the spray control strategy data to obtain the spray equipment control instruction data;

[0182] Specifically, based on the generated sprinkler control strategy data, the system parses specific equipment control instructions. For example, the control strategy includes parameters such as the spraying time period, intensity, coverage, and nozzle position. The system converts these strategies into control instruction data that can be executed by the equipment, such as "start sprinklers A1 and A2 at 9:00 am, with a spraying intensity of 50 liters per minute, covering the height layers 1-3 of area 1, and lasting for 30 minutes". The parsed control instruction data is sent to the control module of the sprinkler equipment.

[0183] Step S42: activating the spray equipment and monitoring the spraying in real time according to the spray equipment control instruction data to obtain real-time monitoring data of the spraying;

[0184] Specifically, after receiving the control command, the spray equipment starts the corresponding nozzle according to the command and starts the spraying operation. For example, in area 1, nozzles A1 and A2 are activated at the same time, and the spraying operation is carried out according to the set spraying intensity and range. During the spraying process, the sensor monitors the environmental changes in real time, such as the increase in humidity, the decrease in dust concentration, and the operating status of the nozzle. The monitoring data is uploaded to the central control system through the wireless network to form real-time monitoring data of the spraying. The data includes information such as timestamp, humidity change curve, dust concentration decrease rate, and equipment operation log.

[0185] Step S43: Analyze the spraying effect of the real-time monitoring data of the spraying to obtain the spraying implementation data for recording and tracing the spraying process.

[0186] Specifically, based on the real-time monitoring data of the spraying, the system analyzes the spraying effect. For example, whether the humidity reaches the target value, whether the dust concentration drops to the standard range, and whether the nozzle operates normally. If the spraying effect is not ideal, such as the humidity does not reach the set value or the dust concentration does not drop enough, the system will record the analysis results and prompt whether it is necessary to adjust the spraying parameters or extend the spraying time. At the same time, the system generates spraying implementation data, which records the results of each spraying in detail, including the humidity increase, dust concentration reduction, spraying coverage and time.

[0187] The data of spraying implementation is stored in the blockchain network to ensure the immutability and traceability of the data. These data contain a complete record of the spraying operation, such as "Area 1, spraying time period: 9:00-9:30, humidity increased from 60% to 75%, and dust concentration decreased from 120µg / m³ to 80µg / m³". Through the traceability function of the blockchain, the spraying records in any time period can be queried to verify whether the spraying operation is carried out as planned.

[0188] Preferably, the present application also provides a blockchain-based ore terminal intelligent spray management system, which is used to execute the blockchain-based ore terminal intelligent spray management method as described above, and the blockchain-based ore terminal intelligent spray management system includes:

[0189] Ore stockpile data acquisition module, used to obtain ore stockpile data;

[0190] The ore stockpile block generation module is used to verify the ore stockpile data through the blockchain consensus mechanism to obtain the ore stockpile block data;

[0191] A spray control strategy generation module is used to generate a spray strategy according to the block data of the ore stockpile yard to obtain spray control strategy data;

[0192] The spray equipment control module is used to control the spray equipment according to the spray control strategy data and obtain the spray implementation data to record and trace the spray process.

[0193] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0194] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A blockchain-based intelligent spray management method for ore terminals, characterized in that: The following steps are involved: Step S1: Acquire ore stockpile data; Step S2: Verify the ore stockpile data through the blockchain consensus mechanism to obtain the ore stockpile block data; Step S3: generating a spraying strategy according to the block data of the ore stockpile yard to obtain spraying control strategy data; Step S4: Control the spraying equipment according to the spraying control strategy data to obtain the spraying implementation data for spraying process recording and tracing operations; The specific step S1 is: Step S11: deploying an ore stockpile monitoring sensor network to cover key areas of the ore stockpile to monitor temperature, humidity, and dust concentration, thereby obtaining real-time collected data; Step S12: performing edge multimodal fusion on the real-time collected data to obtain multimodal data of the ore stockpile yard; Step S13: constructing sequence features for the multimodal data of the ore stockpile to obtain sequence feature data of the ore stockpile; Step S14: encrypting the ore stockpile sequence characteristic data and uploading the data to obtain the ore stockpile data; Wherein step S13 is specifically as follows: Constructing differential sequence features of the multimodal data of the ore stockpile to obtain sequence feature data of the first ore stockpile; Constructing pattern sequence features of the multimodal data of the ore stockpile to obtain sequence feature data of the second ore stockpile; Constructing a multi-sequence weighted graph of the first ore stockpile sequence characteristic data and the second ore stockpile sequence characteristic data to obtain the ore stockpile sequence characteristic data; The differential sequence feature construction is specifically as follows: Dynamically divide the span of multimodal data of ore stockpile to obtain divided sequence data; Perform multimodal cross-difference calculation on the divided sequence data to obtain cross-difference matrix data; Perform differential trend cluster analysis on the cross-difference matrix data to obtain differential sequence cluster data; Perform differential feature extraction on differential sequence clustering data to obtain differential sequence feature data; The differential sequence feature data is embedded into a differential feature sequence to obtain the first ore stockpile sequence feature data.

2. The method according to claim 1, characterized in that The specific construction of pattern sequence features is as follows: Perform mutation point detection on multimodal data of ore stockpile to obtain multimodal mutation point data; Performing event modal mapping on multimodal mutation point data to obtain event modal mapping data; Perform inter-modal pattern association mining on event modal mapping data to obtain modal association data; Reconstructing the mode spectrum of the modal association data to obtain the modal spectrum data; The modal spectrum data is subjected to a pattern dynamic sequence generation to obtain the second ore stockpile sequence characteristic data.

3. The method according to claim 1, characterized in that The construction of the multi-sequence weight graph is as follows: Performing multi-sequence feature fusion initialization on the first ore stockpile sequence feature data and the second ore stockpile sequence feature data to obtain initial feature fusion matrix data; Perform multi-dimensional feature association mapping on the initial feature fusion matrix data to obtain feature association matrix data; Perform sequence difference weight assignment on feature association matrix data to obtain difference weight matrix data; The modal weight map is constructed for the difference weight matrix data to obtain the preliminary ore stockpile sequence characteristic data; Extracting characteristic subgraphs from the preliminary ore stockpile sequence characteristic data to obtain ore stockpile sequence characteristic data; The feature subgraph extraction is as follows: Extract characteristic nodes from the preliminary ore stockpile sequence characteristic data and construct time series characteristics to obtain characteristic node graph data and time series characteristic edge data respectively; Perform cross-modal weight balancing based on feature node graph data and time series feature edge data to obtain cross-modal fusion matrix data; Perform sub-graph aggregation on the cross-modal fusion matrix data to obtain cross-modal sub-graph data; Extracting sub-graph patterns from cross-modal sub-graph data to obtain sub-graph pattern data; The graph features of the cross-modal sub-graph data are embedded according to the sub-graph mode data to obtain the sequence feature data of the ore stockpile.

4. The method according to claim 1, characterized in that Step S2 is specifically as follows: Step S21: starting a consensus voting mechanism for the ore stockpile data to obtain voting status record data; Step S22: Performing data credibility scoring according to the voting status record data to obtain credibility scoring data; Step S23: isolating abnormal data of the ore stockpile data according to the trustworthy scoring data to obtain abnormal isolation record data; Step S24: Generate blockchain blocks based on the trustworthy score data and the abnormal isolation record data to obtain preliminary ore stockpile block data; Step S25: Perform cross-chain verification on the preliminary ore stockpile block data to obtain the ore stockpile block data.

5. The method according to claim 1, characterized in that Step S3 is specifically as follows: Step S31: extracting data according to the block data of the ore stockpile to obtain basic data of the ore stockpile; Step S32: constructing an ore stockpile model based on the ore stockpile basic data to obtain an ore stockpile model; Step S33: performing feature correlation analysis on the ore stockpile model to obtain feature matrix data of the ore stockpile; Step S34: Calculate the spraying demand for the characteristic matrix data of the ore stockpile to obtain the spraying demand data; Step S35: performing spray control simulation on the ore stockpile model according to the spray demand data to obtain spray control simulation data; Step S36: Generate a spray strategy based on the spray control simulation data and the spray demand data to obtain spray control strategy data.

6. The method according to claim 1, characterized in that Step S4 is specifically as follows: Step S41: parsing the equipment control instruction according to the spray control strategy data to obtain the spray equipment control instruction data; Step S42: activating the spray equipment and monitoring the spraying in real time according to the spray equipment control instruction data to obtain real-time monitoring data of the spraying; Step S43: Analyze the spraying effect of the real-time monitoring data of the spraying to obtain the spraying implementation data for recording and tracing the spraying process.

7. A blockchain-based intelligent spray management system for ore terminals, characterized in that: Used to execute the blockchain-based ore terminal intelligent spray management method as claimed in claim 1, the blockchain-based ore terminal intelligent spray management system includes: Ore stockpile data acquisition module, used to obtain ore stockpile data; The ore stockpile block generation module is used to verify the ore stockpile data through the blockchain consensus mechanism to obtain the ore stockpile block data; A spray control strategy generation module is used to generate a spray strategy according to the block data of the ore stockpile yard to obtain spray control strategy data; The spray equipment control module is used to control the spray equipment according to the spray control strategy data and obtain the spray implementation data to record and trace the spray process.

Images