Analysis and decision-making method, device and equipment based on coal quality data and medium

Abstract

The invention relates to the technical field of data processing, in particular to an analysis decision-making method, device and equipment based on coal quality data and a medium. According to the method, the coal quality time sequence knowledge graph fused with the time attribute is constructed, so that traditionally isolated test, production and purchase data are systematically integrated, and the fundamental transformation of coal quality management from scattered information to associated knowledge is realized. On the basis, a stable or fluctuating coal quality group is automatically identified through clustering analysis, and an influence path of process parameters on quality is accurately positioned by utilizing association mining, so that a direct basis is provided for process optimization. Meanwhile, the system can carry out dynamic early warning and intelligent matching recommendation based on the atlas, the core targets of improving the coal quality stability, optimizing the production cost and assisting scientific decision making are finally achieved, and the comprehensive benefits of coal production and utilization are comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to an analysis and decision-making method, apparatus, equipment, and medium based on coal quality data. Background Technology

[0002] The coal industry has accumulated massive amounts of data over its long-term production, quality inspection, and trading processes. This includes coal quality test data (calorific value, sulfur content, ash content, volatile matter, etc.) from various coal mines and batches, production process data, and equipment status data. This data is not only voluminous but also highly dimensional, with complex interrelationships and strong spatiotemporal attributes. Currently, the utilization of this data is significantly insufficient. Quality data, production data, and procurement data belong to different systems, lacking effective integration and failing to be analyzed from a holistic perspective, resulting in data silos. Summary of the Invention

[0003] This invention provides a method, apparatus, equipment, and medium for analysis and decision-making based on coal quality data, in order to solve the problem of data silos caused by the lack of correlation and integration of coal quality data in the prior art.

[0004] In a first aspect, the present invention provides an analysis and decision-making method based on coal quality data. The method includes: acquiring coal quality data from different sources, including testing systems, production systems, and procurement systems; constructing a coal quality time-series knowledge graph with coal batches as core nodes and incorporating time attributes based on the coal quality data; clustering the coal quality time-series knowledge graph to obtain coal quality groups with different patterns; performing association mining on the coal quality groups and obtaining process parameter optimization decisions based on the mining results; and making early warning and multi-source coal blending recommendation decisions based on the coal quality groups.

[0005] This invention constructs a coal quality time-series knowledge graph that integrates time attributes, systematically consolidating traditionally isolated testing, production, and procurement data. This achieves a fundamental shift in coal quality management from fragmented information to interconnected knowledge. Based on this, cluster analysis automatically identifies stable or fluctuating coal quality groups, and association mining precisely pinpoints the impact paths of process parameters on quality, providing direct evidence for process optimization. Simultaneously, the system can provide dynamic early warnings and intelligent blending recommendations based on the graph, ultimately achieving the core objectives of improving coal quality stability, optimizing production costs, and supporting scientific decision-making, thereby comprehensively enhancing the overall benefits of coal production and utilization.

[0006] In one optional implementation, a coal quality time-series knowledge graph is constructed based on the coal quality data, with coal batches as the core nodes and incorporating time attributes. This includes: constructing node types based on the coal quality data, including coal mine nodes, batch nodes, testing indicator nodes, production equipment nodes, and production process nodes; constructing edge relationship types based on the node types, such as batch affiliation to a coal mine, batch indicators, indicator influence indicators, equipment output batches, and process parameter influence indicators; and assigning valid timestamps to the node types and edge relationship types to construct the coal quality time-series knowledge graph with coal batches as the core nodes and incorporating time attributes.

[0007] This invention constructs a time-series knowledge graph centered on coal batches, systematically integrating heterogeneous data from multiple sources, including coal mines, equipment, processes, and testing indicators, and introducing effective timestamps to record dynamic changes. This enables enterprises to shift from static, isolated data records to dynamic, networked correlation analysis. It not only comprehensively traces the spatiotemporal evolution of quality problems and accurately identifies the root causes of processes or equipment affecting coal quality, but also provides quality early warnings and optimized coal blending based on historical time-series patterns. This significantly improves the precision and intelligence of coal quality control, providing strong data-driven support for production and procurement decisions.

[0008] In an optional implementation, before constructing a coal quality time-series knowledge graph with coal batches as the core node and incorporating time attributes, the method further includes: acquiring data of each indicator and each process parameter in the coal quality data using a sliding time window to obtain the time series of the data; determining the correlation coefficient between the time series of two indicators or indicators and process parameters with edge relationships at different lag times; taking the lag time corresponding to the correlation coefficient with the largest absolute value as the influence lag time of the corresponding edge, and taking the corresponding correlation coefficient as the weight of the edge, wherein the influence lag time and weight are updated based on the sliding time window.

[0009] In this invention, the weights and lag times of the correlation edges between indicators and between indicators and process parameters in the knowledge graph are calculated and dynamically updated through a sliding time window. This allows for the precise quantification and characterization of the strength and time delay of causal interactions between different elements. This transforms the constructed time-series knowledge graph from a simple accumulation of static relationships into a precise model that truly reflects the dynamic evolution of coal quality indicators with process and time. This provides a reliable data foundation for subsequent in-depth analyses such as anomaly tracing and trend prediction, greatly improving the accuracy and practicality of correlation analysis.

[0010] In one optional implementation, the coal quality time-series knowledge graph is clustered to obtain coal quality groups with different patterns, including: using a dynamic community discovery method to cluster the coal quality time-series knowledge graph to obtain coal quality groups with different patterns, wherein the batch groups in the coal quality groups come from the same vein or have undergone the same production process.

[0011] In this invention, a dynamic community discovery algorithm is used to intelligently group coal quality time-series knowledge graphs, automatically identifying batches with similar temporal evolution patterns. These groups typically correspond to the same ore vein origin or have undergone the same production process, thus effectively achieving the classification and traceability of complex coal quality data. This method enables researchers and managers to more clearly understand the characteristics, origins, and variation patterns of different coal qualities, providing precise data support for subsequent coal quality traceability, production process optimization, and rational resource allocation, significantly improving the intelligence level of related research, production, and management in the coal industry.

[0012] In one optional implementation, the coal quality group is subjected to association mining, and process parameter optimization decisions are obtained based on the mining results. This includes: based on the coal quality group, using a time-series association path mining method to traverse the edges in the coal quality time-series knowledge graph where the timestamps satisfy causal relationships, and determining the critical path based on weights and impact lag times; and analyzing the process parameters in the critical path to obtain process parameter optimization suggestions.

[0013] This invention utilizes temporal correlation path mining technology to automatically traverse and filter key causal paths connecting changes in quality groups and production parameters. This not only enables accurate identification of hidden patterns in massive, dynamic industrial data, but more importantly, it pinpoints core process steps that significantly impact final coal quality and exhibit clear lag. Based on this, direct and specific decision-making support can be provided for production optimization, such as clearly indicating the adjustment of a specific parameter of particular equipment and its optimal timing. This drives the shift from traditional experience-driven to data-driven production processes, effectively improving process stability, efficiency, and product quality.

[0014] In one optional implementation, early warning and multi-source coal blending recommendation decisions based on coal quality groups include: predicting the index values ​​of future times in the coal quality time-series knowledge graph using a pre-trained temporal graph neural network model based on the coal quality groups, and determining whether to issue an early warning based on the prediction results; constructing a blending model based on the coal quality groups by minimizing the total cost corresponding to the blending ratio of different candidate coal batches and the index constraints; and solving the blending model to obtain multi-source blending recommendation results.

[0015] This invention elevates coal quality management from a passive response to a proactive optimization by integrating early warning and blending decision-making. It utilizes a time-series graph neural network to dynamically predict knowledge graphs, enabling early warning of potential quality problems and effectively reducing production risks. Simultaneously, a blending optimization model is constructed with cost minimization as the objective, automatically calculating the optimal multi-source coal blending scheme while strictly meeting various coal quality constraints. This not only ensures stable coal quality and processes but also significantly reduces raw material procurement and production costs through refined blending, achieving synergistic optimization of quality control and economic benefits.

[0016] In an optional implementation, the method further includes: storing the coal quality time-series knowledge graph and displaying it through a visual interface; and querying the visualization interface using a preset query language and preset conditions.

[0017] This invention transforms complex, multi-dimensional coal quality data, process parameters, and their dynamic relationships into an intuitive and visual network view by presenting a coal quality time-series knowledge graph in a graphical interface and supporting interactive queries. This allows data analysts to quickly understand the inherent connections and evolution trends between data without requiring specialized technical backgrounds. They can also directly retrieve and locate the required information using preset query conditions (such as time range and key indicators), greatly improving data retrieval efficiency and analytical insight, lowering the decision-making threshold, and providing a convenient and efficient visual analysis tool for coal quality management and process optimization.

[0018] Secondly, the present invention provides an analysis and decision-making device based on coal quality data. The device includes: a data acquisition module for acquiring coal quality data from different sources, including sampling systems, sample preparation systems, testing systems, production systems, and procurement systems; a knowledge graph construction module for constructing a coal quality time-series knowledge graph with coal batches as core nodes and incorporating time attributes based on the coal quality data; a clustering module for clustering the coal quality time-series knowledge graph to obtain coal quality groups with different patterns; a first decision module for performing association mining on the coal quality groups and obtaining process parameter optimization decisions based on the mining results; and a second decision module for making early warning and multi-source coal ratio recommendation decisions based on the coal quality groups.

[0019] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the analysis and decision-making method based on coal quality data as described in the first aspect or any corresponding embodiment.

[0020] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the analysis and decision-making method based on coal quality data according to the first aspect or any corresponding embodiment thereof.

[0021] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the analysis and decision-making method based on coal quality data as described in the first aspect or any corresponding embodiment thereof. Attached Figure Description

[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating the analysis and decision-making method based on coal quality data according to an embodiment of the present invention; Figure 2 This is a structural block diagram of an analysis and decision-making device based on coal quality data according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0024] As described in the background section, coal quality data from different sources lacks effective correlation and integration, making it impossible to analyze from a holistic perspective. Specifically, existing analyses are mostly limited to querying and statistically analyzing single reports, batches, or indicators, lacking the ability to uncover complex relationships across multiple indicators, batches, and time periods. For example, they cannot answer questions such as "Will the increase in ash content in Mine A lead to a continuous decrease in its calorific value?" or "What changes in process parameters are the main reasons for the recent abnormal volatile matter content in batch B coal?" Furthermore, existing decision-making relies heavily on historical experience and static rules, failing to utilize deep correlations between data for predictive analysis and optimization decisions, such as dynamically providing the optimal coal blending scheme with the lowest cost.

[0025] Knowledge graph technology, with its powerful semantic expression and associative reasoning capabilities, has been widely applied in fields such as healthcare and finance. Time series analysis technology excels at processing timestamped data. Therefore, this embodiment combines the two to construct a coal quality time series knowledge graph that deeply integrates the time dimension. Based on this, a complete set of association analysis and decision support methods is developed to break down data silos, deeply mine the spatiotemporal correlation patterns hidden behind massive amounts of coal quality data, and transform them into executable intelligent decisions, comprehensively improving the refined management and intelligent level of coal enterprises.

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0028] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0029] According to an embodiment of the present invention, an embodiment of an analysis and decision-making method based on coal quality data is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] This embodiment provides an analysis and decision-making method based on coal quality data. Figure 1 This is a flowchart of an analysis and decision-making method based on coal quality data according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain coal quality data from different sources, including testing systems, production systems, and procurement systems.

[0031] Specifically, this embodiment acquires coal quality data from various sources, such as testing systems, production systems, and procurement systems, resulting in multi-source heterogeneous data. The production system includes sampling, sample preparation, and crushing processes, and the data acquired from it includes batch, equipment, and process data. The testing system includes coal analysis to obtain coal performance data. This performance data includes various physical and chemical indicators that measure coal quality, such as, but not limited to, calorific value (heat value), sulfur content, ash content, volatile matter, moisture, fixed carbon, and Hardgrove Grindability Index. The ERP system (Enterprise Resource Planning) includes procurement processes, through which procurement information can be obtained.

[0032] Step S102: Construct a coal quality time-series knowledge graph based on the coal quality data, with coal batches as the core nodes and incorporating time attributes. Specifically, after acquiring the coal quality data, the coal quality time-series knowledge graph can be constructed through steps such as entity recognition, relation extraction, and timestamp alignment. Entity recognition refers to identifying entity data in the coal quality data to obtain nodes such as batches, testing indicators, equipment, and processes. Relationship extraction refers to analyzing whether there are any relationships between the identified entities and forming edges between nodes with relationships. Simultaneously, time attributes are assigned to the acquired nodes and edges; for example, a batch node has its production date. It should also be noted that the coal quality time-series knowledge graph constructed in this embodiment uses coal batches as the core nodes, that is, it extends outward from the coal batch as the core to construct relationships between any batch and nodes such as indicators, equipment, and processes.

[0033] Step S103 involves clustering the coal quality time-series knowledge graph to obtain coal quality groups with different patterns. These coal quality groups include fluctuating batch groups and stable batch groups. Specifically, since the coal quality time-series knowledge graph is batch-based, clustering the knowledge graph achieves the clustering of batches and their associated nodes. Each clustered coal quality group includes one or more batches and their corresponding associated nodes. Furthermore, this clustering process is based on the characteristics of the associated nodes of each batch, grouping batches with the same or identical associated node characteristics into one group. This enables the identification of coal quality groups with similar spatiotemporal evolution patterns; for example, a particular group suggests that it may originate from the same ore vein or have undergone the same production process.

[0034] Furthermore, for different coal quality groups, the group can be classified as a fluctuating or stable batch group based on the characteristics of the associated nodes of batches within the group. Other types of batch groups can also be categorized, such as ordinary groups. A fluctuating batch group indicates that the associated nodes of batches within the group exhibit significant fluctuations, such as large fluctuations in batch-related indicators over time. A stable batch group indicates that the associated nodes of batches within the group exhibit relatively small fluctuations, while the fluctuation of an ordinary group falls somewhere in between.

[0035] Step S104 involves performing correlation mining on the coal quality groups and obtaining process parameter optimization decisions based on the mining results. Specifically, correlation mining algorithms can be used to analyze the node characteristics of different coal quality groups, such as fluctuating batch groups and stable batch groups, to determine how to optimize their process parameters and reduce their volatility.

[0036] Step S105: Make a multi-source coal blending recommendation decision based on coal quality groups. Specifically, for fluctuating batch groups, due to their large fluctuations, they are designated as key monitoring targets. That is, the node data of fluctuating batch groups can be analyzed to determine whether to issue an early warning. In addition, when blending multi-source coal, stable batch groups can be given priority, such as increasing their proportion.

[0037] This embodiment provides an analysis and decision-making method based on coal quality data, which includes the following steps: Step S201: Obtain coal quality data from different sources, including testing systems, production systems, and procurement systems; for details, please refer to [link to relevant documentation]. Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0038] Step S202: Based on the coal quality data, construct a coal quality time-series knowledge graph with coal batches as the core nodes and integrating time attributes.

[0039] Specifically, step S202 includes: Step S2021: Based on the coal quality data, construct node types including coal mine nodes, batch nodes, laboratory indicator nodes, production equipment nodes, and production process nodes. The attributes of coal mine nodes include mine name, geographical location, and coal seam attributes; the attributes of batch nodes include batch number, production date, and sampling date; the attributes of laboratory indicator nodes include indicator name, indicator value, and testing date; the attributes of production equipment nodes include equipment ID and status; and the attributes of production process nodes include process parameter name and parameter value. It should be noted that in practical applications, the relevant attribute data of coal mine nodes, batch nodes, laboratory indicator nodes, production equipment nodes, and production process nodes can be obtained as the data for the corresponding nodes.

[0040] Step S2022: Based on the node types, construct edge relationship types for batch-belonging coal mines, batch-possessing indicators, indicators influencing indicators, equipment-output batches, and process parameter-influenced indicators. Specifically, after determining the node types, based on the batch nodes, obtain the associated nodes of the batch nodes, such as coal mines, indicators, and equipment, thereby forming equilateral relationships for batch-belonging coal mines, batch-possessing indicators, and equipment-output batches. Simultaneously, for these associated nodes, it is necessary to further determine their associated nodes. For example, if a batch possesses a certain indicator, then determine which other indicators this indicator is related to, or which process parameters this indicator is related to, thereby forming equilateral relationships for indicators influencing indicators and process parameter-influenced indicators.

[0041] Step S2023: Use a sliding time window to obtain the data of each indicator and each process parameter in the coal quality data to obtain the time series of the data; specifically, in the above-mentioned edge relationships, the edge relationships of batch belonging to coal mine, batch having indicators, and equipment producing batch are fixed relationships, but the edge relationships of indicator affecting indicator and process parameter affecting indicator need to measure the degree of their influence, that is, the dynamic correlation between the two nodes in the edge relationship.

[0042] In this embodiment, the relevant data of indicators and process parameters are first obtained from the coal quality data. It should be noted that each node in the coal quality time-series knowledge graph constructed above includes a time attribute. That is, for any node, it may include data obtained at multiple times. For example, for a certain indicator of a certain batch node, it may include indicator values ​​obtained on different test dates. These indicator values ​​are all stored in the knowledge graph. Therefore, when using a sliding time window to obtain the data of indicators and process parameters, the time series of the data can be obtained, such as the data on the change of the indicator value of a certain indicator over time in the past 90 days.

[0043] Step S2024: The correlation coefficient between two indicators with a side relationship, or between an indicator and a process parameter at different lag times.

[0044] In this context, lag time represents the time difference between the impact of a change in one indicator on another (or the impact of process parameters on another indicator). For example, a change in crushing process parameters (particle size) today might affect the calorific value of coal one day later, in which case the lag time is 1 day. After obtaining the time series of data through a sliding time window, the correlation coefficient between two nodes with an edge relationship can be calculated within the window. For example, the correlation coefficient (such as the Pearson correlation coefficient) between the sequences of indicator A and indicator B at different lag times (e.g., 0, 1, 2, ... days) can be calculated. Specifically, the time series of indicator A is aligned with the time series of indicator B starting at different lag points, and the correlation coefficient is calculated.

[0045] Step S2025 involves using the lag time corresponding to the correlation coefficient with the largest absolute value as the influence lag time of the corresponding edge, and using the corresponding correlation coefficient as the weight of the edge. The influence lag time and weight are updated based on a sliding time window. Specifically, after calculating the correlation coefficient, this embodiment selects the lag time corresponding to the correlation coefficient with the largest absolute value as the current influence lag time between the two indicators, and uses this coefficient as the weight of the edge. The above process can be repeated periodically (e.g., daily) using the latest sliding window data to achieve dynamic updates of the edge weights and lag times.

[0046] Step S2026: Assign valid timestamps to the node types and edge relationship types to construct a coal quality time-series knowledge graph with coal batches as the core nodes and incorporating time attributes. The valid timestamps include the effective time and the expiration time. After constructing the node types and edge types in the above manner and calculating the dynamic correlation of edge relationships, the coal quality time-series knowledge graph with coal batches as the core nodes and incorporating time attributes is constructed based on these node types and edge relationship types.

[0047] Step S203: Cluster the coal quality time series knowledge graph to obtain coal quality groups with different patterns, including fluctuating batch groups and stable batch groups.

[0048] Specifically, step S203 includes: Step S2031: The coal quality time-series knowledge graph is clustered using a dynamic community discovery method to obtain coal quality groups with different patterns. The batches in the coal quality groups come from the same vein or have undergone the same production process.

[0049] Specifically, the dynamic community detection method used in this embodiment includes algorithms such as the dynamic Louvain algorithm and temporal hierarchical clustering. During clustering, algorithms can be used to process the constructed coal quality temporal knowledge graph or subgraphs within the graph arranged according to time slices. Simultaneously, the algorithm considers the temporal similarity of node attributes (indicator values) and the topological structure of the graph (connections between batches through indicators, processes, etc.) for clustering. This clustering process can identify batch groups whose indicator change patterns are highly consistent within a specific time period. These groups may indicate that they originate from the same high-quality coal seam or have undergone the same optimized set of production process parameters.

[0050] Step S204: Perform correlation mining on the coal quality groups and obtain process parameter optimization decisions based on the mining results.

[0051] Specifically, step S204 includes: Step S2041: Based on the coal quality group, the temporal correlation path mining method is used to traverse the edges in the coal quality temporal knowledge graph where the timestamps satisfy the causal relationship, and the critical path is determined based on the weight and the impact lag time. Specifically, for different coal quality groups, the temporal correlation path mining method can also be used to mine the correlation paths to find the commonalities, thereby providing a data foundation for subsequent process optimization.

[0052] This temporal correlation path mining method can employ path search algorithms in graph theory, such as Dijkstra's algorithm. In algorithmic mining, the effectiveness of a path depends not only on topological connectivity but also on the continuity of temporal attributes between nodes in the edge relationships (i.e., the time of the cause node must precede the time of the result node, and the difference must be within a reasonable range). Simultaneously, edge weights (influence strength), time delays, and other attributes are considered to evaluate the importance of the path.

[0053] Specifically, when using the time-series correlation path mining method, staff can analyze the abnormal starting points of a given fluctuating batch group. Alternatively, any point within the fluctuating batch group can be directly selected as the starting point. During mining, starting from the starting point, the algorithm traverses the time-series graph along allowed relational edges (such as "indicator influencing indicator" or "process parameter influencing indicator") in a reverse (or forward) direction, selecting only edges whose timestamps conform to a causal relationship (cause preceding effect). During the traversal, the algorithm records the paths reaching various nodes and their cumulative weights (or total time delays). Finally, it outputs several paths (i.e., critical paths) that meet the conditions (the endpoint is usually a process parameter node) and have the highest scores (such as the largest cumulative weight or the smallest total time delay) as the most likely source tracing paths.

[0054] In addition, the same processing can be applied to stable batch groups using this time-series correlation path mining method, and the groups mined from the stable batch groups can be used as a reference for subsequent process parameter optimization.

[0055] Step S2042: Based on the process parameters in the critical path, analyze the process parameters to obtain process parameter optimization suggestions. Specifically, when analyzing the critical path, analyze several paths mined from the fluctuating batch group to determine their commonalities, such as the existence of a process parameter with "crushing particle size > 8mm". Then, the crushing particle size process parameter can be optimized, referring to the process parameter settings in the stable batch group. For example, if the critical path analysis in the stable batch group reveals a process parameter with "crushing particle size < 5mm", the proposed process parameter optimization is to standardize the crusher particle size and strictly control it to < 5mm. When analyzing the mined critical path, statistical analysis such as regression analysis or machine learning analysis such as decision tree analysis can be used. This embodiment does not specifically limit the specific analysis method.

[0056] Step S205: Make early warning and multi-source coal ratio recommendation decisions based on coal quality groups.

[0057] Specifically, step S205 includes: Step S2051: Based on the coal quality group, a pre-trained temporal graph neural network model is used to predict the indicator values ​​of future times in the coal quality time-series knowledge graph, and a warning is issued based on the prediction results. Specifically, after clustering to obtain a fluctuating batch group, this group can be used as a key monitoring target. In this embodiment, a pre-trained temporal graph neural network (T-GNN) model is used to predict the indicator values ​​of the fluctuating batch group in the coal quality time-series knowledge graph. If the predicted value exceeds the threshold, the problem can be detected in advance.

[0058] It should be noted that the time-series graph neural network model in this embodiment can predict not only the indicator values ​​within fluctuating batches, but also the indicator values ​​of other batches in the coal quality time-series knowledge graph. For example, the time-series graph neural network model can predict the trend of a certain indicator for a future batch of a certain coal mine by aggregating the historical characteristics of a node itself and the information of its neighboring nodes (such as other batches belonging to the same coal mine). Furthermore, to achieve focused monitoring of fluctuating batches, the prediction frequency for fluctuating batches can be higher than that for other batches. For example, the time-series graph neural network model can predict the indicator values ​​of fluctuating batches once a day, while predicting the indicator values ​​of other batches every preset number of days, such as five days.

[0059] For this time-series graph neural network model, training is required before prediction. Training data comes from snapshots of a historically constructed coal quality time-series knowledge graph. For example, using graph data at time points [t0, t1, ..., t_n] as input, the model predicts the node index value at time point [t_{n+1}], and the predicted value is compared with the true value to calculate the loss. The training data includes both normal and abnormal historical periods, enabling the model to learn different patterns. The loss function is typically mean squared error (MSE), and the optimizer (such as Adam) adjusts the model parameters through backpropagation until the model's prediction accuracy reaches the required level.

[0060] Step S2052: Based on coal quality groups, a blending model is constructed to minimize the total cost corresponding to the blending ratios of different candidate coal batches and to meet the indicator constraints. Specifically, the total cost minimization can be expressed as MinΣ(xi×unit price i), where xi represents the blending ratio of the i-th candidate coal batch, and unit price i represents the unit price of the i-th candidate coal batch. When determining the total cost through the blending ratios of different candidate coal batches, it is also necessary to identify which coal quality groups these batches belong to and assign them a certain risk coefficient. For example, the risk coefficient for fluctuating batch groups is higher, while the risk coefficient for stable batch groups is lower. Thus, an objective function for minimizing the total cost is constructed based on the blending ratio, unit price, and risk coefficient. It should be noted that the unit price can be the procurement cost, or it can include transportation and storage costs, etc.

[0061] Simultaneously, the objective function needs to satisfy certain constraints, such as a total proportion of 100%, setting upper and lower limits for certain indicators, and the maximum usable quantity for certain batches. Specific constraints can be determined based on actual circumstances. The constructed objective function and corresponding constraints constitute the proportioning model.

[0062] Step S2053: Solve the proportioning model to obtain multi-source proportioning recommendations. Specifically, during the solution process, optimization solvers or related algorithms from relevant technologies can be used. This embodiment does not specify the specific solution process. By solving the proportioning model, the proportioning results for different coal batches can be obtained.

[0063] Step S206: Store the coal quality time-series knowledge graph and display it through a visual interface; query the graph using a preset query language and preset conditions. Specifically, the constructed coal quality time-series knowledge graph can be stored in a graph database that supports time-series attributes (such as Nebula Graph with Time-to-Live). A dedicated query language is also designed to support time-series graph queries such as "Query all indicators with a correlation coefficient greater than 0.7 and a lag time of 1 day during Q3 of 2023 that are related to calorific value indicators."

[0064] Furthermore, this coal quality time-series knowledge graph can be displayed and interacted with through a visual interface. Specifically, the visual interface intuitively displays the complex relationships between coal mines, batches, and indicators in the form of a graph, and dynamically displays its evolution history through a time slider. It supports interactive clicking, exploration, and mining by users, forming a closed loop of "visual analysis - hypothesis testing - decision making." Among them, "hypothesis testing" means that users can verify their business conjectures through the system's interactive functions. For example, if a user hypothesizes that "the coal quality of mine A has declined because of the use of equipment B," they can focus on viewing the nodes and paths related to mine A and equipment B on the visual interface, and use the actual relationships and data in the graph to confirm or refute this hypothesis.

[0065] The present invention has the following advantages: Cognitive Depth: Moving from “single-point” statistics to “networked” relational cognition, revealing complex spatiotemporal relational patterns that traditional methods could not discover.

[0066] Forward-looking decision-making: It has realized the transformation from post-event reporting to pre-event prediction and in-event optimization, such as early warning of coal quality trends and recommendation of optimal coal blending schemes, which has created significant economic benefits for enterprises.

[0067] Precise source tracing: It can quickly and accurately pinpoint the root cause of quality problems, whether they stem from changes in coal mine strata or fluctuations in production processes, greatly improving the efficiency of problem solving.

[0068] Knowledge accumulation: Expert experience and data patterns are accumulated in the form of a computable knowledge graph, forming the "intelligent brain" of the enterprise and reducing the reliance on the experience of individual experts.

[0069] As one or more specific application embodiments of the present invention, taking a power plant as an example, the analysis and decision-making method based on coal quality data is described as follows: A power plant has found that the calorific value of coal delivered to the plant has been fluctuating significantly recently, putting pressure on cost control.

[0070] 1. Knowledge Graph Query. Operators can query the calorific value index nodes and their associated information for all recent batches of coal delivered to the plant through the system's visual interface, thereby obtaining the coal quality time-series knowledge graph constructed for the corresponding time period.

[0071] 2. Dynamic Community Discovery. A dynamic community discovery method is used to cluster the queried coal quality time-series knowledge graph, automatically identifying two distinct communities (groups). Group 1 has a stable high calorific value, while Group 2 has a large fluctuation in calorific value and a lower mean.

[0072] 3. Association Mining. Using time-series association path mining, analysis of the two groups revealed that batches in group 1 primarily originated from coal mine X and were associated with the process parameter "crushed particle size < 5mm"; batches in group 2 originated from multiple coal mines including Y and Z and were associated with the process parameter "crushed particle size > 8mm". The edge in the graph showing "crushed particle size -> calorific value" indicates a strong negative correlation between particle size and calorific value with a one-day lag.

[0073] 4. Intelligent Decision-Making. Based on the above analysis, decision-making suggestions are provided: 4.1 Warning: There is a risk that the coal in coal mines Y and Z may have a low calorific value utilization rate due to substandard crushing particle size.

[0074] 4.2 Recommended Coal Blending. It is recommended to purchase more coal from Mine X and blend it with coal from Mine Y at a ratio of 6:4. This can reduce costs by X yuan / ton while still meeting calorific value requirements.

[0075] 4.3 Process optimization. It is recommended that the sample preparation plant standardize the particle size of the crusher and strictly control it to <5mm.

[0076] In this embodiment, the power plant adopted the coal blending recommendations, which improved the stability of the calorific value of the coal delivered to the plant and effectively reduced the overall fuel cost.

[0077] This embodiment also provides an analysis and decision-making device based on coal quality data. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0078] This embodiment provides an analysis and decision-making device based on coal quality data, such as... Figure 2 As shown, it includes: The data acquisition module 21 is used to acquire coal quality data based on different sources, including sampling systems, sample preparation systems, testing systems, production systems, and procurement systems. Knowledge graph construction module 22 is used to construct a coal quality time-series knowledge graph based on the coal quality data, with coal batches as the core nodes and integrating time attributes; Clustering module 23 is used to cluster the coal quality time-series knowledge graph to obtain coal quality groups with different patterns; The first decision module 24 is used to perform correlation mining on the coal quality group and obtain process parameter optimization decisions based on the mining results; The second decision module 25 is used for early warning and multi-source coal ratio recommendation decisions based on coal quality groups.

[0079] The coal quality data-based analysis and decision-making device provided in this embodiment of the invention can execute the coal quality data-based analysis and decision-making method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the various modules and units described above are the same as in the corresponding embodiments described above, and will not be repeated here.

[0080] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0081] The following is a detailed reference. Figure 3 This diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 11, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 12 or a program loaded from memory 18 into random access memory (RAM) 13. The RAM 13 also stores various programs and data required for the operation of the electronic device. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0082] Typically, the following devices can be connected to I / O interface 15: input devices 16 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 17 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 18 including, for example, magnetic tapes, hard disks, etc.; and communication devices 19. Communication device 19 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0083] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 19, or installed from a memory 18, or installed from a ROM 12. When the computer program is executed by the processor 11, it performs the functions defined in the coal quality data-based analysis and decision-making method of the embodiments of the present invention.

[0084] Figure 3The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0085] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the coal quality data-based analysis and decision-making method shown in the above embodiments is implemented.

[0086] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0087] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for analysis and decision-making based on coal quality data, characterized in that, The method includes: Coal quality data is obtained from various sources, including testing systems, production systems, and procurement systems. Based on the coal quality data, a coal quality time-series knowledge graph is constructed with coal batches as the core nodes and integrating time attributes. Clustering is performed on the coal quality time-series knowledge graph to obtain coal quality groups with different patterns; The coal quality groups are correlated and analyzed, and process parameter optimization decisions are made based on the analysis results. Early warning and multi-source coal blending recommendation decisions are based on coal quality groups.

2. The method according to claim 1, characterized in that, Based on the aforementioned coal quality data, a coal quality time-series knowledge graph is constructed, with coal batches as the core nodes and incorporating time attributes, including: Based on the coal quality data, a node type is constructed, including coal mine nodes, batch nodes, laboratory indicator nodes, production equipment nodes, and production process nodes. Based on the node types, construct edge relationship types for batch-belonging coal mines, batch-have indicators, indicators affecting indicators, equipment output batches, and process parameter-affected indicators; Assign valid timestamps to the node types and edge relationship types to construct a coal quality time-series knowledge graph with coal batches as the core nodes and integrating time attributes.

3. The method according to claim 2, characterized in that, Before constructing a coal quality time-series knowledge graph with coal batches as the core nodes and incorporating time attributes, the method further includes: A sliding time window is used to acquire data on each index and each process parameter in the coal quality data to obtain the time series of the data. The correlation coefficient between two indicators with a marginal relationship, or between an indicator and a process parameter at different lag times; The lag time corresponding to the correlation coefficient with the largest absolute value is taken as the influence lag time of the corresponding edge, and the corresponding correlation coefficient is taken as the weight of the edge. The influence lag time and weight are updated based on the sliding time window.

4. The method according to claim 1, characterized in that, Clustering the coal quality time-series knowledge graph yields coal quality groups with different patterns, including: The coal quality time-series knowledge graph is clustered using a dynamic community discovery method to obtain coal quality groups with different patterns. The batches in the coal quality groups come from the same vein or have undergone the same production process.

5. The method according to claim 3, characterized in that, The coal quality groups are correlated and analyzed, and process parameter optimization decisions are made based on the analysis results, including: Based on the coal quality group, the temporal association path mining method is used to traverse the edges in the coal quality temporal knowledge graph where the timestamps satisfy the causal relationship, and the critical path is determined based on the weight and the impact lag time. Based on the analysis of the process parameters in the critical path, optimization suggestions for process parameters are obtained.

6. The method according to claim 2, characterized in that, Early warning and multi-source coal blending recommendation decisions based on coal quality groups include: Based on coal quality groups, a pre-trained temporal graph neural network model is used to predict the index values ​​of future times in the coal quality temporal knowledge graph, and a warning is issued based on the prediction results. Based on the coal quality population, a blending model is constructed to minimize the total cost corresponding to the blending ratio of different candidate coal batches and to meet the index constraints. Solving the ratio model yields multi-source ratio recommendation results.

7. The method according to claim 1, characterized in that, The method further includes: The coal quality time-series knowledge graph is stored and displayed through a visual interface; Use the preset query language and preset conditions to query on the visual interface.

8. A coal quality data-based analysis and decision-making device, characterized in that, The device includes: The data acquisition module is used to acquire coal quality data from different sources, including sampling systems, sample preparation systems, testing systems, production systems, and procurement systems. The knowledge graph construction module is used to construct a coal quality time-series knowledge graph based on the coal quality data, with coal batches as the core nodes and integrating time attributes. The clustering module is used to cluster the coal quality time-series knowledge graph to obtain coal quality groups with different patterns. The first decision module is used to perform correlation mining on the coal quality group and obtain process parameter optimization decisions based on the mining results; The second decision-making module is used for early warning and multi-source coal ratio recommendation decisions based on coal quality groups.

9. An electronic device, characterized in that, include: A memory and a processor are interconnected, the memory storing computer instructions, and the processor executing the computer instructions to perform the analysis and decision-making method based on coal quality data as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the analysis and decision-making method based on coal quality data as described in any one of claims 1 to 7.

Images