Intelligent fuel data management method and system based on data platform

Abstract

This application provides a fuel data intelligent management method and system based on a data platform, relating to the field of production data management technology. It decomposes the fuel management cycle and analyzes management objectives to obtain data factors for those objectives; establishes data acquisition pathways based on these data factors; acquires real-time updated data through these pathways and locates the fuel management cycle, constructing management cycle data blocks; performs management cycle characteristic analysis based on these data blocks to obtain cycle fuel characteristics; and conducts full-cycle supply imbalance analysis based on these cycle fuel characteristics to identify imbalance nodes and their characteristics, performing balance compensation to obtain a full-cycle balance management strategy. This application solves the technical problem that traditional management methods are cumbersome in their processing and lack systematic analysis tools for the entire cycle, resulting in an inability to provide accurate data support for fuel supply decision analysis. It improves the efficiency and accuracy of fuel data processing and enhances the stability and continuity of fuel supply throughout the entire lifecycle.

Description

Technical Field

[0001] This application relates to the field of production data management technology, specifically to a method and system for intelligent fuel data management based on a data platform. Background Technology

[0002] The fuel supply for thermal power generation requires precise data support at every stage, from procurement and transportation to storage and consumption. The accuracy of fuel data management directly impacts the stability of fuel supply and the efficiency of energy production. Existing fuel data management methods often involve cumbersome processes in data collection, cleaning, and processing. Managers need to manually collect various types of data and filter and process it using traditional algorithms, lacking clear goal orientation, resulting in redundant data and some data that is out of touch with actual management needs. Furthermore, traditional methods typically only analyze data for a single management stage, such as simple statistics on coal procurement, inventory, and consumption, lacking a comprehensive analysis of the correlations between different stages of the entire supply cycle. Therefore, this data management approach makes it difficult to effectively monitor and predict the entire fuel supply chain throughout its entire lifecycle, failing to promptly detect potential supply imbalances or anomalies, and thus unable to provide accurate data support for fuel supply decision-making, thereby affecting the stability and sustainability of fuel supply. Summary of the Invention

[0003] This application provides a fuel data intelligent management method and system based on a data platform, which solves the technical problems of traditional fuel data management methods, such as cumbersome data processing flow and lack of systematic analysis methods for the fuel management cycle, which makes it impossible to provide accurate data support for fuel supply decision analysis. It achieves the technical effect of improving fuel data processing efficiency and accuracy, thereby enhancing the stability and continuity of fuel supply throughout the entire life cycle.

[0004] In view of the above problems, this application provides a fuel data intelligent management method based on a data platform. The method includes: decomposing the fuel management cycle, analyzing the management objectives of the fuel management cycle, and obtaining data factors of the management objectives; establishing a data acquisition path based on the data factors, wherein the data acquisition path is mapped and associated with the data factors, and the data acquisition path uses the data factors as data acquisition indexes, and the data acquisition path is used to establish a connection between the data platform and the data source platform; obtaining real-time updated data from the data source platform through the data acquisition path, locating the fuel management cycle based on the data factors, and constructing a management cycle data block and storing it in the data platform; performing feature analysis of the management cycle based on the management cycle data block to obtain cycle fuel characteristics; performing a full-cycle supply imbalance analysis based on the supply balance impact relationship of the fuel management cycle and the cycle fuel characteristics to obtain imbalance nodes and imbalance characteristics; and performing balance compensation based on the imbalance nodes and imbalance characteristics to obtain a full-cycle balance management strategy.

[0005] On the other hand, this application also provides a fuel data intelligent management system based on a data platform. The system includes: a management target parsing module, used to decompose the fuel management cycle, parse the management target of the fuel management cycle, and obtain data factors of the management target; a data acquisition path establishment module, used to establish a data acquisition path based on the data factors, wherein the data acquisition path is mapped and associated with the data factors, and the data acquisition path uses the data factors as data acquisition indexes, and the data acquisition path is used to establish a connection between the data platform and the data source platform; a data update and storage module, used to obtain real-time updated data from the data source platform through the data acquisition path, locate the fuel management cycle based on the data factors, construct a management cycle data block, and store it in the data platform; a cycle feature analysis module, used to perform feature analysis of the management cycle based on the management cycle data block to obtain cycle fuel features; a supply imbalance analysis module, used to perform full-cycle supply imbalance analysis based on the supply balance impact relationship of the fuel management cycle and the cycle fuel features to obtain imbalance nodes and imbalance features; and a balance compensation module, used to perform balance compensation based on the imbalance nodes and imbalance features to obtain a full-cycle balance management strategy.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] By decomposing the fuel management cycle and analyzing its management objectives, data factors for these objectives are obtained. This provides a clear objective and direction for data collection throughout the fuel data management process, ensuring the relevance and effectiveness of data acquisition. A data acquisition pathway is established based on these data factors and connected to the data source platform, enabling real-time data acquisition and updates, thus improving the efficiency and accuracy of data processing. Real-time updated data is obtained through this pathway, and management cycle data blocks are constructed and stored in the data platform, providing a complete and orderly data foundation for subsequent analysis and facilitating a comprehensive understanding of the data situation within the fuel management cycle. In-depth analysis of the integrated management cycle data blocks identifies key characteristics of each management cycle, yielding cycle-specific fuel characteristics. These characteristics reflect important information such as the state and patterns of fuel within the management cycle, providing a basis for subsequent supply imbalance analysis. Based on the obtained cycle-specific fuel characteristics and their impact on supply balance, a full-cycle supply imbalance analysis is conducted to identify imbalance nodes and characteristics, thereby accurately identifying problems in the fuel supply process and providing crucial information for balance compensation. By performing balance compensation based on imbalance nodes and imbalance characteristics, a full-cycle balance management strategy is obtained, enabling dynamic adjustment of the supply process and ensuring the balance and stability of fuel supply.

[0008] In summary, this application decomposes the fuel management cycle and analyzes target data factors, establishes data acquisition pathways to obtain data and constructs management cycle data blocks, then conducts in-depth analysis to obtain cycle fuel characteristics, further performs full-cycle supply imbalance analysis, and finally obtains a balance management strategy based on the imbalance situation. This series of steps can accurately collect data related to management objectives, improve the efficiency and accuracy of data collection, and also enable timely understanding of fuel characteristics within the management cycle. It allows for the identification of supply imbalance problems from a full-cycle perspective and the development of effective balance management strategies to address imbalances, thereby ensuring the stability and continuity of fuel supply and reducing the risks caused by supply imbalances.

[0009] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the intelligent fuel data management method based on a data platform provided in an embodiment of this application.

[0011] Figure 2 This is a flowchart illustrating the process of obtaining data factors for management objectives in the fuel data intelligent management method based on a data platform provided in this application embodiment.

[0012] Figure 3 This is a schematic diagram illustrating the process of obtaining imbalance nodes and imbalance characteristics in the fuel data intelligent management method based on a data platform provided in this application embodiment.

[0013] Figure 4 This is a schematic diagram of the structure of a fuel data intelligent management system based on a data middle platform, provided in an embodiment of this application.

[0014] Explanation of reference numerals in the attached diagram: Management target analysis module 10, data acquisition path establishment module 20, data update and storage module 30, periodic characteristic analysis module 40, supply imbalance analysis module 50, balance compensation module 60. Detailed Implementation

[0015] This application provides a fuel data intelligent management method and system based on a data platform, which solves the technical problems of traditional fuel data management methods, such as cumbersome data processing flow and lack of systematic analysis methods for the fuel management cycle, which makes it impossible to provide accurate data support for fuel supply decision analysis. It achieves the technical effect of improving fuel data processing efficiency and accuracy, thereby enhancing the stability and continuity of fuel supply throughout the entire life cycle.

[0016] Example 1, as Figure 1 As shown in the embodiment of this application, a fuel data intelligent management method based on a data platform is provided, the method comprising:

[0017] Step S1: Decompose the fuel management cycle, analyze the management objectives of the fuel management cycle, and obtain the data factors of the management objectives.

[0018] Specifically, the fuel management cycle refers to different management intervals in fuel supply chain management, divided according to time or stage. Management objectives refer to the specific goals to be achieved within each fuel cycle, such as sufficient supply, reasonable inventory, and cost control. Data factors refer to various data elements or variables closely related to management objectives, such as inventory levels, demand forecasts, and supplier delivery times.

[0019] The fuel management cycle is divided into detailed phases based on different stages of fuel supply, including the procurement cycle (from fuel demand forecasting to supplier selection and purchase order generation), the transportation cycle (the logistics process of fuel from suppliers to the plant), the storage cycle (the management of fuel storage in the plant), and the usage cycle (fuel blending, unit operation, and fuel consumption). Based on management experience and data analysis, key objectives for each management cycle are determined. These objectives can be specific requirements regarding fuel consumption, inventory levels, supply chain efficiency, etc. Next, data analysis techniques, such as data mining and statistical analysis, are used to identify the data points most relevant to these objectives from historical data; these data points are called data factors. For example, if the management objective is to reduce fuel consumption, then data factors might include fuel utilization rate, equipment efficiency, and operating conditions. The identified data factors more accurately reflect the characteristics and needs of each cycle, providing a more precise foundation for subsequent data collection and analysis, making the entire fuel data management more aligned with actual business processes.

[0020] Step S2: Based on the data factors, establish a data acquisition path. The data acquisition path is mapped and associated with the data factors, and the data acquisition path uses the data factors as the data acquisition index. The data acquisition path is used to establish a connection between the data middle platform and the data source platform.

[0021] Specifically, a data acquisition channel is a specific transmission channel established from the data source to the data platform to acquire relevant data for each period, ensuring the real-time flow and updating of data. First, the data source platforms need to be identified; these platforms may include sensors, ERP systems, warehouse systems, logistics management systems, etc. Then, based on the data factors required for management objectives, multiple data acquisition channels are built. These channels can collect data from the data source platforms in real time and transmit it to the data platform for further analysis and processing. There is a mapping relationship between the data factors and data acquisition channels for each management period. By using the data factors, the corresponding data acquisition channel can be located, thereby collecting data from different platforms to the data platform. This ensures that the acquired data is relevant to management objectives, avoids cumbersome data filtering steps, and improves data collection efficiency.

[0022] Step S3: Obtain real-time updated data from the data source platform through the data acquisition channel, locate the fuel management cycle according to the data factors, and construct a management cycle data block to store in the data platform.

[0023] Specifically, real-time updated data refers to the latest data obtained from the data source platform. This data typically changes in real time and reflects the actual state of the current fuel supply chain. A management cycle data block refers to a data unit that integrates all relevant data for a management cycle, facilitating subsequent analysis and processing. The data platform is a system used to store, process, and analyze data collected from various data sources; all management cycle data blocks are ultimately stored in this platform. The data acquisition pathway continuously collects data from the data source platform, such as warehouse inventory, transportation status, and fuel consumption. Based on data factors, the corresponding management cycle is determined, and the data is categorized and stored in the appropriate area of ​​the data platform. This data is organized into data blocks, each corresponding to a management cycle and containing all relevant data factors within that cycle.

[0024] By collecting and managing the construction of periodic data blocks in real time, it is possible to ensure that data from different periods are accurately classified and stored. The resulting managed periodic data blocks can comprehensively and accurately reflect the actual situation of each period, providing a rich and orderly data foundation for subsequent analysis of each period.

[0025] Step S4: Perform feature analysis of the management cycle based on the management cycle data block to obtain cycle fuel characteristics.

[0026] Specifically, cyclical fuel characteristics are unique attributes of fuel within each management cycle, derived from analyzing data blocks within that cycle. Examples include supplier price fluctuation trends during the procurement cycle and transportation stability during the transportation cycle. In-depth analysis of data blocks within the management cycle identifies key characteristics and patterns within the cycle. Features related to fuel supply in each cycle are extracted to provide a basis for subsequent decision-making. For example, for data blocks in the procurement cycle, data analysis software can be used to analyze and derive features such as the average value and fluctuation range of supplier quotes through statistical analysis methods; these constitute the cyclical fuel characteristics of the procurement cycle. In the transportation cycle, analysis of data such as transportation location and speed, for example using trajectory analysis algorithms, can yield characteristics such as transportation timeliness and route stability. For the storage cycle, analysis of inventory data and quality inspection data can yield characteristics such as inventory turnover rate and fuel quality stability. For the usage cycle, analysis of fuel consumption and unit operating parameters can yield cyclical fuel characteristics such as fuel utilization efficiency.

[0027] By deeply mining the cyclical fuel characteristics from the data blocks of each cycle, we can clearly understand the state and characteristics of fuel in each cycle, providing detailed evidence for the analysis of supply imbalance throughout the entire cycle and helping to grasp the fuel management situation more comprehensively.

[0028] Step S5: Based on the supply balance impact relationship of the fuel management cycle, perform a full-cycle supply imbalance analysis according to the cycle fuel characteristics to obtain imbalance nodes and imbalance characteristics.

[0029] Specifically, the supply balance impact refers to the interrelationships between factors in different cycles throughout the fuel management process that affect the fuel supply balance. For example, supplier selection during the procurement cycle affects transportation arrangements during the transportation cycle, which in turn affects inventory levels during the storage cycle and the stability of fuel supply during the usage cycle. Imbalance nodes refer to specific points or connections in the entire fuel management process where supply imbalances occur between one or more cycles. Imbalance characteristics manifest at these imbalance nodes as phenomena such as supply shortages, inventory buildup, and excessively high costs.

[0030] Based on the impact of supply and demand balance, and combined with the cyclical fuel characteristics of each period, this analysis identifies potential imbalance points and links, and determines the characteristics exhibited at these imbalance points, such as declining inventory and transportation delays. For example, if the cyclical fuel characteristics of the procurement cycle show continuously rising supplier prices and unstable supply, and the transportation cycle shows poor on-time delivery, then the procurement and transportation links can be identified as potential imbalance points, with unstable supply and untimely delivery being the characteristics of imbalance. Through imbalance analysis, potential risk points in fuel management can be identified in advance, providing early warnings for subsequent balance management and reducing fluctuations in fuel supply.

[0031] Step S6: Perform balance compensation based on the imbalance nodes and imbalance characteristics to obtain a full-cycle balance management strategy.

[0032] Specifically, the full-cycle balance management strategy is a set of management plans and measures that cover all cycles of fuel management to maintain the balance of fuel supply, aiming to restore the balance of fuel supply across all cycles.

[0033] Based on the identified imbalance points and characteristics, corresponding compensation measures are formulated, including adjusting production plans, changing inventory strategies, optimizing logistics routes, or negotiating with suppliers to ensure supply stability. Based on the results of the full-cycle supply balance analysis, detailed balance management strategies are developed to optimize all aspects of fuel procurement, storage, transportation, and use to ensure fuel supply stability. By developing a full-cycle balance management strategy, imbalances in the fuel management process can be comprehensively adjusted and optimized, ensuring fuel supply balance across different cycles, improving the overall efficiency and stability of fuel management, and reducing operating costs.

[0034] Furthermore, such as Figure 2 As shown, step S1 in this embodiment includes:

[0035] Step S11: Obtain positive and negative samples for the fuel management cycle. The positive samples are cases where the evaluation result of the fuel management cycle reaches the target threshold, and the negative samples are cases where the evaluation result does not reach the target threshold.

[0036] Step S12: Extract management target events based on the positive and negative examples, and perform impact factor analysis on the management target events to obtain positive and negative example factor analysis results.

[0037] Step S13: Using the management target event as the alignment condition, perform factor alignment and fusion of the positive example factor analysis results and the negative example factor analysis results, take the management target event as the management target, integrate the corresponding data types of the fused influencing factors, and obtain the data factors of the management target.

[0038] Specifically, positive examples are cases where the evaluation results meet pre-set target thresholds during the fuel management cycle. For example, in the fuel management of thermal power plants, if the set target thresholds are no more than three fuel supply interruptions and costs controlled within budget, then fuel management cycle cases that meet these conditions are positive examples. Negative examples, on the other hand, are cases where the evaluation results do not meet the target thresholds during the fuel management cycle. These include cases where the number of fuel supply interruptions exceeds three or costs exceed budget. The target threshold is a set standard used to measure whether fuel management objectives have been achieved.

[0039] First, determine the criteria for evaluating the success of a fuel management cycle, i.e., the target threshold. The target threshold could be an upper limit on the number of fuel supply interruptions, a cost limit, etc. Then, filter historical fuel management data to identify cases that meet the definitions of positive and negative samples. This filtering process can use data mining tools, such as SQL queries, to extract relevant data from the database. For example, by writing SQL queries, positive and negative samples can be selected based on predefined evaluation indicators (such as whether costs are within budget).

[0040] Management objective events refer to specific events related to management objectives within the fuel management cycle, such as bidding events in fuel procurement and delivery events in transportation. These events have a direct or indirect impact on management objectives. Positive example factor analysis results are obtained by analyzing the influencing factors of management objective events in the positive example sample, encompassing various influencing factors in cases where management objectives were successfully achieved. Negative example factor analysis results are obtained by analyzing the influencing factors of management objective events in the negative example sample, reflecting the influencing factors in cases where management objectives were not achieved.

[0041] For each positive and negative sample, in-depth analysis is conducted to extract events related to fuel management objectives, i.e., management objective events. Key factors influencing these management objective events are identified, and the positive and negative factor analysis results are determined. For example, factors in positive samples might be "timely supplier delivery" and "inventory management system optimization," while factors in negative samples might be "transportation delays due to weather" and "insufficient supplier production capacity." The analysis process can utilize tools such as cause-and-effect diagrams or fault trees.

[0042] Using the management objective event as the alignment condition, the influencing factors in the positive and negative example factor analysis results are compared one by one. For example, under the management objective event of procurement bidding, the supplier reputation in the positive and negative example factor analysis results is compared. Then, similar influencing factors are merged, removing duplicates and redundancies. For cases where different data types represent the same influencing factor, data types are integrated. For example, if supplier reputation is represented by a rating in the positive example factor analysis results and by a score in the negative example factor analysis results, they need to be integrated into a unified data type. Through factor alignment and fusion, a comprehensive set of data factors is formed, which can comprehensively reflect the various factors affecting the achievement of fuel management objectives.

[0043] By merging factors, a more accurate and comprehensive set of influencing factors can be obtained, providing a precise and unified data foundation for subsequent fuel management data collection and analysis, thereby improving the accuracy and efficiency of the entire fuel management data processing.

[0044] Furthermore, in step S12, the management target events are extracted based on the positive and negative examples, and the process further includes:

[0045] Step S121: Divide the fuel management cycle into operation steps, obtain the target events generated by each operation step, and summarize the target event set of operation steps.

[0046] Step S122: Use the set of target events for the operation steps to perform a missing traversal on the management target events to obtain the missing step target events.

[0047] Step S123: Supplement the management target event using the missing step target event.

[0048] Specifically, a step-generated target event is an event related to the management objective that occurs during the execution of each operational step. For example, in the bidding process, successfully attracting a certain number of qualified suppliers to participate in the bidding is a step-generated target event. The operational step target event set is the collection of target events generated by all operational steps. Based on the actual business logic of the fuel management cycle, each link or stage in the fuel management cycle is further broken down into specific operational steps. For example, the procurement cycle can be broken down into multiple specific steps such as demand forecasting, supplier selection, and purchase order generation. For each operational step, the target events that should be generated under normal operating conditions are analyzed. For example, in the vehicle dispatching step, timely allocation of suitable transport vehicles is a target event. The target events of each operational step are then aggregated to form the operational step target event set. Through operational step segmentation and target event aggregation, the target events of each link in the fuel management cycle can be comprehensively reviewed, providing a detailed basis for subsequent supplementation and improvement of management target events.

[0049] Using the set of target events for operational steps as a reference, the extracted management target events are iterated and checked. Each event in the set of target events for operational steps is compared with the management target events to identify target events that are missing from the management target events but exist in the set of target events for operational steps. These are defined as missing step target events. For example, a periodic inventory count event in the fuel storage management cycle, which is not in the management target events, is a missing step target event.

[0050] Based on the identified missing step target events, corresponding events are retrieved from historical data to supplement the extracted management target events. The supplemented events are then reintegrated to form a complete set of management target events. This ensures that key target events for each operational step are included in the management target events, providing comprehensive data support. For example, if the real-time fuel quality monitoring event during the fuel usage cycle is found to be a missing step target event, this event is retrieved from historical data and added to the management target events.

[0051] By supplementing the missing step target events, the management target events have been improved, enabling them to more comprehensively reflect all aspects of the fuel management cycle. This provides a more complete foundation for subsequent analysis of influencing factors, thereby more accurately identifying the various factors affecting the achievement of management objectives.

[0052] Furthermore, in step S12, influence factor analysis is performed on the management target event to obtain positive example factor analysis results and negative example factor analysis results, including:

[0053] Step S124: Perform event correlation analysis on the management target events to determine intermediate events and basic events.

[0054] Step S125: Decompose the event factors layer by layer according to the order from the intermediate events to the basic events, and determine the event influencing factors and their influencing relationships.

[0055] Step S126: Construct an event parse tree according to the hierarchical relationship of intermediate events - basic events - event influencing factors.

[0056] Step S127: Based on the event impact factors and their impact relationships, verify the event occurrence probability layer by layer from bottom to top according to the event parsing tree according to the logic gate relationship. When the probability threshold requirement is reached, determine the event impact factors and their impact relationships, and obtain the factor analysis results.

[0057] Specifically, intermediate events are those events in the causal chain of the management target event that are both influenced by and affect other events, occupying an intermediate stage. For example, in fuel supply management, a transportation scheduling event is affected by fuel procurement volume and, in turn, fuel delivery time; therefore, it is considered an intermediate event. Basic events are the first events to occur within the management target event and have an initial impact on other events. For instance, a fuel demand forecasting event is a basic event in fuel management, affecting subsequent procurement, transportation, and a series of other events.

[0058] For management target events, data mining techniques are used for analysis. For example, by analyzing historical fuel management data, the chronological order of events and the correlation coefficients between them can be examined. If two events frequently occur simultaneously, or if one event follows another with a higher probability of occurrence, then a correlation exists between the two events. Through event correlation analysis, intermediate events and foundational events are identified, revealing the internal logical structure of management target events. This provides a clear framework for subsequent event factor decomposition, helping to more effectively identify the various factors influencing management target events.

[0059] Event influencing factors are the various factors that influence the occurrence and outcome of an event, identified during the event factor decomposition process. For example, in a transportation scheduling event, the number of transport vehicles and driver availability are event influencing factors. Following the order from intermediate events to basic events, each event is factored. Through data analysis, historical cases, and expert knowledge, the root causes of each event are analyzed layer by layer to identify the influencing factors for each event and analyze the relationships between these factors and the event. For example, in analyzing the intermediate event of transportation scheduling, a direct positive correlation is found between the number of transport vehicles and the success of the scheduling; that is, the more vehicles, the higher the probability of successful scheduling. Through layer-by-layer factor decomposition, the specific causes of each event can be comprehensively identified, and the interrelationships between influencing factors can be revealed, providing detailed evidence for further factor modeling and event prediction.

[0060] An event parse tree is a graphical data structure used to represent the hierarchical relationships between events. The hierarchical relationship refers to the superior-inferior relationship between events and factors. In a parse tree, basic events are located at the root, intermediate events at the branches, and event factors at the ends. Based on the previously determined intermediate events, basic events, and event influencing factors and their relationships, an event parse tree is constructed according to the hierarchical relationship of intermediate events – basic events – event influencing factors, visually demonstrating the hierarchical structure and interrelationships between events. Taking supply management in fuel management as an example, transportation scheduling (an intermediate event) is a branch node of the tree, fuel demand forecasting (a basic event) is a child node under the transportation scheduling node, and event influencing factors such as the number of transport vehicles and driver availability are leaf nodes under the fuel demand forecasting node. The event parse tree presents the complex relationships of management target events in an intuitive tree structure, clearly showing the hierarchical relationships between various events and influencing factors. This helps to grasp the overall structure of management target events and provides a clear framework for subsequent event probability verification.

[0061] Logical gate relationships are rules used in an event parse tree to represent the logical relationships between event influencing factors, base events, and intermediate events. A probability threshold is a pre-set probability value; when the verified probability of an event occurring reaches this value, the event influencing factor is considered to have sufficient influence on that event.

[0062] Starting from the leaf nodes (event influence factors) of the event parse tree, the probability of each node occurring at the next higher level is calculated based on logical gate relationships. For example, if the occurrence of a basic event requires two event influence factors to be satisfied simultaneously (AND relationship), the probability of the basic event is calculated based on the individual probabilities of these two event influence factors. The calculated probabilities are then passed up layer by layer until the probability of intermediate events is calculated. This process can use probabilistic calculation models, such as Bayesian network models, for probability calculation. When the calculated event occurrence probability reaches a pre-set probability threshold, it is determined that the event influence factor and its influence relationship have a significant impact on the management target event, and this is used as the factor analysis result. By verifying the event occurrence probability layer by layer from the bottom up, reasonable event influence factors and their influence relationships can be determined based on actual data and logical relationships, ensuring the accuracy and reliability of the factor analysis results and providing a scientific basis for subsequent fuel management decisions.

[0063] Furthermore, in step S127, when the probability threshold requirement is not met, the method further includes:

[0064] Step S127-1: Obtain an abnormal event, which is an intermediate or basic event in which the logic gate relationship verification result does not reach the probability threshold.

[0065] Step S127-2: Extract historical event record dataset based on the abnormal event, perform semantic mining on the historical event record dataset, and establish a factor relationship graph for semantic association between response factors.

[0066] Step S127-3: Based on the factor relationship map, mine the dominant and latent factors.

[0067] Step S127-4: Fit the correlation and influence relationship between the dominant factor, the latent factor and the abnormal event based on the historical event record dataset.

[0068] Step S127-5: Add the associated influence relationship to the logical gate relationship, and perform bottom-up reverse layer-by-layer verification based on the event parsing tree until the probability of the event occurs reaches the probability threshold requirement.

[0069] Specifically, during the probability verification process in step S127, if the calculated probability of an intermediate or basic event does not meet the probability threshold requirement, the event is marked as an abnormal event. These abnormal events, under normal logical relationships, do not have the expected probability of occurrence, and there may be special influencing factors that require further analysis.

[0070] Historical event record datasets are collections of data containing records of various past events related to fuel management. These records can contain detailed information about various aspects of fuel procurement, transportation, storage, and use. Factor relationship graphs are graphs constructed through semantic mining to display the semantic relationships between factors (event influencing factors), visually presenting the semantic connections between different factors. Accurately identifying anomalous events provides clear targets for subsequent in-depth analysis, helping to find potential reasons why the probability of event occurrence does not meet expectations, thereby further refining the analysis of event influencing factors and their relationships.

[0071] A dataset of historical event records related to the anomalous events was retrieved from the database. For the textual information in the dataset (such as event descriptions and notes), natural language processing techniques were used for semantic mining analysis to uncover potential influencing factors and their semantic relationships. For example, word vector models were used to convert the text into vector representations, and then algorithms such as cluster analysis and association rule mining were used to extract the semantic relationships between factors, constructing a factor relationship graph. The factor relationship graph reflects the correlation and influence between various factors, providing a structured diagram for subsequent factor analysis.

[0072] Dominant factors are those factors in historical event record datasets that can be directly observed to have a clear correlation with anomalous events through methods such as semantic mining. For example, in fuel transportation anomalies, frequent vehicle malfunction records constitute a dominant factor. Latent factors are those that have a potential correlation with anomalous events but are not easily discovered directly in the dataset and require in-depth analysis of semantic relationships to uncover. For example, in the same transportation anomaly, road construction information around the transportation route may be hidden in the event description; semantic mining reveals its correlation with the transportation anomaly, making it a latent factor. Based on the constructed factor relationship graph, dominant and latent factors are identified by analyzing indicators such as the connection strength and semantic relevance between factors. Factors with high connection strength and direct semantic relevance are identified as dominant factors; factors with weak connection strength but related to the anomalous event through multi-step semantic association are identified as latent factors. The analysis process can use graph analysis algorithms, such as the PageRank algorithm, to measure the connection strength between factors and assist in the discovery of dominant and latent factors. Identifying both dominant and latent factors helps to consider the factors influencing anomalous events more comprehensively, going beyond just those that are easily found on the surface, and provides richer material for further analysis of the causes of anomalous events.

[0073] Statistical analysis and data fitting are performed on historical event record datasets. For example, for abnormal fuel transportation events, with extended transportation time as the abnormal event and the number of vehicle malfunctions (dominant factor) and road construction conditions (latent factor) as independent variables, a regression model (such as a linear regression model) is established to fit the correlation between these factors and the extended transportation time. By fitting the correlation, the influence of dominant and latent factors on abnormal events can be quantified, making subsequent probability verification more accurate.

[0074] The obtained correlation relationships are added to the original logic gate relationships to reconstruct the event parse tree. Then, following the bottom-up, reverse layer-by-layer verification method in step S127, the probability of event occurrence is recalculated. For example, if the original logic gate relationships did not consider the impact of road construction conditions (latent factors) on transportation time extension (abnormal event), after adding the correlation relationships, the probability of the abnormal event of transportation time extension is calculated again from the event impact factors until the probability reaches the required threshold. By continuously adjusting the logic gate relationships and re-verifying the event occurrence probabilities, the analysis of event impact factors and their influence relationships can be gradually improved, making the final results more consistent with the actual situation and improving the accuracy and reliability of fuel management event analysis.

[0075] Furthermore, such as Figure 3 As shown, step S5 in this embodiment includes:

[0076] Step S51: Set the periodic balance target relationship and penalty conditions for the fuel management cycle.

[0077] Step S52: Analyze the characteristic relationships between fuel management cycles based on the cycle balance target relationship.

[0078] Step S53: Construct an evaluation function based on the feature relationships and penalty conditions.

[0079] Step S54: Use the evaluation function to evaluate the balance target relationship of the periodic fuel characteristics and obtain the periodic node evaluation value.

[0080] Step S55: Based on the periodic node evaluation value, locate the imbalance node, and combine the periodic balance target relationship to locate the imbalance characteristics.

[0081] Specifically, the cyclical balance target relationship is a goal-oriented relationship that is expected to be established between different cycles in the fuel management cycle in order to achieve overall balance. The penalty condition is a condition or rule set to measure the degree of deviation when the actual situation in the fuel management cycle deviates from the cyclical balance target relationship. This can be a specific numerical limit, time limit, or resource limit, etc.

[0082] Based on the overall goals of fuel management and the needs and expectations of each link throughout the management cycle, a cyclical balance target relationship is established. This cyclical balance target relationship can be dynamically adjusted according to actual needs. For example, based on historical data and the company's production plan, a reasonable ratio between the procurement volume of the procurement cycle and the usage volume of the usage cycle is determined. Penalty conditions are set according to the company's internal cost accounting system and risk management requirements. For example, based on considerations of warehousing costs and safety stock, penalty conditions corresponding to upper and lower limits of inventory during the storage cycle are set. Setting cyclical balance target relationships and penalty conditions provides standards and basis for subsequent analysis and evaluation, helps to standardize fuel management processes, promptly identify and correct potential imbalances, and improve the efficiency and effectiveness of fuel management.

[0083] Characteristic relationships, based on cyclical balance target relationships, describe the interrelationships between various characteristics exhibited during the actual operation of different fuel management cycles. For example, price fluctuations in the procurement cycle may influence transportation volume selection characteristics in the transportation cycle; this is a characteristic relationship. Based on the previously established cyclical balance target relationships, the actual operational data of each fuel management cycle are analyzed. Data analysis methods, such as correlation analysis and regression analysis, can be employed. Taking the procurement and transportation cycles as examples, the characteristic relationship between them is determined by analyzing the correlation between procurement prices and transportation volumes in historical data. By analyzing characteristic relationships, a deeper understanding of the intrinsic connections between different fuel management cycles can be achieved, providing necessary input for constructing evaluation functions and helping to more accurately assess the balance of the fuel management cycles.

[0084] An evaluation function is a mathematical function constructed based on characteristic relationships and penalty conditions, used to quantitatively evaluate the balance of the fuel management cycle. The evaluation function is constructed based on the analytically obtained characteristic relationships and the set penalty conditions. For example, the evaluation function can be... Where, ω i The weight of the i-th feature reflects its importance in the evaluation function. i It is the actual observed or calculated value of the i-th feature. i目标 It is the target value of the i-th feature. (x) i -x i目标 ) 2 Used to measure the magnitude of the deviation between the feature value and the target value, the squared form ensures that the deviation value is non-negative and that larger deviations will be amplified. R(x) j F is a function related to penalty conditions, representing the penalty value for certain specific conditions (e.g., fuel shortage or excess) in the cycle. n is the total number of features, and m is the total number of penalty conditions. (x)The smaller the value, the closer it is to the equilibrium target relationship. The evaluation function provides a quantitative evaluation standard for the balance of the fuel management cycle, enabling convenient and quick evaluation of fuel management under different conditions, and providing a powerful tool for identifying imbalance nodes and imbalance characteristics.

[0085] The cycle node evaluation value is a numerical value obtained by evaluating the cycle fuel characteristics of each fuel management cycle using an evaluation function. This value reflects the performance of each cycle in balancing the target relationship. By substituting the cycle fuel characteristic data of each fuel management cycle into the evaluation function, the cycle node evaluation value for each cycle is calculated. By calculating the cycle node evaluation value, the balance status of each fuel management cycle can be quantitatively assessed, providing direct data for locating imbalance nodes and helping to accurately identify problematic links in the fuel management process.

[0086] The calculated evaluation values ​​of each cycle node are compared and analyzed. If the evaluation value of a certain cycle node is significantly lower than that of other cycles or lower than a pre-set threshold, then this cycle may be an imbalance node. Then, the specific imbalance characteristics of this imbalance node are further analyzed in conjunction with the cycle balance target relationship. For example, based on the balance target relationship between the procurement cycle and other cycles, if it is found that the procurement volume is far lower than the demand for usage, then "insufficient procurement volume" is the imbalance characteristic of this imbalance node (procurement cycle). By locating the imbalance nodes and imbalance characteristics in the fuel management process, a clear direction is provided for subsequent balancing and compensation measures, which helps to improve the overall balance and efficiency of fuel management.

[0087] Furthermore, step S6 in this embodiment of the application also includes:

[0088] Step S61: Based on the characteristics of the cyclical fuel, predict the supply and demand of the long and short term, and establish the supply time series chain and the demand time series chain.

[0089] Step S62: Based on the connection relationship of the fuel management cycle, perform time alignment of the supply time sequence chain and the demand time sequence chain to construct the supply and demand time sequence chain between cycle nodes.

[0090] Step S63: Based on the supply and demand time series chain, predict the supply and demand balance relationship and obtain the predicted imbalance time series nodes.

[0091] Step S64: Based on the predicted imbalance time series nodes, according to the time series distance and the supply time series characteristics of the periodic fuel characteristics, perform time series constraint analysis to obtain the minimum adjustment time window.

[0092] Step S65: Use the minimum adjustment time window as a constraint to perform balance compensation on the unbalanced timing nodes.

[0093] Specifically, by analyzing historical data and relevant factors, the supply and demand of fuel are estimated over longer periods (e.g., one month, one quarter) and shorter periods (e.g., three days, one week). These estimated long-term and short-term supply and demand quantities are then arranged chronologically to form supply and demand time-series chains. The supply time-series chain reflects the quantity and source of fuel supply at different points in time. The demand time-series chain displays the quantity and purpose of fuel demand at different points in time. Forecasting supply and demand can be accomplished using forecasting algorithms (such as time series analysis and regression analysis). For example, if fuel consumption data from previous years is available, the fuel demand for the next year can be predicted by fitting an ARIMA model. For short-term forecasts, recent realities such as current production plans and equipment operating status can be considered, using methods such as moving averages or exponential smoothing. By forecasting both long-term and short-term supply and demand, a more accurate future view of fuel management can be provided, helping to identify potential supply imbalances and providing data support for subsequent adjustments and compensation.

[0094] Based on the connectivity of the fuel management cycle, the correspondence between each node in the supply and demand time series chains is determined to ensure the matching of supply and demand information at the same point in time. After the supply and demand time series chains are aligned, a supply-demand time series chain is formed, which accurately reflects the changes in the fuel supply and demand relationship over time at each cycle node.

[0095] In supply and demand balance forecasting, imbalance time series nodes are identified as cyclical nodes where supply and demand are out of balance at a future point in time. For each cyclical node in the supply and demand time series chain, the quantitative relationship between supply and demand is compared. A balance threshold can be set; when the difference between supply and demand exceeds this threshold, an imbalance is identified. For example, if the difference between supply and demand is greater than 10% of total demand, the node is considered imbalanced. By traversing all nodes in the supply and demand time series chain, nodes that meet the imbalance criteria are identified; these are the predicted imbalance time series nodes.

[0096] Time series distance is the time interval between predicted imbalance nodes and other relevant nodes in a supply and demand time series chain. Supply time series characteristics are the properties of fuel supply over time, such as periodicity, seasonality, and volatility. The minimum adjustment time window is the shortest time range required to balance and compensate imbalance nodes under various time series constraints.

[0097] First, determine the temporal distance between the predicted imbalance node and other relevant nodes. For example, in fuel supply management, if an imbalance is predicted for a storage cycle node within the next month, it's necessary to analyze the time interval between this node and the procurement cycle node (supply source). Then, analyze the supply temporal characteristics of the cycled fuel, and resolve the temporal constraints by establishing a mathematical model or rule set to determine the minimum adjustment time window. For instance, a linear programming model can be built based on historical data, using supply temporal characteristics and temporal distances as constraints, with the objective of minimizing the adjustment time, thus obtaining the minimum adjustment time window.

[0098] The minimum adjustment time window is used as a constraint to adjust imbalanced time series nodes. If the imbalanced time series node is a situation of insufficient supply, measures such as increasing procurement volume and speeding up transportation can be taken within the minimum adjustment time window. For example, if it is predicted that fuel inventory will be insufficient for a certain storage cycle, and the minimum adjustment time window is one week, then additional purchase orders can be arranged or transportation routes can be optimized to improve transportation efficiency within this week.

[0099] The above steps, by analyzing the supply and demand balance relationship, predict the timing of imbalances, which can help identify potential supply and demand imbalances in the fuel management cycle in advance. This allows for preventative measures to be taken to balance and compensate for imbalances, thereby improving the stability and efficiency of fuel management.

[0100] In summary, the fuel data intelligent management method based on a data platform provided in this application has the following technical effects:

[0101] By decomposing the fuel management cycle and analyzing management objectives, fuel supply management is divided into multiple management cycles, identifying and focusing on resolving the most critical issues in each cycle. Clearly defined management objectives ensure targeted data collection and analysis for each cycle, improving the accuracy and effectiveness of data processing. Establishing data collection pathways and data factor mappings makes the connection between data source platforms and the data platform more efficient, ensuring the automatic collection of key data factors related to management objectives from different data source platforms. This avoids the tedious process of manual data collection and filtering in traditional methods, improving the efficiency and accuracy of data collection. Real-time updated data from source platforms is obtained through data collection pathways, and the fuel management cycle is located based on data factors. Cycle management data blocks are constructed, integrating data from multiple data sources to form a detailed data view of the management cycle, providing a high-quality data foundation for subsequent feature analysis and decision-making. Feature analysis of the management cycle data blocks reveals the cyclical characteristics of fuel supply, providing a scientific basis for the stability and risk management of each stage of fuel supply. Combined with full-cycle analysis, potential supply imbalance nodes and their characteristics can be identified, thus providing an early warning mechanism for balance compensation and risk control. Finally, based on the imbalance points and their characteristics, a full-cycle balance management strategy is formulated. This strategy allows for targeted adjustments to fuel supply plans and management measures, ensuring the balance and stability of the entire supply chain and avoiding the risks of oversupply or shortage.

[0102] Overall, the embodiments of this application improve the efficiency and accuracy of data processing by periodizing fuel management, automating data acquisition pathways, and extracting periodic features and analyzing supply imbalances, ensuring accurate decision-making throughout the entire management cycle. This method can acquire and analyze fuel supply chain data in real time, providing managers with reliable decision support, timely identifying potential risks, and implementing dynamic adjustments. This not only improves the management efficiency of the fuel supply chain but also enhances the ability to cope with market fluctuations and supply imbalances, ensuring the stability and continuity of fuel supply throughout its entire lifecycle.

[0103] Example 2, as Figure 4 As shown in the embodiment of this application, a fuel data intelligent management system based on a data middle platform is provided, the system comprising:

[0104] The management target analysis module 10 is used to decompose the fuel management cycle, analyze the management targets of the fuel management cycle, and obtain the data factors of the management targets.

[0105] The data acquisition path establishment module 20 is used to establish a data acquisition path based on the data factors. The data acquisition path is mapped and associated with the data factors, and the data acquisition path uses the data factors as data acquisition indexes. The data acquisition path is used to establish a connection between the data middle platform and the data source platform.

[0106] The data update and storage module 30 is used to obtain real-time updated data from the data source platform through the data acquisition path, locate the fuel management cycle according to the data factors, and construct a management cycle data block to store in the data platform.

[0107] The cycle feature analysis module 40 is used to perform feature analysis of the management cycle based on the management cycle data block to obtain cycle fuel characteristics.

[0108] The supply imbalance analysis module 50 is used to perform full-cycle supply imbalance analysis based on the supply balance impact relationship of the fuel management cycle and the fuel characteristics of the cycle, and to obtain imbalance nodes and imbalance characteristics.

[0109] The balance compensation module 60 is used to perform balance compensation based on the imbalance nodes and imbalance characteristics to obtain a full-cycle balance management strategy.

[0110] Furthermore, in this embodiment of the application, the management target parsing module 10 is also used to perform the following steps:

[0111] Obtain positive and negative examples of the fuel management cycle. Positive examples are cases where the evaluation result of the fuel management cycle reaches the target threshold, and negative examples are cases where the evaluation result does not reach the target threshold. Extract management target events from the positive and negative examples respectively, and perform impact factor analysis on the management target events to obtain positive and negative factor analysis results. Using the management target events as alignment conditions, perform factor alignment and fusion of the positive and negative factor analysis results, taking the management target events as the management target, and integrating the corresponding data types of the fused impact factors to obtain the data factors of the management target.

[0112] Furthermore, after extracting the management target events based on the positive and negative examples, the management target parsing module 10 is also used to perform the following steps:

[0113] The fuel management cycle is divided into operation steps to obtain the target events generated by each operation step, and the operation step target event set is summarized; the management target events are traversed by the operation step target event set to obtain the missing step target events; the management target events are supplemented by the missing step target events.

[0114] Furthermore, in this embodiment of the application, the management target parsing module 10 is also used to perform the following steps:

[0115] An event correlation analysis is performed on the management target events to identify intermediate events and basic events. Event factors are decomposed layer by layer according to the order from intermediate events to basic events to determine event influencing factors and their relationships. An event parse tree is constructed according to the hierarchical relationship of intermediate events – basic events – event influencing factors. Based on the event influencing factors and their relationships, the probability of event occurrence is verified layer by layer from bottom to top according to the event parse tree using logical gate relationships. When the probability threshold requirement is reached, the event influencing factors and their relationships are determined, and the factor analysis results are obtained.

[0116] Furthermore, when the probability threshold requirement is not met, the management target parsing module 10 is also used to perform the following steps:

[0117] An anomalous event is identified, which is an intermediate or basic event where the logic gate relationship verification result does not reach the probability threshold. Based on the anomalous event, a historical event record dataset is extracted, and semantic mining is performed on the historical event record dataset to establish a factor relationship graph to reflect the semantic associations between factors. Based on the factor relationship graph, dominant and latent factors are mined. The association and influence relationships between the dominant and latent factors and the anomalous event are fitted according to the historical event record dataset. These association and influence relationships are added to the logic gate relationship, and a bottom-up, reverse, layer-by-layer verification is performed based on the event parse tree until the event occurrence probability reaches the probability threshold requirement.

[0118] Furthermore, in this embodiment of the application, the supply imbalance analysis module 50 is also used to perform the following steps:

[0119] Set the periodic balance target relationship and penalty conditions for the fuel management cycle; analyze the characteristic relationships between fuel management cycles based on the periodic balance target relationship; construct an evaluation function based on the characteristic relationships and penalty conditions; use the evaluation function to evaluate the periodic fuel characteristics based on the balance target relationship to obtain the periodic node evaluation value; locate the imbalance node based on the periodic node evaluation value, and locate the imbalance characteristics based on the periodic balance target relationship.

[0120] Furthermore, in this embodiment of the application, the balance compensation module 60 is also used to perform the following steps:

[0121] Based on the cyclical fuel characteristics, short-term and long-term supply and demand are predicted, and supply and demand time-series chains are established. Based on the connection relationships of the fuel management cycle, the supply and demand time-series chains are time-aligned to construct supply and demand time-series chains between cycle nodes. Supply and demand balance relationships are predicted based on the supply and demand time-series chains to obtain predicted imbalance time-series nodes. Based on the predicted imbalance time-series nodes, time-series constraint conditions are analyzed according to time-series distance and the supply time-series characteristics of the cyclical fuel characteristics to obtain the minimum adjustment time window. The minimum adjustment time window is used as a constraint condition to perform balance compensation on the imbalance time-series nodes.

[0122] Through the foregoing detailed description of the fuel data intelligent management method based on a data platform, those skilled in the art can clearly understand that the fuel data intelligent management system based on a data platform in this embodiment, as corresponding to the system disclosed in Embodiment 2, has corresponding functional modules and beneficial effects as it corresponds to the method disclosed in Embodiment 1. For relevant details, please refer to the description in the method section.

[0123] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily 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 this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A fuel data intelligent management method based on a data platform, characterized in that: The method includes: Decompose the fuel management cycle, analyze the management objectives of the fuel management cycle, and obtain the data factors of the management objectives; A data acquisition path is established based on the data factors. The data acquisition path is mapped and associated with the data factors, and the data acquisition path uses the data factors as the data acquisition index. The data acquisition path is used to establish a connection between the data platform and the data source platform. The data acquisition channel obtains real-time updated data from the data source platform, the fuel management cycle is located based on the data factors, and a management cycle data block is constructed and stored in the data platform. Based on the management cycle data blocks, perform feature analysis of the management cycle to obtain cycle fuel characteristics; Based on the supply balance impact relationship of the fuel management cycle, a full-cycle supply imbalance analysis is performed according to the cycle fuel characteristics to obtain imbalance nodes and imbalance characteristics; Based on the imbalance nodes and imbalance characteristics, balance compensation is performed to obtain a full-cycle balance management strategy; Analyze the management objectives of the fuel management cycle to obtain the data factors of the management objectives, including: Obtain positive and negative examples of the fuel management cycle. The positive examples are cases where the evaluation result of the fuel management cycle reaches the target threshold, and the negative examples are cases where the evaluation result does not reach the target threshold. Based on the positive and negative examples, management target events are extracted, and influence factor analysis is performed on the management target events to obtain positive and negative example factor analysis results. Using the management target event as the alignment condition, the positive example factor analysis results and the negative example factor analysis results are fused together, the management target event is taken as the management target, and the corresponding data types of the fused influencing factors are integrated to obtain the data factors of the management target. Based on the positive and negative examples, the management target events are extracted respectively, followed by: The fuel management cycle is divided into operation steps to obtain the target events generated by each operation step, and the target event set of operation steps is summarized. The missing step target events are obtained by performing a missing traversal on the management target events using the target event set of the operation steps. The missing step target event is used to supplement the management target event; Influence factor analysis was performed on the management target event to obtain positive and negative factor analysis results, including: Perform event correlation analysis on the aforementioned management target events to identify intermediate events and basic events; According to the order from the intermediate events to the basic events, the event factors are decomposed layer by layer to determine the event influencing factors and their influencing relationships. Construct an event parse tree according to the hierarchical relationship of intermediate events - basic events - event influencing factors; Based on the event impact factors and their relationships, the probability of event occurrence is verified layer by layer from bottom to top according to the event parse tree in accordance with the logical gate relationship. When the probability threshold requirement is reached, the event impact factors and their relationships are determined, and the factor analysis results are obtained.

2. The intelligent fuel data management method based on a data platform as described in claim 1, characterized in that, When the probability threshold requirement is not met, the method further includes: Obtain abnormal events, which are intermediate or basic events where the logic gate relationship verification result does not reach the probability threshold; Based on the abnormal events, a historical event record dataset is extracted, and semantic mining is performed on the historical event record dataset to establish a factor relationship graph to reflect the semantic associations between factors. Based on the aforementioned factor relationship graph, dominant and latent factors are identified. The correlation and influence relationship between the dominant factor, the latent factor, and the anomalous event is fitted based on the historical event record dataset. The associated influence relationship is added to the logical gate relationship, and a bottom-up reverse layer-by-layer verification is performed based on the event parsing tree until the probability of the event occurs reaches the probability threshold requirement.

3. The intelligent fuel data management method based on a data platform as described in claim 1, characterized in that, Based on the supply balance impact relationship of the aforementioned fuel management cycle, a full-cycle supply imbalance analysis is performed according to the cycle fuel characteristics to obtain imbalance nodes and imbalance characteristics, including: Set the cyclical balance target relationship and penalty conditions for the fuel management cycle; Based on the described periodic balance target relationship, analyze the characteristic relationships between fuel management cycles; Construct an evaluation function based on the aforementioned characteristic relationships and penalty conditions; The evaluation function is used to evaluate the balance target relationship of the periodic fuel characteristics to obtain the periodic node evaluation value; Based on the periodic node evaluation value, the imbalance node is located, and the imbalance characteristics are located by combining the periodic balance target relationship.

4. The intelligent fuel data management method based on a data platform as described in claim 3, characterized in that, Also includes: Based on the characteristics of the cyclical fuel, long-term and short-term supply and demand are predicted, and supply and demand time series chains are established. Based on the connection relationship of the fuel management cycle, the supply time sequence chain and the demand time sequence chain are aligned to construct the supply and demand time sequence chain between cycle nodes. Based on the supply and demand time series chain, the supply and demand balance relationship is predicted to obtain the predicted imbalance time series node; Based on the predicted imbalance time series nodes, the time series constraints are analyzed according to the time series distance and the supply time series characteristics of the periodic fuel characteristics to obtain the minimum adjustment time window; The minimum adjustment time window is used as a constraint to balance and compensate for unbalanced time sequence nodes.

5. A fuel data intelligent management system based on a data middle platform, characterized in that: The system is used to execute the fuel data intelligent management method based on a data platform as described in any one of claims 1-4, including: The management cycle decomposition module is used to decompose the fuel management cycle, analyze the management objectives of the fuel management cycle, and obtain the data factors of the management objectives. A data acquisition path establishment module is used to establish a data acquisition path based on the data factors. The data acquisition path is mapped and associated with the data factors, and the data acquisition path uses the data factors as data acquisition indexes. The data acquisition path is used to establish a connection between the data middle platform and the data source platform. The data update and storage module is used to obtain real-time updated data from the data source platform through the data acquisition channel, locate the fuel management cycle according to the data factors, and construct a management cycle data block to store in the data platform. The cycle feature analysis module is used to perform feature analysis of the management cycle based on the management cycle data block to obtain cycle fuel characteristics. The supply imbalance analysis module is used to perform full-cycle supply imbalance analysis based on the supply balance impact relationship of the fuel management cycle and the fuel characteristics of the cycle, and to obtain imbalance nodes and imbalance characteristics. The balance compensation module is used to perform balance compensation based on the imbalance nodes and imbalance characteristics to obtain a full-cycle balance management strategy.

Images