Gas pipe network operation data intelligent management system

Abstract

The invention discloses an intelligent management system for operation data of a gas pipe network, and relates to the related field of natural gas data management, and the system comprises a multi-dimensional dynamic monitoring module which carries out the multi-dimensional dynamic monitoring of a target gas pipe network; the first factor set acquisition module acquires a first factor set of a first fault and acquires first weight distribution; a first factor parameter set acquisition module traverses a first factor set in real-time operation information to obtain a first factor parameter set, and first fitness is obtained by combining first weight distribution; the intelligent fault prediction module takes the real-time fitness set as input information of an intelligent fault prediction model; if the target prediction fitness does not conform to the preset fine granularity threshold value, the first early warning signal sending module sends out early warning; and the maintenance and inspection pause management module performs maintenance and inspection pause management. The technical problems of low response speed and inaccurate fault prediction in existing gas pipe network operation management are solved, and the technical effects of increasing the response speed and improving the prediction accuracy are achieved.

Description

Technical Field

[0001] This application relates to the field of natural gas data management, and in particular to an intelligent management system for the operation data of gas pipe networks. Background Art

[0002] In the process of urbanization, as an important part of urban infrastructure, the safe and stable operation of gas pipe networks is directly related to the energy supply for residents' lives and industrial production. With the rapid development of technologies such as the Internet of Things, big data, and artificial intelligence, the intelligent management of gas pipe networks has become an important means to improve the operation efficiency of pipe networks and ensure the safety of pipe networks. Traditional gas pipe network management mainly relies on manual inspections. Potential faults are discovered by regularly checking pipe network facilities and recording operation data. This method is time-consuming and laborious, and is easily affected by human factors, making it difficult to discover and handle faults in a timely manner. Some management systems attempt to perform fault prediction based on historical fault data and expert experience. However, due to the lack of scientific data analysis models and algorithm support, the prediction results are often inaccurate and difficult to provide effective guidance for pipe network maintenance.

[0003] In the related technologies at the present stage, there are technical problems of slow response speed and inaccurate fault prediction in the operation management of gas pipe networks. Summary of the Invention

[0004] This application provides an intelligent management system for the operation data of gas pipe networks. By adopting the technical means of obtaining the operation information of the target gas pipe network in real time, determining the relevant factor set of the first fault and assigning weights to it, searching for these factors in the real-time operation information, calculating the fitness by combining the weights, inputting the fitness set into the intelligent fault prediction model to obtain the prediction result, if the prediction result does not meet the threshold, sending out a warning signal, and based on the warning signal, suspending and performing maintenance inspections on the pipe network, etc., the technical effects of accelerating the response speed and improving the prediction accuracy are achieved.

[0005] The present application provides an intelligent management system for the operation data of a gas pipeline network, including: a multi-dimensional dynamic monitoring module for performing multi-dimensional dynamic monitoring on a target gas pipeline network to obtain real-time operation information; a first factor set acquisition module for acquiring a first factor set of a first fault and obtaining a first weight assignment of the first factor set based on a predetermined weight assignment strategy; a first factor parameter set acquisition module for traversing the first factor set in the real-time operation information to obtain a first factor parameter set and combining the first weight assignment to obtain a first fitness; an intelligent fault prediction module for using the real-time fitness set formed based on the first fitness as input information of an intelligent fault prediction model and obtaining output information through the intelligent fault prediction model; a first warning signal sending module for sending a first warning signal if a target prediction fitness in the output information does not meet a predetermined fine-grained threshold; and a suspension maintenance management module for performing suspension maintenance management on the target gas pipeline network based on the first warning signal.

[0006] In a possible implementation manner, the first factor set acquisition module includes: a historical fault data set acquisition unit for acquiring a historical fault data set of a gas pipeline network of the same type as the target gas pipeline network; a first data set acquisition unit for traversing and matching the first fault in the historical fault data set to obtain a first data set corresponding to the first fault; a first fault tree acquisition unit for forming a first training data group based on the first data set and analyzing the first training data group to obtain a first fault tree; a first fault tree inspection unit for forming a first inspection data group based on the first data set and inspecting the first fault tree according to the first inspection data group to obtain a first accuracy rate; a second training data group screening unit for screening out data in the first training data group that does not meet a predetermined accuracy rate threshold with the first accuracy rate as a constraint as a second training data group; a first data volume judgment unit for obtaining a first data volume of the second training data group and judging whether the first data volume is less than a predetermined threshold; a target fault tree determination unit for, if the first data volume is less than the predetermined threshold, taking the first fault tree as a target fault tree; and a first factor set formation unit for obtaining a first minimum cut set of the first fault according to the target fault tree and forming the first factor set of the first fault according to the first minimum cut set.

[0007] In a possible implementation, the first factor set acquisition module further includes: a second fault tree acquisition unit, configured to analyze the second training data set to obtain a second fault tree if the first data volume is greater than or equal to the predetermined threshold; a second fault tree verification unit, configured to form a second verification data set based on the first data set and verify the second fault tree according to the second verification data set to obtain a second accuracy rate; a third training data set screening unit, configured to screen out the data in the second training data set that does not meet the predetermined accuracy rate threshold with the second accuracy rate as a constraint, as the third training data set; a second data volume judgment unit, configured to obtain a second data volume of the third training data set and judge whether the second data volume is less than the predetermined threshold; a fault tree merging unit, configured to merge the first fault tree and the second fault tree to obtain the target fault tree if the second data volume is less than the predetermined threshold.

[0008] In a possible implementation, the first factor set acquisition module includes: a first historical factor parameter set matching unit, configured to match a first historical factor parameter set corresponding to the first factor set in the first data set; a correlation analysis unit, configured to perform a correlation analysis on the first historical factor parameter set and the first fault to obtain a first correlation index set; a first weight assignment acquisition unit, configured to use the normalized first correlation index set as the corresponding weight of the first factor set according to the predetermined weight assignment strategy to obtain the first weight assignment.

[0009] In a possible implementation, the system further includes: a first initial undirected structure diagram drawing module, configured to draw a first initial undirected structure diagram of the first fault according to the first factor set and the first weight assignment; a dimensionality reduction processing module, configured to introduce a matrix eigenvalue decomposition mechanism to perform dimensionality reduction processing on the first initial undirected structure diagram to obtain a first low-dimensional undirected structure diagram; a first factor set adjustment module, configured to adjust the first factor set according to the first low-dimensional undirected structure diagram to obtain a first target factor set and obtain a first target weight assignment of the first target factor set; a first fitness adjustment module, configured to adjust the first fitness based on the first target factor set and the first target weight assignment.

[0010] In a possible implementation, the system further includes: a judgment module, configured to judge whether the first fitness meets a predetermined coarse-grained threshold; a second warning signal sending module, configured to send a second warning signal if the first fitness does not meet the predetermined coarse-grained threshold; an operation and maintenance management module, configured to perform operation and maintenance management on the first fault of the target gas pipeline network based on the second warning signal.

[0011] In a possible implementation manner, the second warning signal sending module includes: a signal strength setting unit, configured to make the second warning signal weaker in strength than the first warning signal.

[0012] In a possible implementation manner, the intelligent fault prediction module includes: a machine learning unit, configured to perform machine learning on a first historical data group formed by a first historical factor parameter set and a first historical fitness in the first data set to obtain a first fitness predictor; a meta-predictor acquisition unit, configured to perform machine learning on a second historical data group formed by a first predicted fitness output by the first fitness predictor and the historical fitness of the same type of gas pipeline network to obtain a meta-predictor; and an integration building unit, configured to use the first fitness predictor as a primary predictor and perform integrated building with the meta-predictor to obtain the intelligent fault prediction model.

[0013] It is intended to propose an intelligent management system for gas pipeline network operation data through this application. The target gas pipeline network is dynamically monitored in multiple dimensions by a multi-dimensional dynamic monitoring module to obtain real-time operation information. A first factor set of a first fault is obtained by a first factor set acquisition module, and a first weight assignment of the first factor set is obtained based on a predetermined weight assignment strategy. A first factor parameter set is obtained by traversing the first factor set in the real-time operation information by a first factor parameter set acquisition module, and a first fitness is obtained by combining the first weight assignment. The intelligent fault prediction module uses a real-time fitness set formed based on the first fitness as input information of the intelligent fault prediction model, and obtains output information through the intelligent fault prediction model. If the target predicted fitness in the output information does not meet a predetermined fine-grained threshold, a first warning signal is sent by a first warning signal sending module, and based on the first warning signal, a suspension maintenance management module performs suspension maintenance management on the target gas pipeline network, achieving the technical effects of accelerating the response speed and improving the prediction accuracy. Description of the Drawings

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments of the present invention will be briefly introduced below. Structure diagrams are used in this application to illustrate the system composition of the embodiments of this application. It should be understood that the operations before or below do not necessarily need to be executed precisely in sequence. On the contrary, according to needs, they can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0015] Figure 1 It is a schematic structural diagram of an intelligent management system for gas pipeline network operation data provided by an embodiment of this application.

[0016] Figure 2Schematic diagram of the first factor set acquisition module in an intelligent management system for gas pipeline network operation data provided by an embodiment of the present application.

[0017] Explanation of reference numerals: multi-dimensional dynamic monitoring module 10, first factor set acquisition module 20, first factor parameter set acquisition module 30, intelligent fault prediction module 40, first warning signal sending module 50, suspension maintenance management module 60. Detailed implementation manners

[0018] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically gives the detailed implementation manners of the present application.

[0019] In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0020] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The terms "first / second" involved are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0021] An embodiment of the present application provides an intelligent management system for gas pipeline network operation data, as Figure 1 shown. The system includes: A multi-dimensional dynamic monitoring module 10, configured to perform multi-dimensional dynamic monitoring on a target gas pipeline network to obtain real-time operation information.

[0022] Specifically, the multi-dimensional dynamic monitoring module 10 utilizes sensor networks, Internet of Things (IoT) technology, remote communication technology, etc. to conduct real-time monitoring of the target gas pipeline network. The monitoring content includes but is not limited to pipeline pressure, flow rate, temperature, leakage detection (such as using acoustic or optical sensors), valve status, etc. The data is transmitted to the central processing unit via wireless or wired means to form real-time operation information.

[0023] The first factor set acquisition module 20 is used to acquire the first factor set of the first fault and obtain the first weight assignment of the first factor set based on a predetermined weight assignment strategy.

[0024] Specifically, the first factor set acquisition module 20 determines a set of key factors (the first factor set) that may cause the first fault (a specific fault) based on historical fault data. According to a predetermined weight assignment strategy (methods and rules for determining factor weights, which can be based on statistical analysis, machine learning algorithms, etc.), weights are assigned to each factor to reflect its importance for the occurrence of the fault.

[0025] As Figure 2 shown, in a possible implementation, the first factor set acquisition module 20 includes: a historical fault data set acquisition unit for acquiring the historical fault data set of the same type of gas pipeline network of the target gas pipeline network; a first data set acquisition unit for traversing and matching the first fault in the historical fault data set to obtain the first data set corresponding to the first fault; a first fault tree acquisition unit for forming a first training data group based on the first data set and analyzing the first training data group to obtain a first fault tree; a first fault tree inspection unit for forming a first inspection data group based on the first data set and inspecting the first fault tree according to the first inspection data group to obtain a first accuracy rate; a second training data group screening unit for screening out the data in the first training data group that does not meet a predetermined accuracy threshold with the first accuracy rate as a constraint as the second training data group; a first data volume judgment unit for obtaining the first data volume of the second training data group and judging whether the first data volume is less than a predetermined threshold; a target fault tree determination unit for, if the first data volume is less than the predetermined threshold, taking the first fault tree as the target fault tree; a first factor set formation unit for obtaining the first minimum cut set of the first fault according to the target fault tree and forming the first factor set of the first fault according to the first minimum cut set.

[0026] Specifically, the historical fault dataset acquisition unit obtains the historical fault dataset of gas pipelines of the same type as the target gas pipeline network from the storage system through methods such as database query, file reading, or API interface call. These datasets contain detailed information such as fault type, occurrence time, fault location, impact range, repair measures, etc. The first dataset acquisition unit traverses the historical fault dataset and uses a matching algorithm (such as string matching, regular expression matching, etc.) to find historical records related to the first fault (i.e., the currently concerned fault type), and forms the first dataset. That is, the first dataset is a data set containing historical fault records related to the first fault.

[0027] The first fault tree acquisition unit uses the fault tree analysis (FTA) method to construct the first training data group (the data set for constructing the first fault tree) based on the first dataset, and forms the first fault tree through the connection relationships of logic gates (such as AND gates, OR gates) and events (basic events, intermediate events, top events). The first fault tree is a tree diagram describing the causes of the first fault and their logical relationships. Among them, fault tree analysis is a fault analysis method based on a logical diagram, which is used to identify various possible causes leading to system failures and their combinations.

[0028] The first fault tree verification unit extracts part of the data from the first dataset to form the first verification data group, and constructs a verification scenario based on each record in the first verification data group. This scenario can simulate the occurrence process of an actual fault, including the cause of the fault, triggering conditions, impact range, etc. Input each verification scenario into the verification logic, and the verification logic will verify the cause of the fault and the logical relationship in the scenario according to the logical structure of the first fault tree, including the operation of logic gates, the triggering order of events, probability calculation, etc. Compare the output result of the verification logic with the actual record in the first verification data group. If the output result is consistent with the actual record, it means that the first fault tree is accurate in this scenario, and it is recorded as "correct"; if not, it is recorded as "wrong". Count the number of "correct" records and the total number of records, and calculate the ratio between the two to obtain the first accuracy rate. That is, the first accuracy rate is the correct prediction ratio of the first fault tree on the verification data group.

[0029] The second training data set screening unit screens out the data in the first training data set that does not meet the predetermined accuracy threshold with the first accuracy as the constraint condition, and uses it as the second training data set. Specifically, a predetermined accuracy threshold is set, the data items in the first training data set are traversed, the accident tree verification accuracy corresponding to each item of data is calculated, and the data items with an accuracy lower than the predetermined threshold are screened out to form the second training data set. The first data volume judgment unit calculates the first data volume of the second training data set and compares it with a predetermined threshold (the lowest standard for judging whether the data volume is sufficient). If the first data volume is less than the predetermined threshold, it is considered that the accuracy of the first accident tree is high enough and can be directly used as the target accident tree. The target accident tree determination unit directly uses the first accident tree as the target accident tree, and the target accident tree is the accident tree finally determined for fault prediction and analysis.

[0030] The first factor set construction unit obtains the first minimum cut set of the first fault (that is, the minimum set of events that cause the first fault to occur) according to the target accident tree, and constructs the first factor set of the first fault according to the first minimum cut set. That is, each event in the first minimum cut set represents a potential fault factor, and the first factor set is a set of fault factors composed of the events in the first minimum cut set. This implementation method accurately identifies the key factors leading to the first fault by obtaining and analyzing the historical fault data set and combining the accident tree method, improving the accuracy of the prediction model.

[0031] In a possible implementation manner, the first factor set obtaining module 20 further includes: a second accident tree obtaining unit, configured to, if the first data volume is greater than or equal to the predetermined threshold, analyze the second training data set to obtain a second accident tree; a second accident tree verification unit, configured to construct a second verification data set based on the first data set and verify the second accident tree according to the second verification data set to obtain a second accuracy; a third training data set screening unit, configured to screen out the data in the second training data set that does not meet the predetermined accuracy threshold with the second accuracy as the constraint, and use it as the third training data set; a second data volume judgment unit, configured to obtain the second data volume of the third training data set and judge whether the second data volume is less than the predetermined threshold; an accident tree merging unit, configured to, if the second data volume is less than the predetermined threshold, merge the first accident tree and the second accident tree to obtain the target accident tree.

[0032] Specifically, when the first data volume is greater than or equal to a predetermined threshold, it indicates that the first fault tree may not be accurate enough or there are situations that need further optimization. At this time, the second fault tree acquisition unit analyzes the second training data set (i.e., the data set in the first training data set that does not meet the predetermined accuracy threshold), and uses a method similar to that of the first fault tree acquisition unit to obtain the second fault tree. The second fault tree verification unit constructs a second verification data set (a data set used to verify the accuracy of the second fault tree, which does not overlap with the first verification data set) based on the first data set, and verifies the second fault tree using the verification logic to obtain the second accuracy rate. The third training data set screening unit screens the data in the second training data set that does not meet the predetermined accuracy threshold with the second accuracy rate as a constraint condition to obtain the third training data set. The second data volume judgment unit calculates the second data volume of the third training data set and compares it with the predetermined threshold. When the second data volume is less than the predetermined threshold, it is considered that the screened data is already small enough, and both the first fault tree and the second fault tree contain valuable information. At this time, the fault tree merging unit merges the first fault tree and the second fault tree to obtain the target fault tree. The merging process includes the merging of logic gates, the integration of basic events, etc. This implementation method can better capture the diversity and complexity of fault occurrence by constructing and merging multiple fault trees, thereby improving the accuracy of fault prediction.

[0033] In a possible implementation manner, the first factor set acquisition module 20 includes: a first historical factor parameter set matching unit, configured to match the first historical factor parameter set corresponding to the first factor set in the first data set; a correlation analysis unit, configured to perform a correlation analysis on the first historical factor parameter set and the first fault to obtain a first correlation index set; a first weight assignment acquisition unit, configured to use the predetermined weight assignment strategy to use the normalized first correlation index set as the corresponding weight of the first factor set to obtain the first weight assignment.

[0034] Specifically, the first historical factor parameter set matching unit searches for the historical factor parameter set that matches the first factor set in the first data set. By traversing the first data set, records containing all the factors in the first factor set are found, and the factor parameter values in these records are extracted to form the first historical factor parameter set. The correlation analysis unit uses statistical analysis methods, such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc., to perform correlation analysis on the first historical factor parameter set and the first fault, calculates the correlation index between each factor and the first fault, and forms the first correlation index set with the calculated correlation indices. The first correlation index set is normalized, such as min-max normalization, Z-score normalization, etc., so that its value range is between 0 and 1 (or other specified value ranges). The first weight assignment obtaining unit directly uses the normalized correlation index as the weight according to the predetermined weight assignment strategy, or makes further adjustments to obtain the first weight assignment. The first weight assignment is the weight value assigned to each factor in the first factor set. This implementation method can assign reasonable weights to each factor by extracting factor parameters related to the fault from historical data and performing correlation analysis, which helps the intelligent fault prediction model to more accurately identify the patterns and trends of fault occurrence and improve the accuracy of prediction.

[0035] The first factor parameter set obtaining module 30 is configured to traverse the first factor set in the real-time operation information to obtain the first factor parameter set, and combine the first weight assignment to obtain the first fitness.

[0036] Specifically, the first factor parameter set obtaining module 30 searches for the parameter values corresponding to the first factor set in the real-time operation information to form the first factor parameter set. According to the weight of each factor, the first fitness is calculated. The first fitness is used to measure the matching degree between the current state and the first fault occurrence condition, and is a weighted sum or weighted product, which is used to quantify the proximity between the current state and the fault occurrence.

[0037] In a possible implementation manner, the system further includes: a first initial undirected structure diagram drawing module, configured to draw a first initial undirected structure diagram of the first fault according to the first factor set and the first weight assignment; a dimensionality reduction processing module, configured to introduce a matrix eigenvalue decomposition mechanism to perform dimensionality reduction processing on the first initial undirected structure diagram to obtain a first low-dimensional undirected structure diagram; a first factor set adjustment module, configured to adjust the first factor set according to the first low-dimensional undirected structure diagram to obtain a first target factor set, and obtain a first target weight assignment for the first target factor set; a first fitness adjustment module, configured to adjust the first fitness based on the first target factor set and the first target weight assignment.

[0038] Specifically, the first initial undirected structure diagram drawing module, based on the first factor set and its weight assignment (i.e., the first weight assignment), takes the factors as nodes and the weights as the association degrees between nodes (which can be direct connections or weighted connections) to draw the first initial undirected structure diagram of the first fault. Among them, the undirected structure diagram is a data structure in graph theory used to represent the undirected relationship between nodes. The dimensionality reduction processing module performs matrix processing (such as adjacency matrix) on the first initial undirected structure diagram, and then applies the matrix eigenvalue decomposition mechanism for dimensionality reduction processing to reduce the complexity of the data while retaining key information. The result after dimensionality reduction is the first low-dimensional undirected structure diagram. Among them, matrix eigenvalue decomposition is a mathematical method used to decompose a matrix into the form of eigenvectors and eigenvalues for operations such as dimensionality reduction and data compression.

[0039] The first factor set adjustment module, based on the first low-dimensional undirected structure diagram, identifies the key factors (i.e., nodes with larger weights or higher connection degrees) that have a greater impact on the first fault, and the redundant factors (i.e., nodes with smaller weights or lower connection degrees). According to this information, the first factor set is adjusted, retaining the key factors and removing the redundant factors to obtain the first target factor set, and the weight assignment of the first target factor set (i.e., the first target weight assignment) is recalculated. The first fitness adjustment module traverses the first target factor parameter set in the real-time operation information according to the first target factor set, and recalculates the first target fitness in combination with the first target weight assignment. This implementation method, by introducing graph theory and the matrix eigenvalue decomposition mechanism, conducts a more in-depth analysis and optimization of the first factor set, thereby identifying the key factors that have the greatest impact on the first fault and recalculating the fitness. This new fitness value can more accurately reflect the actual situation of the first fault in the target gas pipeline network, thus improving the accuracy of prediction.

[0040] In a possible implementation manner, the system further includes: a judgment module for judging whether the first fitness meets a predetermined coarse-grained threshold; a second warning signal sending module for sending a second warning signal if the first fitness does not meet the predetermined coarse-grained threshold; and an operation and maintenance management module for performing operation and maintenance management on the first fault of the target gas pipeline network based on the second warning signal.

[0041] Specifically, the judgment module receives the first fitness value from the first factor parameter set acquisition module 30 and compares it with a predetermined coarse-grained threshold. The coarse-grained threshold is a preset threshold for evaluating the risk of the first failure (such as leakage, explosion, etc.) of the gas pipeline network, which is set based on historical data or industry standards and is different from the fine-grained threshold (for the overall failure risk). If the judgment module determines that the first fitness does not meet the predetermined coarse-grained threshold, it indicates that there is a risk of the first failure in the target gas pipeline network. At this time, the second warning signal sending module triggers a warning signal. The second warning signal is a signal used to prompt the gas pipeline network management personnel to pay attention to the risk of the first failure and is different from the first warning signal for the overall failure. Based on the second warning signal, the operation and maintenance management module starts the corresponding operation, maintenance, and management processes, that is, it does not need to suspend the target gas pipeline network, and conducts maintenance and repair under normal operating conditions, including measures such as increasing the monitoring frequency, adjusting the operating parameters, and arranging professional personnel for on-site inspections, to ensure the safe and stable operation of the gas pipeline network. Compared with the overall failure risk assessment, the coarse-grained risk assessment mechanism for specific failures is more targeted and can more accurately reflect the operating status of specific parts or components in the pipeline network. This implementation method can issue a warning signal in a timely manner when there is a risk of a specific failure in the gas pipeline network by setting a coarse-grained threshold and making a judgment, thereby avoiding or reducing the impact of possible failures on the operation of the pipeline network and improving the safety of the pipeline network.

[0042] In a possible implementation manner, the second warning signal sending module includes: a signal intensity setting unit, configured to make the second warning signal weaker in intensity than the first warning signal.

[0043] Specifically, inside the second warning signal sending module, the signal intensity setting unit sets the intensity of the second warning signal to be weaker than the first warning signal according to a preset rule or strategy to achieve the distinction of the warning signal levels. This implementation method can more effectively distinguish the urgency and importance of failures by setting warning signals with different intensities, which helps the management personnel to prioritize the processing of more urgent or important failures according to the intensity of the warning signal, improving the efficiency and accuracy of failure response.

[0044] The intelligent fault prediction module 40 is configured to use the real-time fitness set formed based on the first fitness as the input information of the intelligent fault prediction model and obtain the output information through the intelligent fault prediction model.

[0045] Specifically, the intelligent fault prediction module 40 uses machine learning or deep learning models (such as neural networks, support vector machines) as the intelligent fault prediction model, and the intelligent fault prediction model is used to predict the possibility of fault occurrence. For each type of fault (including the first fault), the fitness value is calculated according to its real-time factor parameter set and weight. These fitness values are combined into a real-time fitness set. Taking the real-time fitness set as the input, the model outputs the prediction result, including the target prediction fitness, indicating the possibility of fault occurrence within a future period of time.

[0046] In a possible implementation, the intelligent fault prediction module 40 includes: a machine learning unit, configured to perform machine learning on a first historical data group formed by a first historical factor parameter set and a first historical fitness in the first dataset to obtain a first fitness predictor; a meta-predictor acquisition unit, configured to perform machine learning on a second historical data group formed by the first predicted fitness output by the first fitness predictor and the historical fitness of the same type of gas pipeline network to obtain a meta-predictor; and an integrated construction unit, configured to use the first fitness predictor as a primary predictor and integrate it with the meta-predictor to obtain the intelligent fault prediction model.

[0047] Specifically, the machine learning unit extracts the first historical factor parameter set and the corresponding first historical fitness from the first dataset to form a first historical data group. Using the method of supervised learning, the machine learning model is trained with the first historical data group to obtain a first fitness predictor. Among them, each fault type corresponds to a first fitness predictor.

[0048] The meta-predictor acquisition unit also uses the method of supervised learning, with the predicted fitness corresponding to multiple fault types as the input and the overall fault fitness as the output. It is a higher-level prediction task, that is, to predict the prediction result of the fitness predictor. The meta-predictor acquisition unit matches the multiple first predicted fitnesses output by multiple first fitness predictors with the historical fitness of the same type of gas pipeline network to form a second historical data group. Among them, the historical fitness can be the overall fault fitness after aggregation or averaging processing. Using the second historical data group to train the meta-learning model to obtain a meta-predictor. The meta-predictor is used to learn how to combine the outputs of multiple fitness predictors to more accurately predict the overall fault fitness.

[0049] The integrated construction unit uses the method of model integration, taking the outputs of multiple primary predictors (i.e., the first fitness predictors) as the input of the meta-predictor, and the output of the meta-predictor as the final output of the intelligent fault prediction model. This implementation can fully utilize the correlation information between different fault types and improve the prediction accuracy by training a separate predictor for each fault type and integrating it with the meta-predictor.

[0050] The first warning signal issuing module 50 is configured to issue a first warning signal if the target prediction fitness in the output information does not meet a predetermined fine-grained threshold.

[0051] Specifically, the first warning signal issuing module 50 compares the target prediction fitness with a predetermined fine-grained threshold, which is a threshold set based on historical data and risk assessment for determining whether a warning needs to be issued. If the target prediction fitness does not meet (e.g., is higher than) the threshold, the warning mechanism is triggered to issue a first warning signal, which is a warning signal to notify relevant personnel to take actions.

[0052] The suspension maintenance management module 60 is configured to perform suspension maintenance management on the target gas pipeline network based on the first warning signal.

[0053] Specifically, the suspension maintenance management module 60 responds to the first warning signal and automatically starts the suspension maintenance management process, that is, suspends use and conducts maintenance inspections, including measures such as emergency valve closure, dispatching repair teams, and notifying relevant departments, to ensure that the faults are properly handled. In the embodiments of the present application, the running information of the target gas pipeline network is obtained in real time, the relevant factor set of the first fault is determined and weights are assigned to it, these factors are searched in the real-time running information, the fitness is calculated in combination with the weights, the fitness set is input into the intelligent fault prediction model to obtain a prediction result. If the prediction result does not meet the threshold, a warning signal is issued, and based on the warning signal, technical means such as suspending and repairing the pipeline network are carried out, achieving the technical effects of accelerating the response speed and improving the prediction accuracy.

[0054] Although various references are made to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or the server. The included individual units and modules are only divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for facilitating mutual distinction and do not limit the protection scope of the present invention.

[0055] The above specific embodiments do not constitute a limitation on the protection scope of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the protection scope of this application. In some cases, the actions or steps recited in this application can be executed in a different order from that in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A gas network operation data intelligent management system, characterized in that: include: Multi-dimensional dynamic monitoring module, used to conduct multi-dimensional dynamic monitoring of the target gas pipeline network to obtain real-time operation information; A first factor set acquisition module, configured to acquire a first factor set of a first fault, and acquire a first weight allocation of the first factor set based on a predetermined weight allocation strategy; A first factor parameter set acquisition module, configured to traverse the first factor set in the real-time operation information to obtain a first factor parameter set, and obtain a first fitness in combination with the first weight distribution; An intelligent fault prediction module, used to use the real-time fitness set formed based on the first fitness as input information of an intelligent fault prediction model, and obtain output information through the intelligent fault prediction model; A first warning signal issuing module, configured to issue a first warning signal if the target prediction fitness in the output information does not meet a predetermined fine-grained threshold; A maintenance suspension management module is used to perform maintenance suspension management on the target gas pipeline network based on the first warning signal.

2. A gas network operation data intelligent management system according to claim 1, characterized in that: The first factor set acquisition module includes: A historical fault data set acquisition unit, used to acquire a historical fault data set of a gas pipeline network of the same type as the target gas pipeline network; A first data set acquisition unit, configured to traverse and match the first fault in the historical fault data set to obtain a first data set corresponding to the first fault; A first fault tree acquisition unit, configured to form a first training data group based on the first data set, and analyze the first training data group to obtain a first fault tree; a first fault tree verification unit, configured to form a first verification data group based on the first data set, and verify the first fault tree according to the first verification data group to obtain a first accuracy rate; A second training data group screening unit, configured to screen the data in the first training data group that does not meet a predetermined accuracy threshold using the first accuracy rate as a constraint, as a second training data group; a first data volume determination unit, configured to obtain a first data volume of the second training data group, and determine whether the first data volume is less than a predetermined threshold; a target fault tree determining unit, configured to use the first fault tree as a target fault tree if the first data amount is less than the predetermined threshold; The first factor set forming unit is used to obtain a first minimum cut set of the first fault according to the target fault tree, and to form the first factor set of the first fault according to the first minimum cut set.

3. A gas network operation data intelligent management system as claimed in claim 2, characterized in that: The first factor set acquisition module further includes: A second fault tree acquisition unit, configured to analyze the second training data set to obtain a second fault tree if the first data volume is greater than or equal to the predetermined threshold; a second fault tree verification unit, which forms a second verification data group based on the first data set, and verifies the second fault tree according to the second verification data group to obtain a second accuracy rate; A third training data group screening unit, configured to screen the data in the second training data group that does not meet the predetermined accuracy threshold using the second accuracy rate as a constraint, as a third training data group; a second data volume determination unit, which obtains a second data volume of the third training data group and determines whether the second data volume is less than the predetermined threshold; The fault tree merging unit is used to merge the first fault tree and the second fault tree to obtain the target fault tree if the second data amount is less than the predetermined threshold.

4. A gas network operation data intelligent management system as claimed in claim 2, characterized in that: The first factor set acquisition module includes: A first historical factor parameter set matching unit, configured to match a first historical factor parameter set corresponding to the first factor set in the first data set; A correlation analysis unit, configured to perform a correlation analysis on the first historical factor parameter set and the first fault to obtain a first correlation index set; The first weight allocation acquisition unit is used to use the normalized first correlation index set as the corresponding weight of the first factor set to obtain the first weight allocation according to the predetermined weight allocation strategy.

5. A gas network operation data intelligent management system as claimed in claim 4, characterized in that: The system further comprises: A first initial undirected structure graph drawing module, configured to draw a first initial undirected structure graph of the first fault according to the first factor set and the first weight distribution; A dimensionality reduction processing module, used for introducing a matrix feature decomposition mechanism to perform dimensionality reduction processing on the first initial undirected structure graph to obtain a first low-dimensional undirected structure graph; A first factor set adjustment module, configured to adjust the first factor set according to the first low-dimensional undirected structure graph to obtain a first target factor set, and acquire a first target weight allocation of the first target factor set; The first fitness adjustment module is used to adjust the first fitness based on the first target factor set and the first target weight distribution.

6. A gas network operation data intelligent management system according to claim 1, characterized in that: The system further comprises: A judging module, used for judging whether the first fitness meets a predetermined coarse-grained threshold; A second warning signal issuing module, configured to issue a second warning signal if the first fitness does not meet the predetermined coarse-grained threshold; An operation and maintenance management module is used to perform operation and maintenance management on the first fault of the target gas pipeline network based on the second early warning signal.

7. A gas network operation data intelligent management system as claimed in claim 6, characterized in that: The second warning signal issuing module includes: A signal strength setting unit is used to set the strength of the second warning signal to be weaker than that of the first warning signal.

8. The intelligent management system for gas network operation data according to claim 2, characterized in that: The intelligent fault prediction module comprises: a machine learning unit, configured to perform machine learning on a first historical data group formed based on a first historical factor parameter set and a first historical fitness in the first data set to obtain a first fitness predictor; a meta-predictor acquisition unit, configured to perform machine learning on a second historical data set formed by a first predicted fitness output by the first fitness predictor and a historical fitness of the same type of gas pipeline network to obtain a meta-predictor; An integrated construction unit is used to use the first fitness predictor as a primary predictor and integrate it with the meta-predictor to obtain the intelligent fault prediction model.

Images