Gas pipe network leakage event information analysis method and device

Abstract

The invention provides a gas pipe network leakage event information analysis method and device, which are used for carrying out information acquisition and analysis on existing gas pipe network leakage events, classifying leakage characteristics of the existing leakage events and associating the leakage characteristics with accident causes. According to the method, pre-judgment is carried out on an unrepaired leakage accident point through analysis of actual existing similar cases and analysis of first-aid repair measures, so that targeted repair of the leakage point is facilitated, and the first-aid repair efficiency is improved; and a correlation analysis result of leakage characteristics and accident causes can be fed back to a construction link, and the leakage accident rate can be effectively reduced through targeted prevention of main causes.

Description

Technical Field

[0001] The present invention relates to the field of computers, and in particular to a method and device for analyzing information on gas pipeline network leakage events. Background Art

[0002] Properly handling gas pipeline leakage incidents is the key to avoiding gas pipeline leakage accidents. Improving emergency repair efficiency and carrying out targeted prevention from the construction stage are two common means to reduce the incidence of gas leakage incidents. Preliminary analysis and judgment of possible leakage accidents can help improve emergency repair efficiency; and finding out the most likely factors leading to pipeline leakage and carrying out targeted prevention during pipeline laying can help reduce the gas leakage accident rate. Both are important measures to effectively consolidate gas safety.

[0003] At present, the gas company has established a construction and emergency repair ledger for the gas pipeline network under its jurisdiction. However, the existing ledger is basically management data, with problems such as incomplete data information, unclear failure locations, and no service environment information. The types of items collected by each unit, especially accident descriptions and emergency repair measures, are different, which affects the scientificity and completeness of the collection of leakage data information. In addition, there is currently a lack of further analysis of the collected information, which makes it impossible to achieve effective use of the information. Summary of the invention

[0004] The present invention aims to provide a method and device for analyzing information of gas pipeline leakage events that overcome the above-mentioned problems or at least partially solve the above-mentioned problems.

[0005] To achieve the above object, the technical solution of the present invention is specifically implemented as follows:

[0006] One aspect of the present invention provides a method for analyzing information of a gas pipeline network leakage event, comprising:

[0007] Establish a failure case information database for typical parts of gas pipelines, which includes: basic information of leakage events, leakage characteristic information, damage characteristic information, original pipeline engineering information, environmental characteristic information, pipe sampling and testing information, and repair and treatment information;

[0008] Establishing an emergency repair decision-making support system according to the basic information of the leakage event, the leakage characteristic information, the repair processing information, the damage characteristic information, the original engineering information of the pipeline and the environmental characteristic information;

[0009] The emergency repair auxiliary decision-making system performs correlation analysis on the leakage characteristic information and the repair processing information to obtain the repair means corresponding to each leakage characteristic information;

[0010] Establishing a defect cause analysis system according to the environmental characteristic information, the pipe sampling test information and the damage characteristic information;

[0011] The defect cause analysis system performs correlation analysis on the damage characteristic information and the preset causes to determine the causes corresponding to each defect characteristic information;

[0012] Establishing a pipeline network adverse environment geographic information system based on the environmental characteristic information, the original pipeline engineering information and the pipe sampling test information;

[0013] The pipeline network adverse environment geographic information system is used to determine whether the accident is caused by subsidence. If so, the pipelines with the risk of subsidence will be marked; if not, the PE aging sensitive substance content at the location will be marked.

[0014] Optionally, the establishment of the emergency repair auxiliary decision-making system includes: establishing a multi-task decision tree model; the establishment of the defect cause analysis system includes: establishing a multi-task decision tree model;

[0015] Where: The input of the multi-task decision tree model is a sample X of input features i , expressed as: X i ={x1,x2,...,x k ,...,x N}, where x k is the kth feature, which is a discrete feature or a continuous feature, and N is the number of features; the output is a task set: Y = {y1, y2, y3}, y1 is the plugging method, y2 is the determination of the repair method, and y3 is the determination of the cause of the accident.

[0016] Optionally, establishing a multi-task decision tree model includes:

[0017] Create a root node;

[0018] For each node, split according to the preset rules; the preset rules include: traversing all features, calculating the multi-task information gain of each feature, selecting the feature and splitting point corresponding to the maximum information gain, splitting, and dividing the data into sub-nodes;

[0019] Splitting is performed in each child node according to the preset rule until a stop condition is met, thereby establishing the multi-task decision tree model.

[0020] Optionally, calculating the multi-task information gain of each feature includes:

[0021] If the feature x k If is a discrete feature, then assume that it has Q different values. Based on the different values, all samples D are divided into Q completely disjoint subsets:

[0022] D=D1∪D2∪...∪D Q

[0023] Then we get the information gain:

[0024]

[0025] Among them, H(D) is all the samples in the sample D, and these samples in task y i The Gini coefficient calculated from the true label on i =y1, y2 or y3:

[0026]

[0027] Among them, |y i | is the number of all possible labels, p z is the probability of taking z as the sample label in D;

[0028] If the feature x k If it is a continuous feature, determine the optimal split point s and divide the sample into two parts N l and N r , we get the information gain:

[0029]

[0030] The three information gains IG(x k ) is summed as the feature x k Multi-task information gain at any node.

[0031] Optionally, marking the pipeline at risk of subsidence includes:

[0032] Retrieve the original engineering information and environmental characteristics of the leaking pipes that have subsided, and coordinate with the existing pipe network diagram and map to mark the pipes with subsidence risks in the geographic information system;

[0033] The marking of the PE aging-sensitive substance content at the location includes: collecting environmental characteristic information of the leakage event and pipe sampling test information, sampling the soil after the leakage is handled, analyzing the content of PE aging-sensitive substances and the aging conditions of pipes and fittings in the soil, and marking them in the geographic information system.

[0034] One aspect of the present invention provides a gas pipeline network leakage event information analysis device, comprising:

[0035] An information base establishment module is used to establish a failure case information base for typical parts of gas pipelines, the information base includes: basic information of leakage events, leakage characteristic information, damage characteristic information, original pipeline engineering information, environmental characteristic information, pipe sampling and testing information and repair and treatment information;

[0036] The first establishment module is used to establish an emergency repair auxiliary decision-making system according to the basic information of the leakage event, the leakage characteristic information, the repair processing information, the damage characteristic information, the original engineering information of the pipeline and the environmental characteristic information;

[0037] A first analysis module is used to perform a correlation analysis on the leakage feature information and the repair processing information through the emergency repair auxiliary decision system to obtain a repair method corresponding to each leakage feature information;

[0038] A second establishing module is used to establish a defect cause analysis system according to the environmental characteristic information, the pipe sampling test information and the damage characteristic information;

[0039] A second analysis module is used to perform a correlation analysis between the damage characteristic information and a preset cause through the defect cause analysis system to determine the cause corresponding to each defect characteristic information;

[0040] The third establishment module is used to establish a pipeline network adverse environment geographic information system based on the environmental characteristic information, the original pipeline engineering information and the pipe sampling test information;

[0041] The third analysis module is used to determine whether the accident is caused by sedimentation through the pipeline network adverse environment geographic information system. If so, the pipeline with the risk of sedimentation will be marked; if not, the PE aging sensitive substance content at the location will be marked.

[0042] Optionally, the first establishing module establishes the emergency repair auxiliary decision system by: establishing a multi-task decision tree model; the second establishing module establishes the defect cause analysis system by: establishing a multi-task decision tree model;

[0043] Where: The input of the multi-task decision tree model is a sample X of input features i , expressed as: X i ={x1,x2,...,x k ,...,x N}, where x k is the kth feature, which is a discrete feature or a continuous feature, and N is the number of features; the output is a task set: Y = {y1, y2, y3}, y1 is the plugging method, y2 is the determination of the repair method, and y3 is the determination of the cause of the accident.

[0044] Optionally, the first establishing module or the second establishing module establishes the multi-task decision tree model in the following manner:

[0045] Create a root node;

[0046] For each node, split according to the preset rules; the preset rules include: traversing all features, calculating the multi-task information gain of each feature, selecting the feature and splitting point corresponding to the maximum information gain, splitting, and dividing the data into sub-nodes;

[0047] Splitting is performed in each child node according to the preset rule until a stop condition is met, thereby establishing the multi-task decision tree model.

[0048] Optionally, the first building module or the second building module calculates the multi-task information gain of each feature in the following manner:

[0049] If the feature x k If is a discrete feature, then assume that it has Q different values. Based on the different values, all samples D are divided into Q completely disjoint subsets:

[0050] D=D1∪D2∪...∪D Q

[0051] Then we get the information gain:

[0052]

[0053] Among them, H(D) is all the samples in the sample D, and these samples in task y i The Gini coefficient calculated from the true label on i =y1, y2 or y3:

[0054]

[0055] Among them, |y i | is the number of all possible labels, p z is the probability of taking z as the sample label in D;

[0056] If the feature x k If it is a continuous feature, determine the optimal split point s and divide the sample into two parts N l and N r , we get the information gain:

[0057]

[0058] The three information gains IG(x k ) is summed as the feature x kMulti-task information gain at any node.

[0059] Optionally, the third analysis module marks the pipelines with the risk of subsidence in the following manner: calling the original engineering information and environmental characteristic information of the leaking pipes with subsidence, coordinating with the existing pipe network diagram and map, and marking the pipelines with the risk of subsidence in the geographic information system;

[0060] The third analysis module marks the PE aging-sensitive substance content at the location in the following manner: by collecting environmental characteristic information of the leakage event and pipe sampling test information, sampling the soil after the leakage is handled, and analyzing the content of PE aging-sensitive substances and the aging conditions of pipes and fittings in the soil, and marking them in the geographic information system.

[0061] It can be seen that, through the gas pipeline leakage event information analysis method and device provided by the present invention, information collection and analysis of existing gas pipeline leakage events are performed, leakage characteristics of existing leakage events are classified, and leakage characteristics are associated with the causes of accidents. Preliminary evaluation of the leakage accident points that have occurred but not repaired is carried out through the analysis of actual similar cases and the analysis of emergency repair measures, which is more conducive to the targeted repair of the leakage points and improves the efficiency of emergency repair; the correlation analysis results of the leakage characteristics and the causes of the accident can be fed back to the construction link, and the leakage accident rate can be effectively reduced through targeted prevention of the main causes. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0063] Figure 1 A flow chart of a method for analyzing information of a gas pipeline network leakage event provided by an embodiment of the present invention;

[0064] Figure 2 A logic block diagram of a gas network leakage event information analysis system provided by an embodiment of the present invention;

[0065] Figure 3 A flow chart of calculating information gain provided by an embodiment of the present invention;

[0066] Figure 4 A logic block diagram of a pipe network adverse environment geographic information system provided by an embodiment of the present invention;

[0067] Figure 5 A schematic diagram of the structure of a gas pipeline network leakage event information analysis device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0069] Figure 1 A flow chart of a method for analyzing information of a gas pipeline leakage event provided by an embodiment of the present invention is shown. Figure 2 The logic block diagram of the gas pipeline leakage event information analysis system provided by the embodiment of the present invention is shown in FIG. Figure 1 and Figure 2 The gas pipeline network leakage event information analysis method provided by the embodiment of the present invention includes:

[0070] S1. Establish a failure case information database of typical parts of gas pipelines, which includes: basic information of leakage events, leakage characteristic information, damage characteristic information, original pipeline engineering information, environmental characteristic information, pipe sampling and testing information and repair and treatment information.

[0071] Specifically, this step corresponds to the data collection information base part.

[0072] Through preliminary research, the present invention has reorganized the information categories and items that need to be collected for accident analysis, and the online system is filled in uniformly. For pipelines that leak on site, soil sampling and pipe and fitting sampling are also required. The specific categories collected are:

[0073] Basic information of the leakage incident, including the location and year and month of the incident.

[0074] Leakage characteristic information, including leakage location and leakage degree.

[0075] And damage characteristics information. Including the damage degree of pipes (parts).

[0076] Original pipeline engineering information, including pipeline pressure level, pipe outer diameter, and the short-term, medium-term, and long-term operation of the pipe.

[0077] Environmental characteristic information. Including pipeline burial depth, soil moisture content, soil description, geological settlement information, PE aging sensitive substance content analysis results (ketone alcohol organic matter, ester organic matter, phenol organic matter, ether organic matter, amine organic matter, surfactant), and location status information of other municipal facilities (such as gas pipelines, etc.).

[0078] Pipe sampling and testing information, including pipe types, destructive test results, and non-destructive test results.

[0079] Repair and treatment information, including temporary leakage control measures and repair measures taken.

[0080] S2, perform data analysis, including:

[0081] S21, establishing an emergency repair decision-making support system based on the basic information of the leakage event, the leakage characteristic information, the repair processing information, the damage characteristic information, the original pipeline engineering information and the environmental characteristic information; performing a correlation analysis on the leakage characteristic information and the repair processing information through the emergency repair decision-making support system to obtain the repair means corresponding to each leakage characteristic information.

[0082] Specifically, the present invention establishes an emergency repair decision-making support system by analyzing the basic information of the leakage event, leakage feature information, repair and processing information, environmental feature information, damage feature information, and original pipeline engineering information. By analyzing the correlation between leakage features and leakage repair measures, the most commonly used and most effective repair means for each leakage feature are obtained, providing decision-making support for the repair team.

[0083] In the emergency repair auxiliary decision-making system of the present invention, the following are input: basic information, leakage characteristic information, repair processing information, environmental characteristic information, damage characteristic information, and pipeline engineering information.

[0084] Output:

[0085] 1. Leakage blocking methods. Options: tape wrapping, pressure reduction, clamping and air cutting;

[0086] 2. Repair methods. Options: pressure tapping, saddle tee.

[0087] S22, establishing a defect cause analysis system based on the environmental characteristic information, the pipe sampling test information and the damage characteristic information; performing a correlation analysis between the damage characteristic information and the preset cause through the defect cause analysis system to determine the cause corresponding to each defect characteristic information.

[0088] Specifically, the present invention analyzes the environmental characteristic information, pipe sampling test information, and damage characteristic information of the leakage event, and performs correlation analysis between several damage characteristics and several possible causes to find out the most likely cause corresponding to each defect characteristic, which helps to carry out targeted prevention from the source during the construction phase and reduce the accident rate.

[0089] In the defect cause analysis system of the present invention, the following are input: damage characteristic information, environmental characteristic information, and pipe test sampling information.

[0090] Output: Cause of the accident. Options: External force damage, aging failure, incomplete welding, resistor wire misalignment, holes, over-welding.

[0091] S23, establishing a pipeline network adverse environment geographic information system based on the environmental characteristic information, the pipeline original project information and the pipe sampling test information; judging whether the accident is caused by subsidence through the pipeline network adverse environment geographic information system, if so, marking the pipeline with the risk of subsidence; if not, marking the PE aging sensitive substance content at the location.

[0092] Specifically, the present invention analyzes the two environmental factors that mainly affect PE pipeline leakage: sedimentation and aging by analyzing the environmental information of the leakage event and the pipe sampling test information. Figure 4 The present invention determines whether the accident is caused by settlement. If so, the pipeline with the risk of settlement will be marked (all pipelines under the same project construction will be marked as having the risk of settlement); if not, the content of PE aging-sensitive substances at the location will be marked.

[0093] As an optional implementation of the embodiment of the present invention, marking the pipeline with the risk of subsidence includes:

[0094] The original engineering information and environmental characteristic information of the leaking pipe that has subsided are called up, and the pipeline with the risk of subsidence is marked in the geographic information system in coordination with the existing pipeline network diagram and map; the marking of the PE aging sensitive substance content at the location includes: collecting the environmental characteristic information of the leakage event and the pipe sampling test information, sampling the soil after the leakage is handled, and analyzing the content of PE aging sensitive substances and the aging of pipes and fittings in the soil, and marking them in the geographic information system.

[0095] Specifically, settlement is closely related to construction units. Therefore, the present invention calls the original engineering data and environmental characteristic information of the leaking pipes that have settled, and cooperates with the existing pipeline network diagram and map to mark the pipelines with the risk of settlement in the geographic information system, so as to provide a reference for the cause analysis of the leakage incidents that may occur again here and the reasonable laying layout of the subsequent pipeline network.

[0096] In addition, the aging of PE gas pipeline performance is strongly correlated with the content of certain substances in the soil. The present invention collects environmental information of the leakage event and pipe sampling information, and samples the soil after the leakage is treated. The content of PE aging-sensitive substances and the aging of pipes and fittings in the soil is analyzed, and marked in the geographic information system, which serves as a reference for the cause analysis of the leakage event that may occur again and the reasonable layout of the subsequent pipe network.

[0097] As an optional implementation of the embodiment of the present invention, the establishment of the emergency repair auxiliary decision system includes: establishing a multi-task decision tree model; the establishment of the defect cause analysis system includes: establishing a multi-task decision tree model; wherein: the input of the multi-task decision tree model is a sample X of the input feature i , expressed as: X i ={x1,x2,...,x k ,...,x N}, where x k is the kth feature, which is a discrete feature or a continuous feature, and N is the number of features; the output is a task set: Y = {y1, y2, y3}, y1 is the plugging method, y2 is the determination of the repair method, and y3 is the determination of the cause of the accident.

[0098] Specifically, the present invention uses a multi-task decision tree that supports both discrete and continuous features to perform analysis. The emergency repair decision support system and the defect cause analysis system in the present invention have three classification tasks: determining the plugging method y1, determining the repair method y2, and determining the cause of the accident y3. Therefore, for the three classification tasks, the sample X containing all the input features i , which can be expressed as:

[0099] X i ={x1,x2,...,x k ,...,x N}

[0100] x k is the kth feature, which may be a discrete feature or a continuous feature, and N is the number of features. For the task set:

[0101] Y={y1,y2,y3}

[0102] Among them, y1, y2, and y3 are all classification tasks. i In the case of three labels, a multi-task decision tree model can be obtained.

[0103] As an optional implementation of an embodiment of the present invention, the establishment of a multi-task decision tree model includes: establishing a root node; splitting each node according to a preset rule; the preset rule includes: traversing all features, calculating the multi-task information gain of each feature, selecting the feature and splitting point corresponding to the maximum information gain, splitting, and dividing the data into child nodes; splitting in each child node according to the preset rule until the stopping condition is met, and establishing the multi-task decision tree model.

[0104] Specifically, the process of establishing a multi-task decision tree in the present invention includes:

[0105] Step 1: Create a root node.

[0106] Step 2: For each node, traverse all features (including discrete features and continuous features), calculate the multi-task information gain of each feature, select the feature and split point corresponding to the maximum information gain, split, and divide the data into sub-nodes.

[0107] Step 3: After dividing the data into child nodes in step 2, continue to recursively split in each child node according to the rules in step 2 until the stopping condition is met (such as the number of node samples is less than a preset value, or the information gain is less than a preset value).

[0108] As an optional implementation of the embodiment of the present invention, the calculating the multi-task information gain of each feature includes:

[0109] If the feature x k If is a discrete feature, then assume that it has Q different values. Based on the different values, all samples D are divided into Q completely disjoint subsets:

[0110] D=D1∪D2∪...∪D Q

[0111] Then we get the information gain:

[0112]

[0113] Among them, H(D) is all the samples in the sample D, and these samples in task y i The Gini coefficient calculated from the true label on i =y1, y2 or y3:

[0114]

[0115] Among them, |y i | is the number of all possible labels, p z is the probability of taking z as the sample label in D;

[0116] If the feature x k If it is a continuous feature, determine the optimal split point s and divide the sample into two parts N l and N r , we get the information gain:

[0117]

[0118] The three information gains IG(x k ) is summed as the feature x k Multi-task information gain at any node.

[0119] Specifically, see Figure 3, the present invention information IG M The calculation methods include:

[0120] For x k , if it is a discrete feature, assuming it has Q different values, then all samples D can be divided into Q completely disjoint subsets based on different values:

[0121] D=D1∪D2∪...∪D Q

[0122] Then we get the information gain:

[0123]

[0124] Among them, Dr represents the rth subset, H() is a method for calculating the Gini coefficient, and H(D) is the Gini coefficient calculated by all samples in D and the true labels of these samples on task y1 or task y2 or y3. Taking task y1 as an example:

[0125]

[0126] Among them, |y1| is the number of all possible labels, p z is the probability of taking z as the sample label in D.

[0127] For x k If it is a continuous feature, we need to find an optimal split point s to divide the sample into two parts N l and N r , where N l and N r For two disjoint sets, the information gain can be obtained:

[0128]

[0129] At this point, the three IGs obtained from the three tasks (x k ) numerical sum, directly as x k Multi-task information gain at any node.

[0130] After building the decision tree model according to the above steps, any sample can be classified on three tasks.

[0131] It can be seen that the gas pipeline leakage event information analysis method provided by the embodiment of the present invention can pre-evaluate the leakage accident points that have occurred but not repaired through the analysis of actual similar cases and the analysis of emergency repair measures, which is more conducive to the targeted repair of the leakage points and improves the emergency repair efficiency; the correlation analysis results between the leakage characteristics and the causes of the accident can be fed back to the construction link, and the leakage accident rate can be effectively reduced through targeted prevention of the main causes.

[0132] The present invention has the following beneficial effects:

[0133] 1. In view of the problem of inconsistent and incomplete information collection in the existing emergency repair ledgers and construction ledgers, the information data of gas pipeline leakage incidents that need to be collected are re-planned, and a typical failure case information database of gas pipelines is established based on the re-organized collected data to provide data support for the analysis part of the system. The re-collected part mainly includes pipe information such as the manufacturer and batch of the leaking pipe section; response information such as the time and location of the case emergency repair, emergency repair measures, etc.; construction information such as the construction unit responsible for this place during pipeline laying and the time of commissioning; leakage damage characteristics such as burial depth, leakage location, and damage degree; environmental geological information such as soil physical and chemical composition, location of adjacent municipal facilities, and geological settlement.

[0134] 2. In order to improve the efficiency of emergency repair of gas pipeline leakage incidents, the leakage feature information and environmental information of the leakage incidents in the information database are correlated and analyzed to find out the environmental features and defect features with the greatest correlation for each type of leakage feature. The leakage features reflect the defect features, and the most likely leakage features are pre-judged before excavation and repair when the leakage occurs, so that the maintenance team can carry out targeted treatment when rushing to the scene, which is conducive to controlling the gas leakage as soon as possible. According to the actual defect characteristics and response information, the correlation between the defect characteristics and related repair methods is analyzed, and the most effective corresponding means or repair methods under specific defect characteristics are found. Before the repair team rushes to the scene, the repair measures to be taken are pre-judged, which is conducive to targeted repair of leakage incidents and improves the efficiency of emergency repairs.

[0135] 3. For the two environmental characteristics of geological subsidence and PE gas pipeline aging, a gas pipeline subsidence geographic information system and a PE aging sensitive substance geographic information system are established. Using the geological subsidence information collected during the leakage event, according to the existing rules, the nearby subsidence risk points are inferred based on the point, and the map and the existing pipeline network distribution map are superimposed to mark the pipeline network with subsidence risk. Through the soil physical and chemical information collected during the leakage event, the content of PE gas pipeline aging sensitive substances in the soil environment of the accident pipeline is analyzed in the laboratory, and the leakage point is marked on the map, superimposed with the existing pipeline network distribution map, and the distribution of existing PE gas pipeline aging sensitive substances is intuitively reflected. It is conducive to ensuring the comprehensiveness and accuracy of judgment, and also provides a reference for the reasonable transformation and laying of subsequent pipeline networks.

[0136] 4. In order to carry out targeted prevention from the construction perspective and reduce the pipeline leakage rate, the leakage characteristics in the information database are matched with the defect characteristics, and then the defect characteristics are correlated with possible causes. The causes are deduced from the defect characteristics to find out the most likely cause of each leakage characteristic. By statistics and summarizing the possible causes, it is conducive to more targeted control from the construction source, so as to carry out targeted prevention of the causes with the highest occurrence rate.

[0137] Figure 5 The schematic diagram of the structure of the gas pipeline leakage event information analysis device provided by the embodiment of the present invention is shown. The gas pipeline leakage event information analysis device applies the above method. The following is only a brief description of the structure of the gas pipeline leakage event information analysis device. For other matters not covered, please refer to the relevant description in the above gas pipeline leakage event information analysis method. Figure 5 , a gas pipeline network leakage event information analysis device provided by an embodiment of the present invention comprises:

[0138] An information base establishment module is used to establish a failure case information base for typical parts of gas pipelines, the information base includes: basic information of leakage events, leakage characteristic information, damage characteristic information, original pipeline engineering information, environmental characteristic information, pipe sampling and testing information and repair and treatment information;

[0139] The first establishment module is used to establish an emergency repair auxiliary decision-making system according to the basic information of the leakage event, the leakage characteristic information, the repair processing information, the damage characteristic information, the original engineering information of the pipeline and the environmental characteristic information;

[0140] A first analysis module is used to perform a correlation analysis on the leakage feature information and the repair processing information through the emergency repair auxiliary decision system to obtain a repair method corresponding to each leakage feature information;

[0141] A second establishing module is used to establish a defect cause analysis system according to the environmental characteristic information, the pipe sampling test information and the damage characteristic information;

[0142] A second analysis module is used to perform a correlation analysis between the damage characteristic information and a preset cause through the defect cause analysis system to determine the cause corresponding to each defect characteristic information;

[0143] The third establishment module is used to establish a pipeline network adverse environment geographic information system based on the environmental characteristic information, the original pipeline engineering information and the pipe sampling test information;

[0144] The third analysis module is used to determine whether the accident is caused by sedimentation through the pipeline network adverse environment geographic information system. If so, the pipeline with the risk of sedimentation will be marked; if not, the PE aging sensitive substance content at the location will be marked.

[0145] As an optional implementation of the embodiment of the present invention, the first establishment module establishes the emergency repair auxiliary decision system by: establishing a multi-task decision tree model; the second establishment module establishes the defect cause analysis system by: establishing a multi-task decision tree model;

[0146] Where: The input of the multi-task decision tree model is a sample X of input features i , expressed as: X i ={x1,x2,...,x k ,...,x N}, where x k is the kth feature, which is a discrete feature or a continuous feature, and N is the number of features; the output is a task set: Y = {y1, y2, y3}, y1 is the plugging method, y2 is the determination of the repair method, and y3 is the determination of the cause of the accident.

[0147] As an optional implementation manner of the embodiment of the present invention, the first establishing module or the second establishing module establishes the multi-task decision tree model in the following manner:

[0148] Create a root node;

[0149] For each node, split according to the preset rules; the preset rules include: traversing all features, calculating the multi-task information gain of each feature, selecting the feature and splitting point corresponding to the maximum information gain, splitting, and dividing the data into sub-nodes;

[0150] Splitting is performed in each child node according to the preset rule until a stop condition is met, thereby establishing the multi-task decision tree model.

[0151] As an optional implementation manner of the embodiment of the present invention, the first establishment module or the second establishment module calculates the multi-task information gain of each feature in the following manner:

[0152] If the feature x k If is a discrete feature, then assume that it has Q different values. Based on the different values, all samples D are divided into Q completely disjoint subsets:

[0153] D=D1∪D2∪...∪D Q

[0154] Then we get the information gain:

[0155]

[0156] Among them, H(D) is all the samples in the sample D, and these samples in task y i The Gini coefficient calculated from the true label on i =y1, y2 or y3:

[0157]

[0158] Among them, |y i | is the number of all possible labels, p z is the probability of taking z as the sample label in D;

[0159] If the feature x k If it is a continuous feature, determine the optimal split point s and divide the sample into two parts N l and N r , we get the information gain:

[0160]

[0161] The three information gains IG(x k ) is summed as the feature x k Multi-task information gain at any node.

[0162] As an optional implementation of the embodiment of the present invention, the third analysis module marks the pipeline with the risk of settlement in the following manner: calling the original engineering information and environmental characteristic information of the leaking pipe with settlement, cooperating with the existing pipe network diagram and map, and marking the pipeline with the risk of settlement in the geographic information system;

[0163] The third analysis module marks the PE aging-sensitive substance content at the location in the following manner: by collecting environmental characteristic information of the leakage event and pipe sampling test information, sampling the soil after the leakage is handled, and analyzing the content of PE aging-sensitive substances and the aging conditions of pipes and fittings in the soil, and marking them in the geographic information system.

[0164] It can be seen that the gas pipeline leakage event information analysis device provided by the embodiment of the present invention can pre-evaluate the leakage accident points that have occurred but not repaired through the analysis of actual similar cases and the analysis of emergency repair measures, which is more conducive to the targeted repair of the leakage points and improves the emergency repair efficiency; the correlation analysis results between the leakage characteristics and the causes of the accident can be fed back to the construction link, and the leakage accident rate can be effectively reduced through targeted prevention of the main causes.

[0165] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for analyzing information of gas pipeline leakage events, characterized in that: include: Establish a failure case information database for typical parts of gas pipelines, which includes: basic information of leakage events, leakage characteristic information, damage characteristic information, original pipeline engineering information, environmental characteristic information, pipe sampling and testing information, and repair and treatment information; Establishing an emergency repair decision-making support system according to the basic information of the leakage event, the leakage characteristic information, the repair processing information, the damage characteristic information, the original engineering information of the pipeline and the environmental characteristic information; The emergency repair auxiliary decision-making system performs correlation analysis on the leakage characteristic information and the repair processing information to obtain the repair means corresponding to each leakage characteristic information; Establishing a defect cause analysis system according to the environmental characteristic information, the pipe sampling test information and the damage characteristic information; The defect cause analysis system performs correlation analysis on the damage characteristic information and the preset causes to determine the causes corresponding to each defect characteristic information; Establishing a pipeline network adverse environment geographic information system based on the environmental characteristic information, the original pipeline engineering information and the pipe sampling test information; The pipeline network adverse environment geographic information system is used to determine whether the accident is caused by subsidence. If so, the pipelines with the risk of subsidence will be marked; if not, the PE aging sensitive substance content at the location will be marked.

2. The method according to claim 1, characterized in that The establishment of the emergency repair auxiliary decision-making system includes: establishing a multi-task decision tree model; the establishment of the defect cause analysis system includes: establishing a multi-task decision tree model; Where: The input of the multi-task decision tree model is a sample X of input features i , expressed as: X i ={x1,x2,...,x k ,...,x N }, where x k is the kth feature, which is a discrete feature or a continuous feature, and N is the number of features; the output is a task set: Y = {y1, y2, y3}, y1 is the plugging method, y2 is the determination of the repair method, and y3 is the determination of the cause of the accident.

3. The method according to claim 2, characterized in that The multi-task decision tree model is established including: Create a root node; For each node, split according to the preset rules; the preset rules include: traversing all features, calculating the multi-task information gain of each feature, selecting the feature and splitting point corresponding to the maximum information gain, splitting, and dividing the data into sub-nodes; Splitting is performed in each child node according to the preset rule until a stop condition is met, thereby establishing the multi-task decision tree model.

4. The method according to claim 3, characterized in that The calculation of the multi-task information gain of each feature includes: If the feature x k If is a discrete feature, then assume that it has Q different values. Based on the different values, all samples D are divided into Q completely disjoint subsets: D=D1∪D2∪...∪D Q Then we get the information gain: Among them, H(D) is all the samples in the sample D, and these samples in task y i The Gini coefficient calculated from the true label on i =y1, y2 or y3: Among them, |y i | is the number of all possible labels, p z is the probability of taking z as the sample label in D; If the feature x k If it is a continuous feature, determine the optimal split point s and divide the sample into two parts N l and N r , we get the information gain: The three information gains IG(x k ) is summed as the feature x k Multi-task information gain at any node.

5. The method according to claim 1, characterized in that The marking of pipelines with the risk of subsidence includes: Retrieve the original engineering information and environmental characteristics of the leaking pipes that have subsided, and coordinate with the existing pipe network diagram and map to mark the pipes with subsidence risks in the geographic information system; The marking of the PE aging-sensitive substance content at the location includes: collecting environmental characteristic information of the leakage event and pipe sampling test information, sampling the soil after the leakage is handled, analyzing the content of PE aging-sensitive substances and the aging conditions of pipes and fittings in the soil, and marking them in the geographic information system.

6. A gas pipeline network leakage event information analysis device, characterized in that: include: An information base establishment module is used to establish a failure case information base for typical parts of gas pipelines, the information base includes: basic information of leakage events, leakage characteristic information, damage characteristic information, original pipeline engineering information, environmental characteristic information, pipe sampling and testing information and repair and treatment information; The first establishment module is used to establish an emergency repair auxiliary decision-making system according to the basic information of the leakage event, the leakage characteristic information, the repair processing information, the damage characteristic information, the original engineering information of the pipeline and the environmental characteristic information; A first analysis module is used to perform a correlation analysis on the leakage feature information and the repair processing information through the emergency repair auxiliary decision system to obtain a repair method corresponding to each leakage feature information; A second establishment module is used to establish a defect cause analysis system according to the environmental characteristic information, the pipe sampling test information and the damage characteristic information; A second analysis module is used to perform a correlation analysis between the damage characteristic information and a preset cause through the defect cause analysis system to determine the cause corresponding to each defect characteristic information; The third establishment module is used to establish a pipeline network adverse environment geographic information system based on the environmental characteristic information, the original pipeline engineering information and the pipe sampling test information; The third analysis module is used to determine whether the accident is caused by sedimentation through the pipeline network adverse environment geographic information system. If so, the pipeline with the risk of sedimentation will be marked; if not, the PE aging sensitive substance content at the location will be marked.

7. The device according to claim 6, characterized in that The first establishment module establishes the emergency repair auxiliary decision system by: establishing a multi-task decision tree model; the second establishment module establishes the defect cause analysis system by: establishing a multi-task decision tree model; Where: The input of the multi-task decision tree model is a sample X of input features i , expressed as: X i ={x1,x2,...,x k ,...,x N }, where x k is the kth feature, which is a discrete feature or a continuous feature, and N is the number of features; the output is a task set: Y = {y1, y2, y3}, y1 is the plugging method, y2 is the determination of the repair method, and y3 is the determination of the cause of the accident.

8. The device according to claim 7, characterized in that The first establishing module or the second establishing module establishes the multi-task decision tree model in the following manner: Create a root node; For each node, split according to the preset rules; the preset rules include: traversing all features, calculating the multi-task information gain of each feature, selecting the feature and splitting point corresponding to the maximum information gain, splitting, and dividing the data into sub-nodes; Splitting is performed in each child node according to the preset rule until a stop condition is met, thereby establishing the multi-task decision tree model.

9. The device according to claim 8, characterized in that The first building module or the second building module calculates the multi-task information gain of each feature in the following manner: If the feature x k If is a discrete feature, then assume that it has Q different values. Based on the different values, all samples D are divided into Q completely disjoint subsets: D=D1∪D2∪...∪D Q Then we get the information gain: Among them, H(D) is all the samples in the sample D, and these samples in task y i The Gini coefficient calculated from the true label on i =y1, y2 or y3: Among them, |y i | is the number of all possible labels, p z is the probability of taking z as the sample label in D; If the feature x k If it is a continuous feature, determine the optimal split point s and divide the sample into two parts N l and N r , we get the information gain: The three information gains IG(x k ) is summed as the feature x k Multi-task information gain at any node.

10. The device according to claim 6, characterized in that The third analysis module marks the pipelines with the risk of subsidence in the following manner: calling the original engineering information and environmental characteristic information of the leaking pipes with subsidence, coordinating with the existing pipe network diagram and map, and marking the pipelines with the risk of subsidence in the geographic information system; The third analysis module marks the PE aging-sensitive substance content at the location in the following manner: by collecting environmental characteristic information of the leakage event and pipe sampling test information, sampling the soil after the leakage is handled, and analyzing the content of PE aging-sensitive substances and the aging conditions of pipes and fittings in the soil, and marking them in the geographic information system.

Images