A digital twin construction method for pipeline lightweight operation and maintenance system

Abstract

The present application relates to the field of intelligent operation and maintenance technology, and specifically to a method for constructing a digital twin of a pipeline lightweight operation and maintenance system, the method comprising: collecting various operation data of each operation and maintenance node on an oil and gas transportation pipeline at each collection time, and obtaining various event data of each operation and maintenance node; determining the feature characterization capability of various operation data of each operation and maintenance node with respect to each event data based on the difference between various operation data before and after the warning time in each event data of each operation and maintenance node; determining the event influence degree and chain node set of each operation and maintenance node with respect to each event data; determining the event contribution degree of various operation data of each operation and maintenance node with respect to each event data; determining the principal component set and lightweight operation data of the pipeline lightweight operation and maintenance system, and constructing a digital twin of the pipeline lightweight operation and maintenance system. The present application aims to improve the efficiency of digital twin construction.

Description

Technical Field

[0001] The present application relates to the field of intelligent operation and maintenance technology, and specifically to a method for constructing a digital twin of a pipeline lightweight operation and maintenance system. Background Art

[0002] Digital twin technology forms digital twins by building digital mirrors of things in the virtual world, reflecting the development and changes of things in the physical world in the virtual world, and can provide decision-making guidance for physical entities through the evolution of digital twins in the virtual world. By using the full life cycle data of the pipeline, through real-time data update, dynamic data adjustment, data analysis and other intelligent technologies, a digital twin of the pipeline lightweight operation and maintenance system is built to achieve digital and intelligent management of the operation, maintenance, inspection and transformation of the oil and gas transportation pipeline throughout its life cycle, which helps to optimize the use and maintenance strategy of the oil and gas transportation pipeline and extend the service life of the oil and gas transportation pipeline.

[0003] In the process of building a digital twin, there is data information on the entire business and process of the pipeline lightweight operation and maintenance system. The data volume is huge and the data dimensions are multi-dimensional. It is difficult to conduct effective data governance and convert it into a decision-making basis for the use of the pipeline lightweight operation and maintenance system, and it is impossible to form lightweight data assets. At this stage, the use of PCA principal component analysis can extract the main features from the original full life cycle data and reduce redundant information, but the extracted principal components may be difficult to explain their meaning. The Apriori algorithm mines correlations based on data sets that frequently appear in the full life cycle data. It has strong interpretability, but it has high requirements for the input data type. Summary of the invention

[0004] In view of the above, it is necessary to provide a digital twin construction method for a pipeline lightweight operation and maintenance system, which improves the efficiency of digital twin construction compared with the traditional digital twin construction method.

[0005] The digital twin construction method of a pipeline lightweight operation and maintenance system of the present application adopts the following technical solution:

[0006] An embodiment of the present application provides a method for constructing a digital twin of a pipeline lightweight operation and maintenance system, the method comprising the following steps:

[0007] Collect various operation data of each operation and maintenance node on the oil and gas transportation pipeline at each collection time, and obtain various event data of each operation and maintenance node. The event data includes the event type, warning time and the operation and maintenance node itself when the operation and maintenance node appears in the alarm state;

[0008] Based on the difference between various operation data before and after the warning time in any event data of any operation and maintenance node, determine the feature characterization capability of various operation data of any operation and maintenance node with respect to any event data; based on the feature characterization capability, determine the event impact degree of any operation and maintenance node with respect to any event data;

[0009] Recording the event type to which any event data belongs as the event type to be analyzed, and determining a chain node set of any operation and maintenance node with respect to any event data based on the position distance relationship between the operation and maintenance nodes and the distribution of event impact degrees of any operation and maintenance node with respect to all event data belonging to the event type to be analyzed;

[0010] Determining the event contribution of any one type of operation data of any operation and maintenance node with respect to any one event data based on the feature characterization capability of any one type of operation data of each operation and maintenance node in the chain node set with respect to any one event data;

[0011] Based on the distribution of event contributions of various operation data of each operation and maintenance node to all event data belonging to various event types, the principal component set of the pipeline lightweight operation and maintenance system is determined;

[0012] Based on the frequency of occurrence of various operating data in the principal component set, the lightweight operating data of the pipeline lightweight operation and maintenance system is determined, and a digital twin of the pipeline lightweight operation and maintenance system is constructed.

[0013] In one embodiment, the process of determining the feature characterization capability is as follows:

[0014] Arrange various operation data of any operation and maintenance node in preset time periods before and after the warning time in any event data thereof in time sequence to form a pre-event data sequence and a post-event data sequence of any event data of any operation and maintenance node;

[0015] Arrange the data in the pre-preset time interval in the data sequence before each event in time sequence to form each normal reference sequence;

[0016] Taking the preset time interval as the interval length, the data sequence after each event is equally divided into data subsequences;

[0017] The distance between each pre-event data sequence and the corresponding post-event data sequence of any operation and maintenance node is recorded as the state distance;

[0018] Calculate the distance between any normal reference sequence of any operation and maintenance node and each data subsequence of the corresponding post-event data sequence, and arrange them in time sequence to form an abnormal state change sequence of any operation and maintenance node;

[0019] The trend test algorithm is used to obtain the trend slope of each state abnormal change sequence;

[0020] The product of the state distance and the trend slope is used as the feature characterization capability of various operation data of any operation and maintenance node with respect to any event data.

[0021] In one embodiment, the event impact degree is the cumulative sum of the feature characterization capabilities of all types of operation data of any operation and maintenance node with respect to any event data.

[0022] In one embodiment, the process of determining the chain node set is:

[0023] The shortest paths from any operation and maintenance node to the remaining operation and maintenance nodes form a node distance set of any operation and maintenance node;

[0024] Determining a chain influence scale of any operation and maintenance node on any event data based on a distribution of event influence degrees of any operation and maintenance node on all event data belonging to the event type to be analyzed;

[0025] The operation and maintenance nodes corresponding to the data that is less than or equal to the chain influence scale in the node distance set form a chain node set of any operation and maintenance node with respect to any event data.

[0026] In one embodiment, the process of determining the linkage impact scale is:

[0027] Calculate the mean of the event impact of any operation and maintenance node on all event data belonging to the event type to be analyzed, recorded as the impact mean, and calculate the sum of the impact mean and a preset value greater than 0;

[0028] The ratio of the event impact degree of any operation and maintenance node on any event data to the sum value is calculated, and a calculation result of rounding up the ratio is used as the chain impact scale of any operation and maintenance node on any event data.

[0029] In one embodiment, the expression of the event contribution is:

[0030] In the formula, Indicates the event contribution of the mth operation data of the operation and maintenance node p to the event data s; represents the feature characterization capability of the mth type of operation data of operation and maintenance node a regarding event data s in the chain node set of operation and maintenance node p regarding event data s; A represents the total number of operation and maintenance nodes in the chain node set of operation and maintenance node p regarding event data s; β represents a preset weight greater than 1, Φ represents the set of operation and maintenance nodes in the chain node set of operation and maintenance node p regarding event data s that have generated event data of the same event type as event data s within the preset time period after the warning time t; wherein t represents the warning time of event data s occurring in operation and maintenance node p.

[0031] In one embodiment, the process of determining the principal component set is:

[0032] The average of the event contributions of any one type of operation data of all operation and maintenance nodes to all event data belonging to various event types is used as the contribution of any one type of operation data to various event types;

[0033] Obtaining a segmentation threshold of contribution of all types of operation data to any event type, and forming a principal component of any event type from all types of operation data greater than the segmentation threshold;

[0034] The principal components of all event types are combined into the principal component set of the pipeline lightweight operation and maintenance system.

[0035] In one embodiment, the process of determining the lightweight operation data is as follows:

[0036] Get all frequent item sets of the principal component set;

[0037] The ratio of the support of various operation data in the principal component set to the total number of event types of the oil and gas transportation pipeline is used as the principal component proportion of various operation data;

[0038] Based on the principal component proportions of all types of operating data in each frequent item set, the candidate item set of the pipeline lightweight operation and maintenance system is determined;

[0039] Based on the total number of types of operating data in frequent itemsets and the number of frequent itemsets in candidate itemsets, the lightweight operating data of the pipeline lightweight operation and maintenance system is determined.

[0040] In one embodiment, the process of determining the candidate item set is as follows:

[0041] The cumulative sum of the principal component proportions of all types of running data in each frequent item set is taken as the selection possibility of each frequent item set;

[0042] All frequent item sets with a selection probability greater than or equal to the preset selection threshold are selected to form the candidate item set of the pipeline lightweight operation and maintenance system.

[0043] In one of the embodiments, when the number of frequent item sets in the candidate item set is greater than 1, the frequent item set containing the smallest total number of operating data types in the candidate item set is used as the lightweight operating data of the pipeline lightweight operation and maintenance system; otherwise, the frequent item set with the greatest possibility is selected as the lightweight operating data of the pipeline lightweight operation and maintenance system.

[0044] This application has at least the following beneficial effects:

[0045] This application comprehensively considers the difference between the operation data of the operation and maintenance node before and after the abnormal behavior occurs and the long-term change trend characteristics after the abnormal behavior occurs under abnormal behavior conditions, builds feature characterization capabilities, improves the accuracy of the evaluation of the characterization capability of any operation data of the operation and maintenance node with respect to event data, and helps to improve the accuracy of subsequent lightweight operation data screening;

[0046] Furthermore, the ability and universality of the abnormal behavior characterization of different event types of the pipeline lightweight operation and maintenance system by operating data are deeply mined to obtain the principal components of different event types, avoiding the problem of poor interpretability of the principal components caused by PCA principal component analysis, while meeting the input data requirements of the Apriori algorithm;

[0047] Furthermore, the Apriori algorithm is used to reduce the dimensionality of multi-dimensional operation data to obtain high-value data assets required for the pipeline lightweight operation and maintenance system. Lightweight operation data is used to replace multi-dimensional operation data to improve the efficiency of digital twin construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 A flowchart of the steps of a method for constructing a digital twin of a pipeline lightweight operation and maintenance system provided in this application;

[0050] Figure 2 A schematic diagram of the process for determining the feature characterization capability;

[0051] Figure 3 It is a schematic diagram of the process of determining the chain node set;

[0052] Figure 4 Schematic diagram of the process of determining lightweight operating data. DETAILED DESCRIPTION

[0053] In the description of the embodiments of the present application, words such as "exemplary", "or", "for example" and the like are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "or", "for example" and the like is intended to present related concepts in a concrete manner.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. It should be understood that, unless otherwise specified, " / " means or.

[0055] It should also be noted that the terms "first" and "second" in the present application are used to distinguish similar objects rather than to describe a specific order or sequence.

[0056] The following is a detailed description of a specific scheme of a method for constructing a digital twin of a pipeline lightweight operation and maintenance system provided by the present application in conjunction with the accompanying drawings.

[0057] An embodiment of the present application provides a method for constructing a digital twin of a pipeline lightweight operation and maintenance system. Specifically, the following method for constructing a digital twin of a pipeline lightweight operation and maintenance system is provided. Figure 1 , the method comprises the following steps:

[0058] Step 1: Collect various operation data of each operation and maintenance node on the oil and gas transportation pipeline at each collection time, and obtain various event data of each operation and maintenance node.

[0059] An operation and maintenance node is set up at every preset distance on the oil and gas transportation pipeline to collect operation and maintenance data. The operation and maintenance data includes operation data and event data. Among them, the operation data includes: arranging pressure sensors, flow meters and temperature and humidity sensors, vibration sensors, and acoustic sensors on the operation and maintenance nodes to collect operation data of the pressure and flow rate inside the pipeline and the temperature, humidity, vibration signal, and audio signal outside the pipeline respectively, and then obtaining the operation data of the power-on potential, protection current, natural potential, DC current density, and AC interference voltage through the cathodic protection system closest to the operation and maintenance node. Among them, the event data includes: the event type of the alarm state of the operation and maintenance node, the warning time, and the operation and maintenance node itself.

[0060] It should be noted that the collected operation and maintenance data is the data of the oil and gas transportation pipeline from the time it was put into use to the present.

[0061] In this embodiment, the value of the preset distance is 500m. The value of the preset distance is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0062] In this embodiment, when collecting operation data, the collection time interval is 30s. The collection time interval is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0063] To ensure the accuracy and consistency of operation and maintenance data, Apache Spark technology is used to clean the operation and maintenance data.

[0064] Any event data is recorded as s, and the event data s includes the event type r, the alarm time t, and the operation and maintenance node p of the operation and maintenance node p. Taking the mth type of operation data as an example, all the mth type of operation data of the operation and maintenance node p in the preset time period before the alarm time t are arranged in time sequence to form the pre-event data sequence H of the operation and maintenance node p about the event data s m , all the mth type of operation data of the operation and maintenance node p within the preset time period after the warning time t are arranged in time sequence to form the post-event data sequence h of the operation and maintenance node p about the event data s m It should be noted that: when the time before or after the warning time t is less than the preset time period, the data filling method is used to fill the mth type of operation data.

[0065] In this embodiment, the preset time period is 1 hour. The value of the preset time period is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0066] In this embodiment, the mean filling method is used for data filling. As other implementation methods, on the basis of achievable data filling, the implementer may use other existing technologies for data filling, such as mode filling, median filling, etc., and this application does not impose any special restrictions.

[0067] Step 2: Determine the feature characterization capability of the various operation data of any operation and maintenance node with respect to any event data based on the difference between the various operation data before and after the warning time in any event data of any operation and maintenance node; and determine the event impact degree of any operation and maintenance node with respect to any event data based on the feature characterization capability.

[0068] When abnormal behavior occurs in the pipeline operation and maintenance node, such as cathodic protection system failure, flow abnormality, etc., the operation and maintenance node will generate event data corresponding to the abnormal behavior. Before the pipeline operation and maintenance node has abnormal behavior and generates event data, the operation and maintenance node may be in normal operation, and the operation data can be used as the benchmark data when the pipeline is operating normally. After the pipeline operation and maintenance node has abnormal behavior and generates event data, the operation and maintenance node is in an abnormal state. Therefore, the difference between the operation data before and after the pipeline abnormal behavior can effectively reflect the change in the pipeline operation state of the operation and maintenance node.

[0069] Before the abnormal behavior of the pipeline operation and maintenance node occurs, the operation and maintenance node may be in normal operating conditions. The earlier the operation data before the warning time of the operation and maintenance node, the more it can be used as the benchmark data for normal operation of the pipeline. This application takes the event data s as an example to obtain the pre-event data sequence H m The data in the preset time interval before and after the event are arranged in time sequence to form a normal benchmark sequence. m The data is equally divided into subsequences. The purpose of dividing the data subsequences is to perform local analysis on the operation data after the abnormal behavior occurs, because the response mechanism of the pipeline lightweight operation and maintenance system may be insensitive, which may lead to further deterioration of the abnormal behavior of the operation and maintenance nodes.

[0070] In this embodiment, the preset time interval is 10 minutes. The value of the preset time interval is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0071] Based on the difference between the mth type of operation data of the operation and maintenance node p before and after the warning time t, the feature representation capability of the mth type of operation data of the operation and maintenance node p with respect to the event data s is determined. The specific process is as follows:

[0072] The distance between the pre-event data sequence and the post-event data sequence of the operation and maintenance node p is recorded as the state distance;

[0073] Calculate the distance between the normal baseline sequence of the operation and maintenance node p and each data subsequence of the post-event data sequence, and arrange them in time sequence to form the abnormal state change sequence of the operation and maintenance node p;

[0074] The trend test algorithm is used to obtain the trend slope of the abnormal state change sequence;

[0075] The product of the state distance and the trend slope is used as the feature characterization capability of the m-th operation data of the operation and maintenance node p with respect to the event data s.

[0076] In this embodiment, the distance between the sequences is the DTW distance. As other implementation methods, on the basis of being able to measure the distance between the two sequences, the implementer may adopt other existing technologies to obtain the distance between the two sequences, such as the Euclidean distance, the Manhattan distance, etc., and this application does not impose any special restrictions.

[0077] In this embodiment, the Theil-Sen Median slope estimation method is used to obtain the trend slope. As other implementation methods, based on the obtainable trend slope, the implementer can use other existing technologies to obtain the trend slope, such as the Mann-Kendall trend test, Sen's Slope test, etc. This application does not impose any special restrictions.

[0078] It should be noted that: the larger the state distance, the more likely the mth operation data is to undergo a state mutation when the operation and maintenance node p exhibits abnormal behavior, the greater the impact of the abnormal behavior on the mth operation data, the more the mth operation data can reflect the abnormal behavior characteristics of the operation and maintenance node p, the stronger the representation ability of the mth operation data, and the greater the feature representation ability; when the difference between the mth operation data of the operation and maintenance node p and the normal operating baseline data when the pipeline is operating normally gradually increases with time, the greater the trend slope, the more the mth operation data can reflect the long-term deterioration state characteristics of the abnormal behavior of the pipeline, and the greater the feature representation ability of the mth operation data. The flow chart for determining the feature representation ability is as follows: Figure 2 shown.

[0079] The feature representation capability of the mth type of operation data of the operation and maintenance node p with respect to the event data s is used to calculate the feature representation capability of various operation data of the operation and maintenance node p with respect to the event data s. The greater the feature representation capability, the more the operation data can reflect the abnormal behavior characteristics of the event data s.

[0080] Calculate the cumulative sum of the characteristic representation capabilities of all types of operation data of operation and maintenance node p on event data s as the event impact degree of operation and maintenance node p on event data s. The greater the event impact degree, the more operation data caused by event data s to mutate and deteriorate, and the greater the impact on the normal operation status of the pipeline.

[0081] Step 3: Based on the location distance relationship between the operation and maintenance nodes and the distribution of the event impact degree of any operation and maintenance node on all event data belonging to the event type to be analyzed, determine the chain node set of any operation and maintenance node on any event data.

[0082] Oil and gas transportation through pipelines is a dynamic process. The oil and gas transportation in the entire pipeline network affects each other. The fluid flow in the pipeline is constrained by the laws of fluid dynamics. When an operation and maintenance node at a certain location behaves abnormally, this influence may spread along the pipeline and have a chain effect on other upstream and downstream operation and maintenance nodes. The more serious the chain effect of the abnormal behavior of the operation and maintenance node on the entire pipeline lightweight operation and maintenance system, the more other upstream and downstream operation and maintenance nodes will be affected.

[0083] Based on the positions of all operation and maintenance nodes on the pipeline, and according to the upstream and downstream relationships of the pipeline operation and maintenance nodes, the upstream operation and maintenance node is used as the parent node, the downstream operation and maintenance node is used as the child node, and the upstreammost operation and maintenance node is used as the root node to construct the pipeline operation and maintenance node tree. The shortest path from the operation and maintenance node p to the remaining operation and maintenance nodes in the pipeline operation and maintenance node tree is calculated, and the node distance set of the operation and maintenance node p is formed.

[0084] Based on the above analysis, the chain impact scale of the operation and maintenance node p on the event data s is determined, and the expression is:

[0085] Where η p,s represents the chain impact scale of the operation and maintenance node p on the event data s; w p,s is the event impact degree of operation and maintenance node p on event data s; U is the mean event impact degree of operation and maintenance node p on all event data belonging to event type r; is a rounding up function; γ represents a value preset to be greater than 0, the purpose of which is to avoid the denominator being 0. The value of γ is preset manually and can be set by the implementer. In this embodiment, the value of γ is 0.1.

[0086] It should be noted that: The larger it is, the more serious the chain effect of the abnormal behavior of the operation and maintenance node on the entire pipeline lightweight operation and maintenance system. The more other upstream and downstream operation and maintenance nodes are affected, the larger the chain effect scale.

[0087] The operation and maintenance nodes corresponding to the data that is less than or equal to the chain impact scale in the node distance set of the operation and maintenance node p are used to form the chain node set of the operation and maintenance node p regarding the event data s. The schematic diagram of the process of determining the chain node set is as follows: Figure 3 shown.

[0088] Step 4: determining the event contribution of any type of operation data of any operation and maintenance node with respect to any event data based on the feature characterization capability of any type of operation data of each operation and maintenance node in the chain node set with respect to any event data.

[0089] For pipeline lightweight operation and maintenance systems, pipeline operation data has high dimensions, making it difficult to conduct effective data governance, and requiring a large amount of resources to build digital twins. The purpose of this application is to obtain operation data that has a high contribution to event data, convert multi-dimensional operation data into lightweight data assets, reduce operation and maintenance costs, and improve the efficiency of data twin construction.

[0090] In the pipeline lightweight operation and maintenance system, each event type corresponds to an abnormal behavior of the pipeline. The abnormal behavior generated by a certain operation and maintenance node may cause abnormal behavior in other operation and maintenance nodes. In the long-term operation and maintenance process of the pipeline, the contribution of any operation data to the abnormal behavior generated first is analyzed. In order to facilitate the lightweight operation and maintenance system of the pipeline to respond to and manage the abnormal behavior generated later, the operation and maintenance node that generates the abnormal behavior later should be given a higher weight.

[0091] The same calculation method as that for the feature characterization capability of various operation data of the operation and maintenance node p with respect to the event data s is used to calculate the feature characterization capability of various operation data of any operation and maintenance node with respect to the event data s.

[0092] Based on the above analysis, the event contribution of the mth operation data of the operation and maintenance node p to the event data s is determined as follows:

[0093] In the formula, Indicates the event contribution of the mth operation data of the operation and maintenance node p to the event data s; It represents the characteristic characterization capability of the mth type of operation data of the operation and maintenance node a regarding the event data s in the chain node set of the operation and maintenance node p regarding the event data s; A represents the total number of operation and maintenance nodes in the chain node set of the operation and maintenance node p regarding the event data s; β represents a weight preset to be greater than 1, and the value of β is preset manually and can be set by the implementer. In this embodiment, the value of β is 2; Φ represents the set of operation and maintenance nodes that have generated event data of the same event type as the event data s within the preset time period after the warning time t in the chain node set of the operation and maintenance node p regarding the event data s.

[0094] It should be noted that: when a∈Φ, in order to facilitate the subsequent response and governance of the pipeline lightweight operation and maintenance system to abnormal behaviors, a is given a higher weight. The more the m-th type of operation data can capture the abnormal behavior characteristics of the pipeline lightweight operation and maintenance system caused by the event data s, the higher the event contribution. The bigger.

[0095] Step 5: Based on the distribution of event contributions of various operation data of each operation and maintenance node to all event data belonging to various event types, a principal component set of the pipeline lightweight operation and maintenance system is determined.

[0096] All event data of the pipeline lightweight operation and maintenance system are divided based on event types. Taking event type r as an example, all event data belonging to event type r are combined into event data set k. r . All the operation and maintenance nodes’ mth type of operation data are related to the event data set k r The mean of the event contribution of all event data in is taken as the contribution of the mth type of operation data to event type r. The segmentation threshold of the contribution of all types of operation data to event type r is obtained, and all types of operation data with a contribution greater than the segmentation threshold are used to form the principal component of event type r. Among them, each type of operation data in the principal component can capture the abnormal behavior characteristics of event type r. As the principal component of event type r, it can serve as the decision-making basis for the subsequent pipeline lightweight operation and maintenance system.

[0097] In this embodiment, the Otsu threshold method is used to obtain the segmentation threshold. As other implementation methods, based on the available segmentation threshold, the implementer can use other existing technologies to obtain the segmentation threshold, such as histogram analysis, quantile method, etc., and this application does not impose any special restrictions.

[0098] According to the acquisition method of the principal component of event type r, the principal component of each event type is obtained. The principal components of all event types are combined into a principal component set of the pipeline lightweight operation and maintenance system.

[0099] Step 6: Based on the frequency of occurrence of various operating data in the principal component set, determine the lightweight operating data of the pipeline lightweight operation and maintenance system, and build a digital twin of the pipeline lightweight operation and maintenance system.

[0100] The Apriori algorithm is used to obtain all frequent item sets of the principal component set. Among them, the combination of operating data in the frequent item set can synergistically reflect the abnormal behavior characteristics of different event types, making it convenient to mine the abnormal behavior of the pipeline lightweight operation and maintenance system from the combination of operating data in the frequent item set.

[0101] In this embodiment, the minimum support value of the Apriori algorithm is 0.5R, where R is the total number of event types of the oil and gas transportation pipeline. The value of the minimum support is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0102] Based on the above analysis, the main component proportion of the mth type of operation data is determined, and the expression is:

[0103] Where Y m represents the proportion of the principal component of the mth running data; x m is the support of the mth type of operation data in the principal component set; R is the total number of event types of the oil and gas transportation pipeline. The calculation of the support is a well-known technology and will not be repeated in this application.

[0104] It should be noted that: The larger it is, the more universal it is to capture abnormal behavior features from different event types based on the m-th type of operation data, and the more it can serve as a decision-making basis for the subsequent pipeline lightweight operation and maintenance system. The m-th type of operation data should be selected to realize the lightweight of the pipeline lightweight operation and maintenance system and improve the efficiency of digital twin construction and updating.

[0105] The cumulative sum of the principal component proportions of all types of operating data in each frequent item set is taken as the selection possibility of each frequent item set, and all frequent item sets with a selection possibility greater than or equal to the preset selection threshold are used to form the candidate item set of the pipeline lightweight operation and maintenance system.

[0106] In this embodiment, the value of the preset selection threshold is 0.8. The value of the preset selection threshold is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0107] When the number of frequent item sets in the candidate set is greater than 1, the total number of types of operation data contained in each frequent item set is calculated, and the frequent item set with the smallest total number of types in the candidate set is used as the lightweight operation data of the pipeline lightweight operation and maintenance system; otherwise, the frequent item set with the greatest possibility is selected as the lightweight operation data of the pipeline lightweight operation and maintenance system. The flowchart for determining lightweight operation data is shown in the figure below. Figure 4 shown.

[0108] Lightweight operation data has excellent abnormal behavior characterization capabilities for different event types in the pipeline lightweight operation and maintenance system, and can serve as the basis for subsequent pipeline lightweight operation and maintenance system decision-making. Lightweight operation data is a high-value digital asset for building digital twins. It can replace multi-dimensional operation data and greatly improve the efficiency of digital twin construction.

[0109] Obtain the CAD drawings, pipeline materials, physical dimensions and geographic locations of each operation and maintenance node in the oil and gas transmission pipeline, and use the three-dimensional modeling method to construct a three-dimensional scene model of the pipeline lightweight operation and maintenance system. And according to the three-dimensional scene model and lightweight operation data of the pipeline lightweight operation and maintenance system, use the three-dimensional engine to render and mount the three-dimensional scene model, give the virtual entity the same function as the corresponding real entity, and can also be responsible for the three-dimensional CAD direction of the pipeline lightweight operation and maintenance system. Modeling editing functions such as model design, transformation, geometric calculation, etc. can be realized through the three-dimensional modeling engine to realize the visualization of the digital twin, thereby improving the efficiency and quality of the operation and maintenance of the oil and gas transmission pipeline. Among them, the specific construction of the digital twin belongs to the known technology, and this application will not repeat it.

[0110] In summary, this application comprehensively considers the difference between the operation data of the operation and maintenance nodes before and after the abnormal behavior occurs and the long-term change trend characteristics after the abnormal behavior occurs under abnormal behavior conditions, builds feature characterization capabilities, improves the accuracy of the evaluation of the characterization capability of any operation data of the operation and maintenance node with respect to event data, and helps to improve the accuracy of subsequent lightweight operation data screening;

[0111] Furthermore, the ability and universality of the abnormal behavior characterization of different event types of the pipeline lightweight operation and maintenance system by operating data are deeply mined to obtain the principal components of different event types, avoiding the problem of poor interpretability of the principal components caused by PCA principal component analysis, while meeting the input data requirements of the Apriori algorithm;

[0112] Furthermore, the Apriori algorithm is used to reduce the dimensionality of multi-dimensional operation data to obtain high-value data assets required for the pipeline lightweight operation and maintenance system. Lightweight operation data is used to replace multi-dimensional operation data to improve the efficiency of digital twin construction.

[0113] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

[0114] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the basic features of the present application. Therefore, no matter from which point of view, the above embodiments of the present application should be regarded as exemplary and non-restrictive.

Claims

1. A method for constructing a digital twin of a pipeline lightweight operation and maintenance system, characterized in that: The method comprises the following steps: Collect various operation data of each operation and maintenance node on the oil and gas transportation pipeline at each collection time, and obtain various event data of each operation and maintenance node. The event data includes the event type, warning time and operation and maintenance node itself when the operation and maintenance node appears in the alarm state; Based on the difference between various operation data before and after the warning time in any event data of any operation and maintenance node, determine the feature characterization capability of various operation data of any operation and maintenance node with respect to any event data; based on the feature characterization capability, determine the event impact degree of any operation and maintenance node with respect to any event data; Recording the event type to which any event data belongs as the event type to be analyzed, and determining a chain node set of any operation and maintenance node with respect to any event data based on the position distance relationship between the operation and maintenance nodes and the distribution of event impact degrees of any operation and maintenance node with respect to all event data belonging to the event type to be analyzed; Determining the event contribution of any one type of operation data of any operation and maintenance node with respect to any one event data based on the feature characterization capability of any one type of operation data of each operation and maintenance node in the chain node set with respect to any one event data; Based on the distribution of event contributions of various operation data of each operation and maintenance node to all event data belonging to various event types, the principal component set of the pipeline lightweight operation and maintenance system is determined; Based on the frequency of occurrence of various operating data in the principal component set, the lightweight operating data of the pipeline lightweight operation and maintenance system is determined, and a digital twin of the pipeline lightweight operation and maintenance system is constructed.

2. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 1, characterized in that: The process of determining the feature characterization capability is as follows: Arrange various operation data of any operation and maintenance node in preset time periods before and after the warning time in any event data thereof in time sequence to form a pre-event data sequence and a post-event data sequence of any operation and maintenance node regarding any event data; Arrange the data in the pre-preset time interval in the data sequence before each event in time sequence to form each normal reference sequence; Taking the preset time interval as the interval length, the data sequence after each event is equally divided into data subsequences; The distance between each pre-event data sequence and the corresponding post-event data sequence of any operation and maintenance node is recorded as the state distance; Calculate the distance between any normal reference sequence of any operation and maintenance node and each data subsequence of the corresponding post-event data sequence, and arrange them in time sequence to form an abnormal state change sequence of any operation and maintenance node; The trend test algorithm is used to obtain the trend slope of each state abnormal change sequence; The product of the state distance and the trend slope is used as the feature characterization capability of various operation data of any operation and maintenance node with respect to any event data.

3. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 1, characterized in that: The event impact degree is the cumulative sum of the feature representation capabilities of all types of operation data of any operation and maintenance node with respect to any event data.

4. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 1, characterized in that: The process of determining the chain node set is as follows: The shortest paths from any operation and maintenance node to the remaining operation and maintenance nodes form a node distance set of any operation and maintenance node; Determining a chain influence scale of any operation and maintenance node on any event data based on a distribution of event influence degrees of any operation and maintenance node on all event data belonging to the event type to be analyzed; The operation and maintenance nodes corresponding to the data that is less than or equal to the chain influence scale in the node distance set form a chain node set of any operation and maintenance node with respect to any event data.

5. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 4, characterized in that: The process of determining the chain effect scale is as follows: Calculate the mean of the event impact of any operation and maintenance node on all event data belonging to the event type to be analyzed, recorded as the impact mean, and calculate the sum of the impact mean and a preset value greater than 0; The ratio of the event impact degree of any operation and maintenance node on any event data to the sum value is calculated, and a calculation result of rounding up the ratio is used as the chain impact scale of any operation and maintenance node on any event data.

6. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 2, characterized in that: The expression of the event contribution is: In the formula, Indicates the event contribution of the mth operation data of the operation and maintenance node p to the event data s; represents the feature representation capability of the mth operation data of operation and maintenance node a in the chain node set of operation and maintenance node p regarding event data s; A represents the total number of operation and maintenance nodes in the chain node set of operation and maintenance node p regarding event data s; β represents a preset weight greater than 1, Φ represents the set of operation and maintenance nodes that have generated event data of the same event type as the event data s, and the chain nodes of the operation and maintenance node p regarding the event data s are concentrated within the preset time period after the warning time t; wherein t represents the warning time when the event data s occurs in the operation and maintenance node p.

7. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 1, characterized in that: The process of determining the principal component set is as follows: The average of the event contributions of any one type of operation data of all operation and maintenance nodes to all event data belonging to various event types is used as the contribution of any one type of operation data to various event types; Obtaining a segmentation threshold of contribution of all types of operation data to any event type, and forming a principal component of any event type from all types of operation data greater than the segmentation threshold; The principal components of all event types are combined into the principal component set of the pipeline lightweight operation and maintenance system.

8. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 1, characterized in that: The process of determining the lightweight operation data is as follows: Get all frequent item sets of the principal component set; The ratio of the support of various operation data in the principal component set to the total number of event types of the oil and gas transportation pipeline is used as the principal component proportion of various operation data; Based on the principal component proportions of all types of operating data in each frequent item set, the candidate item set of the pipeline lightweight operation and maintenance system is determined; Based on the total number of types of operating data in frequent itemsets and the number of frequent itemsets in candidate itemsets, the lightweight operating data of the pipeline lightweight operation and maintenance system is determined.

9. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 8, characterized in that: The process of determining the candidate item set is as follows: The cumulative sum of the principal component proportions of all types of running data in each frequent item set is taken as the selection possibility of each frequent item set; All frequent item sets with a selection probability greater than or equal to the preset selection threshold are selected to form the candidate item set of the pipeline lightweight operation and maintenance system.

10. The method for constructing a digital twin of a pipeline lightweight operation and maintenance system according to claim 9, characterized in that: When the number of frequent itemsets in the candidate item set is greater than 1, the frequent item set with the smallest total number of operating data types in the candidate item set is used as the lightweight operating data of the pipeline lightweight operation and maintenance system; otherwise, the frequent item set with the greatest possibility is selected as the lightweight operating data of the pipeline lightweight operation and maintenance system.

Images