A method, system, and equipment for detecting anomalies in refined oil pipelines based on ranking networks.

Abstract

The present invention relates to a method, system and device for detecting abnormalities of a refined oil pipeline based on a sorting network. The method comprises establishing a simulation pipeline model; simulating the operation of various normal and abnormal working conditions based on the simulation pipeline model to obtain the import and export pressure time series data of each station in the simulation pipeline model; and building a refined oil pipeline abnormality detection model based on a sorting network based on the pressure time series data for detecting the pipeline. The model of the present invention has high accuracy and strong versatility for abnormality detection of refined oil pipelines. At the same time, the model transforms the previous recognition problem into a sorting problem, which greatly improves the detection accuracy and fault tolerance of the model for abnormal situations.

Description

Technical Field

[0001] The present invention relates to a method, system and equipment for detecting abnormality of a finished oil pipeline based on a sorting network, and relates to the technical field of pipeline detection. Background Art

[0002] As one of the main ways to transport oil products, the safety of the finished oil pipeline transportation process is related to the users and even the national energy development. During the transportation process, there are various interferences in the pipeline (such as frequent changes in working conditions caused by scheduling plans). Due to factors such as pipeline damage, equipment failure, improper operation, etc., abnormal situations often occur, seriously affecting the normal production work on site and causing huge economic losses. At present, the on-site monitoring of working conditions mainly relies on manual identification and monitoring, which is not only time-consuming and labor-intensive, but also easy to misjudge the type of working conditions. Therefore, it is of great significance to realize pipeline anomaly detection in real time, accurately and quickly.

[0003] At present, the research on abnormal detection of refined oil pipelines mainly focuses on pipeline leakage. The method based on numerical simulation is a relatively sensitive method and is often used for pipeline anomaly detection. The principle of the method based on numerical simulation is to use fluid mass, momentum and energy equations to simulate the flow and pressure in the pipeline, and compare the predicted values ​​with the measured data to determine the leak and describe the leak characteristics. De Sousa et al. used ANSYS Fluent to study the impact of oil leakage on pressure and flow characteristics. The results obtained revealed how the leak affects the pressure and flow rate near the leak area. Molina-Espinosa et al. conducted numerical modeling supported by pipeline leakage physical experiments and studied the transient model of incompressible flow in a short pipe with a leak. The results obtained showed that there was a good correlation between the simulation and experimental data in terms of pressure drop near the leak. Zhang et al. proposed a liquid pipeline leak detection and location model based on hydrothermal transient analysis, and then used an improved particle swarm (PSO) for optimization. Although the numerical simulation method can calculate the parameters of pipeline leakage more accurately, its model has high requirements on the calculation speed and scale. Moreover, as the length of the pipeline increases, the calculation parameters will also increase, and the calculation speed will decrease accordingly. At the same time, the numerical simulation method cannot be updated dynamically in real time, and cannot detect whether there is any abnormality in the pipeline in real time.

[0004] Intelligent methods based on algorithms and data-driven are currently popular anomaly detection methods. Gong Jun et al. proposed a pipeline leakage detection model based on principal component analysis and RBF neural network. Li et al. established an innovative model for pipeline mutation, using nonlinear time series based on BP neural network to detect leakage. Kang et al. proposed a method that combines one-dimensional convolutional neural network and support vector machine for leak detection. Fukuda uses statistical analysis technology and pressure gradient method to detect leaks of smaller scale leaks. Although the methods based on intelligent algorithms and data-driven make full use of the operating data of the pipeline SCADA system, most of them only perform anomaly detection from the data level, ignoring the topological structure and hydraulic characteristics of the pipeline, and not considering its spatiotemporal characteristics. Finally, the accuracy of abnormal condition detection is low and the false alarm rate is high. In practical applications, the fault tolerance and recognition accuracy of commonly used models will be very low, and it is impossible to make judgments on unknown abnormal conditions. In addition, most detection methods are only targeted at a few specific working conditions and cannot effectively identify other abnormal conditions. Summary of the invention

[0005] In view of the above problems, the object of the present invention is to provide a refined oil pipeline anomaly detection method, system and device based on a sorting network, which can improve the accurate detection of pipeline anomalies.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a method for detecting abnormalities in a refined oil pipeline based on a sorting network, comprising:

[0008] Establish a simulation pipeline model;

[0009] Based on the simulation pipeline model, the operation of various normal and abnormal working conditions is simulated to obtain the import and export pressure time series data of each station in the simulation pipeline model;

[0010] Based on the pressure time series data, a refined oil pipeline anomaly detection model based on sorting network is built to detect the pipeline.

[0011] The method for detecting abnormality of a refined oil pipeline based on a sorting network, further comprising:

[0012] Deform the pressure time series;

[0013] Define p groups of normal working conditions as positive samples, define q groups of normal working conditions and abnormal working conditions as negative samples, pair the positive samples with the negative samples, and each of the two groups of samples is paired only once;

[0014] After pairing is completed, p×q groups of training combination sample pairs are obtained, and their label forms are defined;

[0015] By using a dual-branch artificial neural network with shared weights, the characteristics of the pressure time series are extracted and scored, and the sorting model is built through training.

[0016] The method for detecting anomalies in a refined oil pipeline based on a sorting network further defines, when constructing training combination sample pairs, the label of the sample pair consisting of normal working conditions is defined as 0, and the label of the sample pair consisting of normal working conditions and abnormal working conditions is defined as 1.

[0017] The method for detecting abnormality in a refined oil pipeline based on a sorting network further includes sending an abnormal operating condition label or a normal operating condition label to the network each time during the training process. The learning network attempts to make the score of the abnormal label higher than the normal one. The learning process is as follows:

[0018] P ij =σ(s i -s j )

[0019] s i =f(x i ,w)

[0020] s j =f(x j ,w)

[0021] Among them, P ij represents the probability of the category to which the working condition to be tested belongs, σ(·) represents the sigmoid function, x i and x j represents the pressure feature extracted from the combined sample, w represents the weight of the proposed sorting network, and s i and j Represents the score obtained by the sorting network.

[0022] The method for detecting abnormalities in a refined oil pipeline based on a sorting network further establishes a simulation pipeline model using SPS simulation software.

[0023] In a second aspect, the present invention further provides a refined oil pipeline anomaly detection system based on a sorting network, the system comprising:

[0024] A pipeline model building unit is configured to build a simulation pipeline model;

[0025] The working condition simulation unit is configured to simulate the operation of various normal and abnormal working conditions based on the simulation pipeline model, and obtain the inlet and outlet pressure time series data of each station in the simulation pipeline model;

[0026] The model building unit is configured to build a refined oil pipeline anomaly detection model based on a sorting network based on pressure time series data to detect the pipeline.

[0027] In a third aspect, the present invention further provides an electronic device, comprising at least a processor and a memory, wherein a computer program is stored in the memory, and the processor executes the method when running the computer program.

[0028] In a fourth aspect, the present invention further provides a computer storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions can be executed by a processor to implement the described method.

[0029] The present invention adopts the above technical solution, which has the following advantages:

[0030] 1. Previous research on pipeline anomaly detection mainly applied classification models to detect from the perspective of a single abnormal situation. At the same time, due to the large dimension and noise of pipeline data, the accuracy of abnormal state recognition results is not high, which can easily lead to missed reports and false reports in on-site monitoring work, affecting the normal production of pipelines. The present invention mainly considers the abnormality detection of finished oil pipelines from the two perspectives of normal working conditions and abnormal working conditions, mainly detecting whether abnormal working conditions occur, and providing effective guidance for timely discovery of on-site abnormal conditions;

[0031] 2. The model of the present invention has high accuracy and strong versatility for abnormal detection of finished oil pipelines. At the same time, the model transforms the previous recognition problem into a sorting problem, which greatly improves the detection accuracy and fault tolerance of the model for abnormal situations;

[0032] In summary, the present invention can quickly assist on-site staff in conducting investigations and guide on-site safe operation monitoring and management when an abnormal situation occurs in a finished oil pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Throughout the accompanying drawings, the same reference numerals are used to represent the same components. In the accompanying drawings:

[0034] Figure 1 The overall flow chart of the method of the present invention is as follows;

[0035] Figure 2 A pipeline simulation model of the present invention;

[0036] Figure 3 The oil product pipeline anomaly detection model based on the sorting network of the present invention;

[0037] Figure 4 It is the recognition result of each model on the simulation pipeline 1 of the present invention;

[0038] Figure 5 It is the recognition result of each model on the simulation pipeline 2 of the present invention;

[0039] Figure 6 This is a structural diagram of an electronic device of the present invention. DETAILED DESCRIPTION

[0040] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0041] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used herein may also be meant to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The method steps, processes, and operations described herein are not interpreted as necessarily requiring them to be performed in the specific order described or illustrated, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps may be used.

[0042] Although the terms first, second, third, etc. can be used in the text to describe multiple elements, components, regions, layers and / or sections, these elements, components, regions, layers and / or sections should not be limited by these terms. These terms can only be used to distinguish an element, component, region, layer or section from another region, layer or section. Unless the context clearly indicates, terms such as "first", "second" and other numerical terms do not imply order or sequence when used in the text. Therefore, the first element, component, region, layer or section discussed below can be referred to as the second element, component, region, layer or section without departing from the teaching of the example embodiments.

[0043] For ease of description, spatially relative terms may be used herein to describe the relationship of one element or feature relative to another element or feature as shown in the figures, such as "inside", "outside", "inner side", "outer side", "below", "above", etc. Such spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation depicted in the figures.

[0044] With the rapid development of data-driven technology, analysis methods based on field data have gradually become one of the research hotspots of anomaly detection technology. If the time series and physical space characteristics of pipeline working condition switching data can be fully explored, various working condition information can be uniformly processed, and the import and export pressure sequences of each station can be comprehensively analyzed, the abnormal detection of finished oil pipelines can be realized, and the staff can be assisted to respond to pipeline abnormalities in a timely manner to reduce on-site production losses. The abnormal detection of finished oil pipelines provided by the present invention first uses the real parameters of the pipeline and ancillary equipment to simulate pipeline modeling, simulate the operation of various normal and abnormal working conditions; then based on the simulated pipeline data, the temporal and spatial characteristics and hydraulic characteristics of the pressure of each station of the pipeline are considered to construct a pressure time series; finally, a sorting model based on an artificial neural network is built to detect the working condition to be tested. The present invention converts the working condition classification problem into a sorting problem, so that the model has a higher tolerance for abnormal working condition detection, and can make effective judgments on unknown abnormal conditions, which has important guiding significance for the safe operation monitoring of the on-site finished oil pipeline.

[0045] Embodiment 1

[0046] like Figure 1 As shown, the method for detecting abnormalities in a refined oil pipeline based on a sorting network provided in this embodiment includes the following steps:

[0047] S1. Use the real parameters of pipelines and auxiliary equipment to establish a simulation pipeline model

[0048] Specifically, this embodiment decomposes the finished oil pipeline system into multiple sub-units, including the first station A, an intermediate station B and the terminal station C. By modeling the real pipeline, the corresponding parameters of the equipment are collected: the rated head, rated power, rated flow, valve opening, oil physical properties, pipeline length, friction coefficient, etc. of the oil pump, and the SPS (Stoner Pipeline Simulator) simulation software is used to establish a simulation model of the finished oil pipeline, and it is used to simulate normal and abnormal working conditions. After the simulation, the inlet and outlet pressure data of the three stations are obtained. These inlet and outlet pressure data are corresponding to the working condition type, and a simulation working condition database is established to facilitate the verification of the abnormal detection model.

[0049] like Figure 2As shown, the first station A of the pipeline simulation model includes an oil depot, an oil pump, and a valve; the intermediate station B includes an oil depot, an oil pump, and a valve; the terminal station C includes an oil depot and a valve. The three stations are connected by two pipelines to form an integrated finished oil transportation system.

[0050] S2. Simulate the operation of various normal and abnormal conditions based on the simulation pipeline model

[0051] The established pipeline simulation model is used to simulate common normal and abnormal conditions, and the inlet and outlet pressure time series data of each station in the pipeline are obtained. Among them, the operating condition according to the scheduling plan is called the normal operating condition, which includes steady-state conditions and unsteady-state conditions; abnormal conditions refer to abnormal conditions that occur in the system outside the scheduling plan. Considering the complex hydraulic characteristics and spatiotemporal characteristics of the pipeline system, the inlet and outlet pressure series of each station are exported as a 4×N (N represents the length of the pressure series) sequence matrix:

[0052]

[0053] In the formula, is the first station exit pressure, The pressure of entering the intermediate station, is the exit pressure of the intermediate station, is the entry pressure of the last station, and t is the time scale of the sequence matrix.

[0054] S3. Based on the pressure time series data, a refined oil pipeline anomaly detection model based on a sorting network is built to detect the working conditions to be tested.

[0055] like Figure 3 As shown, the refined oil pipeline anomaly detection model based on the sorting network in this embodiment adopts a dual-branch ANN (Artificial Neural Networks) with shared weights. There is a connection between the two networks and the weights are shared, so that the features and similarities and differences between the two samples in the sample pair can be captured at the same time. The steps include:

[0056] Firstly, the pressure time series matrix is ​​deformed and converted into a one-dimensional sequence form of 1×4N to facilitate the input and training of the model.

[0057] Then, p groups of normal working conditions are defined as positive samples, q groups of normal working conditions and abnormal working conditions are defined as negative samples, and the positive samples and negative samples are paired with each other, and each of the two groups of samples is paired only once. After the pairing is completed, p×q groups of combined sample pairs are obtained, and their label forms are defined. When constructing the training combined samples, the label of the sample pair composed of normal working conditions is defined as 0, and the label of the sample pair composed of normal working conditions and abnormal working conditions is defined as 1.

[0058] Finally, the features of the pressure sequence are extracted with the help of ANN, and scores are given. The sorting model is built through training, and the sorting model can be applied to anomaly detection in finished oil pipelines.

[0059] Specifically, during the training process, the combined sample pairs send an abnormal condition label or a normal condition label to the network each time, and the learning network attempts to make the score of the abnormal label higher than the normal one. The learning process is:

[0060] P ij =σ(s i -s j ) (2)

[0061] s i =f(x i ,w) (3)

[0062] s j =f(x j ,w) (4)

[0063] Among them, P ij represents the probability of the category to which the working condition to be tested belongs, σ(·) represents the sigmoid function, x i and x j represents the pressure feature extracted from the combined sample, w represents the weight of the proposed sorting network, and s i and j Represents the score obtained by the sorting network.

[0064] S4. Set evaluation indicators for model evaluation

[0065] In the field of machine learning, accuracy, precision, recall, and F1score are often used to measure the recognition performance of the model.

[0066] Table 1 Confusion matrix

[0067]

[0068]

[0069]

[0070]

[0071]

[0072] Wherein, Accuracy: represents the proportion of correctly predicted working conditions to the total working conditions; Precision: represents the proportion of predicted working conditions that are actually positive (T); Recall: represents the proportion of positive working conditions that are truly predicted; F1score: comprehensively represents the evaluation results of Precision and Recall.

[0073] Two simulation pipelines are used as research objects to verify the effectiveness of the sorting network-based anomaly detection model of this embodiment.

[0074] First, the equipment parameters and oil properties of two real finished oil pipelines were collected, and the SPS simulation models were established based on them. Based on the common working conditions, the corresponding non-steady-state and abnormal working conditions of the two simulated pipelines were simulated. The pressure series matrix with a dimension of 4×N was derived from the simulation model, and then it was converted into a one-dimensional time series form to complete the preliminary processing of the pressure data. The data set of the simulation pipeline is shown in Table 2. The processed data set is divided into a training set and a validation set, with the training set accounting for 80% and the validation set accounting for 20%.

[0075] Table 2 Simulation data set

[0076]

[0077] The training set and validation set are introduced into the trained sorting model to realize the anomaly detection of the finished oil pipeline. To prove the superiority of the proposed anomaly detection model based on the sorting network, it is compared with common machine learning models such as ANN (artificial neural network), DT (decision tree), RF (random forest), KNN (nearest neighbor), SVM (support vector machine), GB (gradient boosting). The recognition results of the anomaly detection model on the simulated pipeline 1 dataset are shown in Figure 2. Figure 4 As shown. Figure 4 It can be seen that compared with other machine learning models, the proposed sorting model has the best recognition effect on the simulation data set, with its accuracy, precision, recall and F1score reaching 98.7%, 98.7%, 98.6% and 98.5% respectively. The reason is that while maintaining the time series characteristics of the original data set, the present invention uses a dual network structure to comprehensively mine the potential information of the pressure data, enhances the stability of the model in a probabilistic sorting manner, and realizes the abnormal detection of the finished oil pipeline.

[0078] To verify the generalization performance of the proposed sorting model, we tested it again on the dataset of simulation pipeline 2. We processed the dataset of simulation pipeline 2 in the same way, substituted it into the proposed sorting model and other machine learning models, and compared the results of various indicators. The final recognition results are shown in Figure 2. Figure 5As shown in the figure, the performance of the proposed sorting model is still the best, with accuracy, precision, recall and F1score reaching 98.9%, 98.9%, 98.5% and 98.5% respectively. At the same time, the recognition effect of the proposed model on the two simulated pipelines is similar, which shows that the model has strong generalization ability and can be applied to the on-site pipeline operation safety monitoring work.

[0079] Embodiment 2

[0080] The above-mentioned embodiment 1 provides a refined oil pipeline anomaly detection method based on a sorting network, and correspondingly, this embodiment provides a refined oil pipeline anomaly detection system based on a sorting network. The system provided in this embodiment can implement the refined oil pipeline anomaly detection method based on a sorting network of embodiment 1, and the system can be implemented by software, hardware, or a combination of software and hardware. For the convenience of description, the functions of this embodiment are divided into various units and described separately. Of course, the functions of each unit can be implemented in the same or more software and / or hardware during implementation. For example, the system may include integrated or separate functional modules or functional units to perform the corresponding steps in each method of embodiment 1. Since the system of this embodiment is basically similar to the method embodiment, the description process of this embodiment is relatively simple, and the relevant parts can refer to the partial description of embodiment 1. The embodiment of the refined oil pipeline anomaly detection system based on a sorting network provided by the present invention is only schematic.

[0081] The oil product pipeline anomaly detection system based on the sorting network provided in this embodiment includes:

[0082] A pipeline model building unit is configured to build a simulation pipeline model;

[0083] The working condition simulation unit is configured to simulate the operation of various normal and abnormal working conditions based on the simulation pipeline model, and obtain the inlet and outlet pressure time series data of each station in the simulation pipeline model;

[0084] The model building unit is configured to build a refined oil pipeline anomaly detection model based on a sorting network based on pressure time series data to detect the working condition to be tested.

[0085] Embodiment 3

[0086] This embodiment provides an electronic device corresponding to the method for detecting anomalies in a refined oil pipeline based on a sorting network provided in the first embodiment. The electronic device may be an electronic device for a client, such as a mobile phone, a laptop computer, a tablet computer, a desktop computer, etc., to execute the method in the first embodiment.

[0087] like Figure 6As shown, the electronic device includes a processor, a memory, a communication interface and a bus, and the processor, the memory and the communication interface are connected through the bus to complete mutual communication. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Component (EISA) bus, etc. The memory stores a computer program that can be run on the processor, and when the processor runs the computer program, it executes the refined oil pipeline anomaly detection method based on the sorting network provided in the first embodiment. Those skilled in the art can understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present application, and does not constitute a limitation on the computing device to which the scheme of the present application is applied. The specific computing device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0088] In some implementations, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), optical disk and other media that can store program codes.

[0089] In some other implementations, the processor may be a central processing unit (CPU), a digital signal processor (DSP), or other general-purpose processors of various types, which are not limited herein.

[0090] Embodiment 4

[0091] The method for detecting anomalies in a refined oil pipeline based on a sorting network of the first embodiment may be specifically implemented as a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing the method for detecting anomalies in a refined oil pipeline based on a sorting network of the first embodiment.

[0092] Computer readable storage media can be tangible devices that hold and store instructions used by instruction execution devices. Computer readable storage media can be, for example, but not limited to, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any combination thereof.

[0093] Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. In the description of this specification, the description of the reference terms "one embodiment", "some implementations", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiment of this specification. In this specification, the schematic representation of the above terms does not necessarily target the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples without contradiction.

[0094] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0095] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0096] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting abnormalities in a refined oil pipeline based on a sorting network, characterized in that include: Establish a simulation pipeline model; Based on the simulation pipeline model, the operation of various normal and abnormal working conditions is simulated to obtain the import and export pressure time series data of each station in the simulation pipeline model; Based on the pressure time series data, a refined oil pipeline anomaly detection model based on a sorting network is built to detect the pipeline to be tested; among them: The construction of the refined oil pipeline anomaly detection model based on the sorting network includes: Deform the pressure time series; Define p groups of normal working conditions as positive samples, define q groups of normal working conditions and abnormal working conditions as negative samples, pair the positive samples with the negative samples, and each of the two groups of samples is paired only once; After pairing is completed, p×q groups of training combination sample pairs are obtained, and their label forms are defined; By using a dual-branch artificial neural network with shared weights to extract the features of the pressure time series and give a score, the construction of a refined oil pipeline anomaly detection model based on a sorting network is completed through training; During the training process, the training combination sample sends an abnormal condition label or a normal condition label to the network each time, and the learning network tries to make the score of the abnormal label higher than the normal one. The learning process is: P ij =σ(s i -s j ) s i =f(x i ,w) s j =f(x j ,w) Among them, P ij represents the probability of the category to which the working condition to be tested belongs, σ(·) represents the sigmoid function, x i and x j represents the pressure feature extracted from the combined sample, w represents the weight of the proposed sorting network, and s i and j Represents the score obtained by the sorting network.

2. The method for detecting abnormalities in a refined oil pipeline based on a sorting network according to claim 1, characterized in that: When constructing training combination sample pairs, the label of the sample pair consisting of normal working conditions is defined as 0, and the label of the sample pair consisting of normal working conditions and abnormal working conditions is defined as 1.

3. The method for detecting abnormality in a refined oil pipeline based on a sorting network according to claim 1, characterized in that: The simulation pipeline model was established using SPS simulation software.

4. A refined oil pipeline anomaly detection system based on a sorting network, characterized in that: The system includes: A pipeline model building unit is configured to build a simulation pipeline model; The working condition simulation unit is configured to simulate the operation of various normal and abnormal working conditions based on the simulation pipeline model, and obtain the inlet and outlet pressure time series data of each station in the simulation pipeline model; The model building unit is configured to build a refined oil pipeline anomaly detection model based on a sorting network based on pressure time series data to detect the pipeline to be tested; wherein, The construction of the refined oil pipeline anomaly detection model based on the sorting network includes: Deform the pressure time series; Define p groups of normal working conditions as positive samples, define q groups of normal working conditions and abnormal working conditions as negative samples, pair the positive samples with the negative samples, and each of the two groups of samples is paired only once; After pairing is completed, p×q groups of training combination sample pairs are obtained, and their label forms are defined; By using a dual-branch artificial neural network with shared weights to extract the features of the pressure time series and give a score, the construction of a refined oil pipeline anomaly detection model based on a sorting network is completed through training; During the training process, the training combination sample sends an abnormal condition label or a normal condition label to the network each time, and the learning network tries to make the score of the abnormal label higher than the normal one. The learning process is: P ij =σ(s i -s j ) s i =f(x i ,w) s j =f(x j ,w) Among them, P ij represents the probability of the category to which the working condition to be tested belongs, σ(·) represents the sigmoid function, x i and x j represents the pressure feature extracted from the combined sample, w represents the weight of the proposed sorting network, and s i and j Represents the score obtained by the sorting network.

5. An electronic device, comprising at least a processor and a memory, wherein a computer program is stored in the memory, wherein: When the processor runs the computer program, the method according to any one of claims 1 to 3 is implemented.

6. A computer storage medium, characterized in that: Computer-readable instructions are stored thereon, and the computer-readable instructions can be executed by a processor to implement the method according to any one of claims 1 to 3.

Images