A pipeline safety feature extraction method based on Gram angle field

Abstract

The present invention discloses a pipeline safety feature extraction method based on a Gram angle field, which relates to the technical field of pipeline monitoring, and solves the technical problems that it is difficult to convert pre-processed time series data into images, and it is difficult to extract and analyze pipeline parameters based on the converted images; then, it is difficult to analyze pipeline operation anomalies; finally, it is difficult to comprehensively analyze pipeline safety feature values ​​​​and set thresholds to evaluate pipeline safety levels. The present invention includes the following modules: a data acquisition module, a data processing module, a parameter extraction module, a data analysis module, and a safety assessment module. The pre-processed time series data is converted into an image using the Gram angle field method, and then the pipeline parameters are extracted and analyzed. According to the image, the pipeline operation anomalies are analyzed to monitor the safety status of the pipeline in real time; the pipeline safety feature values ​​are analyzed based on various data and the pipeline safety level is evaluated to ensure the safe operation of the pipeline.

Description

Technical Field

[0001] The invention belongs to the field of pipeline monitoring, and in particular is a pipeline safety feature extraction method of a Gram angle field. Background Art

[0002] With the development of industrial automation, pipelines are important energy transmission facilities, and their safe operation is of vital importance. However, due to factors such as long-term operation, environmental corrosion, and material aging, the safety of pipelines is seriously threatened. Traditional pipeline safety monitoring relies on manual inspections, which is inefficient and difficult to detect early hidden dangers.

[0003] The existing pipeline safety feature extraction method of the Gram Corner field has the following problems: first, it is difficult to convert the preprocessed time series data into images, and it is difficult to extract and analyze pipeline parameters based on the converted images; second, it is difficult to analyze the operation anomalies of the pipeline; finally, it is difficult to comprehensively analyze the pipeline safety feature values ​​​​and set thresholds to evaluate the pipeline safety level. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a pipeline safety feature extraction method of the Gram angle field, which is used to solve the following technical problems:

[0005] First, it is difficult to convert the preprocessed time series data into images, and it is difficult to extract and analyze pipeline parameters based on the converted images; then, it is difficult to analyze pipeline operation anomalies; finally, it is difficult to comprehensively analyze the pipeline safety characteristic values ​​and set thresholds to evaluate the pipeline safety level.

[0006] To solve the above problems, the first aspect of the present invention provides a pipeline safety feature extraction method of Gram angle field, comprising the following modules:

[0007] Data collection module: collects time series data during pipeline operation;

[0008] Data processing module: pre-process the collected time series data;

[0009] Parameter extraction module: convert the preprocessed time series data into images using the Gram angular field method; extract pipeline parameters based on the converted results;

[0010] Data analysis module: Calculate pipeline variation coefficient, variable speed pressure value and pipeline bending influence rate based on pipeline parameter analysis; Analyze pipeline operation anomalies based on images converted from Gram angle field; Calculate pipeline operation anomaly index based on analysis results;

[0011] Safety assessment module: Comprehensively analyze and calculate pipeline safety characteristic values ​​to evaluate the safety level of the pipeline.

[0012] As a further solution of the present invention: collecting time series data during pipeline operation includes the following steps:

[0013] Pressure sensors, flow meters and temperature sensors are installed around the pipeline to collect real-time time series data including: pressure data, flow data and temperature data.

[0014] As a further solution of the present invention: preprocessing the collected data includes the following steps:

[0015] By analyzing the formula:

[0016] The collected time series data Normalize to the interval [0, 1] to get the processed time series data ,in represents the data point at the i-th moment, .

[0017] As a further solution of the present invention: converting the pre-processed time series data into an image by using the Gram angular field method comprises the following steps:

[0018] By analyzing the formula:

[0019]

[0020] get The angle between ;in, represents the inner product of vectors, Represents vectors and vector The module length, represents the data point at the i-th moment, represents the data point at the jth moment, , d is the dimension of each vector, and They are vectors and vector The component in the kth dimension;

[0021] By analyzing the formula:

[0022] Get the element values ​​of the Gram angular field matrix , generate an n×n symmetric matrix;

[0023] By analyzing the formula:

[0024] Get the element value of the Gram angular field matrix to generate the pixel value of the image ;in, express The angle value between .

[0025] As a further solution of the present invention: extracting pipeline parameters according to the transformed results includes the following steps:

[0026] Based on the images transformed by the Gram angular field method from the preprocessed time series data, pipeline parameters are extracted using a deep learning model;

[0027] The marked pipeline parameters include: pipeline length, pipeline temperature, pipeline diameter, pipeline cross-sectional area, pipeline fluid density, pipeline volume flow rate, pipeline bending angle and pipeline bend length;

[0028] Input the annotated pipeline parameters into the deep learning model for model training;

[0029] The pipeline parameters are output by transforming the preprocessed time series data through the Gram angular field method.

[0030] As a further solution of the present invention: calculating the pipeline variation coefficient, the speed change pressure value and the pipeline bending influence rate according to the pipeline parameter analysis, including the following steps:

[0031] By analyzing the formula:

[0032] Get the pipeline variation coefficient ,in Indicates the change in length of the pipeline. represents the temperature change, represents the initial pipe length;

[0033] By analyzing the formula:

[0034]

[0035] Get the speed change pressure value , represents the friction coefficient, Indicates the change in length of the pipeline. D represents the change in pipe diameter. represents the fluid density, represents the fluid velocity, Q represents the volume flow rate through the pipe, Indicates the cross-sectional area of ​​the pipe;

[0036] By analyzing the formula:

[0037] The pipe bending influence rate R is obtained, where: Indicates the pipe bending angle, Indicates the pipe bend length.

[0038] As a further solution of the present invention: analyzing the operation abnormality of the pipeline according to the image after the Gram angle field conversion, the following steps are included:

[0039] Based on the images transformed by the Gram angle field, the pipeline operation anomalies are analyzed through the deep learning model;

[0040] Mark m types of pipeline operation anomalies;

[0041] Input the annotated pipeline operation abnormality data into the deep learning model for model training;

[0042] Through the images transformed by the Gram angular field method after preprocessing the time series data, m types of pipeline operation anomalies are identified and output.

[0043] As a further solution of the present invention: calculating the pipeline operation abnormality index according to the analysis results includes the following steps:

[0044] By analyzing the formula:

[0045] The pipeline operation abnormality index K is obtained, where: Indicates The number of times the pipeline operation abnormality occurs. Indicates The weight of pipeline operation abnormality, Indicates the total number of times m types of pipeline operation abnormalities occur. .

[0046] As a further solution of the present invention: comprehensively analyzing and calculating the pipeline risk assessment value to assess the risk status of the pipeline includes the following steps:

[0047] By analyzing the formula:

[0048] Get pipeline safety characteristic value ,in, represents the pipeline variation coefficient, represents the speed change pressure value, R represents the pipeline bending influence rate, K represents the pipeline operation abnormality index, , , and They represent the weights of pipeline variation coefficient, variable speed pressure value, pipeline bending influence rate and pipeline operation abnormality index respectively;

[0049] According to the pipeline safety characteristic value , the safety level of pipelines is divided into three levels:

[0050] when < It is the first level of security;

[0051] When 0.3≤ When <0.7, it is the second-level safety level;

[0052] when When ≥0.7, it is the third level safety level.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] The present invention converts the preprocessed time series data into images by using the Gram angular field method, extracts and analyzes pipeline parameters through the converted images, and realizes accurate extraction and real-time monitoring of pipeline parameters.

[0055] The present invention can analyze the operation abnormality of the pipeline through the converted image, and can monitor the abnormal condition of the pipeline in real time, which is helpful to timely identify the operation abnormality of the pipeline;

[0056] The present invention analyzes pipeline safety characteristic values ​​by comprehensively analyzing various data and setting thresholds to evaluate the safety level of the pipeline, which helps to more accurately understand the safety status of the pipeline and ensure the safe and stable operation of the pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 It is a schematic diagram of the system module framework of the present invention. DETAILED DESCRIPTION

[0059] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1 The first aspect of the present invention provides a pipeline safety feature extraction method of a Gram angle field, comprising the following modules:

[0061] Data collection module: collects time series data during pipeline operation;

[0062] Data processing module: pre-process the collected time series data;

[0063] Parameter extraction module: convert the preprocessed time series data into images using the Gram angular field method; extract pipeline parameters based on the converted results;

[0064] Data analysis module: Calculate pipeline variation coefficient, variable speed pressure value and pipeline bending influence rate based on pipeline parameter analysis; Analyze pipeline operation anomalies based on images converted from Gram angle field; Calculate pipeline operation anomaly index based on analysis results;

[0065] Safety assessment module: Comprehensively analyze and calculate pipeline safety characteristic values ​​to evaluate the safety level of the pipeline.

[0066] Specifically, the Gram angle field is a method for converting time series data into images. In pipeline safety monitoring, the Gram angle field can be used to effectively extract pipeline parameters. The present invention includes the following modules: a data acquisition module, a data processing module, a parameter extraction module, a data analysis module and a safety assessment module. The time series data is collected from the pipeline operation through the data acquisition module. The collected data is preprocessed according to the data processing module to ensure accuracy and reliability. The Gram angle field method is used to convert the time series data into an image through the parameter extraction module, and the pipeline parameters are further analyzed through the data analysis module to calculate the pipeline variation coefficient, the speed change pressure value and the pipeline bending influence rate. The pipeline operation abnormality is analyzed through the converted image to determine the abnormality index. Finally, combined with the safety assessment module, based on all analysis results, the safety characteristic value of the pipeline is analyzed, and a threshold is set to evaluate the pipeline safety level to ensure the safe operation of the pipeline.

[0067] In one embodiment of the present invention, collecting time series data during pipeline operation includes the following steps:

[0068] Pressure sensors, flow meters and temperature sensors are installed around the pipeline to collect real-time time series data including: pressure data, flow data and temperature data.

[0069] Specifically, the pressure sensor is used to measure the pressure change inside the pipeline. The flow meter is used to measure the fluid flow rate or flow rate in the pipeline. The temperature sensor is used to monitor the temperature change of the pipeline and its surrounding environment. The time series data collected by the pressure sensor, flow meter and temperature sensor include pipeline pressure, flow rate and temperature, where time series data refers to a series of observations collected at consecutive time points.

[0070] In one embodiment of the present invention, the collected data is preprocessed, including the following steps:

[0071] By analyzing the formula:

[0072] The collected time series data Normalize to the interval [0, 1] to get the processed time series data ,in represents the data point at the i-th moment, .

[0073] Specifically, by determining the minimum and maximum values ​​in the time series data, the data is normalized using the normalization formula. , after normalization, , rescale all data points to the range of [0, 1]. This means that the minimum value in the original data will be mapped to 0, the maximum value will be mapped to 1, and all other values ​​will be proportionally distributed between these two extreme values. It helps to eliminate the impact of dimension, so that data of different scales can be analyzed under the same standard.

[0074] In one embodiment of the present invention, the pre-processed time series data is converted into an image by using the Gram angular field method, comprising the following steps:

[0075] By analyzing the formula:

[0076]

[0077] get The angle between ;in, represents the inner product of vectors, Represents vectors and vector The module length, represents the data point at the i-th moment, represents the data point at the jth moment, , d is the dimension of each vector, and They are vectors and vector The component in the kth dimension;

[0078] By analyzing the formula:

[0079] Get the element values ​​of the Gram angular field matrix , generate an n×n symmetric matrix;

[0080] By analyzing the formula:

[0081] Get the element value of the Gram angular field matrix to generate the pixel value of the image ;in, express The angle value between .

[0082] Specifically, for time series data, each data point can be regarded as a point in the vector space. In the one-dimensional case, each data point is represented as a scalar; in the multi-dimensional case, each data point is a vector. The cosine value of the angle between the data point at the i-th moment and the data point at the j-th moment is obtained by dividing the vector inner product of the data point at the i-th moment and the data point at the j-th moment by the product of the two vector norms, and the angle value is obtained by the inverse cosine function. The element value of the Gram angular field matrix is ​​obtained by calculating the angle between these vectors. For any two data points at any time and , the angle between them Can be used as an element of the matrix. For n data points, we can get an n×n symmetric matrix, where each element Finally, converting the element values ​​of the Gram angular field matrix into pixel values ​​of the image involves converting the angle values ​​into grayscale values ​​or other color codes, specifically converting the angle values ​​in [0, π] into grayscale values ​​in [0, 255] to obtain the pixel values ​​of the image.

[0083] In one embodiment of the present invention, extracting pipeline parameters according to the transformed result includes the following steps:

[0084] Based on the images transformed by the Gram angular field method from the preprocessed time series data, pipeline parameters are extracted using a deep learning model;

[0085] The marked pipeline parameters include: pipeline length, pipeline temperature, pipeline diameter, pipeline cross-sectional area, pipeline fluid density, pipeline volume flow rate, pipeline bending angle and pipeline bend length;

[0086] Input the annotated pipeline parameters into the deep learning model for model training;

[0087] The pipeline parameters are output by transforming the preprocessed time series data through the Gram angular field method.

[0088] Specifically, on the converted image, the annotated pipeline parameters include: pipeline length, pipeline temperature, pipeline diameter, pipeline cross-sectional area, pipeline fluid density, pipeline volume flow, pipeline bending angle and pipeline bending length. The annotated pipeline parameters are input into the deep learning model for training. The deep learning model has powerful feature learning and extraction capabilities and can automatically learn useful information from a large amount of data. Through training, the model can learn how to extract pipeline parameters from images. The trained deep learning model can extract pipeline parameters and output results. When a new image is input, the model can extract the pipeline parameters in the image based on its annotated pipeline parameters and output them for further analysis.

[0089] In one embodiment of the present invention, the pipeline variation coefficient, the speed change pressure value and the pipeline bending influence rate are calculated according to the pipeline parameter analysis, including the following steps:

[0090] By analyzing the formula:

[0091] Get the pipeline variation coefficient ,in Indicates the change in length of the pipeline. represents the temperature change, represents the initial pipe length;

[0092] By analyzing the formula:

[0093]

[0094] Get the speed change pressure value , represents the friction coefficient, Indicates the change in length of the pipeline. D represents the change in pipe diameter. represents the fluid density, represents the fluid velocity, Q represents the volume flow rate through the pipe, Indicates the cross-sectional area of ​​the pipe;

[0095] By analyzing the formula:

[0096] The pipe bending influence rate R is obtained, where: Indicates the pipe bending angle, Indicates the pipe bend length.

[0097] Specifically, the pipeline variation coefficient is usually used to describe the degree of linear expansion or contraction of the pipeline under temperature changes. The pipeline variation coefficient is obtained by analyzing and calculating the pipeline length and temperature in the pipeline parameters. The variable speed pressure value is calculated based on the flow rate and flow of the fluid in the pipeline, combined with the pipeline length, the change in pipeline diameter, the cross-sectional area and the friction coefficient; the faster the fluid flows in the pipeline, the greater the impact and friction on the pipe wall, which may cause greater pressure changes and affect the flow resistance and pressure loss of the fluid in the pipeline. Different fluids have different flow characteristics and will have different effects on pipeline pressure. At the same time, the variable speed pressure value is also an important basis for evaluating the safety and stability of the pipeline. The pipeline bending influence rate is obtained by calculation and analysis based on the pipeline bending angle and pipeline bending length in the real-time detected pipeline parameters.

[0098] In one embodiment of the present invention, analyzing the operation abnormality of the pipeline according to the image after Gram's angle field conversion includes the following steps:

[0099] Based on the images transformed by the Gram angle field, the pipeline operation anomalies are analyzed through the deep learning model;

[0100] Mark m types of pipeline operation anomalies;

[0101] Input the annotated pipeline operation abnormality data into the deep learning model for model training;

[0102] Through the images transformed by the Gram angular field method after preprocessing the time series data, m types of pipeline operation anomalies are identified and output.

[0103] Specifically, the preprocessed time series data is converted into an image by the Gram angular field method. This conversion method can convert one-dimensional time series data into an image representation, so as to display the data more intuitively. On the converted image, the data of pipeline operation abnormality is marked. The marked pipeline operation abnormality data is input into the deep learning model for training. The trained deep learning model can be used to identify and output pipeline operation abnormalities. When a new image is input, the model can identify pipeline operation abnormalities in the image according to its marked pipeline operation abnormality data, and analyze the results for further analysis. The marked m types of pipeline operation abnormalities include but are not limited to: blockage, leakage, corrosion, wear, abnormal vibration, abnormal temperature, abnormal pressure, abnormal flow, abnormal noise and electrical failure. Specifically, it is manifested as: fluid impurity accumulation leads to obstruction of circulation, connection defects cause resource leakage, material chemical reaction causes wall thickness thinning, high-speed flow scours the inner wall, fluid impact and mechanical vibration induce structural fatigue, temperature fluctuations affect material properties, pressure fluctuations cause deformation or rupture, abnormal increase or decrease in fluid transmission capacity affects transmission efficiency, abnormal noise reflects flow instability, and electrical systems have short circuits, disconnections and overheating. After the model training is completed, the images transformed by the Gram angular field method using the preprocessed time series data are used to identify and output m types of pipeline operation anomalies.

[0104] In one embodiment of the present invention, calculating the pipeline operation abnormality index according to the analysis result includes the following steps:

[0105] By analyzing the formula:

[0106] The pipeline operation abnormality index K is obtained, where: Indicates The number of times the pipeline operation abnormality occurs. Indicates The weight of pipeline operation abnormality, Indicates the total number of times m types of pipeline operation abnormalities occur. .

[0107] Specifically, according to the image transformed by the Gram angular field method based on the preprocessed time series data, m types of pipeline operation anomaly analysis pipeline operation anomaly index are output. The ratio of the number of times the pipeline operation anomaly occurs to the total number of times the pipeline operation anomaly occurs multiplied by its corresponding weight , then divide it by the weighted square sum and calculate the logarithm to get the accurate pipeline operation abnormality index. The ratio of the number of times the pipeline operation abnormality occurs to the total number of times the pipeline operation abnormality occurs is the ratio of the number of times the pipeline operation abnormality occurs to the total number of times the pipeline operation abnormality occurs to the m types is the ratio of the ... The frequency of abnormal operation of the pipeline, The weight of each pipeline operation anomaly is determined based on the impact on pipeline operation safety. The calculation of the square sum of the weights of the denominator avoids the impact of extreme weight values ​​on the final result, ensuring the accuracy of the final result of the pipeline operation anomaly index.

[0108] In one embodiment of the present invention, the comprehensive analysis and calculation of pipeline safety characteristic values ​​to evaluate the safety level of the pipeline includes the following steps:

[0109] By analyzing the formula:

[0110] Get pipeline safety characteristic value ,in, represents the pipeline variation coefficient, represents the speed change pressure value, R represents the pipeline bending influence rate, K represents the pipeline operation abnormality index, , , and They represent the weights of pipeline variation coefficient, variable speed pressure value, pipeline bending influence rate and pipeline operation abnormality index respectively;

[0111] According to the pipeline safety characteristic value , the safety level of pipelines is divided into three levels:

[0112] when < It is the first level of security;

[0113] When 0.3≤ When <0.7, it is the second-level safety level;

[0114] when When ≥0.7, it is the third level safety level.

[0115] Specifically, the pipeline safety characteristic value is calculated by comprehensively analyzing the pipeline variation coefficient, variable speed pressure value, pipeline bending influence rate and pipeline operation abnormality index multiplied by their corresponding weights. , the safety level of pipelines is divided into three levels: primary safety level, secondary safety level and tertiary safety level. Weight of pipeline variation coefficient Assigned as 0.2; weight of the speed change pressure value Assigned as 0.3; weight of pipe bending influence rate Assigned to 0.25; the weight of pipeline operation abnormality index The weight is assigned as 0.25; the determination of weights is a dynamic adjustment process. When it is found that the impact of certain factors in actual operation is greater or less than expected, the weight should be adjusted accordingly. Therefore, regular review and adjustment are required based on real-time feedback and long-term performance evaluation. The safety level of the pipeline is obtained by comparing the pipeline safety characteristic value with the set threshold. When the pipeline safety characteristic value is less than 0.3, it means that the pipeline safety level is the first-level safety level and the pipeline is in a safe situation; when the pipeline safety characteristic value is greater than or equal to 0.3 and less than 0.7, it means that the pipeline safety level is the second-level safety level, the pipeline is relatively safe, and the pipeline is continuously detected in real time; when the pipeline safety characteristic value is greater than or equal to 0.7, it means that the pipeline safety level is the third-level safety level, the pipeline is in an unsafe situation, and the staff should immediately check the pipeline to solve the problem. By analyzing the pipeline safety characteristic value, the safety situation of the pipeline can be understood, ensuring the safe operation of the pipeline.

[0116] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A pipeline safety feature extraction method based on Gram angle field, characterized in that: Includes the following modules: Data collection module: collects time series data during pipeline operation; Data processing module: pre-process the collected time series data; Parameter extraction module: convert the preprocessed time series data into images using the Gram angular field method; extract pipeline parameters based on the converted results; Data analysis module: Calculate pipeline variation coefficient, variable speed pressure value and pipeline bending influence rate based on pipeline parameter analysis; Analyze pipeline operation anomalies based on images converted from Gram angle field; Calculate pipeline operation anomaly index based on analysis results; Safety assessment module: Comprehensively analyze and calculate pipeline safety characteristic values ​​and assess the safety level of pipelines; The pipeline parameters are extracted according to the transformed results, including the following steps: Based on the images transformed by the Gram angular field method from the preprocessed time series data, pipeline parameters are extracted using a deep learning model; The marked pipeline parameters include: pipeline length, pipeline temperature, pipeline diameter, pipeline cross-sectional area, pipeline fluid density, pipeline volume flow rate, pipeline bending angle and pipeline bend length; Input the annotated pipeline parameters into the deep learning model for model training; The pipeline parameters are outputted by transforming the preprocessed time series data into images using the Gram angular field method; The calculation of pipeline variation coefficient, variable speed pressure value and pipeline bending influence rate according to pipeline parameter analysis includes the following steps: By analyzing the formula: ; Get the pipeline variation coefficient ,in Indicates the change in length of the pipeline. represents the temperature change, represents the initial pipe length; By analyzing the formula: ; ； Get the speed change pressure value , represents the friction coefficient, Indicates the change in length of the pipeline. D represents the change in pipe diameter. represents the fluid density, represents the fluid velocity, Q represents the volume flow rate through the pipe, Indicates the cross-sectional area of ​​the pipe; By analyzing the formula: ; The pipe bending influence rate R is obtained, where: Indicates the pipe bending angle, Indicates the pipe bend length.

2. The pipeline safety feature extraction method of the Gram angle field according to claim 1 is characterized in that: Collecting time series data during pipeline operation includes the following steps: Pressure sensors, flow meters and temperature sensors are installed around the pipeline to collect real-time time series data including: pressure data, flow data and temperature data.

3. The pipeline safety feature extraction method of the Gram angle field according to claim 2 is characterized in that: The collected data is preprocessed, including the following steps: By analyzing the formula: ; The collected time series data Normalize to the interval [0, 1] to get the processed time series data ,in represents the data point at the i-th moment, .

4. The pipeline safety feature extraction method of the Gram angle field according to claim 3 is characterized in that: The preprocessed time series data is converted into an image using the Gram angular field method, including the following steps: By analyzing the formula: ; ； get and The angle between ;in, represents the inner product of vectors, Represents vectors and vector The module length, represents the data point at the i-th moment, represents the data point at the jth moment, , d is the dimension of each vector, and They are vectors and vector The component in the kth dimension; By analyzing the formula: ; Get the element values ​​of the Gram angular field matrix , generate an n×n symmetric matrix; By analyzing the formula: ; Get the element value of the Gram angular field matrix to generate the pixel value of the image ;in, express and The angle value between .

5. The pipeline safety feature extraction method of the Gram angle field according to claim 1 is characterized in that: According to the abnormal operation of the image analysis pipeline after Gram's angle field transformation, the following steps are included: Based on the images transformed by the Gram angle field, the pipeline operation anomalies are analyzed through the deep learning model; Mark m types of pipeline operation anomalies; Input the annotated pipeline operation abnormality data into the deep learning model for model training; Through the images transformed by the Gram angular field method after preprocessing the time series data, m types of pipeline operation anomalies are identified and output.

6. The pipeline safety feature extraction method of the Gram angle field according to claim 5 is characterized in that: Calculating the pipeline operation abnormality index based on the analysis results includes the following steps: By analyzing the formula: ; The pipeline operation abnormality index K is obtained, where: Indicates The number of times the pipeline operation abnormality occurs. Indicates The weight of pipeline operation abnormality, Indicates the total number of times m types of pipeline operation abnormalities occur. .

7. The pipeline safety feature extraction method of the Gram angle field according to claim 1 is characterized in that: Comprehensive analysis and calculation of pipeline safety characteristic values ​​to evaluate the pipeline safety level includes the following steps: By analyzing the formula: ; Get pipeline safety characteristic value ,in, represents the pipeline variation coefficient, represents the speed change pressure value, R represents the pipeline bending influence rate, K represents the pipeline operation abnormality index, , , and They represent the weights of pipeline variation coefficient, variable speed pressure value, pipeline bending influence rate and pipeline operation abnormality index respectively; According to the pipeline safety characteristic value , the safety level of pipelines is divided into three levels: when < It is the first level of security; When 0.3≤ When <0.7, it is the second-level safety level; when When ≥0.7, it is the third level safety level.

Images