Natural gas pipeline leakage intelligent monitoring system and method

Abstract

The invention belongs to the technical field of pipelines, and provides an intelligent natural gas pipeline leakage monitoring system and method. The method comprises the following steps: calling monitoring data acquired by multiple types of monitoring sensors corresponding to a potential area and pipeline distribution structure data; acquiring environment data corresponding to the potential area, and obtaining a parameter weight group based on environment data matching; performing prediction processing on the pipeline pressure data set, the pipeline flow data set, the pipeline audio data set and the pipeline temperature data set by using the corresponding convolution model to obtain a leakage region set corresponding to each data set; performing joint probability analysis on each group of leakage areas and leakage probabilities based on the parameter weight group to obtain a plurality of target leakage areas; and outputting the target leakage area. According to the invention, through fusion and intelligent analysis of multi-source data, a leakage area can be accurately positioned; meanwhile, the parameter weights of different sensing data are dynamically adjusted based on the environmental data, the interference influence is reduced, and the prediction accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipelines, and in particular, to an intelligent monitoring system and method for natural gas pipeline leakage. Background Art

[0002] Natural gas is a key component of the energy supply system. With the continuous expansion of the pipe network scale, pipelines face multiple leakage risks such as corrosion and geological disasters. Traditional monitoring methods, such as manual inspections, are inefficient and subjective, making it difficult to achieve real-time early warning; monitoring based on pressure and flow is vulnerable to operating condition fluctuations and has a high false alarm rate; although fiber optic monitoring and acoustic emission monitoring have certain technical advantages, they have problems such as high costs, complex signal processing, and obvious environmental constraints, and are all difficult to meet the accurate monitoring requirements of long-distance and complex pipe networks.

[0003] Therefore, how to improve the accuracy, reliability, and real-time performance of natural gas pipeline leakage monitoring. Summary of the Invention

[0004] In view of this, the present invention provides an intelligent monitoring method, system, electronic device, computer storage medium, and computer program product for natural gas pipeline leakage to solve the above technical problems.

[0005] The present invention discloses an intelligent monitoring method for natural gas pipeline leakage, and the method includes the following steps:

[0006] Determine potential areas with natural gas pipeline leakage risks, and retrieve the monitoring data collected by multiple types of monitoring sensors corresponding to the potential areas and the pipeline distribution structure data; wherein, the monitoring data includes pipeline pressure data sets, pipeline flow data sets, pipeline audio data sets, and pipeline temperature data sets;

[0007] Obtain the environmental data corresponding to the potential area, the environmental data includes sound environmental data and temperature environmental data, and match the parameter weight groups based on the environmental data.

[0008] Use the corresponding convolutional model to perform prediction processing on the pipeline pressure data set, pipeline flow data set, pipeline audio data set, and pipeline temperature data set respectively, and obtain the leakage area sets corresponding to each data set. Each leakage area set includes several groups of leakage areas and leakage probabilities;

[0009] Perform joint probability analysis on each group of the leakage areas and the leakage probabilities based on the parameter weight groups to obtain several target leakage areas, and the leakage probabilities of the target leakage areas are higher than the probability threshold;

[0010] Output the target leakage areas.

[0011] The present invention also discloses an intelligent monitoring system for natural gas pipeline leakage. The system includes multi-type monitoring sensors, at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. The processor executes the computer program to implement the method as described in any of the preceding paragraphs.

[0012] The present invention also discloses an electronic device, including: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. The processor executes the computer program to implement the method as described in any of the preceding items.

[0013] The present invention also discloses a computer storage medium. The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method as described in any of the preceding paragraphs.

[0014] The present invention also discloses a computer program product. When the computer program product runs on a terminal, it enables the terminal to execute to implement the method as described in any of the preceding paragraphs.

[0015] The beneficial effects of the present invention are as follows:

[0016] The solution of the present invention can accurately locate the leakage area by fusing and intelligently analyzing multi-source data. At the same time, it also dynamically adjusts the parameter weights of different sensing data based on environmental data to reduce the influence of interference. It uses a convolutional model to deeply mine data features to improve the prediction accuracy. The joint probability analysis synthesizes the results of multiple data sets to effectively reduce false alarms. It can quickly output the target leakage area to help maintenance personnel dispose of it in time, significantly improving the accuracy, reliability, and real-time performance of monitoring, and reducing energy losses and safety risks. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic flow chart of an intelligent monitoring method for natural gas pipeline leakage disclosed in an embodiment of the present invention;

[0019] Figure 2 It is a schematic structural diagram of an intelligent monitoring system for natural gas pipeline leakage disclosed in an embodiment of the present invention. Detailed Embodiments

[0020] The following specific embodiments illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope protected by the present application.

[0021] In addition, the technical features involved in different implementation manners of the present application described below can be combined with each other as long as they do not conflict with each other.

[0022] As Figure 1 shown, an embodiment of the present invention discloses an intelligent monitoring method for natural gas pipeline leakage. The method includes the following steps:

[0023] S01. Determine the potential areas with natural gas pipeline leakage risks, and retrieve the monitoring data collected by various types of monitoring sensors corresponding to the potential areas and the pipeline distribution structure data. Among them, the monitoring data includes pipeline pressure data sets, pipeline flow data sets, pipeline audio data sets, and pipeline temperature data sets.

[0024] In a natural gas transmission network, not all areas have the same leakage risk. Through multi-dimensional information such as historical leakage data, pipeline service life, geological condition assessment, and third-party construction activity records, the potential areas with leakage risks can be preliminarily delimited using a risk assessment model.

[0025] Once the potential areas are determined, it is necessary to retrieve the data collected by various types of monitoring sensors in these areas. The pipeline pressure data sets are obtained through pressure sensors, which can reflect the real-time changes in the internal pressure of the pipeline. An abnormal sudden drop or fluctuation in pressure represents the existence of a leakage risk. The pipeline flow data sets are collected by flow sensors, and a sudden decrease or abnormal change in flow is closely related to the leakage situation. The pipeline audio data sets are collected by acoustic sensors. When natural gas leaks, sound waves in a specific frequency range will be generated, and these audio data contain the acoustic characteristics of the leakage. The pipeline temperature data sets are recorded by temperature sensors. Natural gas leakage will cause changes in the surrounding environmental temperature, and the temperature data can be used as an auxiliary basis for leakage judgment.

[0026] At the same time, the pipeline distribution structure data, such as pipeline orientation, pipe diameter changes, valve positions, etc., can help to more accurately locate the leakage points. These data are pre-drawn and stored in, for example, a server, and are retrieved.

[0027] S02. Obtain the environmental data corresponding to the potential areas. The environmental data includes sound environmental data and temperature environmental data, and match the parameter weight groups based on the environmental data.

[0028] Environmental factors have an important impact on the accuracy and reliability of monitoring data. Therefore, it is necessary to obtain the sound environment data and temperature environment data of potential areas. Different environmental data will have different degrees of influence on the monitoring data. For example, in a high-noise environment, the credibility of pipeline audio data will decrease, and at this time, the weight of the audio data in the analysis should be reduced; while in an environment with drastic temperature changes, the pipeline temperature data may be interfered, and the corresponding weight also needs to be adjusted. Through the pre-established environment-weight matching database, the real-time obtained environmental data is compared with the templates in the database, so as to match and obtain the parameter weight set applicable to the current environmental conditions, ensuring that in the subsequent analysis process, each monitoring data can be reasonably weighted to improve the accuracy of the analysis results.

[0029] Among them, the sound environment data can be collected by environmental noise monitoring equipment (which should be an equipment array) to understand the daily noise level in this area and whether there are interference factors such as construction noise; the temperature environment data is obtained by environmental temperature sensors (which should be a sensor array) to master the real-time changes of the environmental temperature.

[0030] S03, use the corresponding convolutional model to perform prediction processing on the pipeline pressure dataset, pipeline flow dataset, pipeline audio dataset, and pipeline temperature dataset respectively to obtain the leakage area set corresponding to each dataset. The leakage area set includes several groups of leakage areas and leakage probabilities.

[0031] Convolutional neural networks (CNNs) perform well in processing data with spatial or temporal features and can automatically extract key features from the data. Special convolutional models are trained for different types of monitoring data.

[0032] For the pipeline pressure dataset, each pressure data in the pipeline pressure dataset is sorted into a matrix form according to the time series and input into the convolutional model. The model analyzes and predicts the current pressure data by learning the feature differences between the normal state and the leakage state in the historical pressure data, and outputs a series of possible leakage areas and the leakage probability corresponding to each area; the pipeline flow dataset is processed in a similar way. The convolutional model mines the change patterns of the flow data and identifies the potential leakage areas and probabilities corresponding to abnormal flows; before the pipeline audio dataset is input into the model, it is preprocessed, such as noise reduction, feature extraction, etc. The processed audio feature data is input into the convolutional model, and the convolutional model judges the potential leakage areas and probabilities included in the audio data according to the leakage sound wave features learned during training; the pipeline temperature dataset is constructed into a format suitable for the convolutional model to process, and the convolutional model predicts the possible leakage areas and leakage probabilities according to the temperature change law. It can be understood that each data in the above datasets actually has corresponding spatial position information, so that the location of the leakage area in the pipeline structure can be determined.

[0033] Finally, after each dataset is processed by the corresponding convolutional model, a leakage area set containing several groups of leakage areas and leakage probabilities can be obtained.

[0034] S04, perform joint probability analysis on each group of the leakage areas and the leakage probabilities based on the parameter weight group, and obtain several target leakage areas, where the leakage probability of the target leakage areas is higher than the probability threshold.

[0035] Based on different types of monitoring data, leakage area sets corresponding to each other are obtained respectively by using convolutional models. Each leakage area contains a leakage probability based on a single data type. However, due to the different reliabilities of different monitoring data under different environmental conditions, when performing joint probability analysis, it is necessary to introduce the parameter weight group obtained in step S02 to reflect the credibility differences of each data.

[0036] Each weight in the parameter weight group corresponds to a type of monitoring data. For example, in a high-noise environment, the weight of pipeline audio data may be low, while the weight of pipeline pressure data is relatively high. Based on this, when calculating the joint probability, for the same potential leakage area, the leakage probabilities obtained based on different datasets are weighted according to the corresponding weights. Suppose the leakage probability of area B obtained based on the pipeline pressure dataset is P1, and its corresponding weight is W1; the leakage probability of area B obtained based on the pipeline flow dataset is P2, and the corresponding weight is W2; the leakage probability of area B obtained based on the pipeline audio dataset is P3, and the corresponding weight is W3; the leakage probability of area B obtained based on the pipeline temperature dataset is P4, and the corresponding weight is W4. The final joint leakage probability P of area B is calculated through the weighted joint probability calculation formula.

[0037] Set a reasonable probability threshold. When the joint leakage probability P of a certain area is higher than this probability threshold, it is determined as a target leakage area.

[0038] By introducing the parameter weight group for joint probability analysis, not only is the multi-source data information fully integrated, but also the influence of environmental factors on data reliability is combined, significantly improving the accuracy of leakage area judgment, effectively reducing the false alarm risk, and providing a more reliable basis for subsequent pipeline maintenance work.

[0039] S05, output the target leakage areas.

[0040] After the target leakage areas are determined, the relevant information of these areas, such as specific location coordinates, leakage probability values, etc., is output through system interfaces, text messages, emails, etc., and conveyed to pipeline maintenance personnel in a timely manner. The maintenance personnel can quickly go to the target leakage areas for further inspection and repair based on this information, so as to achieve a rapid response to natural gas pipeline leakage, reduce the energy loss, environmental pollution and safety accident risks that may be caused by leakage.

[0041] The solution of the present invention can accurately locate the leakage area by fusing and intelligently analyzing multi-source data; at the same time, it can also dynamically adjust the parameter weights of different sensing data based on environmental data to reduce the interference effect; use the convolutional model to deeply mine data features to improve the prediction accuracy; and comprehensively analyze the results of multiple data sets through joint probability analysis to effectively reduce false alarms. It can quickly output the target leakage area, assist maintenance personnel in timely handling, significantly improve the accuracy, reliability and real-time performance of monitoring, and reduce energy losses and safety risks.

[0042] Optionally, to determine the potential area with a risk of natural gas pipeline leakage, the monitoring data collected by multiple types of monitoring sensors corresponding to the potential area and the pipeline distribution structure data are retrieved, including:

[0043] The preliminary leakage level corresponding to the potential area is parsed from the monitoring request signal, and the monitoring time span is determined based on the preliminary leakage level; among them, the higher the preliminary leakage level, the more obvious the natural gas leakage in the potential area, and vice versa, the less obvious the natural gas leakage in the potential area.

[0044] The monitoring data collected by multiple types of monitoring sensors corresponding to the potential area and the pipeline distribution structure data are retrieved according to the monitoring time span; among them, the monitoring time span is negatively correlated with the preliminary leakage level.

[0045] In this embodiment, the preliminary leakage level corresponding to the potential area can be parsed from the monitoring request signal. The monitoring request signal may be triggered manually or automatically generated by the system based on preliminary anomaly detection. The preliminary leakage level is a preliminary assessment of the degree of natural gas leakage in the potential area, and its basis may include simple indicators such as the abnormal fluctuation range of pipeline pressure and the sudden change of flow rate (for example, virtual inspection of each pipeline area based on the aforementioned risk assessment model). For example, when the pressure drops sharply within a short period of time exceeding a certain threshold, combined with the abnormal change of flow rate data, it can be preliminarily determined that the preliminary leakage level of the potential area is relatively high, indicating that the natural gas leakage in the potential area is relatively obvious; on the contrary, if there are only slight fluctuations in pressure and flow rate, the preliminary leakage level is relatively low, indicating that the natural gas leakage in the potential area is not obvious.

[0046] Based on the preliminary judgment of the leakage level, determine the monitoring time span. The present invention sets a negative correlation between the preliminary judgment of the leakage level and the monitoring time span, that is, the higher the preliminary judgment of the leakage level, the shorter the monitoring time span; the lower the preliminary judgment of the leakage level, the longer the monitoring time span. The reason for such a setting is that when the leakage is obvious, in order to quickly determine the leakage area and obtain the latest data in a timely manner to support the analysis and decision-making, only the monitoring data within a short period of time needs to be retrieved, such as the pressure, flow rate, audio, and temperature data within the past 1 hour; while when the leakage is not obvious, it may be necessary to capture subtle changes through data with a longer time span, for example, retrieve data within the past 24 hours or even longer, so as to more comprehensively analyze the potential leakage situation.

[0047] Finally, according to the determined monitoring time span, retrieve the monitoring data collected by various types of monitoring sensors corresponding to this potential area and the pipeline distribution structure data.

[0048] Through this method of dynamically adjusting the monitoring time span according to the preliminary judgment of the leakage level to retrieve data in this embodiment, it can not only improve the pertinence and efficiency of data retrieval, but also accurately obtain the required data under different leakage conditions, providing a more effective data basis for subsequent steps such as analyzing using a convolutional model and calculating joint probabilities, thereby improving the accuracy and timeliness of the entire natural gas pipeline leakage intelligent monitoring method.

[0049] Optionally, the matching of the parameter weight group based on the environmental data includes:

[0050] Perform frequency domain analysis on the sound environment data, obtain the noise frequency distribution characteristics, calculate the similarity with each pre-stored typical noise frequency pattern, and obtain the matching target typical noise frequency pattern;

[0051] Compare the temperature environment data with different temperature change scenarios recorded in the database to determine the current target temperature change scenario;

[0052] Based on the target typical noise frequency pattern and the target temperature change scenario, search in the environment-weight matching database to obtain the corresponding parameter weight group; among them, the weights of the pipeline pressure parameter and the pipeline flow parameter included in the parameter weight group are fixed values, and the weights of the pipeline audio parameter and the pipeline temperature parameter are non-fixed values.

[0053] In this embodiment, corresponding parameter weight groups are pre-constructed for different typical noise frequency patterns and temperature change scenarios to construct an environment-weight matching database. Subsequently, based on the collected sound environment data and temperature environment data, the matching parameter weight group can be determined from this environment-weight matching database.

[0054] First, by performing frequency-domain analysis on the sound environment data, the noise frequency distribution characteristics can be obtained. Among them, frequency-domain analysis uses, for example, the Fast Fourier Transform (FFT) to convert the sound environment data in the time domain from the time dimension to the frequency dimension.

[0055] For example, an array of environmental noise monitoring devices collects a 10-second segment of sound environment data, which appears as a voltage signal waveform that varies with time in the time domain. Applying the FFT algorithm to this data converts it into frequency-domain data, obtaining the amplitude distribution of each frequency component. For instance, after analysis, it is found that the amplitude is relatively high in the frequency range of 50Hz - 100Hz, and there is also a certain amplitude in the frequency range of 200Hz - 300Hz. The distribution characteristics of these amplitudes in frequency constitute the noise frequency distribution characteristics of this sound environment data.

[0056] The pre-stored typical noise frequency patterns are summarized from a large amount of historical data. For example, the construction noise pattern has obvious peaks in the range of 100Hz - 500Hz, and the vehicle noise pattern has a relatively high amplitude in the range of 50Hz - 200Hz, etc. Calculate the similarity between the obtained noise frequency distribution characteristics and these typical patterns, for example, using algorithms such as Euclidean distance and cosine similarity. If the calculation result shows the highest similarity with the construction noise pattern, then the construction noise pattern is the matching target typical noise frequency pattern.

[0057] At the same time, the database stores various temperature change scenarios, such as a steady change scenario (the temperature rises or falls slowly within a certain period of time with a small change amplitude), a drastic fluctuation scenario (the temperature rises and falls significantly within a short period of time), a seasonal change scenario (regular temperature changes that conform to the local season characteristics), etc. For example, an array of temperature sensors collects the temperature data of a certain potential area within 1 hour and finds that the temperature rapidly rises from 20°C to 30°C within 1 hour, with a fluctuation amplitude of 10°C. Comparing this data with the scenarios in the database can determine that the change type of the current target temperature environment is a drastic fluctuation scenario.

[0058] Next, in the environment-weight matching database, each combination of a specific target typical noise frequency pattern and a target temperature change scenario corresponds to a set of parameter weight groups. For example, when the target typical noise frequency pattern is the construction noise pattern and the temperature environment change type is the drastic fluctuation scenario, the corresponding parameter weight group is searched in the database. Among them, since construction noise will seriously interfere with the pipeline audio data, the weight of the pipeline audio parameter will be lowered, for example, from 0.3 in the ideal environment to 0.1; the drastic temperature fluctuation will affect the accuracy of the pipeline temperature data, and the weight of the pipeline temperature parameter will also be appropriately reduced, from 0.2 to 0.1; while the weights of the pipeline pressure parameter and the pipeline flow parameter are fixed values (that is, the weights of the two are not affected by the sound and temperature in the environment), for example, 0.4 and 0.4 respectively. In this way, a parameter weight group [0.4, 0.4, 0.1, 0.1] applicable to the current environmental conditions is obtained for subsequent leakage analysis.

[0059] It can be understood that the frequency distribution characteristics under various noise scenarios where a natural gas leakage event occurs are extracted in advance and sorted according to typical noise types (such as construction noise, traffic noise, industrial noise, natural environmental noise, etc.); and, the change trend of the environmental temperature time series is also analyzed correspondingly to determine the temperature change scenario type, such as stable change, drastic fluctuation, seasonal change, etc. Then, for each combination of typical noise frequency pattern and temperature change scenario, the reliability and importance of different monitoring data in leakage detection are analyzed. For example, in a construction noise environment, the pipeline audio data is seriously interfered, and its weight should be greatly reduced; while in a drastic temperature fluctuation scenario, the credibility of the pipeline temperature data decreases, and the weight needs to be adjusted accordingly. Through multiple rounds of data verification, a reasonable parameter weight group can be formulated for each environmental combination, that is, the specific weight values of the pipeline pressure parameter, the pipeline flow parameter, the pipeline audio parameter, and the pipeline temperature parameter are clarified, and the above-mentioned environment-weight matching database is constructed.

[0060] The environment-weight matching database can use a relational database (such as MySQL) or a non-relational database (such as MongoDB) for data storage to ensure efficient query and management of data. At the same time, the database is regularly updated and maintained, incorporating newly collected environmental data and actual monitoring feedback results, and optimizing the weight parameters so that the database can continuously adapt to new environmental changes and monitoring requirements.

[0061] Optionally, the joint probability analysis is performed on each group of the leakage regions and the leakage probabilities based on the parameter weight group to obtain several target leakage regions, including:

[0062] Align the spatial positions of each of the leakage regions in the pipeline distribution structure, and normalize the leakage probability of each data set;

[0063] For the leakage probability of each of the leakage areas in the same pipeline area, a preliminary joint leakage probability is calculated according to the weighted joint probability.

[0064] Calculate the variance of the area sizes of each of the leakage areas in the same pipeline area, match an additional coefficient according to the variance, calculate the product of the preliminary joint leakage probability and the additional coefficient to obtain the target joint leakage probability, and determine as the target leakage areas those leakage areas for which the target joint leakage probability is higher than the probability threshold.

[0065] In this embodiment, in the natural gas pipeline monitoring scenario, the pipeline distribution structure can be abstracted into a spatial model with a specific coordinate system, and the spatial positions in the pipeline network of each leakage area are determined by the starting point and ending point coordinates on the pipeline. Taking a straight pipeline with a length of 100 meters as an example, assume that the leakage area obtained based on the pipeline pressure dataset has a starting point at 20 meters and an ending point at 30 meters, the leakage area obtained based on the pipeline flow dataset has a starting point at 22 meters and an ending point at 32 meters, the leakage area obtained based on the pipeline audio dataset has a starting point at 23 meters and an ending point at 28 meters, and the leakage area obtained based on the pipeline temperature dataset has a starting point at 24 meters and an ending point at 27 meters. By mapping the coordinates of these areas into a unified pipeline space coordinate system, spatial position alignment is achieved.

[0066] Considering the measurement errors and data fluctuations of different datasets, first perform normalization processing on the leakage probability of each dataset (for example, using the min-max normalization method) to eliminate the influence of the dimension and the data fluctuation range, making the calculation results more accurate and reliable.

[0067] For example, the leakage probability range of the pipeline pressure dataset is [0.2, 0.8], the leakage probability range of the pipeline flow dataset is [0.1, 0.7], the leakage probability range of the pipeline audio dataset is [0.3, 0.6], and the leakage probability range of the pipeline temperature dataset is [0.25, 0.55]. Taking the pipeline pressure dataset as an example, for the leakage probability P1 of a certain leakage area therein, the normalization formula is Through this formula, the leakage probabilities of all datasets are mapped to the interval [0, 1] to eliminate the influence of the dimension and the data fluctuation range.

[0068] Each leakage area in the same pipeline area is predicted based on different datasets, that is, based on different datasets, it is characterized that the pipeline area may have a leakage risk. For the leakage probability of each leakage area corresponding to the pipeline area, a weighted joint probability calculation formula is used for joint probability analysis. The weighted joint probability calculation formula is as follows:

[0069]

[0070] In the formula, is the normalized leakage probability of the i-th leakage area in the same pipeline area, i=1, 2, 3, 4, corresponding to the pipeline pressure dataset, pipeline flow dataset, pipeline audio dataset and pipeline temperature dataset respectively; W i is the corresponding weight in the parameter weight group.

[0071] In addition, in natural gas pipeline leakage monitoring, the size range of leakage areas in the same pipeline area based on different monitoring data sets often differs, which is due to the different sensitivity and expression forms of each data set in response to pipeline leakage. For example, pipeline pressure data may show abnormal fluctuations in a specific range when detecting leaks due to factors such as the accuracy of the pressure sensor and the operating conditions of the pipeline; and pipeline flow data is affected by the inertia of the fluid in the pipeline, the characteristics of the flow metering equipment, etc., and its abnormal range deviates from the pressure data. If the joint analysis is based only on the probability of leakage and the differences in the size of the leakage area are ignored, the analysis results may not match the actual leakage situation.

[0072] Therefore, in addition to the different leakage probabilities of the same pipeline area obtained based on different data sets, the present invention also makes appropriate corrections to the preliminary joint leakage probability based on the size range of each leakage area. Specifically:

[0073] Calculate the regional size variance of each leakage area in the same pipeline area. This variance can quantify the degree of difference in the size range of the leakage area. The larger the variance, the more significant the difference in the leakage area size determined by different data sets, which means that the judgment consistency of the leakage area of ​​each data set is low, and there are large differences in the reliability of the data; the smaller the variance, the closer the leakage area size obtained by each data set is, the higher the data consistency, and the stronger the reliability. Through the pre-established mapping relationship, the additional coefficient is obtained according to the variance matching, and the additional coefficient is used to reasonably adjust the preliminary joint leakage probability, so that the final target joint leakage probability is more in line with the actual leakage situation, making up for the shortcomings of relying solely on leakage probability analysis, and improving the accuracy of leakage area judgment. The schematic mapping relationship is shown in the following table:

[0074]

[0075] It can be determined from the above table that the additional coefficient is negatively correlated with the variance. That is, the lower the degree of consistency of the sizes of the four above-mentioned intervals (i.e., the larger the variance), the smaller the additional coefficient is set to appropriately lower the preliminary combined leakage probability corresponding to the pipeline area; conversely, the larger the additional coefficient is set to appropriately increase the preliminary combined leakage probability corresponding to the pipeline area. It can be understood that the additional coefficient of the present invention is only used for fine-tuning, and the additional coefficient should be a very small value, for example, in the interval [0.95, 1.05].

[0076] Optionally, after obtaining the additional coefficient according to the variance matching, the method further includes:

[0077] Retrieve multiple groups of historical leakage detection records of the same pipeline area. The historical leakage detection records include four historical leakage areas and their area sizes from different data sets, and each area size is sufficient for the area size group;

[0078] Calculate the area size diversity index based on each group of the area size groups, and calculate the target additional coefficient by using the area size diversity index and the additional coefficient. The target additional coefficient is used to multiply with the preliminary combined leakage probability to obtain the target combined leakage probability.

[0079] In this embodiment, the additional coefficient obtained in the foregoing manner is actually calculated only based on the variance of a current group of four area sizes. However, the data set collected once is easily distorted by interference factors. Therefore, the present invention further considers the historical leakage detection data of the same pipeline area to further fine-tune the rationality of the additional coefficient to improve its confidence level. Specifically:

[0080] During the operation of the natural gas pipeline, the monitoring system continuously records the relevant data of each leakage detection and stores them in the database. When it is necessary to optimize the additional coefficient, according to the identifier of the same pipeline area detected currently, retrieve multiple groups of historical leakage detection records of the same pipeline area from the database. For example, for the pipeline area numbered GP-001, retrieve the historical data of the past 10 leakage detections.

[0081] Each group of historical leakage detection records contains four historical leakage area information (for the same leakage event) from the pipeline pressure data set, the pipeline flow data set, the pipeline audio data set, and the pipeline temperature data set. Each leakage area has a corresponding area size (which can be determined by calculating the length, area, etc. of the area. For example, for a linear pipeline, the distance between the two end points of the area is used as the area size). These area sizes together form an area size group. For example, in a certain detection record, the sizes of the four historical leakage areas are 5 meters, 6 meters, 4 meters, and 5 meters respectively, forming the area size group [5, 6, 4, 5].

[0082] The present invention further calculates a diversity index exhibited between these groups of region sizes, and the diversity index can be characterized, for example, by a coefficient of variation.

[0083]

[0084] Wherein, CV i is the coefficient of variation of the region sizes corresponding to the group of region sizes of the i-th group, and σ i is the standard deviation of the region sizes of the group of region sizes of the i-th group, and μ i is the average value of the region sizes of the group of region sizes of the i-th group.

[0085] Then, calculate the average value CV avg of these coefficients of variation, or calculate the median of all the coefficients of variation, etc., to obtain the region size diversity index for all groups. This region size diversity index can comprehensively reflect the overall difference characteristics of the leakage region sizes of different data sets in multiple groups of historical data. Compared with single-group calculation, it can more comprehensively reflect the distribution law and diversity of the data.

[0086] In order to make the additional coefficient more accurately reflect the actual leakage situation, the existing additional coefficient is further optimized in combination with the region size diversity index for all groups obtained above. The calculation method of the target additional coefficient can be:

[0087] K target = K×(1 + α×CV avg )

[0088] Wherein, K is the additional coefficient matched according to the variance of the current detection data; α is a weight adjustment factor, for example, 0.2; CV avg is the average value of the four coefficients of variation.

[0089] For example, for the additional coefficient matched according to the variance of the current detection data, the calculated region size diversity index CV avg = 0.12, and the weight adjustment factor α = 0.2, then the target additional coefficient K target = 0.95×(1 + 0.2×0.12) = 0.9728.

[0090] Finally, multiply the calculated target additional coefficient by the preliminary joint leakage probability to obtain the target joint leakage probability, which is used to finally determine whether the pipeline region is the target leakage region. By introducing the region size diversity index obtained by integrating multiple groups of historical data to optimize the additional coefficient, the historical data characteristics can be more comprehensively considered, making the joint probability calculation result more in line with the actual leakage situation, and further improving the accuracy and reliability of natural gas pipeline leakage monitoring.

[0091] Such asFigure 2 As shown in Figure 2 , an embodiment of the present invention also discloses an intelligent monitoring system for natural gas pipeline leakage. The system includes multi-type monitoring sensors, at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. The processor executes the computer program to implement the method as described in the foregoing embodiment.

[0092] Optionally, the multi-type monitoring sensors are pressure sensors, flow sensors, acoustic sensors, and temperature sensors.

[0093] An embodiment of the present invention also discloses an electronic device, including: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. The processor executes the computer program to implement the method as described in the foregoing embodiment.

[0094] An embodiment of the present invention also discloses a computer storage medium. The computer storage medium stores a computer program, and the computer program is executed by a processor to implement the method as described in the foregoing embodiment.

[0095] An embodiment of the present invention also discloses a computer program product. When the computer program product runs on a terminal, it causes the terminal to execute to implement the method as described in the foregoing embodiment.

[0096] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0097] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, thereby the instructions executed on the computer or other programmable apparatus provide steps for realizing the functions specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 steps for realizing the functions specified in one block or a plurality of blocks.

[0099] As mentioned above, it is only a preferred embodiment of the present invention, and is not used to limit the protection scope of the present invention.

Claims

1. An intelligent monitoring method for natural gas pipeline leakage, characterized in that , including the following steps: Determine the potential areas with risks of natural gas pipeline leakage, and retrieve the monitoring data collected by various types of monitoring sensors corresponding to the potential areas and the pipeline distribution structure data; among them, the monitoring data includes pipeline pressure data sets, pipeline flow data sets, pipeline audio data sets, and pipeline temperature data sets; Obtain the environmental data corresponding to the potential areas, where the environmental data includes sound environmental data and temperature environmental data, and match the parameter weight groups based on the environmental data; Use the corresponding convolution model to perform prediction processing on the pipeline pressure data sets, pipeline flow data sets, pipeline audio data sets, and pipeline temperature data sets respectively, and obtain the leakage area sets corresponding to each data set, where each leakage area set includes several groups of leakage areas and leakage probabilities; Perform joint probability analysis on each group of the leakage areas and the leakage probabilities based on the parameter weight groups to obtain several target leakage areas, and the leakage probabilities of the target leakage areas are higher than the probability threshold; Output the target leakage areas.

2. The intelligent monitoring method for natural gas pipeline leakage according to claim 1, wherein: The determination of the potential areas with risks of natural gas pipeline leakage and the retrieval of the monitoring data collected by various types of monitoring sensors corresponding to the potential areas and the pipeline distribution structure data include: Parse the preliminary leakage level corresponding to the potential areas from the monitoring request signal, and determine the monitoring time span based on the preliminary leakage level; among them, the higher the preliminary leakage level, the more obvious the natural gas leakage in the potential area, and vice versa; Retrieve the monitoring data collected by various types of monitoring sensors corresponding to the potential areas and the pipeline distribution structure data according to the monitoring time span; among them, the monitoring time span is negatively correlated with the preliminary leakage level.

3. The intelligent monitoring method for natural gas pipeline leakage according to claim 1, characterized in that: The matching of the parameter weight groups based on the environmental data includes: Perform frequency domain analysis on the sound environmental data, obtain the noise frequency distribution characteristics, calculate the similarity with each pre-stored typical noise frequency pattern, and obtain the matching target typical noise frequency pattern; Compare the temperature environmental data with different temperature change scenarios recorded in the database to determine the current target temperature change scenario; Based on the target typical noise frequency pattern and the target temperature change scenario, search in the environment-weight matching database to obtain the corresponding parameter weight groups; among them, the weights of the pipeline pressure parameters and pipeline flow parameters included in the parameter weight groups are fixed values, and the weights of the pipeline audio parameters and pipeline temperature parameters are non-fixed values.

4. The intelligent monitoring method for natural gas pipeline leakage according to claim 3, characterized in that: The joint probability analysis of each group of the leakage areas and the leakage probabilities based on the parameter weight groups to obtain several target leakage areas includes: Align the spatial positions of each leakage area in the pipeline distribution structure, and normalize the leakage probabilities of each data set; For the leakage probabilities of each leakage area in the same pipeline area, calculate the preliminary joint leakage probability according to the weighted joint probability; Calculate the variance of the area sizes of each of the leakage areas in the same pipeline area, match an additional coefficient according to the variance, calculate the product of the preliminary combined leakage probability and the additional coefficient to obtain the target combined leakage probability, and determine the leakage areas with the target combined leakage probability higher than the probability threshold as the target leakage areas.

5. An intelligent monitoring method for natural gas pipeline leakage according to claim 4, characterized in that: After matching the additional coefficient according to the variance, the method further includes: Retrieve multiple groups of historical leakage detection records of the same pipeline area. The historical leakage detection records include four historical leakage areas from different data sets and their area sizes, and each area size forms an area size group; Calculate an area size diversity index based on each group of the area size groups, and calculate a target additional coefficient by using the area size diversity index and the additional coefficient. The target additional coefficient is used to multiply the preliminary combined leakage probability to obtain the target combined leakage probability.

6. An intelligent monitoring system for natural gas pipeline leakage, the system comprising multi-type monitoring sensors, at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, characterized in that: The processor executes the computer program to implement the method according to any one of claims 1-5.

7. An intelligent monitoring system for natural gas pipeline leakage according to claim 6, characterized in that: The multi-type monitoring sensors are pressure sensors, flow sensors, acoustic sensors, and temperature sensors.

8. An electronic device, comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein: the processor executes the computer program to implement the method according to any one of claims 1-5.

9. A computer storage medium storing a computer program, characterized in that: The computer program is executed by the processor to implement the method according to any one of claims 1-5.

10. A computer program product, comprising a computer program stored on a non-transitory computer-readable medium, characterized in that: When the computer program is executed by the processor, it implements the method according to any one of claims 1-5.

Images