A quick early warning method for chemical product leakage

Abstract

The present application relates to the technical field of data processing, more particularly, the present application relates to a kind of chemical product leakage fast early warning method, the method comprises, a plurality of pressure data in the chemical product conveying pipeline is collected, obtain each pressure data window, then obtain the data fluctuation degree in each pressure data window, the data fluctuation degree in each pressure data window is adjusted, obtain the data potential abnormality degree in each pressure data window, according to the data potential abnormality degree, obtain the adjustment similar tolerance degree between each pressure data subwindow, according to the adjustment similar tolerance degree between each pressure data subwindow, obtain the fuzzy entropy of each pressure data, then the abnormal condition of the data in each pressure data window is judged, to determine whether chemical product conveying pipeline appears leakage, the present application improves the accuracy of abnormal detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing. More particularly, the present application relates to a quick early warning method for chemical product leakage. BACKGROUND

[0002] With the rapid development of the chemical industry, the production, transportation, storage and use of chemical products have become a key link in modern industry. During the transportation of chemical products, especially through pipeline transportation, there are safety hazards such as leakage and rupture. Leakage accidents can cause serious consequences such as environmental pollution, personnel casualties and property losses. Traditional leakage detection methods often have problems such as response lag, high missed detection rate, etc., and cannot effectively avoid or reduce the occurrence of leakage accidents. In the present scheme, the pressure in the pipeline is monitored in real time, and once an anomaly is found, the system can immediately issue a warning and start an automatic alarm program to notify the operator to take timely measures, thereby improving the response speed of the leakage warning, significantly shortening the reaction time after the accident, and reducing the loss.

[0003] The patent application file with the publication number CN115099624A discloses a multi-attribute decision-making system based on intuitionistic fuzzy entropy and interval fuzzy entropy, which includes a decision information acquisition module, a fuzzy entropy calculation module, and a decision result generation module. The decision information acquisition module is used to acquire a plurality of decision information matrices of a target and transmit them to the fuzzy entropy calculation module. The fuzzy entropy calculation module processes the decision information matrices to obtain the hesitation degree weight coefficients of the decision makers and the benefit attribute hesitation degree weight coefficients, and transmits them to the decision result generation module. The decision result generation module calculates the comprehensive evaluation results of each decision information matrix according to the hesitation degree weight coefficients of the decision makers and the benefit attribute hesitation degree weight coefficients, and obtains the optimal decision according to the comprehensive evaluation results of the decision information matrices.

[0004] When the fuzzy entropy method is used for abnormal detection of pressure data, the fuzzy entropy value of each pressure data needs to be calculated, and the abnormal data is detected according to the fuzzy entropy value of each pressure data. The calculation of the fuzzy entropy value of the pressure data is first to obtain the window of the pressure data, to obtain a plurality of sub-windows of the pressure data according to the window of the pressure data, and finally to obtain the mean value of the fuzzy membership degrees between all the sub-windows of the pressure data. However, in the actual scene, the change of the pressure data caused by the gradual increase of the valve in the pipeline is usually similar to the change of the pressure data caused by the pipeline leakage, i.e. both will cause the decrease of the pressure data in the pipeline. Therefore, the existing fuzzy entropy method cannot well distinguish the above-mentioned situations, resulting in false detection. SUMMARY

[0005] In order to solve the problem that the setting of the similar tolerance degree between the pressure data sub-windows when obtaining the fuzzy membership degree between all the sub-windows of the pressure data by the fuzzy entropy algorithm is uniform, resulting in inaccurate calculation of the fuzzy entropy, the present application proposes a chemical product leakage rapid early warning method, which comprises the following steps:

[0006] Collecting a plurality of pressure data to obtain each pressure data window, obtaining the data fluctuation degree in each pressure data window, and obtaining the data potential abnormality degree in each pressure data window:

[0007] ; represents the data potential abnormality degree in the nth pressure data window; represents the data fluctuation degree in the nth pressure data window; represents the smoothness characteristic factor between the maximum value and the minimum value in the nth pressure data window; norm() represents a normalization function; represents the maximum value of the data in the nth pressure data window; represents the numerical value of the nth pressure data;

[0008] obtaining the adjusted similar tolerance degree between each pressure data sub-window , represents the adjusted similar tolerance degree between the nth pressure data sub-window; exp() represents an exponential function with a natural constant as the base; Y represents the preset similar tolerance degree between the pressure data sub-windows;

[0009] According to the adjusted similar tolerance degree, the fuzzy entropy of each pressure data is obtained, and the abnormality of each pressure data window is judged according to the fuzzy entropy.

[0010] The innovation of the present application lies in analyzing the pressure data change characteristics in the pipeline, introducing the data potential abnormality degree in each pressure data window to adjust the similar tolerance degree between the pressure data sub-windows in the fuzzy entropy method, obtaining the adjusted similar tolerance degree between each pressure data sub-window to obtain the fuzzy entropy value of each pressure data, and this adjustment can more accurately reflect the real characteristics of the pressure change, effectively distinguish the pressure change caused by valve adjustment from the abnormal conditions such as pipeline leakage, thereby reducing the occurrence of false detection and significantly improving the accuracy of abnormal detection.

[0011] Preferably, the obtaining of each pressure data window comprises:

[0012] Predefining the number M of sampling time points, and taking each pressure data and the pressure data at the M sampling time points before the sampling time point of each pressure data as each pressure data window.

[0013] Convenient for subsequent analysis of the data fluctuation degree in each pressure data window.

[0014] Preferably, the acquiring the data fluctuation degree in each pressure data window comprises:

[0015] Acquiring a plurality of neighborhood windows of each pressure data window;

[0016] ;

[0017] In the formula, represents the data fluctuation degree in the nth pressure data window; represents the mean value of all data in the nth pressure data window; represents the median of all data in the nth pressure data window; norm() represents a normalization function; represents the minimum value of data in the nth pressure data window; represents the mean value of the minimum values of data in all neighborhood windows of the nth pressure data window.

[0018] It is convenient for subsequent acquisition of the potential abnormality degree of data in each pressure data window according to the data fluctuation degree in each pressure data window.

[0019] Preferably, the acquiring a plurality of neighborhood windows of each pressure data window comprises:

[0020] A neighborhood window number a is preset, for any one pressure data window, the pressure data at M sampling time points before the sampling time point of the first pressure data in the pressure data window is taken as the first neighborhood window of the pressure data window; the pressure data at M sampling time points before the sampling time point of the first pressure data in the first neighborhood window is taken as the second neighborhood window of the pressure data window; and so on until the ath neighborhood window of the pressure data window is acquired.

[0021] Preferably, the acquiring the flat feature factor between the maximum value and the minimum value of data in the pressure data window comprises:

[0022] ;

[0023] In the formula, represents the flat feature factor between the maximum value and the minimum value of data in the nth pressure data window; represents the sampling time interval between the maximum value and the minimum value of data in the nth pressure data window; represents the variance of the absolute value of the difference between adjacent data between the maximum value and the minimum value of data in the nth pressure data window.

[0024] Preferably, the acquiring the fuzzy entropy of each pressure data according to the adjusted similarity tolerance comprises:

[0025] The length of each pressure data sub-window is preset as B, the fuzzy entropy algorithm is used to obtain each sub-window of each pressure data according to each pressure data window and the length of each pressure data sub-window, and the fuzzy entropy of each pressure data is obtained according to each sub-window of each pressure data and the sign similarity tolerance degree between each pressure data sub-window.

[0026] The obtained fuzzy entropy of each pressure data is more accurate.

[0027] Preferably, the abnormal situation of each pressure data window is judged according to the fuzzy entropy, including:

[0028] The linear normalization method is used to normalize the fuzzy entropy of all pressure data to obtain the normalized fuzzy entropy of each pressure data, and if the normalized fuzzy entropy of any pressure data is greater than a fuzzy entropy threshold T2, it indicates that abnormal data appears in the data in the pressure data window, and the chemical product conveying pipeline leaks in the time period corresponding to the data in the pressure data window.

[0029] The abnormal detection result is improved.

[0030] Preferably, the collection of a plurality of pressure data includes:

[0031] Every ten seconds is a sampling time, and a total of three hours are collected, and each time, the pressure sensor is used to collect each pressure data in the chemical product conveying pipeline.

[0032] The present application has the following beneficial effects: the purpose of the present application is to analyze the pressure data change characteristics in the pipeline, and introduce the potential abnormal degree of the data in each pressure data window to adjust the similarity tolerance degree between the pressure data sub-windows in the fuzzy entropy method, obtain the adjusted similarity tolerance degree between the pressure data sub-windows to obtain the fuzzy entropy value of each pressure data, and this adjustment can more accurately reflect the real characteristics of pressure change, effectively distinguish the pressure change caused by valve adjustment from the abnormal situation such as pipeline leakage, thereby reducing the occurrence of false detection, and significantly improving the accuracy of abnormal detection. BRIEF DESCRIPTION OF DRAWINGS

[0033] The above and other objects, features and advantages of the exemplary embodiments of the present application will be more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which several embodiments of the present application are shown by way of example, and wherein like reference numerals refer to like elements throughout. In the drawings:

[0034] Figure 1 is a step flow chart of a chemical product leakage rapid early warning method according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of the present application.

[0036] The specific implementation of the present application will be described in detail below with reference to the drawings.

[0037] Please refer to Figure 1 which shows a step flow chart of a chemical product leakage rapid early warning method provided by an embodiment of the present application. The method comprises the following steps:

[0038] S001, collecting a plurality of pressure data in a chemical product conveying pipeline.

[0039] In the embodiments of the present application, every ten seconds is a sampling time, and a total of three hours are collected. Each time, the pressure data in the chemical product conveying pipeline is collected by a pressure sensor to obtain the pressure data at each sampling time.

[0040] S002, obtaining each pressure data window and then obtaining the data fluctuation degree in each pressure data window.

[0041] It should be noted that when the fuzzy entropy method is used for abnormal detection of pressure data, the fuzzy entropy value of each pressure data needs to be calculated, and abnormal data detection is performed according to the fuzzy entropy value of each pressure data. The calculation of the fuzzy entropy value of the pressure data is to first obtain the window of the pressure data, to obtain a plurality of sub-windows of the pressure data according to the window of the pressure data, and finally to obtain the mean value of the fuzzy membership degree between all the sub-windows of the pressure data. However, in the actual scene, the change of the pressure data caused by the gradual increase of the valve in the pipeline is usually similar to the change of the pressure data caused by the pipeline leakage, that is, both of them will cause the decrease of the pressure data in the pipeline. Therefore, the existing fuzzy entropy method cannot well distinguish the above-mentioned situations, thereby causing false detection.

[0042] In view of the above problems, by analyzing the change characteristics of the pressure data in the pipeline, the potential abnormal degree of the data in each pressure data window is obtained, and the similar tolerance degree between each pressure data sub-window is corrected according to the index, so as to obtain the adjusted similar tolerance degree between each pressure data sub-window, so as to obtain more accurate abnormal detection results subsequently.

[0043] It needs to be further explained that the numerical performance of the data in each pressure data window needs to be analyzed to obtain the data fluctuation degree in each pressure data window. Since the change of the pressure in the pipeline is reduced when there is a leakage in the pipeline, when analyzing this index, the more the data value distribution characteristics in each pressure data window conforms to the negative skewness, and the smaller the minimum value of the data in each pressure data window compared with the average of the minimum values of the data in all neighboring windows of each pressure data window, the lower the numerical performance of the data in this pressure data window, and the more likely the pipeline has a potential leakage risk in the time period corresponding to the data in this pressure data window. The data fluctuation degree in this pressure data window will also be greater.

[0044] In the embodiment of the application, the preset number of sampling time points M=300. In other embodiments, the implementer can preset the value of the sampling number M according to the specific implementation, and take each pressure data and the pressure data at the M sampling time points before the sampling time point of each pressure data as each pressure data window.

[0045] The method for obtaining a plurality of neighboring windows of each pressure data window is as follows:

[0046] A preset number of neighboring windows a is provided. For any pressure data window, the pressure data at the M sampling time points before the sampling time point of the first pressure data in the pressure data window is taken as the first neighboring window of the pressure data window. The pressure data at the M sampling time points before the sampling time point of the first pressure data in the first neighboring window is taken as the second neighboring window of the pressure data window. Similarly, until the ath neighboring window of the pressure data window is obtained. The pressure data in each neighboring window is different, but the number is the same.

[0047] The data fluctuation degree in each pressure data window is obtained as follows:

[0048] ;

[0049] In the formula, represents the data fluctuation degree in the nth pressure data window; represents the average of all data in the nth pressure data window; represents the median of all data in the nth pressure data window; norm() represents a normalization function; represents the minimum value of the data in the nth pressure data window; represents the average of the minimum values of the data in all neighboring windows of the nth pressure data window;

[0050] a value representing a data value distribution of the data in the nth pressure data window, the greater the value, the more likely that some small pressure values in the nth pressure data window exist to lower the mean value, which can indicate that the data value distribution characteristics of the data in the nth pressure data window are more likely to be negatively skewed, and the data value performance of the data in the nth pressure data window is lower, at this time, the pipeline is more likely to have a potential leakage risk in the time period corresponding to the data in the nth pressure data window, and the greater the fluctuation degree of the data in each pressure data window;

[0051] a value representing the difference between the mean value of the minimum values of the data in all adjacent windows of the nth pressure data window and the minimum value of the data in the nth pressure data window, the greater the value, the greater the credibility that the data value distribution characteristics of the data in the nth pressure data window are negatively skewed, at this time, the pipeline is more likely to have a potential leakage risk in the time period corresponding to the data in the nth pressure data window, and the greater the fluctuation degree of the data in each pressure data window.

[0052] S003、adjust the fluctuation degree of the data in each pressure data window to obtain the potential abnormality degree of the data in each pressure data window, and obtain the adjusted similarity tolerance degree between each pressure data sub-window according to the potential abnormality degree of the data.

[0053] It should be noted that the acquisition of the fluctuation degree of the data in each pressure data window is based on the preliminary analysis of whether the data in the pressure data window has a numerical decrease in the pressure data value performance level, but the pressure data change caused by gradually increasing the valve in the pipeline usually has similar characteristics with the pressure data change caused by pipeline leakage, i.e. both will cause the decrease of pressure data in the pipeline, so only by analyzing the data value performance of the pressure data window may not be able to distinguish the above two conditions that cause the decrease of pressure value, which will eventually lead to errors in the abnormal detection result of the pressure data,

[0054] So according to the scene investigation, since the pressure data change trend caused by gradually increasing the valve in the pipeline is different from the pressure data change trend caused by pipeline leakage, the pressure data change caused by gradually increasing the valve in the pipeline is relatively sharp, but it will return to stable in a short time, while the pressure data change caused by pipeline leakage is relatively flat and has persistence, so in this step, the data value change characteristics in each pressure data window are analyzed, and the data potential abnormality degree in each pressure data window is obtained by optimizing the data fluctuation degree in each pressure data window; the greater the difference between the maximum value of the data in the pressure data window and the value of the pressure data, and the more flat the numerical change between the maximum value of the data in the pressure data window and the minimum value of the data, the more likely the minimum value of the data in the pressure data window is the pressure data, so the data change characteristics in the pressure data window are more in line with the characteristics of the pressure value continuously decreasing caused by pipeline leakage, and the data potential abnormality degree in the pressure data window is greater.

[0055] In the embodiment of the application, the flatness characteristic factor between the maximum value and the minimum value of the data in each pressure data window is obtained:

[0056] ;

[0057] In the formula, represents the flatness characteristic factor between the maximum value and the minimum value of the data in the nth pressure data window; represents the sampling time interval between the maximum value and the minimum value of the data in the nth pressure data window; represents the variance of the absolute value of the difference between adjacent data between the maximum value and the minimum value of the data in the nth pressure data window;

[0058] The greater the flatness characteristic factor, the longer the time required for the change from the maximum value of the data to the minimum value of the data in the nth pressure data window, which can indicate that the numerical change between the maximum value and the minimum value of the data in the nth pressure data window is more flat, and the flatness characteristic factor is also greater; The smaller the flatness characteristic factor, the more consistent the pressure data reduction amount between each group of adjacent sampling time points in the interval from the maximum value of the data to the minimum value of the data in the nth pressure data window, which can indicate that the greater the reliability of the numerical change between the maximum value and the minimum value of the data in the nth pressure data window is more flat, and the corresponding flatness characteristic factor is also greater, The greater the flatness characteristic factor, the greater the possibility that the data segment value in the nth pressure data window has persistence and flatness, that is, the change characteristics of the data in the nth pressure data window are more in line with the characteristics of the pressure value continuously decreasing caused by pipeline leakage.

[0059] Obtaining the potential abnormality degree of the data in each pressure data window:

[0060]

[0061] In the formula, represents the potential abnormality degree of the data in the nth pressure data window; represents the fluctuation degree of the data in the nth pressure data window; represents the smoothness feature factor between the maximum value and the minimum value of the data in the nth pressure data window; norm() represents a normalization function; represents the maximum value of the data in the nth pressure data window; represents the numerical value of the nth pressure data;

[0062] The greater, the greater the fluctuation degree of the data in the nth pressure data window, and the greater the potential abnormality degree of the data in the pressure data window, which can be explained from the pressure numerical value performance level. Quantifies the relative size of the maximum value of the data in the nth pressure data window compared to the numerical value of the nth pressure data. The greater the value, the more likely the minimum value of the data in the nth pressure data window is the nth pressure data. The change characteristics of the data in the nth pressure data window are more consistent with the characteristics of the continuous decrease of the pressure value caused by the pipeline leakage, which can indicate that the pipeline is more likely to have a potential leakage risk in the time period corresponding to the data in the nth pressure data window, and the corresponding potential abnormality degree should be greater. The greater, the more consistent the change characteristics of the data in the nth pressure data window with the characteristics of the continuous decrease of the pressure value caused by the pipeline leakage.

[0063] It should be noted that after obtaining the potential abnormality degree of the data in each pressure data window, the similarity tolerance between each pressure data sub-window needs to be adjusted according to the index, and the adjusted similarity tolerance between each pressure data sub-window is obtained. The greater the potential abnormality degree of the data in the pressure data window, the greater the possibility of pressure abnormal data in the pressure data window, which can also indicate that the possibility of leakage risk in the time period corresponding to the data in the pressure data point window is greater. Therefore, in order to more accurately and conveniently identify the leakage risk, the similarity tolerance between the pressure data sub-windows should be smaller when calculating the fuzzy entropy value of the pressure data to ensure that the final calculated fuzzy entropy value is greater, thereby improving the accuracy of abnormal identification.

[0064] In the embodiment of the present application, the adjusted similarity tolerance between each pressure data sub-window is obtained as follows: ​

[0065] ;

[0066] In the formula, represents the adjustment similarity tolerance degree between the nth pressure data sub-window; represents the potential abnormality degree of the data in the nth pressure data window; exp() represents an exponential function with a natural constant as a base; Y represents a preset similarity tolerance degree between pressure data sub-windows; in an embodiment of the present application, Y is preset as 0.6, and in other embodiments, a person skilled in the art can preset the value of Y according to specific implementation manners.

[0067] S004, the fuzzy entropy of each pressure data is obtained according to the adjustment similarity tolerance degree between each pressure data sub-window, and then the abnormality of the data in each pressure data window is judged to determine whether the chemical product conveying pipeline leaks.

[0068] It should be noted that the fuzzy entropy of each pressure data is obtained according to the adjustment similarity tolerance degree between each pressure data sub-window, and then the abnormality of the data in each pressure data window is judged to determine whether the chemical product conveying pipeline leaks, and a warning needs to be given.

[0069] In an embodiment of the present application, the length of each pressure data sub-window is preset as B, the fuzzy entropy algorithm is used, each sub-window of each pressure data is obtained according to the length of each pressure data window and each pressure data sub-window, and the fuzzy entropy of each pressure data is obtained according to each sub-window of each pressure data and the characteristic similarity tolerance degree between each pressure data sub-window;

[0070] The fuzzy entropy threshold T2 is preset as 0.8, and in other embodiments, a person skilled in the art can preset the fuzzy entropy threshold T2 according to specific implementation manners.

[0071] The fuzzy entropy of all pressure data is normalized using a linear normalization method to obtain the normalized fuzzy entropy of each pressure data, and if the normalized fuzzy entropy of any pressure data is greater than the fuzzy entropy threshold T2, it indicates that abnormal data appears in the data in the pressure data window, that is, the data in the window of the pressure data corresponds to a time period in which the chemical product conveying pipeline may leak, and a warning needs to be given as soon as possible and an automatic alarm program needs to be started to notify the operator to take timely measures.

[0072] The above only describes the preferred embodiments of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for rapid early warning of leakage of chemical products, characterized in that, The method comprises the following steps: Collecting a plurality of pressure data to obtain a plurality of pressure data windows; obtaining the data fluctuation degree in each pressure data window; obtaining the data potential abnormality degree in each pressure data window; ; representing the potential abnormality degree of data in the nth pressure data window; representing the fluctuation degree of data in the nth pressure data window; representing the smoothness characteristic factor between the maximum value and the minimum value of data in the nth pressure data window; norm() represents a normalization function; representing the maximum value of data in the nth pressure data window; representing the numerical value of the nth pressure data; obtaining an adjustment similarity tolerance degree between each pressure data sub-window , represents an adjustment similarity tolerance degree between the nth pressure data sub-window; exp() represents an exponential function with a natural constant as a base number; Y represents a preset similarity tolerance degree between pressure data sub-windows Obtaining the fuzzy entropy of each pressure data according to the adjusted similarity tolerance; Judging the abnormality of each pressure data window according to the fuzzy entropy; The data fluctuation degree in each pressure data window is obtained, including: obtaining a plurality of neighborhood windows of each pressure data window; ; wherein, represents the mean value of all data in the nth pressure data window; represents the median of all data in the nth pressure data window; represents the minimum value of data in the nth pressure data window; represents the mean value of the minimum values of data in all neighborhood windows of the nth pressure data window. The acquisition of the smoothness factor between the maximum and minimum values of the data in the pressure data window comprises: wherein represents the sampling time interval between the maximum and minimum values of the data in the nth pressure data window; represents the variance of the absolute values of the differences between adjacent data between the maximum and minimum values of the data in the n pressure data windows. Obtaining the fuzzy entropy of each pressure data according to the adjusted similarity tolerance, comprising: presetting the length of each pressure data sub-window as B; using the fuzzy entropy algorithm to obtain each sub-window of each pressure data according to each pressure data window and the length of each pressure data sub-window; and obtaining the fuzzy entropy of each pressure data according to each sub-window of each pressure data and the physical sign similarity tolerance between each pressure data sub-window. Judging the abnormality of each pressure data window according to the fuzzy entropy, comprising: using a linear normalization method to normalize the fuzzy entropy of all pressure data to obtain the normalized fuzzy entropy of each pressure data; and if the normalized fuzzy entropy of any pressure data is greater than a fuzzy entropy threshold T2, it indicates that there is abnormal data in the data in the pressure data window, and the chemical product conveying pipeline leaks in the time period corresponding to the data in the pressure data window.

2. The method according to claim 1, characterized in that, Obtaining a plurality of pressure data windows, comprising: Presetting the number of sampling time points M, and taking each pressure data and the pressure data at the M sampling time points before the sampling time point of each pressure data as each pressure data window.

3. The method according to claim 1, characterized in that, Obtaining a plurality of neighborhood windows for each pressure data window, comprising: Presetting a neighborhood window number a, for any pressure data window, taking the pressure data at the M sampling time points before the sampling time point of the first pressure data in the pressure data window as the first neighborhood window of the pressure data window; taking the pressure data at the M sampling time points before the sampling time point of the first pressure data in the first neighborhood window as the second neighborhood window of the pressure data window; and so on until the ath neighborhood window of the pressure data window is obtained.

4. The method according to claim 1, characterized in that, Collecting a plurality of pressure data, comprising: Every ten seconds is a sampling time point, a total of three hours, each time collecting each pressure data in the chemical product conveying pipeline through a pressure sensor.

Images