Oil depot operation whole-process data acquisition method based on edge calculation

By constructing the detection sequence and overall difference index of valve clusters in oil depot operations and screening data, the problem of edge computing technology not fully utilized computing power is solved, and more accurate data acquisition and higher quality data transmission are achieved.

CN118885773APending Publication Date: 2024-11-01NAN TONG YANG HONG SHI HUA CHU YUN YOU XIAN GONG SI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411387996.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-08
Publication Date
2024-11-01

AI Technical Summary

Technical Problem

The existing edge computing technology does not fully utilize the computing power when collecting oil depot operation data, resulting in the data transmitted to the central system being complex and may contain abnormal data, affecting the acquisition effect.

Method used

By obtaining the entire process data of the oil depot operation, we construct the time detection sequence, liquid level velocity detection sequence and concentration detection sequence of the valve cluster, calculate the messyness and overall difference index, integrate the identification index, filter out the data to be detected, and use the mutation point detection algorithm and standard index discard abnormal data to achieve more accurate data collection.

Benefits of technology

It improves the quality and effectiveness of data collection, screens out good data, reduces the transmission of abnormal data, and enhances the safety and data processing capabilities of oil depot operations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118885773A_ABST
    Figure CN118885773A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of data processing, in particular to an oil depot operation whole-process data acquisition method based on edge computing, and the method comprises the steps: obtaining a volume data sequence, a valve opening data sequence, a spent time data sequence, a liquid level average rising speed data sequence and an oil gas average concentration data sequence; constructing a time detection sequence, a liquid level speed detection sequence and a concentration detection sequence; calculating the disorder degree of the time detection sequence, the liquid level speed detection sequence and the concentration detection sequence; calculating a first overall difference index, a second overall difference index and a third overall difference index; calculating a screening index; calculating the fluctuation degree of the to-be-detected concentration sequence; constructing a concentration potential sequence and a volume attribution sequence; calculating a corresponding standard index; obtaining an abandoning index, and completing the data collection of the whole process of the oil depot operation. Therefore, quality evaluation is carried out on the initial data, good data and abnormal data are screened out, and the quality and effect of data acquisition are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and specifically to a method for collecting data from the entire process of oil depot operations based on edge computing. Background Art

[0002] In the global energy industry, oil depots are key nodes in the oil supply chain, and their operating efficiency, safety, and environmental protection are crucial to the stable operation of the entire industry chain. Traditional data collection methods often rely on real-time transmission of all sensor data to the central system for processing. However, oil depot operations involve a large amount of sensor data, such as liquid level measurement, temperature detection, gas detection, etc., with large amounts of data and frequent updates; when the amount of data surges, it may exceed the processing capacity of the central system. At this time, if several operating processes have abnormal conditions, the system response time will be extended to detect whether it is abnormal data, or even the system will crash, thereby reducing the safety of oil depot operations. Faced with complex and diverse oil depot operating equipment, numerous operating processes, and real-time monitoring needs, edge computing methods are used in the data collection and processing links of oil and gas operations with their characteristics of real-time monitoring, rapid response, data fusion and cleaning.

[0003] However, when the existing edge computing technology collects data on the process of oil depot operations through equipment, the edge computing equipment only performs preliminary cleaning and preprocessing on the data, and then uploads the data to the central system. Its data collection and processing method is still similar to the traditional sensor collection method, and does not fully utilize the computing power of the edge computing equipment to filter the data, so that the process data transmitted to the central system is still complicated and chaotic, and may contain abnormal data, and the data collection effect is not good. Summary of the invention

[0004] In order to solve the above technical problems, a full-process data collection method for oil depot operations based on edge computing is provided to solve existing problems.

[0005] The solution to the technical problem in this application is to provide a full-process data collection method for oil depot operations based on edge computing, including the following steps: Obtain the full process data of each filling operation, including volume data series, valve opening data series, time spent data series, average liquid level rising speed data series and average oil and gas concentration data series; Constructing the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster; According to the distribution and change trend of each time data in the time detection sequence of each volume cluster of each valve cluster, the disorder of the time detection sequence is calculated; accordingly, the disorder of the liquid level speed detection sequence and the concentration detection sequence of each volume cluster of each valve cluster is calculated respectively; According to the disorder differences of the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster and the data differences of the corresponding acquisition order, the first, second and third overall difference indexes of each volume cluster of each valve cluster are calculated, and the discrimination index of each volume cluster of each valve cluster is obtained after fusion; Based on the discrimination indexes of all volume clusters of each valve cluster, a volume sequence to be detected and a concentration sequence to be detected of each valve cluster are constructed; According to the discrete distribution degree of the concentration sequence to be detected of each valve cluster and the difference change of each concentration data, the fluctuation degree of the concentration sequence to be detected of each valve cluster is calculated; A mutation point detection algorithm is used for the concentration sequence to be detected of each valve cluster, and according to the fluctuation degree of the concentration sequence to be detected of each valve cluster, a concentration potential sequence and a volume attribution sequence of each valve cluster are constructed; The data position of the same collection order of each valve cluster is recorded as each collection order of each valve cluster; according to the corresponding relationship between the volume data, time data and liquid level speed data of each valve cluster in each collection order, the corresponding standard index of each collection order of each valve cluster is calculated; Based on the corresponding standard index, the discard index of each collection order of each valve cluster is obtained to complete the data collection of the entire process of the oil depot operation.

[0006] Preferably, the method for constructing the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster is: A clustering algorithm is used on the valve opening data sequence, and the obtained clusters are recorded as valve clusters; According to the collection order of the valve data in each valve cluster, the volume data, time data, liquid level speed data and concentration data of the corresponding collection order are extracted from the volume data sequence, the time spent data sequence, the average liquid level rise speed data sequence and the oil and gas average concentration data sequence to form the volume cluster sequence, time cluster sequence, speed cluster sequence and concentration cluster sequence of each valve cluster respectively; Clustering algorithm is used to cluster the volume cluster sequence of each valve cluster, and the obtained cluster is recorded as volume cluster; According to the collection order of each volume data in each volume cluster of each valve cluster, the time data, liquid level velocity data and concentration data of the corresponding collection order are extracted from the time cluster sequence, velocity cluster sequence and concentration cluster sequence respectively to form the time detection sequence, liquid level velocity detection sequence and concentration detection sequence of each volume cluster of each valve cluster.

[0007] Preferably, the calculating the disorder of the time detection sequence includes: The order sequence of each volume cluster of each valve cluster is formed by the collection order corresponding to each time data in the time detection sequence of each volume cluster of each valve cluster; Linear fitting is performed on the time detection sequence and order sequence of each volume cluster of each valve cluster to obtain the slope of the straight line; Calculate the coefficient of variation of the time detection series of each volume cluster for each valve cluster; The disorder of the time detection sequence of each volume cluster of each valve cluster is positively correlated with the slope of the straight line and the coefficient of variation.

[0008] Preferably, the calculating of the first, second and third overall difference indexes of each volume cluster of each valve cluster comprises: Calculate the difference between each time data in the time detection sequence of each volume cluster of each valve cluster and the speed data of the corresponding order in the liquid level speed detection sequence, and record the standard deviation of all the differences as the difference fluctuation; The absolute value of the difference between the disorder of the time detection sequence of each volume cluster of each valve cluster and the disorder of the liquid level speed detection sequence is recorded as the disorder difference; The overall difference index between the time detection sequence and the liquid level speed detection sequence in each volume cluster of each valve cluster is recorded as the first overall difference index of each volume cluster of each valve cluster, and the first overall difference index is positively correlated with the difference fluctuation and the disorder difference respectively; For the time detection sequence and the concentration detection sequence, the calculation method of the first overall difference index is adopted to obtain the second overall difference index; accordingly, for the liquid level speed detection sequence and the concentration detection sequence, the third overall difference index is calculated.

[0009] Preferably, the discrimination index of each volume cluster of each valve cluster is a forward fusion of the first overall difference index, the second overall difference index and the third overall difference index.

[0010] Preferably, the method for constructing the volume sequence to be detected and the concentration sequence to be detected of each valve cluster is: Perform threshold segmentation on the discrimination index of all volume clusters of each valve cluster to obtain the optimal threshold; The volume clusters of each valve cluster whose discrimination index is greater than or equal to the optimal threshold are recorded as the clusters to be detected of each valve cluster; Arrange the volume data of all the clusters to be detected of each valve cluster in ascending order according to the corresponding acquisition order to serve as the volume sequence to be detected of each valve cluster; According to the collection order of each volume data in the volume sequence to be detected, the concentration data corresponding to the collection order is extracted from the oil and gas average concentration data sequence as the concentration sequence to be detected for each valve cluster.

[0011] Preferably, the calculation of the fluctuation degree of the concentration sequence to be detected of each valve cluster includes: Obtain all maximum value points and minimum value points of the concentration sequence to be detected of each valve cluster, and respectively construct the maximum value sequence and the minimum value sequence of the concentration sequence to be detected of each valve cluster; The product of the range and the standard deviation of the maximum value sequence is recorded as the first product; The product of the range and the standard deviation of the minimum value sequence is recorded as the second product; The discreteness of the concentration sequence to be detected of each valve cluster is the mean of the first product and the second product; The absolute value of the difference between each concentration data and the previous concentration data in the concentration sequence to be detected of each valve cluster is recorded as the concentration difference value; The absolute difference of the concentration of the concentration sequence to be detected of each valve cluster is the cumulative sum of the concentration difference values ​​of all data concentrations; The fluctuation degree of the concentration sequence to be detected of each valve cluster is positively correlated with the absolute difference in concentration and the discreteness, respectively.

[0012] Preferably, the method for constructing the concentration potential sequence and volume attribution sequence of each valve cluster is: A mutation point detection algorithm is used to obtain all mutation points of the concentration sequence to be detected for each valve cluster; According to the mutation point, all mutation concentration data are arranged from small to large according to the corresponding collection order to form a concentration mutation sequence for each valve cluster; Arrange the remaining concentration data in the concentration sequence to be detected of each valve cluster from small to large according to the corresponding collection order to form a non-concentration mutation sequence of each valve cluster; If each concentration data in the concentration cluster sequence of each valve cluster belongs to the concentration sequence to be detected of the corresponding valve cluster, then the concentration potential problem factor is the fluctuation degree of the concentration sequence to be detected of each valve cluster; if each concentration data in the concentration cluster sequence of each valve cluster does not belong to the concentration sequence to be detected of the corresponding valve cluster, then the concentration potential problem factor is a first preset value, wherein the first preset value is less than the fluctuation degree; If each volume data in the volume cluster sequence of each valve cluster belongs to the volume sequence to be detected, the volume attribution degree is a second preset value; if each volume data in the volume cluster sequence of each valve cluster does not belong to the volume sequence to be detected, the volume attribution degree is a third preset value, wherein the second preset value is greater than the third preset value; The concentration potential problem factors of all the concentration data in the concentration cluster sequence of each valve cluster are used to form a concentration potential sequence of each valve cluster according to the collection order of each volume data in the volume cluster sequence; The same method as the concentration potential sequence is used to construct the volume attribution sequence of each valve cluster based on the volume attribution of all volume data in the volume cluster sequence of each valve cluster.

[0013] Preferably, the step of calculating the corresponding standard index of each acquisition order of each valve cluster comprises: The ratio between each volume data in the volume cluster sequence of each valve cluster and each time data in the corresponding collection order in the time cluster sequence is recorded as the filling speed; All the filling speeds are combined into a filling speed sequence for each valve cluster according to the order of collecting the volume data in the volume cluster sequence; The absolute value of the difference between each filling speed in the filling speed sequence of each valve cluster and the liquid level speed data of the corresponding collection order in the liquid level speed sequence is recorded as the relative speed difference; Calculate the mean of all valve data of each valve cluster and record it as the average valve opening; Recording the ratio of the relative speed difference to the average valve opening as a first ratio; Using the same method as the filling speed sequence, all the first ratios are combined into a ratio sequence for each valve cluster; The sum of each first ratio in the ratio sequence of each valve cluster and each time data of the corresponding collection order in the time cluster sequence is recorded as a comprehensive coefficient; Using the same method as the filling speed sequence, all of the comprehensive coefficients are combined into a comprehensive coefficient sequence for each valve cluster; The data position of the same collection order of each valve cluster is recorded as the collection orders of each valve cluster. The corresponding standard index of each collection order of each valve cluster is positively correlated with the comprehensive coefficient sequence, concentration potential sequence and volume attribution sequence of each valve cluster, the comprehensive coefficients of the same collection order, the potential problem factors of each concentration and the volume attribution degree.

[0014] Preferably, the method of obtaining the discard index of each collection order of each valve cluster to complete the data collection of the entire process of the oil depot operation is as follows: The mean of the corresponding standard index of all acquisition orders of each valve cluster is calculated and recorded as the control index; The absolute value of the difference between the corresponding standard index of each collection order of each valve cluster and the control index is used as the rejection index of each collection order of each valve cluster; The rejection index of each collection order of each valve cluster is recorded as the rejection index of the full process data of each filling operation, and all the rejection indexes are subjected to threshold segmentation to obtain the optimal segmentation threshold; The full-process data with a discard index less than the optimal segmentation threshold is obtained to complete the data collection of the full-process of the oil depot operation.

[0015] This application has at least the following beneficial effects: The present application clusters valve data and volume data, calculates the disorder of the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster, and has the beneficial effect of reflecting the fluctuation of each data in the sequence by calculating the first overall difference index, the second overall difference index and the third overall difference index of each volume cluster of each valve cluster, and determining the discrimination index of each volume cluster of each valve cluster, and has the beneficial effect of analyzing the overall fluctuation difference between sequences and dividing the clusters, so as to accurately process the clusters that may have potential problems; by treating the detection The discrete distribution degree and concentration difference of the concentration data of the cluster are calculated, and the fluctuation degree of the concentration sequence to be detected for each valve cluster is calculated. The beneficial effect is that it can evaluate whether there is abnormal concentration data; through the correspondence between the fluctuation degree and the volume data, time data and speed data, the corresponding standard index of each collection order of each valve cluster is calculated, and the discard index of each collection order of each valve cluster is determined to complete the data collection of the entire process of the oil depot operation. The beneficial effect is that it can perform a more accurate quality assessment on the initial data collected by the edge computing device, screen out good data and abnormal data, and improve the quality and effect of data collection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The full-process data collection method for oil depot operations based on edge computing of the present application is further described in detail below in conjunction with the accompanying drawings.

[0017] Figure 1 A flowchart of the steps of the full-process data collection method for oil depot operations based on edge computing provided in this application.

[0018] Figure 2 A flowchart of the steps of a method for obtaining a discrimination index of each volume cluster of each valve cluster provided in the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of this application more clear, the following is a further detailed description of the edge computing-based oil depot operation full-process data collection method proposed in this application in conjunction with the accompanying drawings and implementation examples. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0020] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0021] See also Figure 1, which shows a flowchart of a method for collecting data from the entire process of oil depot operations based on edge computing provided by an embodiment of the present application, the method comprising the following steps: Step 1, obtain the full process data of each filling operation, including volume data series, valve opening data series, time spent data series, average liquid level rising speed data series and average oil and gas concentration data series.

[0022] The edge computing device is connected with the ultrasonic liquid level sensor and the oil vapor detector to obtain the internal liquid level rising speed data and the oil and gas concentration data in the air (in units) in real time during each tanker filling operation. ), the time interval for collecting the liquid level rise speed and the oil and gas concentration in the air is t, and the collection period of a filling operation is from the beginning to the end of this filling operation, that is, from an empty tanker to a full tanker is a collection period. The mean of the liquid level rise speed data and the mean of the oil and gas concentration data at all collection moments within a collection period are taken as the average liquid level rise speed and the average oil and gas concentration of a filling operation.

[0023] The edge computing device is connected to the tanker vehicle system and the crane monitoring system to obtain the tanker volume data, crane valve opening data (in percentage) and filling time (in seconds) during this collection cycle. In the same collection cycle, the volume of the tanker, the valve opening of the closed loading crane during the filling operation, the filling time, the average liquid level rise speed, and the average oil and gas concentration are used as the full process data of a filling operation. According to the above steps, T filling operations are carried out to obtain the full process data of T filling operations.

[0024] Preferably, as an embodiment of the present application, the collection time interval is t=1s; T=300 filling operations are performed to obtain the full process data of each filling operation.

[0025] Based on the full-process data of T filling operations obtained, wherein each filling operation corresponds to a collection order, the volume data, valve opening data, time spent data, average liquid level rising speed data, and average oil and gas concentration data obtained in each collection order are used to construct a volume data sequence, a valve opening data sequence, a time spent data sequence, an average liquid level rising speed data sequence, and an average oil and gas concentration data sequence, respectively, and the multiple data sequences are normalized.

[0026] Preferably, as an embodiment of the present application, maximum normalization processing is performed on all data sequences to eliminate the influence of different dimensions on the data, wherein the maximum normalization technology is a well-known technology and will not be described in detail here.

[0027] At this point, the volume data sequence, valve opening data sequence, time spent data sequence, average liquid level rising speed data sequence and average oil and gas concentration data sequence are obtained.

[0028] Step 2, construct the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster; calculate the disorder of the time detection sequence of each volume cluster of each valve cluster according to the distribution and change trend of each time data in the time detection sequence of each volume cluster of each valve cluster, and accordingly, calculate the disorder of the liquid level speed detection sequence and the disorder of the concentration detection sequence of each volume cluster of each valve cluster respectively.

[0029] In order to ensure the safety of oil depot operations, strict operating requirements are usually imposed during oil depot operations. Therefore, if the valve opening is consistent, the oil delivery speed is consistent, and the volume of the tanker is consistent, the average rising speed of the liquid level inside the tanker, the time spent on the tanker filling operation, and the oil and gas concentration in the air should be consistent. Therefore, the degree of difference between the data can be used to determine whether the data collected each time is abnormal.

[0030] The data points in the volume data sequence, valve opening data sequence, time spent data sequence, average liquid level rising speed data sequence and average oil and gas concentration data sequence are recorded as volume data, valve data, time data, liquid level speed data and concentration data respectively; The valve opening data sequence is clustered using a clustering algorithm, and the obtained clusters are recorded as valve clusters.

[0031] Preferably, as an embodiment of the present application, a K-means clustering algorithm is used to cluster the valve opening data, and the number K of clusters is obtained by the elbow rule, wherein the K-means algorithm and the elbow rule are well-known technologies and will not be elaborated here.

[0032] It should be understood that this embodiment only provides one clustering algorithm, namely, the K-means clustering algorithm. As other implementation methods, implementers may use other clustering algorithms in the prior art to cluster valve opening data, and this embodiment does not impose any special restrictions on this.

[0033] According to the collection order of the valve data in each valve cluster, the volume data, time data, liquid level speed data and concentration data of the corresponding collection order are extracted from the volume data sequence, the time spent data sequence, the average liquid level rise speed data sequence and the oil and gas average concentration data sequence to form the volume cluster sequence, time cluster sequence, speed cluster sequence and concentration cluster sequence of each valve cluster respectively; Clustering algorithm is used to cluster the volume cluster sequence of each valve cluster, and the obtained cluster is recorded as volume cluster; According to the collection order of each volume data in each volume cluster of each valve cluster, the time data, liquid level speed data and concentration data of the corresponding collection order are respectively extracted from the time cluster sequence, the speed cluster sequence and the concentration cluster sequence to form the time detection sequence, the liquid level speed detection sequence and the concentration detection sequence of each volume cluster of each valve cluster; The order sequence of each volume cluster of each valve cluster is formed by the collection order corresponding to each time data in the time detection sequence of each volume cluster of each valve cluster; Linear fitting is performed on the time detection sequence and order sequence of each volume cluster of each valve cluster to obtain the slope of the straight line; Calculate the coefficient of variation of the time detection series of each volume cluster for each valve cluster; The disorder of the time detection sequence of each volume cluster of each valve cluster is positively correlated with the slope of the straight line and the coefficient of variation.

[0034] It should be understood that a positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. The specific relationship can be a multiplication relationship, an addition relationship, or the idempotence of an exponential function, which is determined by actual application and this application does not impose any special restrictions on this.

[0035] Preferably, as an embodiment of the present application, a K-means clustering algorithm is used to obtain volume clusters for the volume clustering sequence of each valve cluster; linear fitting is performed using the least squares method; and the product of the slope of the straight line and the coefficient of variation is used as the disorder of the time detection sequence of each volume cluster of each valve cluster.

[0036] It should be understood that this embodiment only provides a linear fitting method, namely the least squares method. As other implementation methods, implementers can use other methods of the prior art to perform linear fitting, such as maximum likelihood estimation, Bayesian linear regression, etc. This embodiment does not impose any special restrictions on this.

[0037] It should be noted that during the tanker filling operation, if the valve opening of the closed loading crane pipe is consistent and the volume of the tanker is consistent, the time spent on each filling should be consistent, so the difference between the time data in the time detection sequence should be small, the data fluctuates around the mean, and the overall trend is very stable, so the time detection sequence is less chaotic. At this time, if the slope of the fitting line is larger and the coefficient of variation is larger, it means that the time data in the time detection sequence shows an upward trend, the data fluctuates more, and the possibility of data problems is greater.

[0038] The same calculation method as the disorder of the time detection sequence is used to obtain the disorder of the liquid level velocity detection sequence and the concentration detection sequence of each volume cluster of each valve cluster.

[0039] So far, the disorder of the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster is obtained.

[0040] Step 3: Calculate the first, second and third overall difference indexes of each volume cluster of each valve cluster based on the difference in disorder of the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster and the data difference of the corresponding collection order, and obtain the discrimination index of each volume cluster of each valve cluster after fusion.

[0041] Calculate the difference between each time data in the time detection sequence of each volume cluster of each valve cluster and the speed data of the corresponding order in the liquid level speed detection sequence, and record the standard deviation of all the differences as the difference fluctuation; The absolute value of the difference between the disorder of the time detection sequence of each volume cluster of each valve cluster and the disorder of the liquid level speed detection sequence is recorded as the disorder difference; The overall difference index between the time detection sequence and the liquid level speed detection sequence in each volume cluster of each valve cluster is recorded as the first overall difference index of each volume cluster of each valve cluster, and the first overall difference index is positively correlated with the difference fluctuation and the disorder difference respectively; Preferably, as an embodiment of the present application, the sum of the difference fluctuation and the disorder difference is used as the first overall difference index of each volume cluster of each valve cluster.

[0042] It should be noted that, since the average rising speed of the liquid level is relatively consistent when the valve opening and the volume of the tanker are consistent, the disorder of the liquid level speed detection sequence is relatively small. If the fluctuations of the time detection sequence and the liquid level speed detection sequence are relatively stable, the difference fluctuation is small, and the first overall difference index is small. If one or both sequences are not stable, the first overall difference index is large.

[0043] The same calculation method as the first overall difference index is used to calculate the overall difference index between the time detection sequence and the concentration detection sequence in each volume cluster of each valve cluster, which is recorded as the second overall difference index. Accordingly, the overall difference index between the liquid level speed detection sequence and the concentration detection sequence in each volume cluster of each valve cluster is calculated and recorded as the third overall difference index. The discrimination index of each volume cluster of each valve cluster is a forward fusion of the first overall difference index, the second overall difference index and the third overall difference index.

[0044] It should be understood that the specific relationship that can be represented by forward fusion is a multiplication relationship, an addition relationship, etc., which is determined by actual applications and this application does not impose any special restrictions on this.

[0045] Preferably, as an embodiment of the present application, the sum of the first overall difference index, the second overall difference index and the third overall difference index is used as the discrimination index of each volume cluster of each valve cluster.

[0046] It should be noted that the smaller the discrimination index of each volume cluster of each valve cluster, the more consistent the stability of the corresponding three detection sequences, and the more consistent with the characteristics of good data. The larger the discrimination index of each volume cluster of each valve cluster, the more likely the data is abnormal, and the data fluctuates greatly, so the data needs to be further processed to filter out abnormal data and good data.

[0047] Further, the step flow chart of the method for obtaining the discrimination index of each volume cluster of each valve cluster provided in the present application is as follows: Figure 2 shown.

[0048] At this point, the discrimination index of each volume cluster of each valve cluster is obtained.

[0049] Step 4, construct the volume sequence to be detected and the concentration sequence to be detected for each valve cluster; calculate the fluctuation degree of the concentration sequence to be detected for each valve cluster according to the discrete distribution degree of the concentration sequence to be detected of each volume cluster of each valve cluster and the difference change of each concentration data.

[0050] Perform threshold segmentation on the discrimination index of all volume clusters of each valve cluster to obtain the optimal threshold; The volume clusters of each valve cluster whose discrimination index is greater than or equal to the optimal threshold are recorded as the clusters to be detected of each valve cluster; Arrange the volume data of all the clusters to be detected of each valve cluster in ascending order according to the corresponding acquisition order to serve as the volume sequence to be detected of each valve cluster; According to the collection order of each volume data in the volume sequence to be detected, the concentration data corresponding to the collection order is extracted from the oil and gas average concentration data sequence as the concentration sequence to be detected for each valve cluster; Preferably, as an embodiment of the present application, the Otsu threshold method is used to segment the discrimination indexes of all volume clusters of each valve cluster to obtain the optimal threshold.

[0051] It should be understood that this embodiment only provides a threshold segmentation algorithm, namely, the Otsu threshold method. As other implementation methods, the implementer can use other threshold segmentation algorithms in the prior art to segment the discrimination indexes of all volume clusters, such as the information entropy threshold method, etc. This embodiment does not impose any special restrictions on this.

[0052] It should be noted that, when the valve opening is consistent, if the volume of the tank truck is inconsistent, the time spent on each filling and the average liquid level rise rate will change with the change of the volume of the tank truck. The larger the volume of the tank truck, the longer the time spent on filling and the lower the average liquid level rise rate. The oil and gas concentration data in the air will not change due to the change of the volume of the tank truck. The oil and gas concentration data is only related to the valve opening. When the valve opening data is large, the oil delivery speed is large, and the impact force at the connection between the crane pipe used for loading and the tank truck is large, which may cause the oil and gas concentration in the air to rise; when the valve opening is consistent, if some data values ​​in the oil and gas concentration data are large or small, it means that there is a high possibility that there is a problem in the collected data.

[0053] Based on the above analysis, all the maximum and minimum points of the concentration sequence to be detected of each valve cluster are obtained, and the maximum and minimum sequences of the concentration sequence to be detected of each valve cluster are respectively constructed; The product of the range and the standard deviation of the maximum value sequence is recorded as the first product; The product of the range and the standard deviation of the minimum value sequence is recorded as the second product; The discreteness of the concentration sequence to be detected of each valve cluster is the mean of the first product and the second product; The absolute value of the difference between each concentration data and the previous concentration data in the concentration sequence to be detected of each valve cluster is recorded as the concentration difference value; The absolute difference of the concentration of the concentration sequence to be detected of each valve cluster is the cumulative sum of the concentration difference values ​​of all data concentrations; The fluctuation degree of the concentration sequence to be detected of each valve cluster is positively correlated with the absolute difference in concentration and the discreteness, respectively.

[0054] Preferably, as an embodiment of the present application, the degree of fluctuation of the concentration sequence to be detected of each valve cluster is the product of the absolute difference in concentration and the discreteness.

[0055] It should be noted that, due to the strict operating requirements between the closed loading crane and the tanker, the concentration data monitored during each filling operation should be relatively consistent and have a small fluctuation range. The difference between the internal elements of the corresponding maximum value sequence and minimum value sequence is small, and the discreteness is small. At the same time, since the fluctuation range of the concentration sequence to be detected is small, the concentration sequence to be detected is also small, and the degree of fluctuation of the concentration sequence to be detected is small. If there is an oil and gas leak, damage to the oil vapor detector, etc., which causes the concentration data to increase or decrease abnormally, the discreteness is large, and the concentration difference value of the abnormal concentration data is large, then the degree of fluctuation is large. At this time, if the degree of fluctuation is smaller, the possibility of good concentration data is greater; the greater the degree of fluctuation, the greater the possibility of abnormal concentration data.

[0056] At this point, the fluctuation degree of the concentration sequence to be detected of each valve cluster is obtained.

[0057] Step 5, construct the concentration potential sequence and volume attribution sequence of each valve cluster; record the data position of the same collection order of each valve cluster as each collection order of each valve cluster, and calculate the corresponding standard index of each collection order of each valve cluster according to the corresponding relationship between the volume data, time data and liquid level speed data of each collection order of each valve cluster.

[0058] A mutation point detection algorithm is used to obtain all mutation points of the concentration sequence to be detected for each valve cluster; According to the mutation point, all mutation concentration data are arranged from small to large according to the corresponding collection order to form a concentration mutation sequence for each valve cluster; Arrange the remaining concentration data in the concentration sequence to be detected of each valve cluster from small to large according to the corresponding collection order to form a non-concentration mutation sequence of each valve cluster; If each concentration data in the concentration cluster sequence of each valve cluster belongs to the concentration sequence to be detected of the corresponding valve cluster, then the concentration potential problem factor is the fluctuation degree of the concentration sequence to be detected of each valve cluster; if each concentration data in the concentration cluster sequence of each valve cluster does not belong to the concentration sequence to be detected of the corresponding valve cluster, then the concentration potential problem factor is a first preset value, wherein the first preset value is less than the fluctuation degree; If each volume data in the volume cluster sequence of each valve cluster belongs to the volume sequence to be detected, the volume attribution degree is a second preset value; if each volume data in the volume cluster sequence of each valve cluster does not belong to the volume sequence to be detected, the volume attribution degree is a third preset value, wherein the second preset value is less than the third preset value; Preferably, as an embodiment of the present application, the Pelt algorithm is used for mutation point detection; the first preset value is 1, the second preset value is 3, and the third preset value is 1. As other implementation methods, the implementer can set it according to the actual situation.

[0059] It should be understood that this embodiment only provides a mutation point detection algorithm, namely, the Pelt algorithm. As other implementation methods, implementers can use other mutation point detection algorithms in the prior art to obtain all mutation points of the concentration sequence to be detected for each valve cluster, such as the Bayesian mutation detection algorithm, the Pettitt mutation detection algorithm, etc. This embodiment does not impose any special restrictions on this.

[0060] The concentration potential problem factors of all the concentration data in the concentration cluster sequence of each valve cluster are used to form a concentration potential sequence of each valve cluster according to the collection order of each volume data in the volume cluster sequence; The same method as the concentration potential sequence is used to construct the volume attribution sequence of each valve cluster based on the volume attribution of all volume data in the volume cluster sequence of each valve cluster.

[0061] It should be noted that if the overall fluctuation of the concentration sequence to be detected is small, the potential problem factor of the concentration is small, and the possibility of problems with the concentration data is still small; if the overall fluctuation of the concentration sequence to be detected is large, the potential problem factor of the concentration is large, and the possibility of problems with the concentration data is large. Similarly, if the volume data in the volume cluster sequence of each valve cluster belongs to the volume sequence to be detected, it means that the volume data of the acquisition order is located in the volume cluster with a larger discrimination index, and the possibility of problems with the data is large.

[0062] The ratio between each volume data in the volume cluster sequence of each valve cluster and each time data in the corresponding collection order in the time cluster sequence is recorded as the filling speed; All the filling speeds are combined into a filling speed sequence for each valve cluster according to the order of collecting the volume data in the volume cluster sequence; The absolute value of the difference between each filling speed in the filling speed sequence of each valve cluster and the liquid level speed data of the corresponding collection order in the liquid level speed sequence is recorded as the relative speed difference; Calculate the mean of all valve data of each valve cluster and record it as the average valve opening; Recording the ratio of the relative speed difference to the average valve opening as a first ratio; Using the same method as the filling speed sequence, all the first ratios are combined into a ratio sequence for each valve cluster; The sum of each first ratio in the ratio sequence of each valve cluster and each time data of the corresponding collection order in the time cluster sequence is recorded as a comprehensive coefficient; Using the same method as the filling speed sequence, all of the comprehensive coefficients are combined into a comprehensive coefficient sequence for each valve cluster; The data position of the same collection order of each valve cluster is recorded as the collection orders of each valve cluster. The corresponding standard index of each collection order of each valve cluster is positively correlated with the comprehensive coefficient sequence, concentration potential sequence and volume attribution sequence of each valve cluster, the comprehensive coefficients of the same collection order, the potential problem factors of each concentration and the volume attribution degree.

[0063] Preferably, as an embodiment of the present application, the product of each comprehensive coefficient, each concentration potential problem factor and each volume attribution degree of the same collection order in the comprehensive coefficient sequence, concentration potential sequence and volume attribution sequence of each valve cluster is used as the corresponding standard index of each collection order of each valve cluster.

[0064] It should be noted that in normal data with consistent valve opening, the volume of the tanker, the time spent filling and the average liquid level rising speed have a certain correspondence. Therefore, no matter how the volume of the tanker changes, the filling speed and the average liquid level rising speed are relatively consistent, that is, the relative speed difference is small; and because the valve opening is relatively consistent, the change in volume can also be inferred through the change in filling time, so the filling time and volume are also relatively consistent. Therefore, if the corresponding standard index of each data point of each valve cluster is smaller, it means that the corresponding relationship between volume, time and speed in each collection order is more in line with physical laws, and the possibility of normal data is greater; if the corresponding standard index of each collection order of each valve cluster is larger, it means that it is more likely to be abnormal data.

[0065] At this point, the corresponding standard index of each collection order of each valve cluster is obtained.

[0066] Step 6: Based on the corresponding standard index of each collection order of each valve cluster, the rejection index of each collection order of each valve cluster is obtained to complete the data collection of the entire process of the oil depot operation.

[0067] The mean of the corresponding standard index of all acquisition orders of each valve cluster is calculated and recorded as the control index; The absolute value of the difference between the corresponding standard index of each collection order of each valve cluster and the control index is used as the rejection index of each collection order of each valve cluster; It should be noted that the smaller the rejection index of each collection order of each valve cluster, the smaller the possibility of problems with the data of each collection order, and the greater the possibility of being uploaded to the central system. The larger the rejection index of each collection order of each valve cluster, the greater the difference between the data of each collection order and the data of other collection orders, the greater the possibility of mutation data, and the greater the possibility of problems, and the mutation data needs to be further processed.

[0068] The rejection index of each collection order of each valve cluster is used as the rejection index of the full process data of each filling operation, and all the rejection indexes are segmented by thresholds to obtain the optimal segmentation threshold; The full-process data with an index less than the optimal segmentation threshold is discarded, and the obtained full-process data is uploaded to the central system to complete the collection of high-quality data on the entire process of oil depot operations using edge computing technology.

[0069] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0070] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0071] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the present application. It should be pointed out that, for those of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application without departing from the content of the technical solution of the present application, shall fall within the protection scope of the technical solution of the present application.

Claims

1. The full-process data collection method of oil depot operation based on edge computing is characterized by: The method comprises the following steps: Obtain the full process data of each filling operation, including volume data series, valve opening data series, time spent data series, average liquid level rising speed data series and average oil and gas concentration data series; Constructing the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster; According to the distribution and change trend of each time data in the time detection sequence of each volume cluster of each valve cluster, the disorder of the time detection sequence is calculated; accordingly, the disorder of the liquid level speed detection sequence and the concentration detection sequence of each volume cluster of each valve cluster is calculated respectively; According to the disorder differences of the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster and the data differences of the corresponding acquisition order, the first, second and third overall difference indexes of each volume cluster of each valve cluster are calculated, and the discrimination index of each volume cluster of each valve cluster is obtained after fusion; Based on the discrimination indexes of all volume clusters of each valve cluster, a volume sequence to be detected and a concentration sequence to be detected of each valve cluster are constructed; According to the discrete distribution degree of the concentration sequence to be detected of each valve cluster and the difference change of each concentration data, the fluctuation degree of the concentration sequence to be detected of each valve cluster is calculated; A mutation point detection algorithm is used for the concentration sequence to be detected of each valve cluster, and according to the fluctuation degree of the concentration sequence to be detected of each valve cluster, a concentration potential sequence and a volume attribution sequence of each valve cluster are constructed; The data position of the same collection order of each valve cluster is recorded as each collection order of each valve cluster; according to the corresponding relationship between the volume data, time data and liquid level speed data of each valve cluster in each collection order, the corresponding standard index of each collection order of each valve cluster is calculated; Based on the corresponding standard index, the discard index of each collection order of each valve cluster is obtained to complete the data collection of the entire process of the oil depot operation.

2. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 1 is characterized in that: The construction method of the time detection sequence, liquid level speed detection sequence and concentration detection sequence of each volume cluster of each valve cluster is as follows: A clustering algorithm is used on the valve opening data sequence, and the obtained clusters are recorded as valve clusters; According to the collection order of the valve data in each valve cluster, the volume data, time data, liquid level speed data and concentration data of the corresponding collection order are extracted from the volume data sequence, the time spent data sequence, the average liquid level rise speed data sequence and the oil and gas average concentration data sequence to form the volume cluster sequence, time cluster sequence, speed cluster sequence and concentration cluster sequence of each valve cluster respectively; Clustering algorithm is used to cluster the volume cluster sequence of each valve cluster, and the obtained cluster is recorded as volume cluster; According to the collection order of each volume data in each volume cluster of each valve cluster, the time data, liquid level velocity data and concentration data of the corresponding collection order are extracted from the time cluster sequence, velocity cluster sequence and concentration cluster sequence respectively to form the time detection sequence, liquid level velocity detection sequence and concentration detection sequence of each volume cluster of each valve cluster.

3. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 1 is characterized in that: The calculation of the disorder of the time detection sequence includes: The order sequence of each volume cluster of each valve cluster is formed by the collection order corresponding to each time data in the time detection sequence of each volume cluster of each valve cluster; Linear fitting is performed on the time detection sequence and order sequence of each volume cluster of each valve cluster to obtain the slope of the straight line; Calculate the coefficient of variation of the time detection series of each volume cluster for each valve cluster; The disorder of the time detection sequence of each volume cluster of each valve cluster is positively correlated with the slope of the straight line and the coefficient of variation.

4. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 1 is characterized in that: The calculating of the first, second and third overall difference indexes of each volume cluster of each valve cluster comprises: Calculate the difference between each time data in the time detection sequence of each volume cluster of each valve cluster and the speed data of the corresponding order in the liquid level speed detection sequence, and record the standard deviation of all the differences as the difference fluctuation; The absolute value of the difference between the disorder of the time detection sequence of each volume cluster of each valve cluster and the disorder of the liquid level speed detection sequence is recorded as the disorder difference; The overall difference index between the time detection sequence and the liquid level speed detection sequence in each volume cluster of each valve cluster is recorded as the first overall difference index of each volume cluster of each valve cluster, and the first overall difference index is positively correlated with the difference fluctuation and the disorder difference respectively; For the time detection sequence and the concentration detection sequence, the calculation method of the first overall difference index is adopted to obtain the second overall difference index; accordingly, for the liquid level speed detection sequence and the concentration detection sequence, the third overall difference index is calculated.

5. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 1 is characterized in that: The discrimination index of each volume cluster of each valve cluster is a forward fusion of the first overall difference index, the second overall difference index and the third overall difference index.

6. The method for collecting data from the entire process of oil depot operations based on edge computing according to claim 1, characterized in that: The method for constructing the volume sequence to be detected and the concentration sequence to be detected of each valve cluster is as follows: Perform threshold segmentation on the discrimination index of all volume clusters of each valve cluster to obtain the optimal threshold; The volume clusters of each valve cluster whose discrimination index is greater than or equal to the optimal threshold are recorded as the clusters to be detected of each valve cluster; Arrange the volume data of all the clusters to be detected of each valve cluster in ascending order according to the corresponding acquisition order to serve as the volume sequence to be detected of each valve cluster; According to the collection order of each volume data in the volume sequence to be detected, the concentration data corresponding to the collection order is extracted from the oil and gas average concentration data sequence as the concentration sequence to be detected for each valve cluster.

7. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 1 is characterized in that: The step of calculating the fluctuation degree of the concentration sequence to be detected of each valve cluster comprises: Obtain all maximum value points and minimum value points of the concentration sequence to be detected of each valve cluster, and respectively construct a maximum value sequence and a minimum value sequence of the concentration sequence to be detected of each valve cluster; The product of the range and the standard deviation of the maximum value sequence is recorded as the first product; The product of the range and the standard deviation of the minimum value sequence is recorded as the second product; The discreteness of the concentration sequence to be detected of each valve cluster is the mean of the first product and the second product; The absolute value of the difference between each concentration data and the previous concentration data in the concentration sequence to be detected of each valve cluster is recorded as the concentration difference value; The absolute difference of the concentration of the concentration sequence to be detected of each valve cluster is the cumulative sum of the concentration difference values ​​of all data concentrations; The fluctuation degree of the concentration sequence to be detected of each valve cluster is positively correlated with the absolute difference in concentration and the discreteness, respectively.

8. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 2 is characterized in that: The method for constructing the concentration potential sequence and volume attribution sequence of each valve cluster is as follows: A mutation point detection algorithm is used to obtain all mutation points of the concentration sequence to be detected for each valve cluster; According to the mutation point, all mutation concentration data are arranged from small to large according to the corresponding collection order to form a concentration mutation sequence for each valve cluster; Arrange the remaining concentration data in the concentration sequence to be detected of each valve cluster from small to large according to the corresponding collection order to form a non-concentration mutation sequence of each valve cluster; If each concentration data in the concentration cluster sequence of each valve cluster belongs to the concentration sequence to be detected of the corresponding valve cluster, then the concentration potential problem factor is the fluctuation degree of the concentration sequence to be detected of each valve cluster; if each concentration data in the concentration cluster sequence of each valve cluster does not belong to the concentration sequence to be detected of the corresponding valve cluster, then the concentration potential problem factor is a first preset value, wherein the first preset value is less than the fluctuation degree; If each volume data in the volume cluster sequence of each valve cluster belongs to the volume sequence to be detected, the volume attribution degree is a second preset value; if each volume data in the volume cluster sequence of each valve cluster does not belong to the volume sequence to be detected, the volume attribution degree is a third preset value, wherein the second preset value is greater than the third preset value; The concentration potential problem factors of all the concentration data in the concentration cluster sequence of each valve cluster are used to form a concentration potential sequence of each valve cluster according to the collection order of each volume data in the volume cluster sequence; The same method as the concentration potential sequence is used to construct the volume attribution sequence of each valve cluster based on the volume attribution of all volume data in the volume cluster sequence of each valve cluster.

9. The method for collecting data from the entire process of oil depot operations based on edge computing according to claim 2, characterized in that: The calculating of the corresponding standard index of each acquisition order of each valve cluster includes: The ratio between each volume data in the volume cluster sequence of each valve cluster and each time data in the corresponding collection order in the time cluster sequence is recorded as the filling speed; All the filling speeds are combined into a filling speed sequence for each valve cluster according to the order of collecting the volume data in the volume cluster sequence; The absolute value of the difference between each filling speed in the filling speed sequence of each valve cluster and the liquid level speed data of the corresponding collection order in the liquid level speed sequence is recorded as the relative speed difference; Calculate the mean of all valve data of each valve cluster and record it as the average valve opening; Recording the ratio of the relative speed difference to the average valve opening as a first ratio; Using the same method as the filling speed sequence, all the first ratios are combined into a ratio sequence for each valve cluster; The sum of each first ratio in the ratio sequence of each valve cluster and each time data of the corresponding collection order in the time cluster sequence is recorded as a comprehensive coefficient; Using the same method as the filling speed sequence, all of the comprehensive coefficients are combined into a comprehensive coefficient sequence for each valve cluster; The data position of the same collection order of each valve cluster is recorded as the collection orders of each valve cluster. The corresponding standard index of each collection order of each valve cluster is positively correlated with the comprehensive coefficient sequence, concentration potential sequence and volume attribution sequence of each valve cluster, the comprehensive coefficients of the same collection order, the potential problem factors of each concentration and the volume attribution degree.

10. The method for collecting data of the entire process of oil depot operation based on edge computing according to claim 1, characterized in that: The method of obtaining the discard index of each collection order of each valve cluster and completing the data collection of the entire process of the oil depot operation is as follows: The mean of the corresponding standard index of all acquisition orders of each valve cluster is calculated and recorded as the control index; The absolute value of the difference between the corresponding standard index of each collection order of each valve cluster and the control index is used as the rejection index of each collection order of each valve cluster; The rejection index of each collection order of each valve cluster is recorded as the rejection index of the full process data of each filling operation, and all the rejection indexes are subjected to threshold segmentation to obtain the optimal segmentation threshold; The full-process data with a discard index less than the optimal segmentation threshold is obtained to complete the data collection of the full-process of the oil depot operation.