Product oil ground tank interval tank capacity online correction and verification method and system

Abstract

The invention discloses a product oil ground tank interval tank capacity online correction and verification method and system, and the method comprises the steps: collecting the liquid level of a tank body and the number of oil guns, combining the oil gun error, the oil tank liquid level fluctuation, the temperature influence in the oil tank during oil unloading, the slight influence of external force on the liquid level of the oil tank in daily operation, and the like; the noise points are repeatedly removed through filtering to form a visual map, actual labeling of the noise points is completed in the map, a basic version map is formed, and a new map is formed again after two periods are changed according to the volume of the oil product; repeatedly comparing the determined noisy point set with the visual map of the original tank capacity table to remove the overlapped part, rewriting the spots of the non-overlapped part into a new temporary tank label, and carrying out the next round of verification until the determined noisy point is a suspicious point; and exporting the temporary tank capacity table to form a new formal tank capacity table so as to realize online tank capacity correction and verification. According to the method, a visual map comparison type tank calibration mode is adopted, so that the operability of a station end is improved.

Description

Technical Field

[0001] The invention relates to a method for calibrating the capacity of a gas station ground tank, and in particular to a method and system for online correction and calibration of tank capacity between finished oil ground tanks. Background Art

[0002] At present, one of the methods for implementing ground tank calibration in the industry is to use the closed-business measurement and standard tank model. The disadvantages of the traditional standard tank model are: it is impossible to control the actual quantity of finished oil, it is impossible to correctly measure and calculate financial statements, it is impossible to eliminate false losses, there are oil unloading disputes, and it is impossible to discover hidden oil leakage incidents in time. Finding a scientific and reasonable remote tank calibration method has become a top priority.

[0003] Based on 200-2000 liter standard container tanks and combined with flow meter on-site standard tanks, this method requires coordination and cooperation of gas stations under the premise of ensuring safety. On-site operations at gas stations have potential safety hazards and have a certain impact on gas station operations.

[0004] The current tank calibration technology is unable to detect tank deformation and tank error in a timely manner, cannot distinguish between loss and tank deformation, cannot control the actual amount of refined oil, cannot correctly measure and account for financial statements, cannot eliminate false losses, and has oil unloading disputes. It is imperative to find a scientific and reasonable remote tank calibration method.

[0005] The common calibration schemes for finished oil tanks at home and abroad generally include laser method, artificial geometry method, standard tank method, large flowmeter method, online level meter method, and online tank calibration method. In the process of calculating the volume of oil products, flowmeter is a commonly used measurement tool. It mainly measures the changes in temperature, pressure, viscosity, etc. Therefore, it is difficult to correct the measured values ​​of pressure, volume, temperature, etc. in the later application process. Among them, the first four belong to physical measurement methods. Because they consume manpower and physics and are relative to modern technology, they are backward and cannot be traced. Therefore, they are gradually lagging behind in the current trend of technological development and have a clear trend of elimination. The latter two online calibration methods mainly use algorithm models for precise calculations. The calculation results are often presented in the form of tables or physical databases, with high precision. Based on the data calibration tank volume, network docking services are opened. Oil companies directly provide data docking through the network to quickly solve the problem of inaccurate tank volume at gas stations. This not only changes the complex mode of traditional standard tank on-site operations, but also allows gas stations to calibrate the accurate tank volume without affecting operations and without safety hazards. However, the results still need to be interpreted and reviewed by professionals. The main responsibilities of the station personnel are sales operations or management personnel. They do not have high technical skills or mathematical calculation skills. They cannot directly participate in the entire tank calibration process, and cannot form a certain degree of judgment on the tank calibration process and the final tank label. They cannot form a certain understanding of the interpretation of ground tank warnings or ground tank deformation. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a method and system for online correction and verification of tank capacity between finished oil tanks. This method can solve the problem of direct tank calibration without stopping business, reduce the metering cost caused by tank calibration, and avoid sales losses caused by suspension of business. The present invention adopts a visual graph comparison tank calibration mode based on data tank calibration. It not only has the accuracy and rigor of digital tank calibration, but also relies on artificial intelligence and rich business data samples to form a visualized and concrete tank calibration process and intuitive results. In the actual application process, it reduces the dependence on the experience and technology of the station personnel, greatly improves the operability of the station, and simplifies the precise and complex work of tank calibration into daily inspection work. Bring simplicity and convenience back to the audience group of gas station metering personnel.

[0007] The technical solution of the present invention is:

[0008] A method for online correction and verification of tank capacity of a finished oil tank area, comprising:

[0009] (1) Online data collection automatically calibrates the tank and generates a tank capacity table, which is sent to the station when there is no refueling. Since the data has been repeatedly verified, its accuracy is relatively high. Since there are certain unverifiable areas in the data analysis method, such as the tank bottom liquid level, there are such barriers in the data analysis and verification method. The present invention borrows the unchanged part of the original tank capacity table, and through long-term data learning and calculation of the variation of the tank bottom part, with the help of AI data imaging, while visually comparing, reversely infers and verifies the difficult-to-verify parts, thereby shortening the tank calibration cycle and improving the accuracy of tank calibration;

[0010] (2) By collecting the tank liquid level and the number of oil guns for a long time (such as 2 to 3 months), combined with the difference caused by the oil gun error and the liquid level fluctuation of the oil tank, combined with the temperature impact on the inside of the oil tank during oil unloading, and the slight impact of external forces on the liquid surface of the oil tank during daily operation, the Kalman filter is used to repeatedly remove the noise points formed in the process to form a visual map. The actual annotation of the noise points is completed in the map to form a basic version map. After 2 cycles of oil volume change, a new map is re-formed. This three times is a round. One round of maps is automatically compared and analyzed with subsequent maps. When the noise points overlap, they are regarded as suspicious points. If the noise points do not overlap, they enter the next round. If the overlap exceeds 3 rounds, it is regarded as overlapping. Until the confirmation rate of the formed noise points reaches 95%, the noise points that cannot be verified in the multiple rounds of map comparison are directly eliminated, and finally a confirmed noise point set is formed; the confirmed noise points are formed into spots, and the spots are used as the main basis for the subsequent generation of tank capacity tables.

[0011] (3) Repeatedly compare the confirmed noise point set with the visual map of the original tank capacity table, mark the verified overlapping parts as spots, rewrite the non-overlapping parts into a new temporary tank label, and conduct the next round of verification until the noise points are confirmed to be suspicious points, set them as spots, and redraw them on the map. Finally, recalculate and generate the tank capacity table based on the size spots in the map. Mark the temporary tank capacity table as a temporary tank capacity table, which is used for online correction and verification of the tank capacity between ground tanks. After the verification results are completed, this temporary tank capacity table will be used as the official tank capacity table and will be issued remotely after passing the internal approval process.

[0012] The graph comparison method adopted in the present invention predicts the target tank capacity table through historical data, compares the target tank capacity table with the current volume table to form a visual graph difference, and performs repeated comparisons, aiming to visualize the tank capacity table generation process and reduce its difficulty. Ordinary gas station employees can clearly observe important difference data such as abnormal points, error rates, spots, etc. that appear in the process until a complete visual comparison result is formed. The volume table formed by the comparison result has a strong reference significance and traceability. The graph comparison method reduces the analysis difficulty of the tank calibration process, and tries to enable station personnel to accurately visualize the tank deformation and tank capacity error, so that they can understand and recognize it when an alarm is generated later.

[0013] The advantage of the atlas comparison method is that it shortens the calibration time, which is at least 50% shorter than the full comparison, and has a higher accuracy rate. In addition, due to the large number of ground tanks, the server operation is subject to a large load. The current artificial intelligence framework is relatively complete, the hardware computing power is relatively easy to obtain, and the use of machine learning and deep learning to form a visual atlas saves a lot of manpower and material resources, and has obviously broken through the opacity of the original geometric measurement method, and has a strong visual operability. And because the operation time is segmented, the resource usage can still be further optimized, and the atlas comparison can be stored in the atlas snapshots of each time period, which is conducive to the long-term monitoring and tracing of the deformation of the tank body or the deformation trend analysis based on the atlas in the later stage. It has great analytical value and early warning value in multiple ground tanks, multiple regions, and even multiple regions with similar geological features.

[0014] The present invention uses time series analysis and has the following characteristics:

[0015] Prediction: It is used to predict the future value of the time series. After the tank calibration begins, the volume table is preformed using the data from the past 2-3 months. The preformed volume table is repeatedly compared with the current volume table to generate a difference and noise comparison table.

[0016] Classification: In some cases, time series data can be considered a classification task. RNN can learn the sequence patterns of normal and abnormal heartbeats, forming heartbeat data every 1 minute during the unloading and gun discharge period, and supplemented by supplementary data formed by each fluctuation.

[0017] Anomaly detection: Use RNN technology to timely monitor the operating status of equipment and the measurement errors of the measuring equipment involved in the process, and classify the anomalies themselves.

[0018] Sequence generation: Generate time series data. In natural language processing, RNN (especially its variants such as long short-term memory network LSTM) is widely used for text generation. Subsequently, according to the specific conditions of each station, the feature information introduced by the receiving station is adjusted and optimized for time series data.

[0019] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for online correction and verification of tank capacity between finished oil tank sections of the present invention.

[0020] An online correction and verification system for an online correction and verification method of a tank capacity of a finished oil tank area, comprising:

[0021] The client of the network structure service is deployed on the station side to transmit data, with the purpose of collecting the input and output data of the tank;

[0022] Initialize the basic information of the station, tank, and gun on the server side, and import the current tank capacity table.

[0023] On the client detection device, execute the scanning program, scan the network environment, and verify the transmission protocol.

[0024] Start the data receiving service on the server side to receive the tank-related data from the station side;

[0025] Use client devices to scan the station network structure and transmission protocol to troubleshoot possible anomalies;

[0026] Start the tank calibration command on the server, specify the site and tank number, start the tank calibration process, and manually start the tank calibration action;

[0027] Generate a tank capacity table based on the site and tank number to form a reference for map comparison;

[0028] Repeatedly compare the tank capacity table with the original tank capacity table to find out the differences and form a difference table;

[0029] Enable time series verification for the difference parts, and repeatedly verify and repair the real-time data according to the calculation model until the complete calculation results are generated;

[0030] Conduct overall data review based on calculation results;

[0031] After the review is completed, a new tank capacity table is generated;

[0032] The tank capacity table is sent out on the server side, and the tank capacity table is overwritten on the corresponding tank metering equipment on the station side.

[0033] Compared with the prior art, the present invention has the following advantages:

[0034] (1) The present invention can start the tank calibration process at any time. If the tank calibration process is not started, data precipitation is automatically performed to provide rich data support for the next round of online tank calibration;

[0035] (2) Visualization. Compared with the existing technology, both the geometric measurement method and the online measurement method based on the algorithm group are inaccessible to grassroots employees. It is like a black box. It is impossible to judge its accuracy and completeness, and even it is impossible to judge the deformation of the tank in time. The visualization map analysis method described in this article clearly displays each tank, each round, and each liquid level scale. It does not require any professional knowledge, and can analyze and check the abnormality of the tank. Future warning information can be traced back to the snapshot. The station staff can restore the actual map associated with the warning information through WeChat or App, which is conducive to the station staff to clearly understand the changes in the tank and the abnormal loss.

[0036] (3) Automation: data is always in a rolling state. If a site or carrier has questions about the capacity table of a ground tank, the operating system can join the process or set an expedited priority, and a new tank capacity table can be completed within 1-3 months. Existing tank capacity tables can also be visualized;

[0037] (4) Quick, no need to make an appointment for the closure time, because the closure time appointment needs to go through multiple approval nodes, and the actual closure also needs to take into account the season or the plan issued by the branch, which takes a long time. After using this technology, since there is no additional burden on the station end and no long approval process is required, it can be started with one click and accumulated continuously until the tank capacity table is generated;

[0038] (5) Accurate. After multiple rounds of calibration and data restoration comparison with the same liquid level, same measurement unit, and same temperature, the calibration and verification of the same point has actually gone through more data rounds compared to the shutdown calibration, and the accuracy rate is relatively high. Because the current artificial intelligence technology is relatively mature, the map generation technology is relatively robust, and because of the continuous collection of data in different time periods, the richness of data samples will directly improve the quality of map formation. Therefore, the accuracy of the map and the image recognition process do not require the participation of the entire GPU, and the station-side CPU is used for client rendering. To a certain extent, it reduces the resource allocation of the server and has a high promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of the method of the present invention.

[0040] Figure 2 It is a visual atlas formed according to the density of the tank capacity table of the present invention.

[0041] Figure 3 This is an example of the comparison results of using multiple curves in the tank calibration process. DETAILED DESCRIPTION

[0042] By collecting the tank liquid level and the number of oil guns for a long time, combined with the difference caused by the oil gun error and the level fluctuation of the oil tank, combined with the temperature impact on the inside of the oil tank during oil unloading, and the slight impact of external forces on the liquid surface of the oil tank during daily operation, the Kalman filter is used to repeatedly remove the noise points formed in the process to form a visual map, complete the actual marking of the noise points in the map, and form a basic version map. After 2 cycles of oil volume changes, a new map is re-formed. This is three times a round. One round of maps is automatically compared and analyzed with subsequent maps. When the noise points overlap, they are regarded as suspicious points. If the noise points do not overlap, they enter the next round. If the overlap exceeds 3 rounds, it is regarded as overlapping, until the confirmation rate of the formed noise points reaches 95%. The noise points that cannot be verified in the multi-round map comparison process are directly eliminated. Finally, a set of confirmed noise points is formed. According to the confirmed noise point set and the visual map of the original tank capacity table, the overlapping parts are eliminated, and the non-overlapping parts are rewritten into the new temporary tank label for the next round of verification until the noise points are confirmed to be suspicious points. Export the temporary tank capacity table to form a new formal tank capacity table.

[0043] Visualization of existing tank capacity tables: Import the existing tank capacity tables into the system, or initialize them once into the server system, and form a page visualization map based on the density of the tank capacity tables, such as Figure 2 shown.

[0044] Finding new noise points: Repeated confirmation is performed based on the liquid level information currently transmitted every minute. The liquid level information is divided into rising and falling. A certain set of data is formed after rising and falling. The data set is recalibrated and new noise points are formed at places that are different from the original tank capacity table.

[0045] Noise point confirmation: If the liquid level passes a noise point more than three times and the noise point is found to be abnormal, it will be marked as a noise point.

[0046] Noise point removal: If a suspected noise point is not verified after the liquid surface passes through it twice in a row, then this noise point is considered invalid and will be removed.

[0047] Generate a temporary tank capacity table: During the spectrum comparison process, the overlapping noise points generated in the first round are marked as spots, and the subsequent noise points are also marked as spots after confirmation. The tank capacity table is regenerated based on all the spots. The storage of the tank capacity table and the temporary tank capacity table physical table support subsequent export actions.

[0048] Tank capacity table verification: refers to the reconfirmation of the temporary tank capacity table, the inspection and analysis of the finished oil purchase, sales and inventory data, to obtain data on product quality, volume and density. These data can be used to evaluate the accuracy of the current temporary tank capacity table and provide a reference for further processing and distribution. If there is a significant deviation in the verification spot, it is calculated as an error rate. If the error rate is higher than 5%, the temporary tank capacity table is not recommended for adoption. If the business needs to be replaced urgently, the approval process is used to inform, and the time span of use cannot exceed 3 months. If the tank capacity table is verified, it can be issued.

[0049] The tank calibration process uses a variety of curve comparisons, and the comparison results are as follows: Figure 3 shown.

[0050] like Figure 1 As shown, the implementation process of the present invention is:

[0051] (1) Networking at the station end for data transmission.

[0052] (2) Enable the data receiving service on the server to receive data.

[0053] (3) Use client devices to scan the station network structure and transmission protocol to check for possible anomalies.

[0054] (4) The server initializes basic data, station data, ground tank data, and oil gun configuration data.

[0055] (5) Start the tank calibration command on the server, specify the site and tank number.

[0056] (6) Generate a tank capacity table based on the site and tank number.

[0057] (7) Generate snapshot maps and original tank label maps.

[0058] (8) Noise points are confirmed and reviewed during the map comparison process. The map comparison method is used to repeatedly perform modeling comparisons on various parameters such as color, position, time, and oil grade. Noise point comparisons are performed on three layers: the existing tank table map, the current snapshot map, and the previous snapshot map.

[0059] (9) Generate spot information based on noise points and convert the spots into a tank capacity record table.

[0060] (10) Review and verify according to the new tank capacity table. The corresponding scales of each liquid level in the new tank table are re-verified with the purchase, sales and inventory data. Finally, a new tank table is regenerated based on the "red + black spots", and the real-time data is repeatedly verified and repaired until a complete calculation result is generated. The overall data is reviewed based on the calculation result, and a formal tank capacity table is formed after the review is passed; the meaning of the "red spots + black spots" is: abnormal information found in the original tank table is marked as black spots, and new spots found on the basis of the original tank table are marked as red spots; red spots are used for early warning and tank label correction, and the red spots are automatically stored in black after the tank capacity table is converted into a formal tank capacity table, providing a reference for the next cycle.

[0061] (11) The tank capacity table is sent down on the server side; first, the tank capacity table on the station side is imported, and then data is collected. According to the level of data accumulation, the command for automatically pre-generating the tank capacity table is initiated. After the tank capacity table is generated, it is sent down to the station side and replaces the original tank capacity table.

[0062] As an embodiment, the system includes:

[0063] Deploy a network structure at the station end for data transmission. Configure the network IP on the level meter console, and start the upload program after the client device scans it. Trigger condition: The front-end level meter console has been installed and is working properly. The execution action is to adjust the operation mode.

[0064] Start the data receiving service on the server side to receive data. From the moment the first piece of data is received, until the data volume reaches the level required to generate a tank capacity table, the data is collected and the analysis program is not started. The data volume reaches a certain amount (three months, or the main liquid level marking points have more than three times).

[0065] Use the client device to scan the station network structure and transmission protocol to check for possible anomalies. If anomalies occur, prompt the device and scan the environment again after checking until the environment meets the standards.

[0066] Start the tank calibration command on the server, specify the site and tank number, start the automatic running program, and collect data for trial calculation at 2:30 a.m. every day. If the trial calculation result is completed, update the flag TestCalcu to 1 in the / opt / config.properties configuration file. If it is not completed, try the trial calculation repeatedly.

[0067] Generate a tank capacity table based on the site and tank number. A basic tank capacity table is generated based on the calculation results. After the tank capacity table is generated, snapshot maps of each stage will be repeatedly generated based on this tank capacity table. Read the / opt / config.properties identifier. If TestCalcu is 1, try to generate the tank capacity table. If the generation fails, enter the next cycle. If TestCalcu is 1 and the generation is successful, update Created in the configuration file to the current time, such as: 2024-03-13 14:12:13; subsequent continuous generation will continue to modify and replace the configuration. At the same time, the tank capacity table is generated to the / opt / createdlists / directory with the file name XA31-2.xlsx. This file is only used as a difference reference file. After the difference comparison is completed, it is automatically moved from this directory to / opt / referencelists / .

[0068] Repeatedly compare the tank capacity table with the original tank capacity table map to find the differences. After the differences are formed, a comparison visualization chart is formed on the back end. Repeatedly compare the differences. After the program finds the reference tank capacity table in / opt / createdlists / , it starts the difference analysis program and performs data analysis and repeated verification on the capacity table of the same tank number in / opt / currentlists / .

[0069] Data verification is enabled for the difference parts, and real-time data is repeatedly verified and repaired according to the calculation model until a complete calculation result is generated. After the comparison is completed, a new tank capacity table is formed and stored in the / opt / currentlists / directory. The original tank capacity table in the original directory is placed in the / opt / baklists / directory and automatically renamed with a date suffix.

[0070] According to the calculation results, the overall data is reviewed. At 3:00 a.m. every day, the data in the / opt / currentlists / directory that is no more than 30 days old is repeatedly checked. After the check is completed, the doubts are recorded and the noise is analyzed. Repairs are made once a day for a total of 30 days. After 30 days, this tank capacity table is regarded as the final tank capacity table and waits for subsequent issuance. If it has not been read, the file will be read-only after a maximum of 30 days and will be archived.

[0071] After the review is completed, a new tank capacity list is generated. / opt / currentlists / If it is not issued, no review will be performed after 30 days until new data for the same station and tank is generated.

[0072] The server sends the tank capacity table. After confirming that there is no refueling at the station, the server manually sends the table. After the table is sent, the specific time and sender of the station tank number are recorded in the sending record table.

Claims

1. A method for online correction and verification of tank capacity between finished oil tanks, characterized in that: include: By collecting the tank liquid level and the amount of finished oil emitted by the oil gun for several months, combined with the oil gun error, the liquid level fluctuation of the oil tank, the impact of the temperature fluctuation inside the oil tank during oil unloading on the tank capacity, and the impact of external forces on the liquid level of the oil tank during daily operation, the Kalman filter is used to repeatedly remove the noise points formed in the process to form a visual map, and the actual annotation of the noise points is completed in the map to form a basic version of the map; after two cycles of oil volume changes, a new map is re-formed; this cycle is repeated three times as a round, and each round of the map is automatically compared and analyzed with the subsequent map. When the noise points overlap, they are regarded as suspicious points. If the noise points do not overlap, they enter the next round. If they overlap for more than three rounds, they are regarded as overlapping, until the confirmation rate of the formed noise points reaches a certain threshold value; the noise points that cannot be verified in the multi-round map comparison process are directly eliminated, and finally a confirmed noise point set is formed; Repeatedly compare the confirmed noise point set with the visual map of the original tank capacity table, remove the overlapping parts, rewrite the non-overlapping parts into a new temporary tank label, and perform the next round of verification until the noise point is confirmed to be a suspicious point; The temporary tank capacity table is exported to form a new formal tank capacity table, so as to realize the online correction and verification of the tank capacity between the ground tank sections and issue it as needed.

2. The method according to claim 1, characterized in that: The specific steps include: (1) Visualization of existing tank capacity tables: Import the existing tank capacity tables into the system, or initialize them once into the server system, and generate a page visualization map based on the density of the tank capacity tables; (2) Finding new noise points: Repeated confirmation is performed based on the liquid level information currently transmitted every minute. The liquid level information is divided into rising and falling. After rising and falling, a set of data is formed. The scale map of the data set is re-measured, and the time series cross-section map different from the original tank capacity table forms a new noise point confirmation result; (3) Noise point confirmation: During the liquid level rise and fall process, the liquid level data of each point is continuously collected. The liquid level collection is divided into two communication modes: one is when the liquid level fluctuation exceeds 1 mm, and the other is when the interval is 1 minute. If the liquid level passes through a noise point more than three times and is confirmed to be abnormal, it is marked as a noise point; (4) Noise point elimination: If a suspected noise point is not verified twice in the subsequent liquid level, it is considered invalid and eliminated; (5) Generate a temporary tank capacity table: The overlapping noise points generated in the first round of the spectrum comparison process are marked as spots, and the subsequent noise points are also marked as spots after confirmation. The tank capacity table is regenerated based on all the spots. The tank capacity table is stored in the temporary tank capacity table and supports subsequent export; (6) Tank capacity table verification: By reconfirming the temporary tank capacity table, the refined oil purchase, sales and inventory data are tested and analyzed to obtain the quality, volume and density data of the refined oil. These data are used to evaluate the accuracy of the current temporary tank capacity table and provide a reference for further processing and distribution. If there is a significant deviation in the verification spot, the error rate is calculated. If the error rate is less than 95%, the temporary tank capacity table will not be adopted. In the case of urgent replacement in business, the unadopted temporary tank capacity table will be used through the approval process and the use time span will be determined to be no more than 3 months. If the tank capacity table verification passes, it will be issued.

3. The method according to claim 1 or 2, characterized in that: For the difficult-to-verify area of ​​the tank bottom liquid level, the unchanged part of the original tank capacity table is used, and through long-term data learning and calculation of the variation of the tank bottom part, reverse inference and verification of the difficult-to-verify part.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the method for online correction and verification of tank capacity between finished oil tanks as described in any one of claims 1 to 3 are implemented.

5. An online correction and verification system for an online correction and verification method for tank capacity between finished oil tanks according to any one of claims 1 to 3, characterized in that: include: (1) Networking at the station end for data transmission; (2) Enable data receiving service on the server to receive data; (3) Use client devices to scan the station network structure and transmission protocol to check for possible anomalies; (4) The server initializes basic data, station data, ground tank data, and oil gun configuration data; (5) Start the tank calibration command on the server, specify the site and tank number; (6) Generate tank capacity table based on site and tank number; (7) Generate snapshot maps and original tank label maps; (8) Confirm and review the noise points during the map comparison process, including: using the map comparison method to repeatedly perform modeling comparisons on various parameters such as color, position, time, and oil product label, and compare the noise points on three layers: the existing tank table map, the current snapshot map, and the previous snapshot map; (9) Generate spot information based on the noise points, where new spots are marked in red; spots in the original tank label are marked in black, and finally the spots are converted into a tank capacity record table by adding red and black spots; (10) Review and verify the new tank capacity table, including: re-verify the corresponding scales of each liquid level in the new tank table with the purchase, sales and inventory data, and finally regenerate a new tank table based on the "red spots + black spots". Repeatedly verify and repair the real-time data until a complete calculation result is generated, and review the overall data based on the calculation result. After the review is passed, a formal tank capacity table is formed; the red spots are abnormal information found in the original tank table, and the black spots are new spot marks found on the basis of the original tank table; the red spots are used for early warning and tank label correction. After the tank capacity table is converted into a formal tank capacity table, the red spots are automatically stored as black spots to provide reference for the next cycle; (11) The tank capacity table is sent out on the server side, including: first importing the tank capacity table on the station side, then collecting data, and initiating an instruction to automatically pre-generate the tank capacity table based on the amount of data deposited; after the tank capacity table is generated, it is sent out to the station side and replaces the original tank capacity table at the same time.

6. The system according to claim 5, characterized in that The network structure deployed at the station end includes: Configure the network IP on the level meter console, and start the upload program after the client device scans it; the triggering condition is: the front-end level meter console has been installed and is working normally; the execution action is to adjust the operating mode.

7. The system according to claim 5, characterized in that The data receiving service started on the server side includes: From the time the first data is received until the data volume reaches the level required to generate a tank capacity table, the system is always in the collection state. The analysis program is not started, and graphs are constantly generated and compared. The main process is to generate liquid point position graphs in each time series. The data volume meets the requirement of three months or more liquid level marking points.

8. The system according to claim 5, characterized in that The overall data review based on the calculation results includes: New tank capacity tables no older than 30 days are checked repeatedly every day. After the check is completed, the noise points are recorded on the map and repaired by comparison. The table is verified once a day for a total of 30 days. After 30 days, the table is considered the final one and awaits subsequent issuance. If it has not been read, the table will be read-only after a maximum of 30 days and will be archived.

9. The system according to claim 8, characterized in that The generating of a new tank capacity table comprises: If the final tank capacity table is not issued, no review will be performed after 30 days until new data for the same tank at the same station is generated.

10. The system according to any one of claims 5 to 9, characterized in that: The action of sending the tank capacity table on the server side also includes: After confirming that there is no refueling action at the station end, the server end manually operates to perform the sending action operation; after the sending is completed, the specific sending time and sender information of the station tank number are recorded in the sending record table.

Images