A method and system for flow transfer control of bottled liquefied gas

Abstract

The application provides a kind of bottled liquefied gas flow control method and system, the application is established for each gas cylinder Unique electronic file, and the unique two-dimensional code containing anti-fake encryption information is distributed;At each link of gas cylinder circulation, the effectiveness of the gas cylinder is tested;Based on the forced recycling ratio, the gas cylinder is forced to recycle;And through the risk assessment model, the abnormal situation of the gas cylinder is graded early warning, and the corresponding disposal scheme is pushed;At the same time, through the credit rating system, the operation specification coefficient, the rule deduction times, the early warning disposal and the timeliness index are directly associated with the rating.The application can effectively prevent the circulation of illegal gas cylinder through intelligent verification of the whole link of bottled liquefied gas circulation, such as filling, transportation, supply, distribution, use and recycling, accurate risk early warning based on machine learning and dynamic recycling mechanism, deal with potential hazards in advance, and greatly reduce the probability of safety accidents;At the same time, through the alarm and enterprise deduction mechanism, the behavior of enterprise staff can be effectively constrained.

Description

Technical Field

[0001] This invention relates to the field of liquefied gas safety technology, and in particular to a method and system for controlling the flow of bottled liquefied gas. Background Technology

[0002] Bottled liquefied petroleum gas (LPG), as an essential commodity for people's livelihood, presents extremely high requirements for safety management during its distribution process due to its flammable and explosive properties. Currently, there are many problems in the management of bottled LPG distribution that urgently need to be addressed: First, there is insufficient standardization in operational procedures at each stage. Filling, transportation, and distribution rely on manual operations and lack intelligent verification methods, making it prone to violations such as delivery without completing safety checks, cylinder swapping, and abnormal filling parameters. Second, regulatory methods are lagging behind, relying mostly on post-event spot checks, making it difficult to achieve real-time traceability throughout the entire lifecycle. The ability to predict changes in cylinder condition and potential risks is weak, making it impossible to promptly detect and address potential safety issues. Third, the mandatory recycling mechanism is rigid. Existing technology calculates the recycling ratio based on a single operational indicator and a fixed attenuation coefficient, without considering key variables such as regional risk differences, the condition of the cylinders themselves, and seasonal factors. This leads to a mismatch between the recycling ratio and actual safety needs, either excessively forcing recycling and increasing enterprise operating costs, or insufficient recycling and causing safety hazards.

[0003] Current technologies only manage a single stage, lacking a closed-loop design for the entire lifecycle and exhibiting insufficient control precision and risk prediction capabilities. Therefore, there is an urgent need to develop a bottled liquefied gas flow control solution that integrates intelligent verification, dynamic risk adaptation, multi-entity collaboration, and full-process traceability. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for controlling the flow of bottled liquefied gas cylinders. This invention ensures the safety of cylinder flow by detecting each step of the multi-cylinder flow process. At the same time, based on machine learning, it effectively prevents illegal cylinder flow, addresses potential hazards in advance, and significantly reduces the probability of safety accidents.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for controlling the gas flow of bottled liquefied petroleum gas, comprising the following steps: S1) Establish a unique electronic file for each gas cylinder and assign a unique QR code containing anti-counterfeiting and encryption information; S2) The validity of the gas cylinders is checked at each stage of their circulation. S3) Establish a dynamic ratio adjustment model based on the operation indicators, calculate the mandatory recycling ratio, and carry out mandatory recycling of gas cylinders based on the mandatory recycling ratio; S4) Construct a risk assessment model, and use the trained risk assessment model to issue graded early warnings for abnormal gas cylinder situations and push corresponding handling plans. S5) Establish a credit rating system that directly links the rating to indicators such as the coefficient of work standardization, the number of times violations are penalized, and the timeliness of early warning and handling.

[0006] Preferably, the gas cylinder circulation process includes filling, transportation, supply, distribution, use, and recycling.

[0007] Preferably, the expression for the dynamic scaling adjustment model is: ; In the formula, The mandatory recycling ratio; As the benchmark ratio; This is the work standard coefficient; Attenuation coefficient.

[0008] As a preferred option, a risk assessment model is constructed as follows: S41) Construct a dataset, collect historical data of the entire life cycle of gas cylinders over the past 3-5 years, and preprocess it; S42) Filtering high-quality feature sets based on feature engineering; S43) Construct a risk assessment model based on random forest, gradient boosting tree, and lightweight long short-term memory network; S44) The hyperparameters of the risk assessment model are optimized by using grid search combined with 5-fold cross-validation. Based on the optimized hyperparameters, the risk assessment model is trained by selecting a high-quality feature set. S45) Deploy the optimized risk assessment model to the cloud server, and based on the trained and optimized risk assessment model, classify and assess the abnormal risks of each link of the gas cylinder and take corresponding control measures.

[0009] Secondly, the present invention also provides a bottled liquefied gas flow control system, comprising: The gas cylinder identification module is used to create a unique electronic file for each gas cylinder and assign a unique QR code containing anti-counterfeiting and encryption information. The gas cylinder inspection module is used to inspect the validity of gas cylinders at every stage of their circulation. The mandatory recycling ratio calculation module is used to establish a dynamic ratio adjustment model based on operational indicators and calculate the mandatory recycling ratio. The graded early warning module uses a built-in trained risk assessment model to issue graded early warnings for abnormal gas cylinder situations and pushes corresponding handling plans. The credit rating module establishes a credit rating system that directly links operational compliance coefficients, number of violation deductions, and timeliness of early warning and handling indicators to the rating.

[0010] The beneficial effects of this invention are as follows: 1. This invention establishes an electronic file and QR code for each gas cylinder, thereby facilitating inspection at each subsequent stage and transmitting inspection information at each stage for easy viewing by inspection personnel; 2. This invention effectively prevents the circulation of illegally operated gas cylinders by intelligently verifying the entire process of filling, transporting, supplying, distributing, using, and recycling bottled liquefied gas, providing accurate risk warnings based on machine learning, and establishing a dynamic recycling mechanism. It also addresses potential hazards in advance and significantly reduces the probability of safety accidents. 3. This invention achieves accurate risk prediction through a machine learning risk assessment model, avoiding interference from invalid early warnings. At the same time, through alarms and corporate penalty mechanisms, it can effectively constrain the behavior of corporate employees. 4. This invention introduces a dynamic ratio adjustment mechanism to link the mandatory recycling ratio with operational behavior standard indicators such as filling and delivery ratio and delivery safety inspection ratio, and sets an attenuation coefficient to ensure that the mandatory recycling ratio is accurately matched with actual safety requirements, thereby avoiding increased enterprise operating costs caused by excessive mandatory recycling. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the method in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the process for constructing a risk assessment model in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the system framework in Embodiment 2 of the present invention. Detailed Implementation

[0012] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings: Example 1 like Figure 1 As shown in the figure, this embodiment of a method for controlling the gas flow of bottled liquefied petroleum gas includes the following steps: S1) Establish a unique electronic file for each gas cylinder and assign a unique QR code containing anti-counterfeiting and encryption information; In this embodiment, the electronic file includes information such as the gas cylinder's specifications, manufacturer, production date, inspection cycle, material parameters, historical maintenance records, and serial number.

[0013] S2) The validity of the gas cylinders is checked at each stage of their circulation. The aforementioned gas cylinder circulation process includes filling, transportation, supply, distribution, use, and recycling. Among these: During the cylinder filling process, a smart LPG filling electronic scale scans the cylinder's unique QR code to obtain cylinder information. The cylinder's data is then verified. If the cylinder is within its inspection validity period, not scrapped, empty, and has a unique electronic record, the cylinder verification passes. The smart LPG filling electronic scale then opens the valve to fill the cylinder and generates an unalterable filling record containing the filling amount, filling time, and operator information. Otherwise, if the cylinder fails inspection, the smart LPG filling electronic scale cannot open or restart the valve for filling.

[0014] During the transportation of gas cylinders, transport personnel scan the unique QR code of the gas cylinder through a client app to obtain the cylinder information for verification. If the gas cylinder is a heavy cylinder, has an electronic record, is within the inspection validity period, and has not been scrapped, the inspection is passed, and transportation information including transport vehicle information, transportation time, and transportation route is generated.

[0015] In the gas cylinder supply process, supply station personnel scan the unique QR code of the gas cylinder through the client to obtain the gas cylinder information in order to verify the filling information of the gas cylinder. If the gas cylinder is a heavy cylinder, has an electronic record, is within the inspection validity period, and has not been scrapped, the inspection is passed; and a supply record containing the supply station personnel, supply time, supply station, and inventory changes is generated.

[0016] During the gas cylinder delivery process, delivery personnel scan the unique QR code on the gas cylinder through the client to obtain the cylinder information in order to verify the filling information of the gas cylinder. If the gas cylinder is a heavy cylinder, has an electronic record, is within the inspection validity period, and has not been scrapped, the inspection is passed, and delivery information including gas delivery worker information, delivery time, and delivery user is generated.

[0017] During the gas cylinder usage process, delivery personnel conduct gas leak and interface sealing tests on the gas cylinders, and deliver the gas cylinders that pass the tests to the users. The delivery personnel scan the unique QR code on the gas cylinders, and generate the test results and user confirmation information.

[0018] In the gas cylinder recycling process, delivery personnel collect gas cylinders from users, generate recycling records by scanning the unique QR code on the gas cylinders, and then deliver the gas cylinders to the storage and distribution station.

[0019] S3) Establish a dynamic ratio adjustment model based on the operation indicators, calculate the mandatory recycling ratio, and carry out mandatory recycling of gas cylinders based on the mandatory recycling ratio; In this embodiment, the expression for the dynamic scaling adjustment model is: ; In the formula, The mandatory recycling ratio; As the benchmark ratio; This is the work standard coefficient; Attenuation coefficient.

[0020] In this embodiment, the work standard coefficient The calculation expression is: ; In the formula, , These are weighting coefficients; in this embodiment, ; ; , These are the filling-to-delivery ratio and the delivery-to-security-inspection ratio, respectively.

[0021] The attenuation coefficient Represented as: ; In the formula, Number of months in operation; This is the preset total number of months for the transition period.

[0022] In this embodiment, the filling delivery ratio This is expressed as the ratio of actual deliveries to filling volumes within a certain period, i.e.: ; In the formula, This represents the actual delivery volume within the period. This refers to the filling volume within the cycle. The aforementioned delivery security check ratio This represents the ratio of the number of deliveries that completed in-home security checks to the actual number of deliveries within a certain period, i.e.: ; In the formula, This refers to the number of deliveries that complete in-home security checks within the specified period.

[0023] For example: In the 6th month of operation of a gas company's system ( ),but ; If the monthly filling volume is 1000 bottles, , ,but 8; Since it is less than 1%, the mandatory recycling rate for that month is 1%, meaning that a maximum of 10 gas cylinders (1000×1%) can be forcibly recycled.

[0024] S4) Construct a risk assessment model, and use the trained risk assessment model to issue graded early warnings for abnormal gas cylinder situations, and push corresponding handling plans; such as Figure 2 As shown, the specific steps include the following: S41) Construct a dataset, collect historical data of the entire life cycle of gas cylinders over the past 3-5 years, and preprocess it; The historical data mentioned includes: Basic attribute data of gas cylinders: service life, material type, inspection cycle, factory serial number, historical maintenance records, and cumulative filling times; Operational data in the circulation process: filling pressure, flow rate, temperature, transportation route, pass rate of door-to-door safety inspection, delivery delay time, recycling compliance status, and equipment calibration records; Environmental sensing data: transportation environment temperature and humidity, user environment ventilation conditions, and regional meteorological data; Safety accident and hazard record data: historical leakage incidents, overdue inspection hazards, abnormal filling incidents, transportation violation records, and accident loss level.

[0025] The preprocessing includes: Data cleaning: Statistical methods were used to identify and remove invalid samples with a missing rate of more than 20%. Box plot method and Z-score method were used to correct extreme outliers such as pressure anomalies and temperature anomalies, and to fill in a small number of missing data. Data reduction: Principal component analysis (PCA) is used to reduce data dimensionality, remove redundant features, and retain principal components with a cumulative variance contribution rate of over 90%, thereby improving model training efficiency.

[0026] S42) Filtering high-quality feature sets based on feature engineering; Feature screening: Pearson correlation coefficient analysis was used to analyze the correlation between features and risk events. Principal component analysis (PCA) and recursive feature elimination (RFE) method were combined to screen core indicators (covering four categories: basic cylinder attributes, circulation operations, environmental parameters, and accident history) from the original data and eliminate irrelevant and highly redundant features. Feature expansion: The core features after screening are expanded, including segmented coding of service life (e.g., 0-2 years, 2-4 years, more than 4 years), feature cross-coding of filling pressure and temperature (generating pressure-temperature co-abnormal features), and processing classification features through unique thermal coding to enrich feature dimensions; Feature Validation: The effectiveness of core features is verified through feature importance assessment (using random forest feature importance scores) to ensure the contribution of selected features to risk prediction.

[0027] In this embodiment, the high-quality feature set selected by feature engineering is divided into a training set, a validation set, and a test set in a ratio of 7:2:1.

[0028] S43) Construct a risk assessment model based on random forest, gradient boosting tree, and lightweight long short-term memory network; and integrate the prediction results of each individual model through weighted voting to avoid the limitations of a single model.

[0029] S44) The hyperparameters of the risk assessment model are optimized by using a grid search combined with 5-fold cross-validation. The risk assessment model is trained based on the optimized hyperparameters and by selecting a high-quality feature set. Early stopping is used to prevent model overfitting. Training is stopped when the loss function on the validation set has not decreased for 5 consecutive rounds. S45), Model Performance Evaluation This embodiment uses accuracy, precision, recall, F1 score, ROC curve, and AUC value as evaluation metrics for model performance.

[0030] By analyzing the confusion matrix, we can identify the reasons for model misjudgments, such as specific scenarios where low-risk gas cylinders are misjudged as high-risk (false positive) or high-risk gas cylinders are misjudged as low-risk (false negative), and adjust the feature weights or model hyperparameters accordingly. Iterative optimization: Repeat the feature engineering and model training process until the model achieves a warning accuracy of no less than 95%, a recall of no less than 92%, and an F1 score of no less than 93% on the test set, meeting the actual regulatory requirements for warning accuracy.

[0031] S46) Deploy the optimized risk assessment model to the cloud server, and collect newly added gas cylinder circulation data and risk event records in real time, and update the model parameters regularly using incremental learning.

[0032] A risk assessment model based on training optimization is used to classify and assess the abnormal risks of each link of the gas cylinder and take corresponding control measures.

[0033] In this embodiment, the graded assessment of abnormal risks at each stage of the gas cylinder process includes: General warning: The risk score is 0-30 points, the gas cylinder inspection validity period is more than 3 months and no other abnormalities are found; Medium warning, with a risk score of 30-60 points, including gas cylinder inspection validity period within 1-3 months and no other abnormalities; A serious warning is issued if the risk score is above 60 points, including gas cylinders with an inspection validity period of less than one month, those that have experienced a leakage incident, those that were rectified after failing inspection and were reused, and those that have exceeded the service life limit.

[0034] S5) Establish a credit rating system that directly links operational standardization coefficients, number of violation deductions, and timeliness of early warning and handling indicators to the rating, effectively constraining the behavior of company employees.

[0035] Example 2 like Figure 3As shown, this embodiment provides a bottled liquefied gas flow control system. The system can implement the method of Embodiment 1, and the system includes: The gas cylinder identification module is used to create a unique electronic file for each gas cylinder and assign a unique QR code containing anti-counterfeiting and encryption information. The gas cylinder inspection module is used to inspect the validity of the gas cylinder at each stage of the gas cylinder circulation process. In this embodiment, the gas cylinder circulation process includes filling, transportation, supply, distribution, use, and recycling.

[0036] The mandatory recycling ratio calculation module is used to establish a dynamic ratio adjustment model based on operational indicators and calculate the mandatory recycling ratio. The graded early warning module uses a built-in trained risk assessment model to issue graded early warnings for abnormal gas cylinder situations and pushes corresponding handling plans. The credit rating module establishes a credit rating system that directly links operational compliance coefficients, number of violation deductions, and timeliness of early warning and handling indicators to the rating.

[0037] The embodiments and descriptions above are merely illustrative of the principles and preferred embodiments of the present invention. Various changes and modifications may be made to the present invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed.

Claims

1. A method for controlling the gas flow of bottled liquefied petroleum gas, characterized in that, Includes the following steps: S1) Establish a unique electronic file for each gas cylinder and assign a unique QR code containing anti-counterfeiting and encryption information; S2) The validity of the gas cylinders is checked at each stage of their circulation. S3) Establish a dynamic ratio adjustment model based on the operation indicators, calculate the mandatory recycling ratio, and carry out mandatory recycling of gas cylinders based on the mandatory recycling ratio; S4) Construct a risk assessment model, and use the trained risk assessment model to issue graded early warnings for abnormal gas cylinder situations and push corresponding handling plans. S5) Establish a credit rating system that directly links the rating to indicators such as the coefficient of operational standardization, the number of times violations are penalized, and the timeliness of early warning and handling.

2. The method for controlling the gas flow of bottled liquefied petroleum gas according to claim 1, characterized in that: The gas cylinder circulation process includes filling, transportation, supply, distribution, use, and recycling.

3. The method for controlling the flow of bottled liquefied gas according to claim 2, characterized in that: During the cylinder filling process, a smart LPG filling electronic scale scans the cylinder's unique QR code to obtain cylinder information. The cylinder's data is then verified. If the cylinder is within its inspection validity period, not scrapped, empty, and has a unique electronic file, the cylinder verification passes. The smart LPG filling electronic scale then opens the valve to fill the cylinder and generates an unalterable filling record containing the filling amount, filling time, and operator information. Otherwise, if the cylinder fails the inspection, the smart LPG filling electronic scale cannot open or restart the valve for filling.

4. The method for controlling the gas flow of bottled liquefied petroleum gas according to claim 3, characterized in that: During the transportation of gas cylinders, transport personnel scan the unique QR code of the gas cylinder through a client app to obtain the cylinder information for verification. If the gas cylinder is a heavy cylinder, has an electronic record, is within the inspection validity period, and has not been scrapped, the inspection is passed, and transportation information including transport vehicle information, transportation time, and transportation route is generated.

5. The method for controlling the gas flow of bottled liquefied petroleum gas according to claim 4, characterized in that: In the gas cylinder supply process, supply station personnel scan the unique QR code of the gas cylinder through the client to obtain the gas cylinder information in order to verify the filling information of the gas cylinder. If the gas cylinder is a heavy cylinder, has an electronic record, is within the inspection validity period, and has not been scrapped, the inspection is passed; and a supply record containing the supply station personnel, supply time, supply station, and inventory changes is generated.

6. The method for controlling the gas flow of bottled liquefied petroleum gas according to claim 5, characterized in that: During the gas cylinder delivery process, delivery personnel scan the unique QR code on the gas cylinder through the client to obtain the cylinder information in order to verify the filling information of the gas cylinder. If the gas cylinder is a heavy cylinder, has an electronic record, is within the inspection validity period, and has not been scrapped, the inspection is passed, and delivery information including gas delivery worker information, delivery time, and delivery user is generated.

7. The method for controlling the flow of bottled liquefied gas according to claim 6, characterized in that: During the gas cylinder usage process, delivery personnel conduct gas leak and interface sealing tests on the gas cylinders, and deliver the gas cylinders that pass the tests to the users. The delivery personnel scan the unique QR code on the gas cylinders and generate the test results and user confirmation information. In the gas cylinder recycling process, delivery personnel collect gas cylinders from users, generate recycling records by scanning the unique QR code on the gas cylinders, and then deliver the gas cylinders to the storage and distribution station.

8. The method for controlling the gas flow of bottled liquefied petroleum gas according to claim 1, characterized in that: Step S4) specifically includes the following steps: S41) Construct a dataset, collect historical data of the entire life cycle of gas cylinders over the past 3-5 years, and preprocess it; S42) Filtering high-quality feature sets based on feature engineering; S43) Construct a risk assessment model based on random forest, gradient boosting tree, and lightweight long short-term memory network, and integrate the prediction results of each individual model through weighted voting; S44) The hyperparameters of the risk assessment model are optimized by using grid search combined with 5-fold cross-validation. Based on the optimized hyperparameters, the risk assessment model is trained by selecting a high-quality feature set. S45) Deploy the optimized risk assessment model to the cloud server, and based on the trained and optimized risk assessment model, classify and assess the abnormal risks of each link of the gas cylinder and take corresponding control measures.

9. The method for controlling the gas flow of bottled liquefied petroleum gas according to claim 8, characterized in that: In step S42), high-quality feature sets are selected based on feature engineering, which specifically includes the following steps: Feature selection: Pearson correlation coefficient was used to analyze the correlation between features and risk events. Principal component analysis (PCA) and recursive feature elimination (RFE) method were combined to select core features from the original data. Feature expansion: The core features after screening are expanded, including segmented coding of service life, feature cross-coding of filling pressure and temperature, and processing of classification features through unique thermal coding; Feature validation: The effectiveness of core features is validated through feature importance assessment to ensure the contribution of selected features to risk prediction.

10. A bottled liquefied gas flow control system, characterized in that: The system described herein utilizes the method described in any one of claims 1-9 to achieve pneumatic control of bottled liquefied gas flow, and the system comprises: The gas cylinder identification module is used to create a unique electronic file for each gas cylinder and assign a unique QR code containing anti-counterfeiting and encryption information. The gas cylinder inspection module is used to inspect the validity of gas cylinders at every stage of their circulation. The mandatory recovery ratio calculation module is used to establish a dynamic ratio adjustment model based on operational indicators and calculate the mandatory recovery ratio. The graded early warning module uses a built-in trained risk assessment model to issue graded early warnings for abnormal gas cylinder conditions and pushes corresponding handling plans. The credit rating module establishes a credit rating system that directly links operational compliance coefficients, number of violation deductions, and timeliness of early warning and handling indicators to the rating.

Images