Cloud management platform applied to alcohol-based liquid fuel use monitoring

Abstract

The invention discloses a cloud management platform applied to alcohol-based liquid fuel use monitoring, and relates to the technical field of stove safety, the platform obtains multi-modal data of a fuel storage tank, a liquid supply pipeline, a control valve and an alcohol-based stove through a data acquisition module; the data processing module integrates the multi-modal data into a corresponding data sequence, and obtains a dynamic equipment simulation model in combination with the regional equipment simulation model; the intelligent analysis module outputs leakage risk items and corresponding parameter values based on the graph neural network model; the decision response module locates a leakage source part and generates a safety alarm according to the leakage risk item sequence and a graph neural network model; the state real-time monitoring and fusion analysis of the alcohol-based fuel supply device are realized, the leakage risk is accurately identified and the source part is positioned through the dynamic equipment simulation model and graph neural network modeling, the safety early warning accuracy is effectively improved, and the use safety of the alcohol-based liquid fuel is ensured.

Description

Technical Field

[0001] This invention relates to the field of stove safety technology, specifically to a cloud management platform for monitoring the use of alcohol-based liquid fuels. Background Technology

[0002] Alcohol-based liquid fuels are widely used due to their clean and environmentally friendly characteristics. However, they are volatile, flammable, and explosive, posing a high risk of leakage. Existing technologies generally use a single methanol concentration sensor threshold alarm scheme. In the monitoring platform for alcohol-based liquid fuels, fixed concentration threshold detection does not take into account dynamic environmental changes. In the actual environment, factors such as wind speed, temperature, and pressure can significantly affect the leakage diffusion path and concentration distribution, leading to false alarms for fixed thresholds. This wastes emergency resources and may cause major safety accidents due to missed alarms. Furthermore, existing technologies do not associate sensors, pipelines, and equipment for dynamic risk assessment, resulting in fault location relying on human experience, delayed response, and inability to meet the needs of handling hazardous chemicals.

[0003] How to achieve intelligent monitoring of alcohol-based gas stoves, and accurately identify and trace the source of leak risks are the problems we need to solve. To this end, we now provide a cloud management platform for monitoring the use of alcohol-based liquid fuels. Summary of the Invention

[0004] The purpose of this invention is to provide a cloud management platform for monitoring the use of alcohol-based liquid fuels.

[0005] The objective of this invention can be achieved through the following technical solution: a cloud management platform for monitoring the use of alcohol-based liquid fuels, comprising the following:

[0006] The regional simulation model construction module is used to construct an initial regional equipment simulation model;

[0007] The data acquisition module is used to acquire multimodal data of each component of the alcohol-based gas stove;

[0008] The data processing module is used to process multimodal data to obtain corresponding multimodal data sequences;

[0009] The intelligent analysis module is used to input multimodal data sequences into the constructed graph neural network model and output leakage risk items and corresponding parameter values;

[0010] The decision response module is used to locate the source component of the leak based on the leakage risk item and the corresponding parameter value.

[0011] Furthermore, the process by which the regional simulation model construction module constructs the initial regional equipment simulation model includes:

[0012] The alcohol-based gas stove is composed of several components, including a fuel storage tank, a liquid supply pipeline, a control valve, and a gas stove burner.

[0013] Each component has corresponding basic component information, which includes component type, size, location information, and physical connection relationship with other components;

[0014] Based on the basic information of the components, virtual structural units corresponding to each component are generated, and the basic information of the components is imported into the corresponding virtual structural units. Based on the physical connection relationship of each component, the virtual structural units are topologically connected to obtain the initial regional equipment simulation model of the alcohol-based gas stove.

[0015] Furthermore, the data acquisition module consists of several data acquisition units, and the process by which the data acquisition module acquires multimodal data of each component of the alcohol-based gas stove includes:

[0016] Data acquisition units are deployed in key locations such as fuel storage tanks, liquid supply pipelines, control valves, and alcohol-based gas stoves. Multimodal data of each component during the operation of the alcohol-based gas stove are obtained through the deployed data acquisition units.

[0017] Multimodal data is input into the corresponding virtual structural units in the initial regional equipment simulation model to obtain a dynamic equipment simulation model.

[0018] Furthermore, the process by which the data processing module processes the multimodal data to obtain the corresponding multimodal data sequence includes:

[0019] Set a sliding time window and set several sampling points within the sliding time window. Based on the multimodal data of each component corresponding to each sampling point, obtain the multimodal data sequence corresponding to each component.

[0020] Based on the physical connection between the components, the association relationships between the components are set, and the corresponding association coefficients are set for the components with association relationships.

[0021] Furthermore, the intelligent analysis module inputs the multimodal data sequence into the constructed graph neural network model and outputs the leakage risk term and corresponding parameter values, including the following process:

[0022] The obtained multimodal data sequence is input into the trained graph neural network model, and the graph neural network model outputs the leakage risk item and the corresponding parameter value corresponding to the multimodal data sequence.

[0023] Based on the physical connection relationship of the components corresponding to the leakage risk items, the leakage risk items are sorted, and the sorting result is used as the output of the graph neural network model.

[0024] Furthermore, the process of constructing the graph neural network model by the intelligent analysis module includes:

[0025] Construct a graph neural network model and initialize the parameters of the constructed graph neural network model;

[0026] Collect sample data, which consists of historical operating parameters with different abnormal leaks. The historical operating parameters include leak risk items and corresponding parameter values.

[0027] The collected sample data is divided into training set and test set according to a preset ratio;

[0028] The constructed graph neural network model is trained using the training set to obtain training results. The training results are then verified using the test set. If the risk identification accuracy of the training results meets expectations, the training of the graph neural network model is complete. If the training results do not meet expectations, the model parameters are adjusted and retraining is performed until the training results meet expectations or the number of training iterations reaches the preset number, thus obtaining a trained graph neural network model.

[0029] Furthermore, the process by which the decision response module locates the source component of the leak based on the leakage risk item and the corresponding parameter value includes:

[0030] The components corresponding to leakage risk items are marked as risk components. These risk components are then mapped to corresponding virtual structural units in the dynamic equipment simulation model, which are also marked as risk units. The number of adjacent virtual structural units with physical connections to the risk units is recorded as follows: ;

[0031] like If so, the component corresponding to the risk unit is directly identified as the leakage source component;

[0032] like This indicates that multiple adjacent virtual structural units are simultaneously marked as risk units;

[0033] The source confidence level of each risk unit is obtained by making a correlation judgment based on its corresponding leakage risk item, parameter value and correlation coefficient.

[0034] The component corresponding to the risk unit with the highest source confidence value is marked as the source leakage component;

[0035] After completing the analysis of the current leakage risk item, the next leakage risk item is analyzed, until all project risk items have been analyzed, and all marked leakage source components are output.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows: Based on the basic information and physical connection relationships of the various components of the alcohol-based gas stove, an initial regional equipment simulation model of the alcohol-based gas stove is constructed. Data acquisition units are used to collect relevant multimodal data for each component within the alcohol-based gas stove, and this data is input into the corresponding virtual structural units in the initial regional equipment simulation model to obtain a dynamic equipment simulation model. A data processing module processes the multimodal data to obtain corresponding multimodal data sequences, and the correlation coefficients between the various components are obtained. The multimodal data sequences are input into the constructed graph neural network model, which outputs the leakage risk items and corresponding parameter values ​​corresponding to the multimodal data sequences. These are then sorted, and the sorting results are used as the output of the graph neural network model. Based on the leakage risk items, parameter values, and correlation coefficients, the leakage source component is located, achieving accurate identification and source location of leakage risks in the alcohol-based gas stove. This significantly improves the accuracy and reliability of leakage risk assessment and response, ensuring the safe operation of the alcohol-based gas stove. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0038] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation

[0039] like Figure 1 As shown, a cloud management platform for monitoring the use of alcohol-based liquid fuels includes a cloud management platform, which is communicatively connected to a regional simulation model construction module, a data acquisition module, a data processing module, an intelligent analysis module, and a decision response module.

[0040] The regional simulation model construction module is used to construct an initial regional equipment simulation model;

[0041] The data acquisition module is used to acquire multimodal data of each component of the alcohol-based gas stove;

[0042] The data processing module is used to process multimodal data to obtain corresponding multimodal data sequences;

[0043] The intelligent analysis module is used to input multimodal data sequences into the constructed graph neural network model and output leakage risk items and corresponding parameter values;

[0044] The decision response module is used to locate the source component of the leak based on the leakage risk item and the corresponding parameter value.

[0045] The alcohol-based gas stove is composed of several components, including a fuel storage tank, a liquid supply pipeline, a control valve, and a gas stove burner.

[0046] Each component has corresponding basic component information, which includes component type, size, location information, and physical connection relationship with other components;

[0047] The specific process by which the regional simulation model construction module constructs an initial regional equipment simulation model based on the basic information of the components is as follows:

[0048] Based on the basic information of the components, virtual structural units corresponding to each component are generated, and the basic information of the components is imported into the corresponding virtual structural units. Based on the physical connection relationship of each component, the virtual structural units are topologically connected to obtain the initial regional equipment simulation model of the alcohol-based gas stove.

[0049] It should be further explained that, in the specific implementation process, the data acquisition module consists of several data acquisition units, and the process by which the data acquisition module acquires multimodal data of each component of the alcohol-based gas stove includes:

[0050] Data acquisition units are deployed in key locations such as fuel storage tanks, liquid supply pipelines, control valves, and alcohol-based gas stoves. Multimodal data of each component during the operation of the alcohol-based gas stove are obtained through the deployed data acquisition units.

[0051] Specifically:

[0052] A data acquisition unit deployed in a fuel storage tank is used to collect methanol concentration and tank pressure.

[0053] A data acquisition unit deployed in the liquid supply pipeline is used to collect methanol concentration and pipeline pressure;

[0054] The data acquisition unit deployed on the control valve is used to collect methanol concentration and valve opening / closing status;

[0055] A data acquisition unit deployed at the burner head of the gas stove is used to collect methanol concentration and gas stove operating status;

[0056] Multimodal data is input into the corresponding virtual structural units in the initial regional equipment simulation model to obtain a dynamic equipment simulation model.

[0057] It should be further explained that, in the specific implementation process, the process by which the data processing module processes the multimodal data to obtain the corresponding multimodal data sequence includes:

[0058] Each data acquisition unit is labeled as i, where i = 1, 2, ..., n;

[0059] For the i-th data acquisition unit, with the current time as the end point, a sliding time window of length T is set. This window contains k sampling points with equal time intervals. Each sampling point is labeled and denoted as j, where j = 1, 2, ..., k.

[0060] Generate multimodal data sequences corresponding to each component;

[0061] The data acquisition unit for the corresponding fuel storage tank has the following multimodal data sequence:

[0062] ;

[0063] in This represents the methanol concentration at the time corresponding to the j-th sampling point. This represents the tank pressure at the time corresponding to the j-th sampling point;

[0064] The data acquisition unit corresponding to the liquid supply pipeline has the following multimodal data sequence:

[0065] ;

[0066] in This represents the pipeline pressure at the time corresponding to the j-th sampling point;

[0067] The data acquisition unit corresponding to the control valve has the following multimodal data sequence:

[0068] ;

[0069] in This indicates the valve's on / off state at the time corresponding to the j-th sampling point;

[0070] The data acquisition unit corresponding to the gas stove burner head has the following multimodal data sequence:

[0071] ;

[0072] in This indicates the operating status of the gas stove at the time corresponding to the j-th sampling point;

[0073] Based on the physical connection relationship between each component, the association relationship between each component is set, and the association relationship coefficient is obtained for the components with association relationship.

[0074] It should be further explained that, in the specific implementation process, the intelligent analysis module inputs the multimodal data sequence into the constructed graph neural network model and outputs the leakage risk item and the corresponding parameter value, including the following steps:

[0075] The obtained multimodal data sequence is input into the trained graph neural network model, and the graph neural network model outputs the leakage risk item and the corresponding parameter value corresponding to the multimodal data sequence.

[0076] Based on the physical connection relationship of the components corresponding to the leakage risk items, the leakage risk items are sorted, and the sorting result is used as the output of the graph neural network model.

[0077] It should be further explained that, in the specific implementation process, the process of constructing the graph neural network model by the intelligent analysis module includes:

[0078] Construct a graph neural network model and initialize the parameters of the constructed graph neural network model;

[0079] Collect sample data, which consists of historical operating parameters with different abnormal leaks. The historical operating parameters include leak risk items and corresponding parameter values.

[0080] The collected sample data is divided into training set and test set according to a preset ratio;

[0081] The constructed graph neural network model is trained using the training set to obtain training results. The training results are then verified using the test set. If the risk identification accuracy of the training results meets expectations, the training of the graph neural network model is complete. If the training results do not meet expectations, the model parameters are adjusted and retraining is performed until the training results meet expectations or the number of training iterations reaches the preset number, thus obtaining a trained graph neural network model.

[0082] It should be further explained that, in the specific implementation process, the decision response module's process of locating the leak source component based on the leak risk item and the corresponding parameter value includes:

[0083] The leakage risk items are labeled with numbers based on the sorting results output by the graph neural network model, denoted as n, where n = 1, 2, ..., N;

[0084] The components corresponding to the leakage risk items are marked as risk components;

[0085] Starting with the risk component corresponding to the leakage risk item labeled n=1, it is mapped to the corresponding virtual structural unit in the dynamic equipment simulation model and marked as the risk unit;

[0086] Obtain the adjacent virtual structural units that have a physical connection with the virtual structural unit, and record their number as . ;

[0087] like 0 indicates that no adjacent virtual structural units are marked as risk units, and the component corresponding to the risk unit is directly determined to be the leakage source component;

[0088] like This indicates that multiple adjacent virtual structural units are simultaneously marked as risk units. The adjacent risk units are labeled and denoted as q, where q = 1, 2, ... Q;

[0089] Based on the corresponding leakage risk item, parameter value, and correlation coefficient M, a correlation judgment is made to generate the source confidence level of each risk unit, denoted as . ,Right now:

[0090] ;

[0091] in This represents the parameter value corresponding to the leakage risk item of the nth risk unit; This represents the parameter value corresponding to the leakage risk term of the adjacent risk unit q of the nth risk unit; The coefficient representing the correlation between risk unit n and its adjacent risk unit q;

[0092] Source confidence The component corresponding to the risk unit with the highest value is marked as the source of the leak.

[0093] After completing the analysis of the leakage risk item currently labeled n, let Analyze the next leakage risk item until... After analyzing all project risk items, output all marked leak source components.

[0094] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications or equivalent substitutions made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A cloud management platform for monitoring the use of alcohol-based liquid fuels, characterized in that, include: The regional simulation model building module is used to build the initial regional equipment simulation model; The data acquisition module is used to acquire multimodal data of various components of the alcohol-based gas stove; The data processing module is used to process multimodal data to obtain corresponding multimodal data sequences; The intelligent analysis module is used to input multimodal data sequences into the constructed graph neural network model and output leakage risk items and corresponding parameter values; The decision response module is used to locate the source component of the leak based on the leakage risk item and the corresponding parameter value.

2. The cloud management platform for monitoring the use of alcohol-based liquid fuels according to claim 1, characterized in that, The process by which the regional simulation model construction module constructs the initial regional equipment simulation model includes: The alcohol-based gas stove is composed of several components, including a fuel storage tank, a liquid supply pipeline, a control valve, and a gas stove burner. Each component has corresponding basic component information, which includes component type, size, location information, and physical connection relationship with other components; Based on the basic information of the components, virtual structural units corresponding to each component are generated, and the basic information of the components is imported into the corresponding virtual structural units. Based on the physical connection relationship of each component, the virtual structural units are topologically connected to obtain the initial regional equipment simulation model of the alcohol-based gas stove.

3. The cloud management platform for monitoring the use of alcohol-based liquid fuels according to claim 2, characterized in that, The data acquisition module consists of several data acquisition units. The process by which the data acquisition module acquires multimodal data of each component of the alcohol-based gas stove includes: Data acquisition units are deployed in key locations such as fuel storage tanks, liquid supply pipelines, control valves, and alcohol-based gas stoves. Multimodal data of each component during the operation of the alcohol-based gas stove are obtained through the deployed data acquisition units. Multimodal data is input into the corresponding virtual structural units in the initial regional equipment simulation model to obtain a dynamic equipment simulation model.

4. The cloud management platform for monitoring the use of alcohol-based liquid fuels according to claim 3, characterized in that, The process by which the data processing module processes multimodal data to obtain corresponding multimodal data sequences includes: Set a sliding time window and set several sampling points within the sliding time window. Based on the multimodal data of each component corresponding to each sampling point, obtain the multimodal data sequence corresponding to each component. Based on the physical connection between the components, the association relationships between the components are set, and the corresponding association coefficients are set for the components with association relationships.

5. The cloud management platform for monitoring the use of alcohol-based liquid fuels according to claim 4, characterized in that, The intelligent analysis module inputs multimodal data sequences into the constructed graph neural network model and outputs leakage risk terms and corresponding parameter values. The process includes: The obtained multimodal data sequence is input into the trained graph neural network model, and the graph neural network model outputs the leakage risk item and the corresponding parameter value corresponding to the multimodal data sequence. Based on the physical connection relationship of the components corresponding to the leakage risk items, the leakage risk items are sorted, and the sorting result is used as the output of the graph neural network model.

6. The cloud management platform for monitoring the use of alcohol-based liquid fuels according to claim 5, characterized in that, The process of constructing the graph neural network model by the intelligent analysis module includes: Construct a graph neural network model and initialize the parameters of the constructed graph neural network model; Collect sample data, which consists of historical operating parameters with different abnormal leaks. The historical operating parameters include leak risk items and corresponding parameter values. The collected sample data is divided into training set and test set according to a preset ratio; The constructed graph neural network model is trained using the training set to obtain training results. The training results are then verified using the test set. If the risk identification accuracy of the training results meets expectations, the training of the graph neural network model is complete. If the training results do not meet expectations, the model parameters are adjusted and retraining is performed until the training results meet expectations or the number of training iterations reaches the preset number, thus obtaining a trained graph neural network model.

7. The cloud management platform for monitoring the use of alcohol-based liquid fuels according to claim 6, characterized in that, The process by which the decision response module locates the source component of the leak based on the leak risk item and the corresponding parameter value includes: The components corresponding to leakage risk items are marked as risk components. These risk components are then mapped to corresponding virtual structural units in the dynamic equipment simulation model and marked as risk units. The adjacent virtual structural units with physical connections to the risk units are obtained, and their number is recorded as follows: ; like If so, the component corresponding to the risk unit is directly identified as the leakage source component; like This indicates that multiple adjacent virtual structural units are simultaneously marked as risk units; The source confidence level of each risk unit is obtained by making a correlation judgment based on its corresponding leakage risk item, parameter value and correlation coefficient. The component corresponding to the risk unit with the highest source confidence value is marked as the source leakage component; After completing the analysis of the current leakage risk item, the next leakage risk item is analyzed, until all project risk items have been analyzed, and all marked leakage source components are output.

Images