Hydrogen leakage response method and related device based on decision tree

Abstract

The present invention relates to the technical field of hydrogen leakage, and discloses a hydrogen leakage response method based on a decision tree and a related device, wherein the hydrogen leakage response method based on the decision tree comprises: evaluating and analyzing a basic decision tree according to historical data information and historical response decisions, and adjusting the basic decision tree according to the analysis results to obtain a target decision tree; using a sensor to detect data to obtain current data information, the current data information includes current hydrogen data and current environmental data; using the target decision tree to judge hydrogen leakage according to the current data information, and outputting the corresponding current response decision. The present invention can solve the problem of insufficient flexibility of hydrogen leakage.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen leakage, and in particular to a hydrogen leakage response method based on a decision tree and a related device. Background Art

[0002] Hydrogen leakage is an important safety issue in the production, storage, transportation and use of hydrogen energy. Hydrogen has a very small molecular size and is easy to leak. It is also very easy to diffuse in the air. Once it leaks, it is difficult to detect it in time. In addition, hydrogen is colorless and odorless, and human senses cannot directly detect its leakage, which further increases the difficulty of timely detecting leakage and taking emergency measures.

[0003] The detection and emergency response of hydrogen leaks mainly rely on fixed preset conditions and response strategies. Traditional systems usually use fixed-point hydrogen sensors to monitor the hydrogen concentration in the environment. When the detected concentration exceeds the preset safety threshold, the alarm system and emergency measures are automatically activated, such as starting the ventilation system and closing related valves.

[0004] Although current hydrogen leakage systems can handle hydrogen leakage situations to a certain extent, they usually lack flexibility and adaptability and cannot be dynamically adjusted according to different scenarios and environmental changes. Summary of the invention

[0005] In view of this, the present invention provides a hydrogen leakage response method and related devices based on a decision tree to solve the problem of insufficient flexibility in hydrogen leakage.

[0006] In a first aspect, the present invention provides a hydrogen leakage response method based on a decision tree, which includes: evaluating and analyzing a basic decision tree based on historical data information and historical response decisions, and adjusting the basic decision tree based on the analysis results to obtain a target decision tree; using a sensor to perform data detection to obtain current data information, the current data information includes current hydrogen data and current environmental data; using the target decision tree to make a hydrogen leakage judgment based on the current data information, and outputting a corresponding current response decision.

[0007] In this implementation, the detection mechanism can be automatically and dynamically adjusted based on the historical data information and historical response decision analysis of the basic decision tree. Compared with the traditional static decision tree, the dynamic decision tree can automatically adjust the decision tree based on the results of real-time data and historical data learning, so as to more accurately reflect the current operating environment and leakage conditions and improve the flexibility of hydrogen leakage. At the same time, by implementing a dynamic decision tree and an adaptive learning mechanism, the response strategy can be automatically adjusted based on real-time environmental data and feedback from historical leakage events. This adaptability enables the system to maintain efficient monitoring and response performance under different operating conditions and environmental changes, thereby greatly improving the accuracy and response speed of leak detection.

[0008] In an optional embodiment, the basic decision tree is evaluated and analyzed based on historical data information and historical response decisions, and the basic decision tree is adjusted based on the analysis results. Before obtaining the response decision tree, the process includes: determining decision parameters and response decisions, where the decision parameters include data information that affects the response decisions, and the response decisions include valve opening and closing decisions and ventilation system decisions; determining decision thresholds of multiple basic decision points and the decision logic of each basic decision point based on the decision parameters; and combining the decision thresholds of multiple basic decision points, the decision logic of each basic decision point and the response decisions to construct a basic decision tree.

[0009] In an optional embodiment, evaluating and analyzing the basic decision tree based on historical data information and historical response decisions includes: obtaining real-time data information, and using the basic decision tree to determine the real-time response decision corresponding to the real-time data information; using the historical data information and historical response decisions to train a machine learning model; and evaluating and analyzing the real-time data information and the corresponding real-time response decisions based on the trained machine learning model.

[0010] In an optional embodiment, training a machine learning model using historical data information and historical response decisions includes: obtaining historical data information and historical response decisions, the historical data information including historical detection time, historical hydrogen data, and historical environmental data; performing feature extraction on the historical data information and historical response decisions to obtain multiple feature information, the feature information including the rate of change and acceleration of hydrogen concentration, obtaining the relationship between the historical response decisions and the rate of change of hydrogen concentration, and the relationship between the historical response decisions and the historical environmental data; and training the machine learning model based on the feature information.

[0011] In an optional implementation, the environmental data includes temperature data, humidity data, wind speed and direction data, and air pressure data.

[0012] In an optional embodiment, the machine learning model is a support vector machine model, and training the machine learning model according to feature information includes: calculating performance indicators of the machine learning model based on the feature information, the performance indicators including accuracy, recall rate and F1 score; based on the performance indicators, optimizing the parameters of the machine learning model using cross-validation technology.

[0013] In an optional embodiment, adjusting the basic decision tree according to the analysis results to obtain the target decision tree includes: adjusting the number of basic decision points, adjusting the decision thresholds of the basic decision points, and adjusting one or more of the decision logics of each basic decision point according to the analysis results; and combining the decision thresholds of multiple adjusted basic decision points, the decision logic of each adjusted basic decision point, and the adjusted response decisions to construct the target decision tree.

[0014] In a second aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the decision tree-based hydrogen leakage response method of the above-mentioned first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0015] In a third aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the decision tree-based hydrogen leak response method of the first aspect or any corresponding embodiment thereof.

[0016] In a fourth aspect, the present invention provides a computer program product, comprising computer instructions for causing a computer to execute the decision tree-based hydrogen leakage response method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is a flow chart of a hydrogen leakage response method based on a decision tree according to an embodiment of the present invention;

[0019] Figure 2 is a flow chart of another hydrogen leakage response method based on a decision tree according to an embodiment of the present invention;

[0020] Figure 3is a schematic diagram of a decision tree according to an embodiment of the present invention;

[0021] Figure 4 is a structural block diagram of a hydrogen leakage response device based on a decision tree according to an embodiment of the present invention;

[0022] Figure 5 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] It is critical to ensure safe and effective monitoring and response to hydrogen leaks during the production, storage, transportation and use of hydrogen. Due to the high diffusion rate and colorless and odorless characteristics of hydrogen, its leaks are difficult to detect and deal with in a timely manner, which increases potential safety risks.

[0025] The current hydrogen leak monitoring and emergency response systems mainly face the following technical problems: 1) Insensitive or delayed leak detection: Existing detection systems may not be able to quickly identify leaks in the early stages of hydrogen leaks, resulting in delayed responses to leaks. 2) Lack of integrated real-time monitoring and automatic emergency response: Most systems require human intervention to analyze monitoring data and initiate emergency measures, which increases response time and the risk of human error. 3) Insufficient system integration and data processing capabilities: Existing solutions often lack effective data integration and processing capabilities, making it difficult to quickly extract useful information from large amounts of monitoring data and make accurate emergency decisions based on it.

[0026] According to an embodiment of the present invention, a decision tree-based hydrogen leak response method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] In this embodiment, a hydrogen leakage response method based on a decision tree is provided. Figure 1 is a flow chart of a hydrogen leakage response method based on a decision tree according to an embodiment of the present invention, such as Figure 1 As shown, the process includes the following steps:

[0028] Step S101, evaluating and analyzing the basic decision tree according to historical data information and historical response decisions, and adjusting the basic decision tree according to the analysis results to obtain a target decision tree.

[0029] Decision Tree is a decision analysis method that uses a decision tree to construct a decision tree based on the known probabilities of various situations to obtain the probability that the expected value of the net present value is greater than or equal to zero, evaluate project risks, and determine its feasibility. It is a graphical method that intuitively uses probability analysis.

[0030] The basic decision tree includes multiple basic decision nodes, the decision threshold of each basic decision node, the decision logic of each basic decision point, and the final response decision. Each decision tree branch includes at least one decision node.

[0031] The decision threshold of each basic decision point includes one or more of hydrogen data and environmental data thresholds, wherein the environmental data includes temperature data, humidity data, wind speed and direction data, and air pressure data.

[0032] Among them, temperature affects the diffusion rate and volatility of hydrogen; humidity affects the degree of mixing of hydrogen and air and its distribution in space; wind speed and direction determine the diffusion path and speed after hydrogen leakage, which is crucial to the selection of emergency response measures; air pressure affects the density and leakage rate of the gas.

[0033] Exemplarily, the decision thresholds of a basic decision point include: the hydrogen concentration is greater than the hydrogen threshold, the air pressure value is greater than the air pressure threshold, and the temperature value is greater than the temperature threshold.

[0034] Among them, response decisions include low response, standard response, emergency response, monitoring, alarm, etc.

[0035] According to the historical data information collected in real time, the rationality of the basic decision tree is judged by reasonable response decision. When the basic decision tree is unreasonable, the basic decision tree is adjusted to obtain the target decision tree.

[0036] For example, environmental parameters are incorporated into the assessment of leak severity, and the threshold of the leak response strategy is changed to adapt to different environmental conditions. For example, in the case of high wind speed, even if the hydrogen concentration is low, more stringent response measures may be required.

[0037] Exemplarily, environmental parameters are used as a basis for adjusting decision thresholds, such as adjusting the LEL (Lower Explosive Limit) threshold of hydrogen concentration, to adapt to specific environmental conditions, thereby improving the accuracy and safety of the response.

[0038] Step S102: Utilize sensors to detect data and obtain current data information.

[0039] The current data information includes current hydrogen data and current environmental data, and the current environmental data includes current temperature data, current humidity data, current wind speed and wind direction data, and current air pressure data.

[0040] In one implementation, a hydrogen sensor is used to detect the current hydrogen concentration, a temperature sensor is used to detect the current temperature data, a humidity sensor is used to detect the current humidity data, a wind sensor is used to detect the current wind speed and wind direction data, and an air pressure sensor is used to detect the current air pressure data.

[0041] Step S103, using the target decision tree to make a hydrogen leakage judgment based on the current data information, and outputting the corresponding current response decision.

[0042] At the initial node of the target decision tree, the decision threshold of the current data information is judged, the current data information is compared with the decision threshold, and the decision threshold of the current data information is further judged using the next node that meets the conditions until the final current response decision is output.

[0043] The hydrogen leakage response method based on decision tree provided in this embodiment can automatically and dynamically adjust the detection mechanism according to the historical data information and historical response decision analysis based decision tree. Compared with the traditional static decision tree, the dynamic decision tree can automatically adjust the decision tree according to the results of real-time data and historical data learning, so as to more accurately reflect the current operating environment and leakage conditions, and improve the flexibility of hydrogen leakage. At the same time, by realizing the dynamic decision tree and adaptive learning mechanism, the response strategy can be automatically adjusted according to the feedback of real-time environmental data and historical leakage events. This adaptability enables the system to maintain efficient monitoring and response performance under different operating conditions and environmental changes, thereby greatly improving the accuracy and response speed of leak detection.

[0044] In this embodiment, a hydrogen leakage response method based on a decision tree is provided, which can be used in the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 2 is a flow chart of another hydrogen leakage response method based on a decision tree according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0045] Step S201, evaluating and analyzing the basic decision tree according to historical data information and historical response decisions, and adjusting the basic decision tree according to the analysis results to obtain a target decision tree.

[0046] Specifically, the above step S201 includes:

[0047] Step S2011, constructing a basic decision tree.

[0048] Determine the decision parameters and response decisions. The decision parameters include data information that affects the response decisions. The response decisions include valve opening and closing decisions and ventilation system decisions. Among them, the decision parameters include hydrogen concentration, temperature value, air pressure value, etc.

[0049] The decision thresholds of multiple basic decision points and the decision logic of each basic decision point are determined based on the decision parameters, and the decision thresholds of multiple basic decision points, the decision logic of each basic decision point and the response decision are combined to construct a basic decision tree.

[0050] For example, see Figure 3 , Figure 3 is a schematic diagram of a decision tree according to an embodiment of the present invention. The decision parameter is the hydrogen concentration. Figure 3 As shown:

[0051] The root node of the decision tree is leak detection, which is used to detect whether a hydrogen leak occurs. Hydrogen data from the hydrogen sensor is detected, and the hydrogen data includes the hydrogen concentration and the operating status of the sensor. The decision logic of the root node is used to determine whether it is operating normally, whether there is a fault, and whether there is a hydrogen leak. When the sensor is in operation, the normal monitoring status is maintained; if the sensor status is abnormal, the redundant sensor is started and maintenance is notified. If the hydrogen concentration is less than 5%, go to the next node "No response: normal operation"; if the hydrogen concentration is greater than or equal to 5%LEL or the temperature and pressure suddenly change, and the sensor status is normal, go to the next node "Preliminary Leak Assessment".

[0052] The node "Preliminary Leak Assessment" is used to preliminarily determine the potential severity of the leak. If the hydrogen concentration is between 5% and 15% LEL, go to the next node "Small Leak Response"; if the hydrogen concentration is between 15% and 25% LEL, go to the next node "Medium Leak Response"; if the hydrogen concentration exceeds 25% LEL, go to the next node "Large Leak Response".

[0053] The node "Small Leak Response" is used for specific response measures to small leaks. If the leak detection results show that the hydrogen concentration is between 5% and 10% LEL or the hydrogen concentration rises sharply, the automation will take low response measures and then transfer to the next node "Tracking and Feedback". Among them, the low response measures include activating the local ventilation system to reduce the hydrogen concentration, and issuing warnings to the control center and staff in related areas to ensure that the staff are aware and alert, and continue to monitor the area to track changes in hydrogen concentration. If the leak detection results show that the hydrogen concentration exceeds 10% but does not exceed 15% LEL, the automation will take enhanced monitoring measures, and then transfer to the next node "Tracking and Feedback". Among them, the enhanced monitoring measures include increasing the frequency of hydrogen detection on the basis of low response measures, collecting more environmental data about the leakage area, and evaluating the trend and spread rate of the leakage to determine whether further action is needed.

[0054] For the node "Medium-scale Leak Response", it is used for specific response measures for medium-scale leaks. If the leak detection results show that the hydrogen concentration is between 15% and 20% LEL or the hydrogen concentration rises sharply, the automation takes standard response measures and then transfers to the next node "Safety Assessment". After the node "Safety Assessment", it transfers to the next node "Tracking and Feedback". Among them, the standard response measures include immediately closing the valve closest to the leak point to isolate the hydrogen source, starting the emergency ventilation system to reduce the hydrogen concentration, and notifying the on-site safety personnel and management. Start the emergency lighting and guidance system to guide possible evacuations. If the leak detection results show that the hydrogen concentration exceeds 20% but does not exceed 25% LEL, the automation takes enhanced response measures and then transfers to the next node "Safety Assessment". After the node "Safety Assessment", it transfers to the next node "Tracking and Feedback". Among them, the enhanced response measures include, in addition to implementing all the measures of the standard response, evaluating and closing valves in other potentially affected areas to prevent the spread of the leak. Organize local evacuations, especially if monitoring data show that a rapid increase in hydrogen concentration or a change in wind direction may cause the leak to spread to critical areas.

[0055] For the node "Large-Scale Leak Response", it is used for emergency response measures for large-scale leaks. Automatically take emergency response measures and transfer to the next node "Crisis Management". After the node "Crisis Management", transfer to the next node "Tracking and Feedback". Among them, the emergency response measures include immediately launching a comprehensive evacuation plan and shutting down all operating systems throughout the facility to prevent any potential fire or explosion risks. Automatically or manually start the fire extinguishing system, especially in the leak area and its surrounding areas. Ensure that all personnel evacuate to a safe place quickly and orderly, and notify local emergency services for response and support.

[0056] The node "tracking and feedback" is used for follow-up tracking and system feedback of leakage events; analyzing leakage event data, optimizing decision logic; and adjusting system parameters and strategies based on analysis results. Among them, optimization is used to improve the efficiency and accuracy of the system's response to leakage events, including reducing the error rate of leakage detection and shortening the time from leakage detection to response.

[0057] For the node "Safety Assessment", it is used to assess the environmental safety after a medium-sized leak; continue normal operations after confirming that the environment is safe; if the environment is still at risk, consider further safety measures. Specifically, the environmental safety status after a medium-sized leak is assessed through a decision tree algorithm. The current safety status is assessed taking into account the size of the leak, changes in environmental conditions, and implemented safety measures.

[0058] Continue to monitor hydrogen concentration and environmental parameters to ensure that all parameters remain within safe ranges. Regularly reassess based on dynamic data and changes in the operating environment to ensure continued safety.

[0059] The node "Crisis Management" is used to manage the crisis of large-scale leaks; comprehensively analyze the cause of the incident and evaluate the response effect; and make long-term system and policy improvements based on post-event analysis.

[0060] Step S2012: evaluate and analyze the current data information and the corresponding current response decision based on the trained machine learning model.

[0061] Acquire real-time data information, wherein the real-time data information includes real-time hydrogen data and real-time environmental data, and the real-time environmental data includes one or more of real-time temperature data, real-time humidity data, real-time wind speed and direction data, and real-time air pressure data.

[0062] The basic decision tree is used to determine the real-time response decision corresponding to the real-time data information. The response strategy includes ventilation, valve operation, etc.

[0063] Furthermore, a trained machine learning model is obtained. The method for obtaining the trained machine learning model includes:

[0064] Use historical data information and historical response decisions to train machine learning models.

[0065] In one implementation, historical data information and historical response decisions are obtained, the historical data information including historical detection time, historical hydrogen data and historical environmental data; the environmental data includes temperature data, humidity data, wind speed and direction data and air pressure data, etc.

[0066] Among them, the detection time is used to determine whether the leak is increasing or being effectively controlled. For example, continuous or periodic measurements showing that the hydrogen concentration increases over time indicate that the leak may be worsening; conversely, if the concentration decreases, it may mean that the leak is under control. Records of detection time can be used to optimize the timeliness of emergency response. By analyzing the time from detection to the leakage concentration reaching various thresholds, the response speed and efficiency of emergency measures can be evaluated.

[0067] Among them, using detection time can compare current data with historical data to analyze trends and patterns of leakage events. This comparison can help predict the possible impact of the leakage and develop more effective response strategies.

[0068] Among them, combining environmental parameters and detection time can more accurately assess the potential impact of the leak. Environmental changes may affect the diffusion rate and direction of the leaked gas, and the detection time helps to understand the immediate impact of environmental factors.

[0069] Furthermore, the collected data are cleaned and standardized to remove noise and outliers and ensure data quality.

[0070] Feature extraction is performed on historical data information and historical response decisions to extract features that are helpful for prediction and decision-making, and multiple feature information is obtained. The feature information includes the rate of change and acceleration of hydrogen concentration, the relationship between historical response decisions and the rate of change of hydrogen concentration, and the relationship between historical response decisions and historical environmental data.

[0071] The most influential features are selected through statistical analysis and machine learning techniques, and the machine learning model is trained based on the feature information.

[0072] Specifically, select a suitable machine learning model according to the characteristics of the problem, such as support vector machine (SVM).

[0073] Among them, the core parameters of the support vector machine model include penalty parameter (C), kernel function (Kernel) and gamma (γ).

[0074] The penalty parameter is used to control the degree of penalty for misclassification. A larger C value means strict classification and no more misclassification, which is suitable for the strict requirements of leakage detection. A smaller C value can allow more misclassification and improve the generalization ability of the model.

[0075] Kernel functions include linear kernels and RBF kernels (radial basis functions). The linear kernel is used for linearly separable data. The RBF kernel is suitable for nonlinear problems and can handle irregular data patterns. It is a commonly used kernel function in hydrogen leak detection. The choice of kernel function depends on the distribution of the data and the characteristics of the problem.

[0076] Gamma only needs to be set when using the RBF kernel. It is used to control the range of influence of a single training sample. A larger value means a smaller range of influence, which may lead to overfitting of the model. The best gamma value is selected through cross-validation.

[0077] In one implementation, machine learning model parameters are optimized by cross-validation and other techniques to improve the accuracy and generalization of the model. Cross-validation techniques are used to optimize model parameters, such as using grid search to adjust C and gamma simultaneously.

[0078] Specifically, based on the feature information, the performance index of the machine learning model is calculated, and the performance index is used to evaluate the performance of the machine learning model. Based on the performance index, the parameters of the machine learning model are optimized using cross-validation technology.

[0079] Among them, performance indicators include precision, recall, and F1 score, etc. The recall rate can capture all real leakage events.

[0080] After each response implemented by the system, feedback results are collected, such as the effectiveness of measures, possible false alarms, etc. The model is regularly retrained and fine-tuned based on new data and feedback collected during actual operation. Based on the feedback and continuous learning results of the model, the decision logic of the decision tree and the threshold of hydrogen concentration are dynamically adjusted to adapt to environmental changes and new data patterns. The present invention adopts a continuous learning mechanism to automatically absorb new information and optimize the model over time to ensure that the system can cope with future challenges and unknown leakage situations.

[0081] The trained machine learning model is used to evaluate and analyze the current data information and the corresponding current response decision. Specifically, the trained machine learning model analyzes the current data trends and patterns based on the latest collected data and past historical data.

[0082] The present invention uses machine learning technology, especially support vector machines, to learn and predict leakage risks from complex data related to hydrogen leakage. By continuously collecting system response feedback and new monitoring data, the model can continuously update and optimize itself, ensuring that the decision-making system adapts to new leakage patterns and environmental changes over time. It can more accurately identify real leakage events and reduce false alarms caused by environmental noise or equipment failure.

[0083] Step S2013, adjusting the basic decision tree according to the analysis results to obtain a target decision tree.

[0084] The analysis results are used for historical performance evaluation and pattern recognition. Among them, historical performance evaluation characterizes which parameters in basic decision points, decision thresholds, decision logic and response strategies perform well in actual use and which parameters need further improvement. Pattern recognition Through statistical analysis and pattern recognition, the system can discover new patterns or trends that are not covered by existing decision logic, triggering updates to the decision tree structure, such as adding new decision nodes or adjusting the logic of existing nodes.

[0085] In one implementation, each operation and response generates feedback data, such as the correctness of the operation, false positives and false negatives, and the decision tree structure and parameters are optimized by analyzing the actual decision results.

[0086] Based on these analysis results, nodes (decision points) of the decision tree may be added, deleted, or modified to reflect the latest environmental conditions and system status.

[0087] For example, if a certain threshold frequently leads to false positives, the model will automatically propose to adjust the threshold.

[0088] For example, according to the analysis results, one or more of the number of basic decision points, the decision thresholds of the basic decision points, and the decision logic of each basic decision point are adjusted, and the target decision tree is constructed by combining the decision thresholds of multiple adjusted basic decision points, the decision logic of each adjusted basic decision point, and the adjusted response decisions.

[0089] By comparing the system performance before and after the adjustment, such as reducing the false alarm rate and improving the accuracy of leak detection, the effect of the adjustment can be verified, and further decisions can be made as to whether continuous optimization or adjustment strategies are needed.

[0090] The machine learning model of this application regularly accepts new data input for retraining or fine-tuning. This process is automatically performed by the algorithm without human intervention.

[0091] Through intelligent analysis and data-driven decision-making processes, this application can optimize the activation of emergency measures such as ventilation and valve operation, avoiding unnecessary equipment operation and manpower input. It not only reduces energy and maintenance costs, but also improves the operating efficiency of the entire system.

[0092] Step S202: Utilize sensors to detect data and obtain current data information.

[0093] Among them, the current data information includes current hydrogen data and current environmental data, and the current environmental data includes current temperature data, current humidity data, current wind speed and wind direction data, and current air pressure data.

[0094] In one implementation, a hydrogen sensor is used to detect the current hydrogen concentration, a temperature sensor is used to detect the current temperature data, a humidity sensor is used to detect the current humidity data, a wind sensor is used to detect the current wind speed and wind direction data, and an air pressure sensor is used to detect the current air pressure data.

[0095] Step S203, using the target decision tree to make a hydrogen leakage judgment based on the current data information, and outputting the corresponding current response decision.

[0096] At the initial node of the target decision tree, the decision threshold of the current data information is judged, the current data information is compared with the decision threshold, and the decision threshold of the current data information is further judged using the next node that meets the conditions until the final current response decision is output.

[0097] The hydrogen leak response method based on decision tree of the present application can analyze hydrogen concentration data in real time and automatically execute corresponding safety measures according to the preset safety protocol. It does not rely on preset fixed thresholds and procedures, has strong adaptability to complex situations and environmental changes, and can be dynamically adjusted according to real-time data and leakage scenarios. It can also achieve adaptive learning and continuous optimization, and adjust its performance and response strategy according to new data patterns.

[0098] In a specific embodiment, assuming there is a hydrogen storage facility, the hydrogen leakage response method based on the decision tree of the present application is used for detection and analysis. The sensors in the hydrogen storage facility can measure the hydrogen concentration and environmental data, wherein the environmental data includes temperature data and pressure data.

[0099] (1) Set up a basic decision tree.

[0100] The basic decision tree includes two main decision points. At decision point A: If the hydrogen concentration exceeds 1000ppm, conduct further environmental checks and proceed to decision point B. At decision point B: If the temperature exceeds 25°C, it is judged as a high-risk leak and emergency measures are initiated.

[0101] (2) Adjust the basic decision tree to generate the target decision tree.

[0102] Data collection and analysis: The system continuously monitors and records data. According to the analysis results, at low temperatures (such as below 10°C), even if the hydrogen concentration is slightly higher (such as 1100ppm), the possibility of leakage is very low.

[0103] Online model learning and adjustment: Using online learning algorithms to analyze newly collected data, the system recognizes that under low temperature conditions, the existing hydrogen concentration threshold may be too sensitive, resulting in frequent false alarms.

[0104] Automatic adjustment of the decision tree: Add a new decision point C after decision point A. Decision point C: If the temperature is lower than 10°C, adjust the threshold of hydrogen concentration to 1200ppm.

[0105] Feedback loop optimization: After the system implemented the adjustment, it continued to monitor the results of the decision. It was found that the false alarm rate at low temperatures dropped significantly after the adjustment. Further monitoring data showed that in a high-pressure environment, leaks may occur even if the hydrogen concentration is not high.

[0106] Automatic adjustment of decision tree: Add decision point D: When the pressure exceeds 5 bar, lower the threshold for judging hydrogen concentration as leakage to 800 ppm.

[0107] (3) Use the target decision tree for further testing.

[0108] Through the dynamic adjustment of the decision tree, the decision tree not only reflects the actual leakage risk more accurately, but also significantly reduces false alarms and missed alarms, thereby improving the overall efficiency of the system.

[0109] In this embodiment, a hydrogen leak response device based on a decision tree is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0110] This embodiment provides a hydrogen leakage response device based on a decision tree, such as Figure 4 As shown, Figure 4 is a structural block diagram of a hydrogen leakage response device based on a decision tree according to an embodiment of the present invention, the hydrogen leakage response device comprising:

[0111] The analysis module 401 is used to evaluate and analyze the basic decision tree according to historical data information and historical response decisions, and adjust the basic decision tree according to the analysis results to obtain a target decision tree.

[0112] The detection module 402 is used to perform data detection using sensors to obtain current data information, where the current data information includes current hydrogen data and current environmental data.

[0113] The decision module 403 is used to use the target decision tree to make a hydrogen leakage judgment according to the current data information, and output the corresponding current response decision.

[0114] The further functional description of each of the above modules is the same as that of the above corresponding embodiments and will not be repeated here.

[0115] The decision tree-based hydrogen leak response device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0116] The embodiment of the present invention also provides a computer device having the above Figure 4 A hydrogen leak response device based on a decision tree is shown.

[0117] See also Figure 5 , Figure 5 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 5 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0118] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0119] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0120] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0121] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0122] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.

[0123] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device may be a touch screen.

[0124] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0125] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0126] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A hydrogen leak response method based on a decision tree, characterized in that: The method comprises: Determine decision parameters and response decisions, wherein the decision parameters include data information that affects the response decisions, and the response decisions include valve opening and closing decisions and ventilation system decisions; Determine decision thresholds of a plurality of basic decision points and decision logic of each basic decision point based on the decision parameters, wherein the decision thresholds include a hydrogen data threshold and an environmental data threshold; The decision thresholds of multiple basic decision points, the decision logic of each basic decision point and the response decision are combined to construct a basic decision tree, wherein the root node of the basic decision tree is leakage detection, which is used to detect whether hydrogen leakage occurs; Acquire real-time data information, and use the basic decision tree to determine the real-time response decision corresponding to the real-time data information; use historical data information and historical response decisions to train a machine learning model; evaluate and analyze the real-time data information and the corresponding real-time response decision based on the trained machine learning model; the historical data information includes historical detection time, historical hydrogen data, and historical environmental data; The training of the machine learning model using the historical data information and the historical response decision includes: obtaining the historical data information and the historical response decision; performing feature extraction on the historical data information and the historical response decision to obtain a plurality of feature information, wherein the feature information includes the rate of change and acceleration of hydrogen concentration, and obtaining the relationship between the historical response decision and the rate of change of hydrogen concentration, and the relationship between the historical response decision and the historical environmental data; and training the machine learning model according to the feature information; According to the analysis results, the number of the basic decision points, the decision thresholds of the basic decision points, and the decision logic of each basic decision point are adjusted; the decision thresholds of the multiple adjusted basic decision points, the decision logic of each adjusted basic decision point, and the adjusted response decision are combined to construct a target decision tree; Using sensors to detect data and obtain current data information, wherein the current data information includes current hydrogen data and current environmental data; The target decision tree is used to make a hydrogen leakage judgment based on the current data information, and the corresponding current response decision is output.

2. The hydrogen leakage response method based on decision tree according to claim 1, characterized in that: The environmental data includes temperature data, humidity data, wind speed and direction data, and air pressure data.

3. The hydrogen leakage response method based on decision tree according to claim 1, characterized in that: The machine learning model is a support vector machine model, and training the machine learning model according to the feature information includes: Based on the feature information, calculating performance indicators of the machine learning model, the performance indicators including accuracy, recall and F1 score; Based on the performance indicators, the parameters of the machine learning model are optimized using cross-validation techniques.

4. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the decision tree-based hydrogen leakage response method according to any one of claims 1 to 3 by executing the computer instructions.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the hydrogen leakage response method based on a decision tree according to any one of claims 1 to 3.

6. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the hydrogen leakage response method based on a decision tree according to any one of claims 1 to 3.

Images