An AI intelligent gas valve safety control device and control method

Abstract

The application discloses an AI intelligent gas valve safety control device and a control method, and belongs to the technical field of gas safety control. The device comprises a valve body, a sealing execution mechanism, an AI safety control assembly and a multi-sensing detection unit. A ball valve core is arranged in the valve body. The sealing execution mechanism drives the valve core to open and close through an electric telescopic rod and a transmission structure. The multi-sensing detection unit integrates gas concentration detection, pipeline gas pressure detection and an active sampling structure. The AI safety control assembly is internally provided with an AI calculation module, a communication module and an alarm unit. The application realizes intelligent identification and hierarchical prediction of gas leakage risks through an AI algorithm, reduces the false alarm rate by combining multi-sensing data fusion, executes hierarchical safety control strategies according to different risk levels, automatically closes the valve to cut off the gas source in a high-risk scenario, simultaneously realizes local alarm and remote linkage control, greatly improves the safety protection capability of the gas valve, and effectively reduces the risk of safety accidents caused by gas leakage.

Description

Technical Field

[0001] This invention relates to the field of gas safety control technology, and more specifically, to an AI intelligent gas valve safety control device and control method. Background Technology

[0002] With the widespread use of urban gas, safety accidents such as explosions and poisonings caused by gas leaks are frequent. As a core control component of gas pipelines, the safety performance of gas valves directly affects the life and property safety of users. Most existing gas valves are manually operated, possessing only basic on / off control functions. Some gas valves with automatic shut-off functions can only detect leaks using a single gas sensor and then shut off, exhibiting the following technical shortcomings: Firstly, single-sensor detection is susceptible to environmental interference, resulting in a high false alarm rate. Frequent accidental valve closures can affect normal user operation. Secondly, leak detection is delayed, only triggering actions after a leak occurs and the concentration reaches a threshold, making it impossible to predict risks and prevent them in advance. Third, the control logic is simplistic and cannot implement differentiated control strategies for different leakage scenarios and risk levels, resulting in poor applicability. Fourth, the lack of remote linkage capability means that users cannot monitor valve status and leakage risks in real time, and cannot respond promptly when an accident occurs.

[0003] Therefore, developing a gas valve safety control device and method based on AI intelligent recognition, with leakage risk prediction, hierarchical control, and low false alarm rate is of great practical significance. Summary of the Invention

[0004] To address the shortcomings of existing gas valve technologies, such as insufficient leakage prevention capabilities, high false alarm rates, simplistic control logic, and inability to predict risks, this invention provides an AI-powered intelligent gas valve safety control device and method. This invention achieves accurate identification and tiered prediction of gas leakage risks through multi-sensor data fusion and AI intelligent algorithms. Combined with a tiered control strategy, it enables proactive prevention and emergency response, while simultaneously reducing false alarm rates and enhancing overall safety protection throughout the gas usage process.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an AI intelligent gas valve safety control device, comprising a valve body, a sealing actuator, an AI safety control component, and a multi-sensor detection unit; both ends of the valve body are fixed with docking rings, and the end faces of the docking rings are embedded with mounting sealing rings; a spherical valve core is rotatably disposed inside the valve body, and sealing pipes are connected to both ends of the spherical valve core inside the valve body; the sealing actuator includes a support plate fixed to the top of the valve body, a protective shell fixed above the support plate, a control component disposed inside the protective shell, the control component including a transmission sleeve coaxially fixed with the valve stem of the spherical valve core, a rotating arm fixed to the outer wall of the transmission sleeve, a circular through hole opened on the surface of the rotating arm, and a hinged component to the inner wall of the protective shell. The rotating electric telescopic rod has its output end hinged to a circular through hole. The multi-sensor detection unit includes a gas sensor for detecting gas concentration, a gas pressure and flow sensor for detecting pipeline gas pressure, and a fan for collecting ambient airflow. The gas pressure and flow sensor is installed inside a sealed tube. Both the gas sensor and the fan are fixed to the outer wall of the protective housing, and the fan's airflow direction is towards the sensing end of the gas sensor. The AI ​​safety control component includes a circuit board fixed inside the protective housing. The circuit board integrates an AI computing module, a communication module, a battery, and a buzzer. The gas sensor, gas pressure and flow sensor, rotating electric telescopic rod, fan, buzzer, and communication module are all electrically connected to the AI ​​computing module. The battery powers all electrical components.

[0006] As a further improvement to this technical solution, the AI ​​computing module has a built-in gas leakage risk classification and prediction model. The gas leakage risk classification and prediction model takes real-time gas concentration data collected by gas sensors and real-time pipeline gas pressure fluctuation data collected by gas pressure and flow sensors as inputs, and outputs leakage risk level, leakage type determination results and corresponding control commands.

[0007] As a further improvement to this technical solution, the AI ​​computing module also has a built-in abnormal data filtering and false alarm correction unit. The false alarm correction unit is used to dynamically calibrate the gas concentration detection data by combining ambient temperature, airflow disturbance data, and concentration change rate, and to eliminate abnormal interference data in non-leakage scenarios.

[0008] As a further improvement to this technical solution, the fan adopts an intermittent start-stop working mode. The AI ​​computing module drives the fan to start according to a preset cycle, pushing ambient air to the sensing end of the gas sensor. When the AI ​​computing module detects that the gas concentration exceeds a preset warning threshold, it drives the fan to work continuously.

[0009] As a further improvement to this technical solution, the inner wall of the sealing tube is provided with a sealing lip that fits against the outer wall of the ball valve core. The two sets of sealing tubes are symmetrically distributed at the air inlet and air outlet of the ball valve core. The detection end of the air pressure and flow sensor extends into the sealing cavity between the sealing tube and the ball valve core.

[0010] As a further improvement to this technical solution, the communication module includes a Bluetooth communication unit and an IoT wireless communication unit. The AI ​​computing module communicates bidirectionally with the user's mobile terminal and the community gas safety management platform through the communication module to upload leak alarm information, equipment operating status, and receive remote control commands.

[0011] An AI-powered intelligent gas valve safety control method, applied to the aforementioned AI-powered intelligent gas valve safety control device, includes the following steps: S1 Real-time data acquisition and preprocessing: Ambient gas concentration data is acquired through a gas sensor, and real-time gas pressure data inside the gas pipeline is acquired through a gas pressure and flow sensor. The AI ​​computing module performs filtering, noise reduction, and standardization preprocessing on the acquired raw data; S2 AI-powered intelligent leakage risk identification: The preprocessed concentration and pressure data are input into a pre-trained gas leakage risk classification and prediction model. The model outputs the current leakage risk level and leakage type; S3 Graded safety control execution: The AI ​​computing module executes the corresponding graded control strategy based on the output leakage risk level; S4 Alarm and remote linkage: When the AI ​​computing module identifies a leakage risk, it drives a buzzer to issue a local audible and visual alarm, and simultaneously pushes alarm information and equipment status to the user terminal and management platform through the communication module; S5 Data self-learning and model iteration: The AI ​​computing module records the entire process data of each alarm and control action, and performs incremental training and iterative optimization of the gas leakage risk classification and prediction model by combining user feedback and on-site verification results.

[0012] As a further improvement to this technical solution, in step S2, the leakage risk level includes four levels: no risk, low risk warning, medium risk alarm, and high risk emergency shutdown. The leakage types include internal valve body seal leakage, pipeline joint leakage, pipeline rupture leakage, and leakage due to user misoperation and failure to close the valve.

[0013] As a further improvement to this technical solution, in step S3, the graded control strategy is specifically as follows: when the risk level is determined to be no risk, the valve is maintained in normal working condition, and data acquisition and equipment self-test are performed according to a preset cycle; when the risk level is determined to be low, the fan is driven to work continuously, the sampling frequency is increased, the buzzer is activated for intermittent warning, and a warning prompt is pushed to the user terminal at the same time; when the risk level is determined to be medium, the concentration and pressure data are continuously collected and verified, the buzzer is activated for continuous audible and visual alarm, alarm information is pushed to the user and management platform, and the user's remote control command is awaited; when the risk level is determined to be high, the electric telescopic rod is immediately driven to rotate, and the ball valve core is driven to rotate to a fully closed state through the rotating arm and transmission sleeve, cutting off the gas passage, and audible and visual alarm and remote information push are performed at the same time.

[0014] As a further improvement to this technical solution, step S1 also includes a device self-test step. The AI ​​calculation module performs self-tests on the working status of the rotating electric telescopic rod, gas sensor, gas pressure and flow sensor, buzzer, and communication module according to a preset cycle. If a component failure is detected, a fault alarm message is immediately pushed to the user and management platform.

[0015] Beneficial effects 1. This invention uses a gas leak risk classification and prediction model built into the AI ​​calculation module, combined with dual-dimensional data of gas concentration and pipeline gas pressure, to achieve accurate identification and advance prediction of leak risks. It can not only respond quickly after a leak occurs, but also predict potential leak risks through gas pressure fluctuations and concentration change trends, thus upgrading from "post-event handling" to "pre-event prevention".

[0016] 2. This invention sets up a multi-sensor detection unit, which, together with the active sampling structure of the fan, improves the sensitivity and response speed of gas concentration detection. At the same time, through the false alarm correction unit of the AI ​​calculation module, the detection data is dynamically calibrated in combination with environmental parameters, effectively eliminating false alarms caused by environmental interference, greatly reducing the false alarm rate and the probability of accidental valve closure, and taking into account both safety and user experience.

[0017] 3. This invention designs a graded safety control strategy, which performs differentiated control actions for four levels: no risk, low risk, medium risk, and high risk. It takes into account early warning prompts, user intervention, and emergency shutdown, adapts to different leakage scenarios, avoids one-size-fits-all valve shut-off operations, and improves the applicability of the device.

[0018] 4. This invention integrates a communication module, enabling two-way communication with user terminals and the gas management platform. Users can remotely view valve status, receive alarm information, and perform remote control. At the same time, the management platform can achieve centralized management and control of batch devices, improving the grid-based management capability of gas safety.

[0019] 5. The AI ​​model of this invention has incremental self-learning capability, and can be continuously iterated and optimized based on operating data, user feedback and on-site verification results, so as to continuously improve the accuracy of leakage identification and the precision of risk prediction, and adapt to different usage environments and working conditions. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall structure assembly of the present invention; Figure 2 This is a cross-sectional view of the valve body and sealing actuator of the present invention. Figure 3 This is a schematic diagram of the structural layout of the control component of the present invention; Figure 4 This is a schematic diagram of the installation structure of the multi-sensor detection unit of the present invention; Figure 5 This is an electrical connection framework diagram of the AI ​​safety control component of the present invention; Figure 6 This is a cross-sectional view of the present invention; Figure 7 This is a flowchart of the present invention.

[0021] The labels in the diagram represent the following: 1. Sealing ring installation; 2. Connecting ring; 3. Valve body; 4. Support plate; 5. Protective housing; 6. Control components; 601. Rotating electric telescopic rod; 602. Rotating arm; 603. Circular through hole; 604. Transmission sleeve; 606. Gas sensor; 607. Circuit board; 608. AI computing module; 609. Communication module; 610. Battery; 611. Buzzer; 7. Sealing tube; 8. Gas pressure and flow sensor; 9. Ball valve core. Detailed Implementation

[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] This implementation 1 This embodiment provides an AI-powered intelligent gas valve safety control device, such as... Figures 1-6 As shown, it includes valve body 3, sealing actuator, AI safety control components and multi-sensor detection unit.

[0024] Both ends of the valve body 3 are integrally formed with mating rings 2, which are used for connection and installation with gas pipelines. The end face of the mating ring 2 has an annular groove, and a sealing ring 1 is embedded in the annular groove. The sealing ring 1 is used to ensure the sealing of the pipeline connection and prevent joint leakage. Inside the valve body 3, a ball valve core 9 is rotatably installed via a valve stem. The ball valve core 9 is used to control the opening and closing of the gas passage inside the valve body 3. Sealing pipes 7 are fixedly connected to both ends of the ball valve core 9 inside the valve body 3. The inner wall of the sealing pipe 7 is provided with a sealing lip that fits against the outer wall of the ball valve core 9. The two sets of sealing pipes 7 are symmetrically distributed at the air inlet and air outlet ends of the ball valve core 9 to ensure the sealing performance of the ball valve core 9 in the closed and open states and prevent internal sealing leakage of the valve body.

[0025] The sealing actuator includes a support plate 4 fixed to the top of the valve body 3. A protective housing 5 is bolted to the top of the support plate 4. The protective housing 5 is waterproof and dustproof, used to protect the internal electrical control components. Inside the protective housing 5 is a control component 6, which includes a transmission sleeve 604 coaxially fixed to the valve stem of the ball valve core 9. A rotating arm 602 is welded to the outer wall of the transmission sleeve 604. A circular through hole 603 is formed on the surface of the rotating arm 602. A rotating electric telescopic rod 601 is hinged to the inner wall of the protective housing 5. The output end of the rotating electric telescopic rod 601 is hinged to the circular through hole 603. When the rotating electric telescopic rod 601 extends or retracts, it drives the rotating arm 602 to rotate around the axis of the transmission sleeve 604, thereby driving the valve stem and the ball valve core 9 to rotate through the transmission sleeve 604, achieving automatic opening and closing control of the valve.

[0026] The AI ​​safety control component includes a circuit board 607 fixed inside the protective housing 5. The circuit board 607 integrates an AI computing module 608, a communication module 609, a battery 610, and a buzzer 611. The gas sensor 606, the gas pressure and flow sensor 8, the rotating electric telescopic rod 601, the buzzer 611, and the communication module 609 are all electrically connected to the AI ​​computing module 608. The battery 610 supplies power to all electrical components. The battery 610 can be powered by lithium batteries or alkaline batteries, and also supports external mains power supply to ensure emergency operation capabilities in the event of a power outage.

[0027] The AI ​​computing module 608 incorporates a gas leak risk classification and prediction model. This model is pre-trained based on a large amount of sample data from gas leak scenarios. It takes real-time gas concentration data collected by the gas sensor 606 and real-time pipeline gas pressure fluctuation data collected by the gas pressure and flow sensor 8 as input, and outputs the leak risk level, leak type determination result, and corresponding control commands. Simultaneously, the AI ​​computing module 608 also includes an abnormal data filtering and false alarm correction unit. This unit dynamically calibrates the gas concentration detection data by combining ambient temperature, airflow disturbance data, and concentration change rate, eliminating abnormal interference data from non-leak scenarios, such as concentration anomalies caused by kitchen fumes or temporary gas residue, effectively reducing the false alarm rate.

[0028] The communication module 609 includes a Bluetooth communication unit and an IoT wireless communication unit. The AI ​​computing module 608 communicates bidirectionally with the user's mobile terminal and the community gas safety management platform through the communication module 609. It is used to upload leakage alarm information and equipment operating status, and to receive remote control commands. Users can remotely view valve status, receive alarm information, and remotely control the opening and closing of valves through a mobile APP.

[0029] The working principle of this embodiment is as follows: The device is installed at the inlet of the user's gas pipeline. Under normal use, the AI ​​calculation module 608 collects the ambient gas concentration through the gas sensor 606 and the pipeline gas pressure data through the gas pressure and flow sensor 8 at preset intervals. At the same time, it drives the fan 605 to start and stop intermittently to improve detection sensitivity. When an abnormal gas concentration or pipeline gas pressure fluctuation is detected, the AI ​​calculation module 608 inputs the data into the gas leak risk classification and prediction model to determine the leak risk level and leak type, and executes the corresponding control strategy according to the risk level. When a high-risk leak is determined, the electric telescopic rod 601 is immediately driven to rotate, causing the ball valve core 9 to close, cutting off the gas passage. At the same time, the buzzer 611 is activated for audible and visual alarm, and alarm information is pushed to the user and management platform through the communication module 609, realizing intelligent prevention and emergency handling of gas leaks.

[0030] Example 2 This embodiment provides an AI-powered intelligent gas valve safety control method, applied to the AI-powered intelligent gas valve safety control device described in Embodiment 1, such as... Figure 7 As shown, it includes the following steps: S1 Real-time Data Acquisition and Preprocessing: Ambient gas concentration data is collected through gas sensor 606, and real-time gas pressure data inside the gas pipeline is collected through gas pressure and flow sensor 8. AI calculation module 608 performs filtering, noise reduction, and standardization preprocessing on the collected raw data, eliminating invalid abnormal jump data. At the same time, AI calculation module 608 performs self-checks on the working status of rotating electric telescopic rod 601, gas sensor 606, gas pressure and flow sensor 8, buzzer 611, and communication module 609 according to a preset cycle. If a component failure is detected, a fault alarm information is immediately pushed to the user and management platform to ensure stable operation of the device.

[0031] S2AI Leakage Risk Intelligent Identification: The pre-processed concentration data and gas pressure data are input into a pre-trained gas leakage risk classification and prediction model. The model outputs the current leakage risk level and leakage type. The leakage risk level includes four levels: no risk, low risk warning, medium risk alarm, and high risk emergency shutdown. The leakage type includes leakage of valve body internal seal, leakage of pipeline joint, leakage of pipeline damage, and leakage of leakage due to user misoperation of valve not being closed, so as to achieve accurate classification of leakage scenarios.

[0032] S3 graded safety control execution: The AI ​​calculation module 608 executes the corresponding graded control strategy based on the output leakage risk level. Specifically: when the risk level is determined to be no risk, the valve maintains normal operation and performs data acquisition and equipment self-check according to the preset cycle; when the risk level is determined to be low risk warning level, the sampling frequency is increased, the buzzer 611 is activated for intermittent warning, and a warning prompt is pushed to the user terminal to remind the user to check the gas usage; when the risk level is determined to be medium risk alarm level, the concentration and gas pressure data are continuously collected and verified, the buzzer 611 is activated for continuous audible and visual alarm, alarm information is pushed to the user and management platform, and the user is given time for on-site investigation and intervention by waiting for remote control commands; when the risk level is determined to be high risk emergency shutdown level, the electric telescopic rod 601 is immediately driven to rotate, and the ball valve core 9 is rotated to the fully closed state through the rotating arm 602 and the transmission sleeve 604, cutting off the gas passage and preventing the leakage from expanding from the source, while simultaneously executing audible and visual alarm and remote information push.

[0033] S4 Alarm and Remote Linkage: When the AI ​​computing module 608 detects a leakage risk, it drives the buzzer 611 to issue a local audible and visual alarm, reminding on-site personnel to take timely action. At the same time, it pushes alarm information to the user terminal and management platform through the communication module 609. The alarm information includes the leakage risk level, leakage type, equipment location, valve status, etc. Users can remotely view the on-site situation through the terminal and send remote control commands. The management platform can simultaneously dispatch emergency response resources.

[0034] S5 Data Self-Learning and Model Iteration: The AI ​​computing module 608 records the entire process data of each alarm and control action. Combined with user feedback and on-site verification results, it incrementally trains and iteratively optimizes the gas leak risk classification prediction model, continuously improving the accuracy of model recognition, adapting to different usage environments and operating conditions, and reducing false alarm rate and missed alarm rate.

[0035] This embodiment fully leverages the hardware's AI-powered leak prevention capabilities, with its core advantages revolving around intelligent and precise leak prevention logic: 1. The pre-trained gas leak risk classification and prediction model, with concentration and gas pressure as input, can not only identify leaks that have already occurred, but also predict potential leak risks through gas pressure fluctuation trends and concentration change rates. This upgrades gas safety protection from "post-leakage handling" to "pre-leakage prediction", fundamentally reducing the probability of leak accidents.

[0036] 2. Differentiated control logic is set for four levels: no risk, low risk warning, medium risk alarm, and high risk emergency shutdown. Low risk only provides a warning, medium risk requires user intervention, and high risk requires immediate forced valve shutdown. This completely solves the problem of the traditional gas valve's "one-size-fits-all" approach of shutting off the valve as soon as the risk exceeds the limit. Under the premise of ensuring safety, it greatly reduces the impact of accidental valve shutdown on users' normal lives.

[0037] 3. Set up incremental training and optimization steps for the model, automatically record the full process data of each alarm and control action, and continuously optimize the model by combining user feedback and on-site verification results. The longer the model is used, the stronger its adaptability to users' gas usage habits and installation environment becomes, the higher the accuracy of leak identification becomes, and the lower the false alarm rate and missed alarm rate become.

[0038] 4. During the data acquisition phase, a self-test procedure is set up simultaneously to monitor the working status of sensors, actuators, and communication units in real time, identify equipment faults in advance and issue alarms to avoid protection gaps caused by equipment failure; at the same time, through multi-terminal linkage of local audible and visual alarms, user terminals, and management platforms, a full-link response of on-site reminders, remote notifications, and emergency dispatch is achieved, which greatly improves the efficiency of handling leakage accidents.

[0039] Example 3 This embodiment provides a flow gas safety detection method for the ZXMFSV-6 AI gas safety valve. This method is applied to the AI ​​intelligent gas valve safety control device described in Embodiment 1, and is also compatible with the AI ​​intelligent gas valve safety control method described in Embodiment 2. It achieves accurate detection, dynamic calibration, graded early warning, and closed-loop safety management for all-dimensional abnormal gas flow scenarios. The core functionality relies on the device's built-in MEMS flow sensor, gas pressure and flow sensor 8, temperature sensor, and AI calculation module 608. The specific steps are as follows: After the device is powered on, the AI ​​computing module 608 first performs a full-link initialization of the flow detection system: completing power-on self-tests and zero-point calibrations of the MEMS flow sensor, air pressure flow sensor 8, and temperature sensor, verifying sensor sampling accuracy and communication link stability. If a sensor fault is detected, it immediately pushes fault alarm information to the user terminal and management platform, and simultaneously locks the manual / automatic valve opening and closing permissions to avoid safety risks caused by detection failure. After the self-test is completed, the AI ​​computing module 608 automatically loads the device's preset flow control benchmark parameters, including: default upper limit flow rate of 0.72 m³ / h, default lower limit flow rate of 0. The flow rate is 0.048 m³ / h, the default constant flow time is 60 minutes, and the default valve closing grace time is 20 minutes. Users can also customize parameters within the adjustable range (upper limit flow 0.3~6.0 m³ / h, lower limit flow 0.0~0.12 m³ / h, constant flow time 30~300 minutes, valve closing grace time 10~60 minutes) via device buttons or a remote terminal. Combined with the pipeline reference pressure and ambient reference temperature acquired at the initialization moment, a flow detection temperature and pressure compensation benchmark model is established to adapt to the device's operating environment range of -10℃ to 50℃, eliminating interference from ambient temperature and humidity and pipeline base pressure on flow detection accuracy.

[0040] The AI ​​computing module 608 collects instantaneous and cumulative flow data in the gas pipeline in real time through a MEMS flow sensor at a preset sampling frequency of 10Hz. Simultaneously, it collects real-time pressure data of the pipeline through a gas pressure flow sensor 8 and real-time ambient temperature data through a temperature sensor, forming a multi-dimensional correlation dataset of flow, pressure, and temperature. For the collected raw data, the AI ​​computing module 608 performs filtering, noise reduction, and standardization preprocessing: it removes abnormal jump data and invalid data caused by environmental airflow disturbances during the sensor sampling process. Through the temperature and pressure compensation benchmark model established in step S301, it dynamically calibrates the instantaneous flow data to restore the real gas flow value under standard operating conditions, ensuring the accuracy and stability of the flow detection data. The preprocessed data is synchronously transmitted to the built-in gas flow anomaly identification model and a 128x64 OLED Chinese LCD screen.

[0041] The pre-processed flow, pressure, and temperature data are input into the pre-trained gas flow anomaly identification model. The model combines preset parameters with multi-dimensional data features to accurately identify the type of flow anomaly and simultaneously match the corresponding leakage risk level in Example 2. The specific identification rules and classification logic are as follows: Overcurrent anomaly identification: When the calibrated instantaneous flow rate exceeds the user-set upper limit flow rate for 3 seconds, it is determined to be an overcurrent anomaly and matched with the medium risk alarm level; Minor Leakage Anomaly Identification: When the calibrated instantaneous flow rate is consistently above the lower limit flow rate and below the user's normal minimum gas consumption flow rate, and the pipeline pressure continuously decreases by 1.3~1.8 kPa within 5 minutes, it is determined to be a minor leakage anomaly and matched with a low-risk warning level; Gas usage overtime anomaly identification: When the valve remains open and the gas flow rate remains stable within the normal gas usage range for a continuous period exceeding the constant flow time set by the user, it is determined to be an abnormal gas usage overtime and is matched with a medium-risk alarm level; Pipe rupture anomaly identification: When the calibrated instantaneous flow rate spikes to twice or more than the set upper limit flow rate, and the pipeline pressure suddenly drops, it is judged as a pipe rupture anomaly and matched with a high-risk emergency shutdown level; Normal operating condition determination: When the flow data is within the normal range set by the user and there are no abnormal changes in pressure or temperature, it is determined to be a risk-free normal operating condition.

[0042] The AI ​​computing module 608 outputs the anomaly type and risk level, and executes the corresponding hierarchical closed-loop control strategy, which is linked with the hierarchical control strategy in Example 2. The specific execution logic is as follows: Risk-free normal operating conditions: Maintain the valve in a normal open state, perform data acquisition and equipment self-test according to the preset cycle, and update flow, pressure, temperature and valve status data in real time on the OLED screen simultaneously; Low-risk warning (minor leak anomaly): The buzzer 611 is activated for intermittent local warning, and at the same time, a minor leak warning prompt is pushed to the user terminal through the 4G communication module 609 to remind the user to check the pipeline sealing and not to perform forced valve closure. Medium-risk alarm (overflow abnormality / excessive gas usage abnormality): The buzzer 611 is activated to continuously sound and light alarm, pushes the corresponding alarm information to the user terminal and management platform, and starts the countdown of the valve closing grace period set by the user. During the grace period, the user can manually clear the alarm and adjust the parameters through the device button or remote terminal. If the abnormality is not eliminated after the grace period ends, the electric telescopic rod 601 is immediately driven to rotate, which drives the ball valve core 9 to completely close and cut off the gas passage. High-risk emergency shutdown (pipe burst anomaly): Without waiting for user intervention, immediately drive the rotating electric telescopic rod 601 to perform emergency valve closing action, cut off the gas passage within 1 second, and simultaneously activate continuous audible and visual alarms, push pipe burst emergency alarm information and equipment location to user terminals and community gas safety management platforms, and trigger emergency response procedures in a coordinated manner. The AI ​​computing module 608 records all flow data, user gas usage habits, alarm events, and user feedback during the entire operation of the device, establishing a user-specific gas usage behavior database. Based on the accumulated operational data, the AI ​​computing module 608 performs incremental training and iterative optimization of the gas flow anomaly identification model. It automatically optimizes the flow anomaly judgment threshold based on the user's daily gas flow peaks and usage duration patterns, eliminating interference from normal gas usage fluctuations in non-abnormal scenarios, such as instantaneous flow peaks when starting high-power gas appliances or continuous gas usage during long cooking sessions, further reducing the false alarm rate. Simultaneously, based on user gas usage habits, it automatically generates optimized recommended values ​​for parameters such as flow limit and constant flow time. Users can confirm and modify these values ​​with a single click via local buttons or a remote terminal, balancing safety protection with user convenience in daily use. The AI ​​computing module 608 locally encrypts and stores the entire process data of each flow anomaly detection, alarm trigger, and valve closure action. The stored content includes the time of anomaly occurrence, anomaly type, real-time data of flow, pressure, and temperature, records of control actions, and user handling results, with a storage period of no less than 12 months. At the same time, the communication module 609 synchronously uploads the flow operation data and anomaly event records to the user's mobile terminal and the community gas safety management platform, allowing users to access historical gas usage data and anomaly event records at any time. It also facilitates the gas management department to achieve flow safety monitoring, fault diagnosis, and emergency management of batch equipment, forming a closed loop of full-process flow safety management of "detection, identification, control, traceability, and optimization".

[0043] Those skilled in the art should understand that this invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples of the invention and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An AI-powered intelligent gas valve safety control device, characterized in that, It includes a valve body (3), a sealing actuator, an AI safety control component, and a multi-sensor detection unit; both ends of the valve body (3) are fixed with docking rings (2), and the end face of the docking rings (2) is embedded with a sealing ring (1); a ball valve core (9) is rotatably arranged inside the valve body (3); and both ends of the ball valve core (9) are connected to a sealing tube (7). The sealing actuator includes a support plate (4) fixed to the top of the valve body (3), a protective housing (5) fixed above the support plate (4), a control component (6) is provided inside the protective housing (5), the control component (6) includes a transmission sleeve (604) coaxially fixed with the valve stem of the ball valve core (9), a rotating arm (602) is fixed on the outer wall of the transmission sleeve (604), a circular through hole (603) is opened on the surface of the rotating arm (602), and a rotating electric telescopic rod (601) is hinged to the inner wall of the protective housing (5), and the output end of the rotating electric telescopic rod (601) is hinged to the circular through hole (603); The multi-sensor detection unit includes a gas sensor (606) for detecting the concentration of the gas and a gas pressure and flow sensor (8) for detecting the gas pressure in the pipeline. The gas pressure and flow sensor (8) is installed inside the sealed tube (7). The AI ​​safety control component includes a circuit board (607) fixed inside the protective housing (5). The circuit board (607) integrates an AI computing module (608), a communication module (609), a battery (610), and a buzzer (611). The gas sensor (606), the gas pressure and flow sensor (8), the rotating electric telescopic rod (601), the fan (605), the buzzer (611), and the communication module (609) are all electrically connected to the AI ​​computing module (608). The battery (610) supplies power to each electrical component.

2. The AI ​​intelligent gas valve safety control device according to claim 1, characterized in that, The AI ​​calculation module (608) has a built-in gas leakage risk classification prediction model. The gas leakage risk classification prediction model takes the gas concentration data collected in real time by the gas sensor (606) and the pipeline gas pressure fluctuation data collected in real time by the gas pressure and flow sensor (8) as input, and outputs the leakage risk level, leakage type judgment result and corresponding control command.

3. The AI ​​intelligent gas valve safety control device according to claim 2, characterized in that, The AI ​​computing module (608) also has a built-in abnormal data filtering and false alarm correction unit. The false alarm correction unit is used to dynamically calibrate the gas concentration detection data by combining ambient temperature, airflow disturbance data and concentration change rate, and eliminate abnormal interference data in non-leakage scenarios.

4. The AI ​​intelligent gas valve safety control device according to claim 1, characterized in that, Ambient air is pushed to the sensing end of the gas sensor (606), and the AI ​​calculation module (608) detects that the gas concentration exceeds the preset warning threshold.

5. The AI ​​intelligent gas valve safety control device according to claim 1, characterized in that, The inner wall of the sealing tube (7) is provided with a sealing lip that fits against the outer wall of the ball valve core (9). The two sets of sealing tubes (7) are symmetrically distributed at the air inlet and air outlet of the ball valve core (9). The detection end of the air pressure flow sensor (8) extends into the sealing cavity between the sealing tube (7) and the ball valve core (9).

6. The AI ​​intelligent gas valve safety control device according to claim 1, characterized in that, The communication module (609) includes a Bluetooth communication unit and an IoT wireless communication unit. The AI ​​computing module (608) communicates bidirectionally with the user's mobile terminal and the community gas safety management platform through the communication module (609) to upload leakage alarm information, equipment operating status, and receive remote control commands.

7. A safety control method for an AI intelligent gas valve, applied to the AI ​​intelligent gas valve safety control device according to any one of claims 1-6, characterized in that, Includes the following steps: S1 Real-time data acquisition and preprocessing: Ambient gas concentration data is acquired through gas sensor (606), and real-time gas pressure data inside the gas pipeline is acquired through gas pressure and flow sensor (8). AI calculation module (608) performs filtering, noise reduction and standardization preprocessing on the acquired raw data. S2AI Leakage Risk Intelligent Identification: Input the pre-processed concentration data and gas pressure data into the pre-trained gas leakage risk classification and prediction model, and the model outputs the current leakage risk level and leakage type; S3 graded safety control execution: The AI ​​calculation module (608) executes the corresponding graded control strategy based on the output leakage risk level; S4 Alarm and Remote Linkage: When the AI ​​computing module (608) detects a leakage risk, it drives the buzzer (611) to issue a local audible and visual alarm, and at the same time pushes alarm information and equipment status to the user terminal and management platform through the communication module (609). S5 Data Self-Learning and Model Iteration: The AI ​​computing module (608) records the full-process data of each alarm and control action, and combines user feedback and on-site verification results to incrementally train and iteratively optimize the gas leak risk classification prediction model.

8. The AI ​​intelligent gas valve safety control method according to claim 7, characterized in that, In step S2, the leakage risk level includes four levels: no risk, low risk warning, medium risk alarm, and high risk emergency shutdown. The leakage types include internal valve body seal leakage, pipeline joint leakage, pipeline damage leakage, and leakage caused by user misoperation and failure to close the valve.

9. The AI ​​intelligent gas valve safety control method according to claim 8, characterized in that, In step S3, the graded control strategy is as follows: when the risk level is determined to be no risk, the valve is kept in normal working condition and data acquisition and equipment self-test are performed according to the preset cycle; when the risk level is determined to be low risk warning level, the sampling frequency is increased, the buzzer (611) is activated for intermittent warning, and a warning prompt is pushed to the user terminal at the same time; when the risk level is determined to be medium risk alarm level, the concentration and pressure data are continuously collected and verified, the buzzer (611) is activated for continuous audible and visual alarm, the alarm information is pushed to the user and management platform, and the user is waiting for remote control instructions; when the risk level is determined to be high risk emergency shutdown level, the electric telescopic rod (601) is immediately driven to rotate, and the ball valve core (9) is driven to rotate to the fully closed state through the rotating arm (602) and the transmission sleeve (604), cutting off the gas passage, and simultaneously performing audible and visual alarm and remote information push.

10. The AI ​​intelligent gas valve safety control method according to claim 7, characterized in that, Step S1 also includes a device self-test step. The AI ​​calculation module (608) performs a self-test on the working status of the rotating electric telescopic rod (601), gas sensor (606), gas pressure and flow sensor (8), buzzer (611), and communication module (609) according to a preset cycle. If a component failure is detected, a fault alarm information is immediately pushed to the user and management platform.

Images