Intelligent residual volume management and risk early warning system for anesthetic gas evaporation tank and method thereof

Abstract

The application discloses an intelligent residual amount management and risk early warning system and method for anesthetic gas volatilization tanks. The system comprises a multi-modal fusion sensing module for collecting weight, liquid level, flow, temperature and attitude data; an intelligent data processing and early warning unit for analyzing data, identifying abnormal patterns, calculating residual gas amount and maintenance time through a dynamic risk prediction engine; an interactive early warning and display terminal for displaying information and performing hierarchical early warning through a double-ring interface; and a cloud-edge collaborative communication module for realizing remote data reporting and management. The method comprises multi-source data collection and fusion, residual amount calculation and cross-validation, risk dynamic prediction and identification, hierarchical early warning interaction and cloud-edge collaborative data sharing. The application realizes the transformation from passive monitoring to active early warning, and improves the accuracy and safety of anesthetic management.

Description

Technical Field

[0001] This invention relates to the field of medical equipment technology, specifically to an intelligent residual management and risk warning system and method for anesthetic gas vaporizers. Background Technology

[0002] Anesthetic gas vaporizers are crucial equipment in clinical anesthesia, used to precisely control the concentration of anesthetic gases released, ensuring the safety and effectiveness of patient inhalation. During surgery, anesthesiologists need to monitor the amount of remaining anesthetic in the vaporizer and its estimated duration of use in real time to avoid intraoperative risks due to anesthetic depletion. However, traditional vaporizer monitoring relies mainly on manual observation or simple mechanical indicators, which suffer from low accuracy, slow response, and inability to dynamically predict remaining time, making it difficult to meet the demands of modern precision anesthesia.

[0003] Currently, common methods for monitoring the gas volume in evaporation tanks mainly include single-modal detection methods such as float-type level gauges, weight measurement, or flow accumulation. Float-type level gauges are easily affected by the type of anesthetic, temperature, and tank tilt, resulting in significant measurement errors. While weight measurement can directly reflect the mass of the anesthetic, it is susceptible to interference from factors such as vibration and temperature drift, leading to insufficient long-term stability. Methods relying solely on flow accumulation to estimate remaining gas volume cannot cope with flow fluctuations or sudden leaks. Furthermore, existing systems typically lack multi-sensor data fusion and cross-validation mechanisms, which can lead to incorrect gas volume estimations in the event of a single sensor failure, posing potential safety hazards.

[0004] On the other hand, traditional vaporizers have relatively simple human-machine interaction functions, typically only providing basic liquid level markings or low liquid level alarms, and cannot intuitively display the remaining time or dynamically adjust prediction models. Furthermore, they lack the ability to interact with hospital information systems (HIS) in real time, which is detrimental to centralized monitoring and resource allocation by the anesthesiology department.

[0005] Therefore, there is an urgent need for a high-precision, multi-modal fusion intelligent residual management and risk warning system for anesthetic gas vaporizers with intelligent risk warning capabilities. This system should be able to calculate and cross-verify the remaining gas volume in real time through multi-dimensional sensor data, dynamically predict the sustainable time, and combine real-time flow, pressure, and environmental data to achieve risk warning, intelligent intervention, and remote collaborative management, so as to comprehensively improve the safety, reliability, and management efficiency of the anesthesia process. Summary of the Invention

[0006] To address the aforementioned shortcomings in existing technologies, the present invention aims to provide an intelligent residual management and risk warning system and method for anesthetic gas volatilization tanks, which features multi-source data fusion, intelligent prediction, and proactive risk warning functions.

[0007] The technical solution adopted by the present invention to achieve the above objectives is: an intelligent residual management and risk warning system for an anesthetic gas evaporation tank, comprising: ① A multimodal fusion sensing module is used to collect multidimensional physical parameters and environmental status data of the volatilization tank in real time, including: A high-precision weight sensor is installed in the isolation chamber at the bottom of the evaporation tank to monitor changes in the total mass of the tank in real time. An optical liquid level sensor, which uses the principle of infrared transmission, is installed in a transparent window on the side wall of the tank to detect the liquid level of the anesthetic. A gas flow meter, integrated into the outlet pipeline of the vaporizer, is used to monitor the real-time output flow rate of anesthetic gas; Temperature sensors are placed at key locations inside the tank to collect temperature distribution data inside the tank; A pressure sensor is installed in the sealed cavity of the evaporator and the output pipeline to monitor the stability of the gas pressure; Attitude sensors are used to detect the tilt or vibration of the tank and to compensate for errors in weight and liquid level measurement. An ambient light sensor is used to adjust the detection sensitivity of an optical liquid level sensor to adapt to different lighting conditions.

[0008] ② Intelligent data processing and early warning unit, connected to the multimodal fusion sensing module, includes: An embedded microcontroller is used to receive and fuse multi-source sensor data in real time, and calculate the remaining gas volume and sustaining time with high reliability based on preset anesthetic physical property parameters and adaptive compensation algorithm. The dynamic risk prediction engine integrates sliding time window traffic analysis, trend extrapolation and abnormal pattern recognition algorithms to predict the remaining time in real time and trigger risk assessment when traffic mutations, pressure anomalies or data conflicts are detected. The sensor health monitoring and fault tolerance module diagnoses the working status of each sensor in real time, automatically activates redundant data sources or switches to safe prediction mode in the event of a single point of failure, and records fault logs. Signal conditioning and filtering circuits are used to suppress environmental noise and interference and improve signal quality.

[0009] ③ An interactive early warning and display terminal, connected to the intelligent data processing and early warning unit, includes: The high-definition touch screen features a dynamic dual-ring fusion interface. The outer ring displays the remaining gas percentage with a gradient color ring bar, while the inner ring displays the remaining time and consumption rate with a combination of countdown and trend icons. The bottom of the interface integrates real-time warning indicator lights and an operation guide area. The intelligent sound and light alarm system supports multi-level contextual warnings and can trigger differentiated alarm strategies based on remaining gas volume, consumption trends, and system status, including visual prompts, sound alerts, and tactile feedback. The one-button emergency response panel provides physical buttons for quick silencing, alarm confirmation, and activation of backup gas supply.

[0010] ④ The cloud-edge collaborative communication and management module, connected to the intelligent data processing and early warning unit, includes: The dual-mode wireless transmission unit supports Wi-Fi and Bluetooth 5.0, enabling real-time data synchronization with the hospital information system (HIS), anesthesia workstations, and mobile terminals. Edge computing nodes are used for lightweight local data analysis and real-time decision-making, reducing reliance on the cloud. The risk event real-time reporting link automatically pushes warning information, on-site data snapshots and handling suggestions to designated terminals when a medium or high level warning is triggered or a system anomaly is detected. The remote diagnostic and configuration interface supports remote device status query, parameter calibration, and software upgrades.

[0011] Furthermore, the evaporation vessel adopts an integrated intelligent structural design: The bottom composite sensing chamber adopts a thermal isolation and vibration damping design and integrates weight, temperature and attitude sensors. Quick-release self-sealing air circuit interface, integrating flow meter and airtightness self-test unit, automatically completes positioning and sealing test during connection; The adaptive optics detection window is coated with an intelligent anti-reflective and anti-contamination coating and has an embedded ambient light sensor to achieve dynamic gain adjustment for liquid level detection.

[0012] Furthermore, the dynamic dual-ring fusion interface also includes: The outer ring intelligent chromatographic strip dynamically changes color and flashing frequency according to the remaining percentage and consumption rate; The inner ring prediction display area, except for the remaining time, uses arrow icons to indicate the acceleration or deceleration trend of consumption; The bottom status bar displays the anesthetic type, real-time flow rate, system health score, and network status in a scrolling manner.

[0013] Furthermore, the multi-level contextualized early warning system of the intelligent sound and light alarm system includes: Trend warning (remaining amount < 40% and consumption rate accelerates): The outer ring of the interface flashes yellow slowly, and a prompt message pops up on the screen; Inventory warning (remaining amount < 25%): the outer ring remains yellow and emits intermittent low beeping sounds; Emergency Warning (Remaining amount <10% or abnormal flow): The outer ring flashes red and emits a continuous beeping sound. A replacement guide pops up on the interface. System anomaly warning (sensor failure or severe data conflict): Triggers a red rotating alarm icon, escalates the audible alarm, and automatically sends a remote assistance request.

[0014] Furthermore, the system also includes an adaptive intelligent calibration module, which can automatically or semi-automatically calibrate the weight reference, density parameters and liquid level when the tank is empty, when the anesthetic is changed or during regular maintenance, and has a drift compensation function based on historical data.

[0015] Based on the above system, the present invention also provides a method for intelligent residual management and risk warning of anesthetic gas volatilization tanks, comprising the following steps: Step S1: Synchronous Acquisition and Fusion of Multi-Source Data The multimodal fusion sensing module collects data on weight, liquid level, flow rate, temperature, pressure, attitude, and ambient light in real time. The original data is processed by timestamp alignment, filtering, and normalization.

[0016] Step S2: High-confidence calculation and cross-validation of remaining gas volume The current mass of the anesthetic is calculated based on weight sensor data and empty can baseline values. Based on the density parameter corresponding to the selected anesthetic type, convert mass to volume; The liquid level height and corresponding volume were calculated using data from an optical liquid level sensor and cross-validated with the results of the gravimetric method. If the deviation between the two exceeds the allowable range, the diagnostic process is initiated, and a reliable data source is selected or calibration is triggered according to preset rules.

[0017] Step S3: Dynamic Remaining Time Prediction and Risk Identification A sliding time window algorithm is used to analyze recent traffic sequences and calculate the average consumption rate and trend. By combining the current remaining gas volume with the consumption rate, the remaining sustainable time can be dynamically predicted; Real-time monitoring of flow stability, pressure changes, and consistency of multi-sensor data; identification of abnormal patterns (such as leakage, blockage, and sensor failure). The risk level is assessed based on the identification results, and early warning events are generated.

[0018] Step S4: Contextualized Early Warning Interaction and Information Presentation On the interactive early warning and display terminal, the remaining gas percentage, remaining time, consumption trend and system status are displayed in real time through a dynamic dual-ring fusion interface. When the warning threshold is reached or a risk event is identified, the corresponding level of audible and visual alarm and interface prompts will be triggered according to the preset scenario. Provides quick access to one-click confirmation, mute, or start emergency procedures.

[0019] Step S5: Cloud-edge collaborative data sharing and remote response Through the cloud-edge collaborative communication and management module, the device status, remaining gas volume, prediction time and alarm records are regularly uploaded to the hospital information system or cloud platform. In the event of a medium- or high-risk event, real-time alerts and recommended measures will be automatically sent to the designated anesthesiologist's terminal or the central monitoring station. It supports remote viewing of real-time device data, historical trends, and execution of diagnostic commands.

[0020] Furthermore, the method also includes an adaptive intelligent calibration step: Automatically record the baseline weight value when the can is empty; When changing anesthetics, the type of anesthetic can be selected through the interface or automatically identified, and the corresponding physical property parameters can be loaded. Periodically or during calibration triggers, use optical liquid level data to perform online accuracy verification and compensation for the weight measurement system.

[0021] Furthermore, the method also includes anomaly self-healing and safety fallback steps: When an anomaly is detected in data from a single sensor, the system automatically switches to a backup computing model supported by other sensors and marks that sensor as pending inspection. When multiple sensor data conflict or the system determines that a high-risk state is reached, it automatically switches to a safe mode (such as conservative prediction and increased alarm frequency) and forces manual intervention for inspection. All warning events, mode switches, and operational interventions generate encrypted logs for post-event analysis and accountability.

[0022] The beneficial effects of this invention are: 1. Multi-source fusion and high-reliability measurement: By fusing and cross-validating data from multiple sensors such as weight, liquid level, flow rate, temperature, pressure and attitude, combined with an adaptive compensation algorithm, the accuracy and reliability of residual gas volume measurement are significantly improved, and the risk of false alarms and missed alarms is reduced.

[0023] 2. Intelligent prediction and proactive risk warning: The introduction of a dynamic risk prediction engine not only dynamically calculates the remaining time, but also identifies abnormal consumption patterns (such as leakage, abnormal high flow), sensor failures and environmental interference in real time, realizing the leap from "remaining capacity display" to "risk warning".

[0024] 3. Contextualized Interaction and Emergency Response: The system adopts a dynamic dual-ring fusion interface and a multi-level contextualized alarm system to provide intuitive and hierarchical risk information presentation and quick emergency operation access, thereby improving human-computer interaction efficiency and emergency response speed.

[0025] 4. Cloud-edge collaboration and remote management: Relying on dual-mode communication and edge computing capabilities, it enables real-time reporting of equipment data, remote monitoring and intelligent diagnosis, supports collaborative management and centralized decision-making of multiple devices within the hospital, and improves the overall operation and maintenance level of the anesthesiology department.

[0026] 5. System self-healing and enhanced safety: It has sensor health monitoring, fault-tolerant switching, adaptive calibration and safety fallback mechanism, which can maintain basic functions or safety warnings in the event of some failures or abnormalities, greatly improving the robustness and clinical safety of the system. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the system structure of the present invention; Figure 2 This is a schematic diagram of the dual-ring fusion interface design of the present invention; Figure 3 This is a flowchart of the intelligent margin management and risk warning method of the present invention. Detailed Implementation

[0028] The technical solutions of 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.

[0029] Please see Figure 1-3 A smart residual management and risk warning system for an anesthetic gas vaporization tank, comprising: ① A multimodal fusion sensing module is used to collect multidimensional physical parameters and environmental status data of the volatilization tank in real time, including: The high-precision weight sensor adopts a strain gauge structure, with a range of 0-500g and a resolution of ±0.05g. It is installed in the bottom isolation chamber and is equipped with a silicone and aerogel composite heat insulation layer between itself and the tank. The optical liquid level sensor uses an 850nm wavelength infrared pair, which is installed behind a transparent polycarbonate window on the side wall. The window is coated with an anti-reflective and anti-fouling coating, and integrates an ambient light sensor to achieve automatic gain adjustment. The gas flow meter adopts a thermal mass flow meter with a range of 0-1000 mL / min and a response time of <0.3 seconds, and is integrated into the quick-release interface. The temperature sensor, a PT100 platinum resistance thermometer, is placed at the top, middle, and bottom of the tank to monitor the temperature gradient. The pressure sensor, using MEMS piezoresistive technology, monitors the pressure in the sealed cavity and output pipeline of the tank, with a range of 0-200 kPa. Attitude sensor, integrating a three-axis accelerometer and gyroscope, is used to detect tank tilt and vibration; An ambient light sensor is used to sense the intensity of ambient light during operation and dynamically adjust the emission power of the liquid level sensor.

[0030] ② Intelligent data processing and early warning unit, connected to the multimodal fusion sensing module, includes: The embedded microcontroller uses an ARM Cortex-M7 core with a main frequency of ≥400MHz and is responsible for multi-source data fusion, core algorithm operation and system scheduling. The dynamic risk prediction engine is a software implementation that includes sliding window traffic analysis, Kalman filter-based trend prediction, and an anomaly detection model that combines rules and machine learning. The sensor health monitoring and fault tolerance module continuously monitors the signal quality, output stability, and logic consistency with other sensors of each sensor, enabling fault diagnosis and mode switching. The signal conditioning and filtering circuit adopts a combination of multi-stage active filtering and digital filtering to effectively suppress power frequency, radio frequency and random noise.

[0031] ③ An interactive early warning and display terminal, connected to the intelligent data processing and early warning unit, includes: The high-definition touchscreen display is a 7-inch IPS LCD with a resolution of 1024×600. It uses capacitive touch and displays the dynamic dual-ring fusion interface. The intelligent sound and light alarm system includes a multi-color LED light ring (integrated into the screen frame), a piezoelectric ceramic buzzer, and a linear motor (for tactile feedback). The one-button emergency response panel has three physical buttons, corresponding to "Alarm Confirmation / Mute", "Send Status" and "Emergency Standby".

[0032] ④ The cloud-edge collaborative communication and management module, connected to the intelligent data processing and early warning unit, includes: Dual-mode wireless transmission unit, supporting Wi-Fi 802.11ac and BLE 5.0, which can be automatically selected or used in parallel; Edge computing nodes, which are also microcontrollers, perform data compression, local early warning judgment, and latency tolerance calculation. The risk event real-time reporting link adopts the MQTT protocol. Alarm events include device ID, event level, timestamp, key data snapshot and suggested code; The remote diagnostic interface supports parameter configuration, log download, and firmware OTA upgrades via a secure tunnel.

[0033] The evaporation vessel adopts an integrated intelligent structural design: The bottom composite sensor compartment is CNC machined from aluminum alloy and filled with high-performance aerogel, which effectively isolates the operating room from heat source interference and equipment vibration. The quick-release self-sealing gas circuit interface uses a combination of magnetic positioning (neodymium iron boron magnetic ring) and mechanical locking, triggering an airtightness self-test upon connection (differential pressure method test, threshold <1kPa / min). The adaptive optics detection window features a multi-layer nano-coating on the glass surface, significantly reducing reflection and adhesion, while a photosensitive chip is integrated on the back.

[0034] The dynamic dual-ring fusion interface also includes: The outer ring intelligent chromatographic strip is green under normal conditions, gradually turning yellow, orange, and then red as the balance decreases, and flashing increases when the consumption rate accelerates; Inner loop prediction display area: The central large font displays the remaining time (HH:MM), the smaller font below displays the current average flow rate (mL / min), and the trend arrow on the right indicates the change in consumption rate; Bottom status bar: The left side displays the anesthetic icon and name, the middle displays the real-time flow rate and temperature, and the right side displays the signal strength and system status icons (such as calibration status and network status).

[0035] The multi-level contextualized early warning strategy configuration of the intelligent sound and light alarm system is as follows: Trend warning (remaining amount < 40% and average consumption rate increased by > 20% in the past 5 minutes): slow yellow flashing on the outer ring (1Hz), "Consumption is accelerating, please pay attention" is displayed at the bottom of the screen; Inventory warning (remaining balance < 25%): The outer ring is constantly yellow, the buzzer sounds for 0.5 seconds (70dB) every 10 seconds, and the screen displays the remaining time and is highlighted; Emergency warning (remaining capacity <10%, or flow rate remains at 0 for 30 seconds, or suspected leak detected): The outer ring flashes red rapidly (2Hz), the buzzer sounds continuously (85dB), and a full-screen replacement guide and confirmation button pop up on the screen; System anomaly warning (critical sensor failure, or weight and liquid level volume deviation >10%): The outer ring flashes red and blue alternately, the buzzer emits intermittent and rapid alarms, the screen displays the anomaly code and suggests contacting maintenance, and at the same time automatically sends a remote alarm.

[0036] The adaptive intelligent calibration module operates as follows: 1. When the weight of the can is detected to be stable and close to the preset range for an empty can, the user is prompted to confirm the empty can, and the current weight is automatically recorded as the new baseline value.

[0037] 2. When the weight increases dramatically (e.g., when the canister is replaced with a full one), the system prompts the user to select the type of anesthetic to be infused (e.g., sevoflurane, isoflurane, etc.) via the touchscreen, or to automatically identify it by scanning the drug barcode.

[0038] 3. The system loads the standard density-temperature curve of the anesthetic and performs density compensation based on the current temperature.

[0039] 4. Periodically (e.g., every 24 hours) or whenever the deviation between the liquid level sensor data and the gravimetric data exceeds a threshold (e.g., 3%), initiate the online verification process. Calculate the volume using the liquid level height and compare it with the gravimetric volume. If the deviation persists, automatically fine-tune and compensate the output of the gravimetric sensor and record the compensation log.

[0040] Based on an intelligent residual management and risk warning system for anesthetic gas vaporization tanks, this invention also provides an intelligent residual management and risk warning method for anesthetic gas vaporization tanks. Please refer to [link to relevant documentation]. Figure 3 Specifically, it includes the following steps: Step S1: Synchronous Acquisition and Fusion of Multi-Source Data After the system is powered on and initialized, all sensors synchronously collect data at a set frequency (e.g., 10Hz for weight, 5Hz for flow rate, and 2Hz for liquid level), and the microcontroller performs timestamp alignment.

[0041] The raw data is then smoothed by a software digital filter (such as a moving average or low-pass filter) after passing through a hardware filtering circuit.

[0042] Gradient analysis is performed on the temperature data, and tilt angle is calculated on the attitude data for subsequent compensation.

[0043] Step S2: High-confidence calculation and cross-validation of remaining gas volume Readings from weight sensor Subtract the standard weight of the empty can from the middle The net weight of the anesthetic was obtained. : Based on the current average temperature inside the tank Query density function for selected anesthetic type Calculate the volume using the weight method : Based on the readings of the optical liquid level sensor Tank cross-sectional area model Calculate the volume using the liquid level method : Calculate volume deviation : like If the figure is less than 5%, the data is considered valid, and the final remaining gas volume is... Take the weighted average of the two: If 5% ≤ If the percentage is less than 15%, a suspicious data flag is triggered, sensor diagnostics are initiated, and the system is temporarily... This is the main output.

[0044] like If the gas level is ≥15%, a sensor conflict alarm is triggered, the system switches to safe mode, the remaining gas volume is displayed as "--", and the alarm is highlighted.

[0045] Step S3: Dynamic Remaining Time Prediction and Risk Identification Obtaining continuous flow time series from gas flow meter .

[0046] Apply a sliding time window of length N (e.g., N = minutes) and calculate the moving average of the flow rate within the window. : Calculate the linear regression slope β of the flow rate to determine the consumption trend: Current remaining time The prediction is: The risk prediction engine runs the following monitoring in parallel: a. Traffic mutation monitoring: If the current traffic... A deviation of more than ±50% from the average of the past 1 minute is marked as a "traffic spike" event.

[0047] b. Pressure Leakage Monitoring: If the pressure in the sealed cavity drops below a set threshold (e.g., 5 kPa) within 1 minute and the flow rate is not zero, it is marked as a "suspected leak" event.

[0048] c. Data consistency monitoring: i.e., volume deviation in step S2 monitor.

[0049] Based on the monitored event combinations and their severity, the rule base is invoked to assess the current overall risk level (low, medium, high, urgent).

[0050] Step S4: Contextualized Early Warning Interaction and Information Presentation The processor will calculate (Convert to percentage) Consumption trend icons and risk levels are sent to the display terminal in real time to drive the dynamic dual-ring fusion interface.

[0051] The interface controls the outer ring color and flashing mode, and updates the inner ring time and trend arrows based on the risk level and remaining percentage.

[0052] When the risk assessment results trigger a certain warning level (see warning strategy), control the audible and visual alarm system to execute the corresponding alarm mode, and display the warning text and operation instructions in a designated area of ​​the screen (such as the top banner).

[0053] Users can confirm alarms, mute them, or trigger emergency menus via touchscreen or physical buttons.

[0054] Step S5: Cloud-edge collaborative data sharing and remote response The communication module packages the following data (in JSON format, including device ID and timestamp) and sends it to the preset cloud gateway or hospital intranet server every 30 seconds. When the risk level is "medium" or above, a real-time report (non-periodic) is immediately triggered, and the event.type and suggested.action fields are added to the data packet.

[0055] After receiving the alarm data packet, the cloud server or central monitoring station can highlight the device on the map or list interface, and can choose to push a notification to the handheld terminal of the anesthesiologist in charge of the operating room.

[0056] Authorized users can remotely log in to the system management backend via the hospital intranet or VPN to view the real-time data stream, historical trend charts, and alarm logs of designated devices, and can issue remote diagnostic commands (such as requesting raw sensor data) or schedule maintenance tasks.

[0057] The method also includes an adaptive intelligent calibration step, which is typically initiated automatically by the system or manually triggered in maintenance mode: 1. Empty Tank Calibration: When the system detects that the tank weight remains consistently at an extremely low value for an extended period (e.g., less than 5% of the full tank weight) and the flow meter reading is zero, it determines that the tank may be empty. The system prompts the user on the interface, "Perform empty tank calibration?" After the user confirms, the system records the average weight over the next 10 seconds as the new calibration value. And store.

[0058] 2. Anesthetic Identification and Parameter Loading: When a significant increase in weight is detected (within the range of typical full-can weight), the system prompts "Filling detected, please select anesthetic type." The user selects from the list or scans a barcode, and the system automatically loads the physical property database of that anesthetic, including standard density, boiling point, and saturated vapor pressure curve.

[0059] 3. Online cross-validation and compensation: During normal system operation, whenever the level sensor reading is reliable (e.g., when ambient light is stable), [the following will be implemented / compensated]. and Compare them. If a systematic bias (i.e., a deviation) is found... If the sampling is biased to one side (in multiple samplings), a software compensation algorithm is activated. For example, a compensation factor can be calculated. : In the subsequent gravimetric calculations, Multiply Then, volume conversion is performed. The compensation factor is updated incrementally and has upper and lower limit protection.

[0060] The method also includes abnormal self-healing and safety fallback steps: Single sensor failure: For example, if the system continuously monitors and detects that the signal strength of the optical level sensor remains at 0 or exceeds its range, it is determined to be a failure. The health monitoring module marks this sensor as "faulty" and stops using level data in the remaining gas volume calculation, relying entirely on the gravimetric and flow integral methods. Simultaneously, a level sensor fault icon is displayed on the interface, and all cross-validation and advanced alerts that rely on level data are disabled.

[0061] Multi-source data conflict and safety mode: When the volume deviation calculated by gravimetric and level methods continuously exceeds the emergency threshold (e.g., 20%), and the flow data also fluctuates drastically, the risk prediction engine cannot determine a reliable data source, and the system is deemed to have entered a "high uncertainty" state. The system automatically switches to safety mode: a) The remaining time is displayed as "Estimating" or the calculation result based on the most conservative (maximum) consumption rate is displayed; b) The alarm threshold is automatically increased (e.g., the emergency warning threshold is temporarily increased from 10% to 20%); c) A "System needs to be checked" prompt is displayed in a prominent position on the interface, and manual verification is recommended.

[0062] Logs and Traceability: All sensor status changes, warning triggers, mode switches, user confirmation operations, and remote commands are stored in the device's non-volatile memory as encrypted, timestamped records. These logs can be uploaded periodically or exported via the maintenance interface for device performance analysis, event review, and quality improvement.

[0063] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0064] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A smart residual management and risk early warning system for an anesthetic gas volatilization tank, characterized in that, include: A multimodal fusion sensing module is used to collect multidimensional physical parameters and environmental status data of the volatilization tank in real time; The intelligent data processing and early warning unit is connected to the multimodal fusion sensing module. It is used to receive and fuse multi-source sensor data, calculate the remaining gas volume and sustainable time, and identify abnormal patterns and trigger risk assessment based on the dynamic risk prediction engine. An interactive early warning and display terminal, connected to the intelligent data processing and early warning unit, is used to display the remaining gas volume, remaining time and risk status through a dynamic dual-ring fusion interface, and to execute multi-level contextualized early warnings. The cloud-edge collaborative communication and management module is connected to the intelligent data processing and early warning unit to enable real-time uploading, remote monitoring and collaborative management of device data.

2. The intelligent residual management and risk early warning system for an anesthetic gas volatilization tank according to claim 1, characterized in that: The multimodal fusion sensing module includes: A high-precision weight sensor is installed in the isolation chamber at the bottom of the evaporation tank to monitor changes in the total mass of the tank in real time. An optical liquid level sensor, which uses the principle of infrared transmission, is installed in a transparent window on the side wall of the tank to detect the liquid level of the anesthetic. A gas flow meter, integrated into the outlet pipeline of the vaporizer, is used to monitor the real-time output flow rate of anesthetic gas; Temperature sensors are placed at multiple locations inside the tank to collect temperature distribution data inside the tank; A pressure sensor is installed in the sealed cavity of the evaporator and the output pipeline to monitor the stability of the gas pressure; Attitude sensors are used to detect the tilt or vibration of the tank and compensate for measurement errors. An ambient light sensor is used to adjust the detection sensitivity of an optical liquid level sensor.

3. The intelligent residual management and risk early warning system for an anesthetic gas volatilization tank according to claim 2, characterized in that: The intelligent data processing and early warning unit includes: An embedded microcontroller is used to perform multi-source data fusion, remaining gas volume calculation, and remaining time prediction. The dynamic risk prediction engine integrates sliding time window flow analysis, trend extrapolation, and abnormal pattern recognition algorithms to identify sudden changes in flow, pressure anomalies, and sensor data conflicts in real time. The sensor health monitoring and fault tolerance module is used to diagnose the working status of each sensor and automatically activate redundant data sources or switch to safe prediction mode in the event of a single point of failure. Signal conditioning and filtering circuits are used to suppress environmental noise and interference.

4. The intelligent residual management and risk warning system for an anesthetic gas volatilization tank according to claim 3, characterized in that: The dynamic dual-ring fusion interface includes: The outer ring intelligent chromatographic strip dynamically displays the percentage of remaining gas volume using a gradient color ring strip, with the color and flashing frequency changing with the remaining amount and consumption rate. The inner ring prediction display area shows the remaining time with countdown numbers and uses arrow icons to indicate whether the consumption is accelerating or decelerating. The bottom status bar displays the anesthetic type, real-time flow rate, system health score, and network status in a scrolling manner.

5. The intelligent residual management and risk early warning system for an anesthetic gas volatilization tank according to claim 4, characterized in that: The multi-level contextualized early warning includes: Trend warning is triggered when the remaining amount is below 40% and the consumption rate accelerates; the outer ring of the interface will slowly flash yellow. Inventory warning is triggered when the remaining amount is below 25%, with the outer ring remaining yellow and emitting intermittent beeps; Emergency warning is triggered when the remaining volume is below 10%, abnormal flow is detected, or a suspected leak is detected. The outer ring flashes red and emits a continuous beeping sound. The system provides an anomaly warning, triggered when a sensor malfunctions or there is a serious data conflict. The outer ring flashes red and blue alternately and automatically sends a remote assistance request.

6. The intelligent residual management and risk early warning system for an anesthetic gas volatilization tank according to claim 1, characterized in that: The cloud-edge collaborative communication and management module includes: Dual-mode wireless transmission unit, supporting Wi-Fi and Bluetooth 5.0, for real-time synchronization with hospital information systems and mobile terminals; Edge computing nodes are used for lightweight local data analysis and real-time decision-making. A real-time risk event reporting link is used to automatically push early warning information and handling suggestions when medium- to high-risk events occur; The remote diagnostic and configuration interface supports device status query, parameter calibration, and software upgrade.

7. The intelligent residual management and risk early warning system for an anesthetic gas volatilization tank according to claim 6, characterized in that, It also includes an adaptive intelligent calibration module for: Automatically record the baseline weight value when the can is empty; When changing anesthetics, the system can automatically identify or manually select the type of anesthetic and load the corresponding physical property parameters. Regularly use optical liquid level data to perform online accuracy verification and compensation for the weight measurement system.

8. A method for intelligent residual management and risk early warning of an anesthetic gas volatilization tank, characterized in that, The system applied to any one of claims 1-7 includes the following steps: The multimodal fusion sensing module collects data on weight, liquid level, flow rate, temperature, pressure, attitude, and ambient light in real time. The current mass of the anesthetic is calculated based on the weight sensor data and the empty can baseline value, and then converted into volume according to the anesthetic density parameter; The liquid level height and corresponding volume were calculated using data from an optical liquid level sensor and cross-validated with the results of the gravimetric method. The sliding time window algorithm is used to analyze the flow sequence, calculate the average consumption rate and trend, and dynamically predict the remaining time. Real-time monitoring of flow stability, pressure changes, and consistency of multi-sensor data; identification of abnormal patterns and assessment of risk levels. The remaining gas percentage, remaining time, and risk status are displayed through a dynamic dual-ring fusion interface. When the warning threshold is reached or a risk event is identified, the corresponding level of audible and visual alarm is triggered; Device data is uploaded regularly through the cloud-edge collaborative communication module, and alarms are pushed in real time when medium- or high-risk events occur.

9. The intelligent residual management and risk warning method for an anesthetic gas volatilization tank according to claim 8, characterized in that: The calculation of the remaining gas volume includes the following steps: Readings from weight sensor Subtract the standard weight of the empty can from the middle The net weight of the anesthetic was obtained. : Based on the current average temperature inside the tank Query density function for selected anesthetic type Calculate the volume using the weight method : Based on the readings of the optical liquid level sensor Tank cross-sectional area model Calculate the volume using the liquid level method : Calculate volume deviation : like <5%, final remaining gas volume Take the weighted average of the two: like If the value is ≥5%, initiate the diagnostic process or trigger calibration.

10. The intelligent residual management and risk warning method for an anesthetic gas volatilization tank according to claim 8, characterized in that: It also includes abnormal self-healing and safety fallback steps: When an anomaly is detected in data from a single sensor, the system automatically switches to a backup computing model supported by other sensors. When multiple sensor data conflict or the system determines that a high-risk state is reached, it automatically switches to safe mode, raises the alarm threshold, and prompts for manual intervention. All warning events, mode switches, and operational interventions generate encrypted logs for post-event analysis and tracing.

Images