Automatic inspection method of gas fire extinguishing system based on pressure and bottle group state monitoring
By constructing a multi-dimensional inspection model and a digital twin model for simulation verification, the problem of insufficient accuracy in traditional gas fire extinguishing system inspection methods has been solved, realizing intelligent management and fault diagnosis of the system, and improving the system's safety and management scientificity.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGXI OULIAN EQUIP INSTALLATION PROJECT CO LTD
- Filing Date
- 2026-01-28
- Publication Date
- 2026-05-29
AI Technical Summary
Traditional gas extinguishing systems rely on manual inspection and simple threshold monitoring for inspection. This method cannot effectively distinguish between environmental changes and equipment malfunctions, resulting in insufficient accuracy in fault diagnosis, inability to perform in-depth data mining and real-time environmental correction, and difficulty in meeting the intelligent management needs of modern fire safety.
Construct a multi-dimensional inspection model, generate intelligent tasks, coordinate perception and data fusion, use a digital twin model for simulation verification, generate decision support reports and execute closed-loop actions to achieve multi-source information fusion and in-depth diagnosis.
It enables accurate fault diagnosis and system-level efficiency quantitative assessment of gas fire extinguishing systems, improves operation and maintenance response efficiency and safety, has dynamic adaptive capabilities, and forms a proactive prediction and data-driven management system.
Abstract
Description
Technical Field
[0001] This invention relates to the field of automatic inspection technology for gas fire extinguishing systems, specifically to an automatic inspection method for gas fire extinguishing systems based on pressure and cylinder status monitoring. Background Technology
[0002] Traditional gas fire extinguishing systems rely heavily on periodic manual inspections and simple threshold-based remote monitoring for operation and maintenance. Under current technology, maintenance personnel typically need to periodically visit the site to assess cylinder status through visual inspection and mechanical pressure gauge readings, or rely on monitoring systems equipped with pressure transmitters to receive remote over-limit alarms. However, these methods have systemic flaws and are no longer adequate for the urgent needs of modern high-standard protected environments for intelligent and precise fire safety management. Firstly, the monitoring logic is rudimentary and lacks diagnostic capabilities; most existing solutions only collect a single pressure parameter, and their judgments rely entirely on static pressure. Current systems, based on threshold comparisons, can only output binary conclusions such as "high / low pressure," failing to effectively distinguish between normal pressure changes caused by ambient temperature fluctuations and abnormal pressure decay due to chronic leakage or valve seal failure. They also lack access to multi-dimensional key information such as the opening / closing status of the valve head, the position of the safety pin, and the health of the drive unit. This results in root cause analysis relying entirely on human experience, leading to severely insufficient accuracy and foresight in early warning systems. Existing systems cannot deeply mine pressure time-series data to predict lifespan trends, nor can they dynamically adjust pressure criteria based on real-time ambient temperature. Inspection plans are rigid, and resource allocation is suboptimal. With the implementation of standards such as the "Technical Standard for Internet of Things Systems for Fire Protection Facilities" and the deepening of smart fire protection construction, developing a next-generation intelligent inspection technology capable of multi-source information fusion, deep intelligent diagnosis, system performance simulation, and closed-loop decision execution has become an inevitable direction for overcoming industry bottlenecks and ensuring the fire safety of critical infrastructure. Summary of the Invention
[0003] To solve the above technical problems, the present invention provides the following technical solution: an automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring, comprising the following steps:
[0004] S1. Construct a multi-dimensional inspection model: Configure a multi-dimensional inspection model containing static attributes, dynamic thresholds and correlations for each fire extinguishing agent cylinder group;
[0005] S2, Intelligent Triggering and Task Generation: Dynamically generates a priority inspection task list in response to inspection triggering conditions;
[0006] S3. Collaborative perception and data fusion: Real-time pressure data and multi-source bottle group status data of the target bottle group are collected in parallel, and the data is fused to generate a comprehensive status vector of the bottle group.
[0007] S4. Status Diagnosis and Health Assessment: Input the comprehensive status vector into the bottle group health assessment model, and output the health score and status diagnosis conclusion.
[0008] S5. Compliance and effectiveness verification based on simulation: For bottle groups with abnormal status diagnosis, start the system simulation model to simulate and verify the fire extinguishing effectiveness.
[0009] S6. Generate Decision Support Report and Execute Closed Loop: Generate a decision support report containing health score, diagnostic conclusion, verification results and maintenance suggestions, and execute the corresponding preset closed loop actions.
[0010] Preferably, in step S5, the system simulation model is a digital twin model constructed based on the principles of fluid dynamics and thermodynamics, and the specific process of performing fire extinguishing effectiveness simulation verification includes:
[0011] S5.1 Initialize the simulation scenario: Use the measured state of all current bottle groups or the diagnosed equivalent abnormal state as the initial conditions, and set the environmental parameters and system startup logic.
[0012] S5.2 Execute dynamic simulation solution: Based on the digital twin model, solve the dynamic physical equations of the entire process from the start-up of the cylinder group to the spraying of the extinguishing agent through the pipeline to the diffusion of the protected area, and calculate the pressure and flow distribution in the pipeline and the changes in the concentration of the extinguishing agent at key locations in the protected area over time in real time.
[0013] S5.3 Comparison of key performance indicators: Compare the key performance indicators obtained from simulation calculations with the design specifications. The key performance indicators include at least one or more of the following: extinguishing agent injection time, time to reach design concentration, and design concentration immersion time.
[0014] S5.4 Generate system-level impact assessment conclusions: Based on the comparison results, output an assessment conclusion on whether the system can still meet the design fire extinguishing performance under the current state.
[0015] Preferably, in step S5.1, the method for setting the diagnosed equivalent abnormal state is as follows:
[0016] For bottle groups with abnormal pressure, their current measured pressure value is used as the initial pressure for simulation.
[0017] For bottle groups that are inferred to have leaks, an equivalent flow area corresponding to the inferred leakage rate is set for their valves in the model.
[0018] Preferably, in step S5.2, the dynamic simulation solution process includes the following within each computation time step:
[0019] a. Calculate the outlet mass flow rate of each bottle group based on the instantaneous internal pressure and valve status;
[0020] b. Calculate the dynamic distribution of extinguishing agent in each node and branch of the pipeline network based on fluid network theory;
[0021] c. Calculate the jet flow rate based on the real-time pressure at each nozzle;
[0022] d. Update the remaining amount and pressure status of the extinguishing agent inside the cylinder group based on the total mass of the outflow.
[0023] Preferably, in step S5.3, the simulated concentration-time curve of the extinguishing agent in the protected area is compared with preset concentration thresholds and time thresholds to determine whether they are simultaneously satisfied:
[0024] The time required to reach the fire extinguishing design concentration does not exceed the limit, the immersion time for maintaining the concentration above the fire extinguishing design concentration meets the standard, and the total spraying time complies with the specifications.
[0025] Preferably, in step S1, the dynamic threshold includes a pressure standard threshold range that is corrected in real time based on ambient temperature; the association defines the protection zone, pipeline loop, and logic drive group to which the bottle group belongs.
[0026] Preferably, in step S3, the multi-source bottle group status data includes direct status signals from sensors at the bottle head valve, and valve sealing status indirectly inferred by analyzing the slight pressure changes in the outlet pipe section of the target bottle group during the system quiescent period.
[0027] Preferably, in step S4, the bottle group health assessment model is a machine learning-based model, whose training features include historical pressure sequences, ambient temperature sequences, and historical fault labels, and the health score is a continuous numerical value.
[0028] Preferably, in step S6, the preset closed-loop action includes: updating the state of the digital twin model, triggering a backup bottle group logic switching instruction, or automatically generating and dispatching a maintenance work order to a designated terminal.
[0029] It has the following beneficial effects:
[0030] By integrating multi-dimensional modeling, data intelligence fusion, and dynamic simulation verification based on digital twins, this solution elevates traditional discrete threshold alarms to precise fault diagnosis and system-level quantitative assessment of fire extinguishing effectiveness. It fundamentally solves the core engineering challenge of determining whether local anomalies affect overall functionality. Based on this, the system constructs a complete intelligent closed loop of perception-diagnosis-verification-decision-execution, capable of automatically switching to backup logic and generating optimized maintenance work orders based on diagnostic results, significantly improving operational response efficiency and reliability. Furthermore, the solution possesses dynamic adaptive and continuous learning capabilities; its diagnostic and simulation models can continuously optimize themselves with data accumulation, ultimately forming an intelligent safety management system that becomes increasingly accurate with use. This achieves a leap from passive response to proactive prediction, from isolated handling to system assurance, and from human experience-driven to data intelligence-driven approaches, significantly improving the system's safety, economy, and management scientific rigor. Detailed Implementation
[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below. 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.
[0032] In a first embodiment, the present invention provides a technical solution: an automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring.
[0033] Step 1: System initialization and multi-dimensional inspection model construction. The administrator configures the initial model for the fire suppression system of the computer room in Area A of the data center through the platform interface.
[0034] Static attribute entry: Enter information for 8 heptafluoropropane storage cylinders, including: cylinder number, volume, design filling pressure, filling amount, production date, etc.
[0035] Dynamic threshold setting: Based on the PT curve characteristics of heptafluoropropane, the platform pre-sets a pressure-temperature correspondence table for each bottle group. When the ambient temperature of the computer room is 25℃, the platform automatically calculates that the normal pressure range of bottle P1 should be (4.35±0.2)MPa. This range is the dynamic threshold at the current temperature.
[0036] Association configuration: P1-P4 bottle groups are set as the first drive group to protect sub-area A1; P5-P8 are the second drive group (backup group) to protect the same sub-area A1. At the same time, in the digital twin model of the platform, the pipeline topology diagram from these 8 bottle groups to the protected space of area A1 is accurately drawn, including information such as pipe length, pipe diameter, elbows, and nozzle positions.
[0037] Step 2: Intelligent Triggering and Task Generation
[0038] The system is set to perform a scheduled comprehensive inspection once a day at 02:00. When the platform log shows that the pressure of the P3 cylinder group has dropped by more than 0.05MPa in the past 24 hours, the event-linked inspection is triggered. After analyzing the event, the platform dynamically generates the inspection task for this inspection: priority is given to in-depth inspection of the P3 cylinder group and its associated P1, P2 and P4 cylinder groups. The key monitoring parameter set includes real-time pressure, pressure change rate and valve micro-leakage index.
[0039] Step 3: Collaborative Sensing and Data Fusion. The platform sends instructions to the field to execute the following collaborative data collection:
[0040] Direct data acquisition: The current pressure of bottles P1-P4, the valve position signal of the bottle head valve, and the status of the electromagnetic actuator safety pin can be read immediately through the intelligent monitoring terminal.
[0041] Indirect state inference: The platform automatically issues a command: briefly close the area selection valve in zone A1, so that the pipeline from the bottle group outlet to the valve forms a closed silent pipeline section. In the following 5 minutes, the pressure sensor data at the outlet of bottle P3 is monitored at high frequency. The data analysis module calculates that the pressure in this pipeline section drops from 4.10MPa to 4.098MPa, with a decay rate of 0.024MPa / min, which exceeds the preset sealing threshold.
[0042] Data fusion: The platform fuses the four data points of bottle P3 pressure = 4.10MPa, valve position = closed, safety pin = in position, and inferred leakage rate = 0.024MPa / min into a comprehensive state vector [4.10, 0, 1, 0.024] for subsequent analysis;
[0043] Step 4: State Diagnosis and Health Assessment. Input the comprehensive state vector of bottle P3 into the trained XGBoost health assessment model. After the model runs, the output is:
[0044] Health score: 72 points;
[0045] Condition diagnosis conclusion: The pressure is too low and there is a risk of chronic leakage. Specifically, the pressure is insufficient and the valve seals are suspected of aging.
[0046] Based on this, the platform determined that the P3 bottle group was in an abnormal state and required the initiation of advanced verification.
[0047] Step 5: Based on the compliance and performance verification of the simulation, the platform automatically calls the heptafluoropropane injection CFD-DYN digital twin simulation model built for area A1.
[0048] Simulation initialization: The ambient temperature is set to 25℃, and the initial state of the bottle group is set as follows: P1, P2, and P4 are in normal state; bottle P3 is set to an equivalent leakage abnormal state, the initial pressure is set to 4.10MPa, and a small equivalent leakage orifice is set at the outlet of its bottle head valve.
[0049] Dynamic simulation execution: The model simulates a fire. After a 30-second system delay, cylinder groups P1-P4 are started simultaneously. The model solves the fluid equations within each millisecond-level time step.
[0050] Calculations show that the initial discharge flow rate is lower than the design value due to the lower pressure in bottle P3.
[0051] The simulation of the dynamic distribution of extinguishing agent in the pipeline network shows that some extinguishing agent flows from the normal cylinder group to the area with lower pressure as compensation.
[0052] Calculate the time-varying curve of heptafluoropropane concentration in Zone A1, especially in the core cabinet;
[0053] Performance index comparison: After the simulation, key results are extracted:
[0054] Time to reach the designed fire extinguishing concentration: 38 seconds;
[0055] Soaking time: 7.5 minutes;
[0056] Total spray time: 48 seconds;
[0057] Compared with the design specifications, all indicators still meet the requirements;
[0058] The system-level assessment conclusion is generated: The simulation verification module concludes that under the current state, the anomaly of the P3 bottle group did not cause the overall failure of the fire extinguishing system in Zone A1. The system still has compliant fire extinguishing capabilities, but it is in a degraded operation state with reduced safety margin.
[0059] Step Six: Generate Decision Support Report and Execution Closed Loop. The platform integrates all the above information and automatically generates a fire protection system inspection and decision support report:
[0060] Abstract: An anomaly was found in the P3 bottle group, but the overall system performance verification passed;
[0061] Detailed content: Lists the health score, diagnostic details, simulation verification data, and conclusions of bottle P3;
[0062] Maintenance action recommendation: It is recommended to contact maintenance personnel within 3 calendar days to check the sealing and replenish the pressure of the P3 bottle group. During this period, the system can operate normally, but the second drive group (P5-P8) has been automatically promoted to the preferred logic start group.
[0063] At the same time, the platform automatically executes the following closed-loop actions:
[0064] Update the status label of bottle P3 in the digital twin model of area A1 to degraded operation;
[0065] Send a logic command to the gas extinguishing controller to temporarily switch the main start group of zone A1 from the first drive group (P1-P4) to the second drive group (P5-P8).
[0066] Create a maintenance work order with a priority of "medium" in the work order management system and push it to the mobile terminal APP of the maintenance personnel in the corresponding area;
[0067] Step 7: Continuous learning and model optimization. One week later, maintenance personnel confirmed on-site that the sealing ring of the P3 bottle group bottle head valve was aging. After replacement, the bottle was repressurized. The maintenance results were backed up to the platform via the APP. The platform will use the complete data chain from anomaly diagnosis to on-site confirmation as a new training sample and add it to the background database of the health assessment model for the next round of model iteration training, making future diagnoses more accurate.
[0068] Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art and related fields based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described and explained in the present invention, unless otherwise specified or limited, shall be implemented according to conventional means in the art.
Claims
1. An automatic inspection method for gas fire extinguishing systems based on pressure and cylinder status monitoring, characterized in that, Includes the following steps: S1. Construct a multi-dimensional inspection model: Configure a multi-dimensional inspection model containing static attributes, dynamic thresholds and correlations for each fire extinguishing agent cylinder group; S2, Intelligent Triggering and Task Generation: Dynamically generates a priority inspection task list in response to inspection triggering conditions; S3. Collaborative perception and data fusion: Real-time pressure data and multi-source bottle group status data of the target bottle group are collected in parallel, and the data is fused to generate a comprehensive status vector of the bottle group. S4. Status Diagnosis and Health Assessment: Input the comprehensive status vector into the bottle group health assessment model, and output the health score and status diagnosis conclusion. S5. Compliance and effectiveness verification based on simulation: For bottle groups with abnormal status diagnosis, start the system simulation model to simulate and verify the fire extinguishing effectiveness. S6. Generate Decision Support Report and Execute Closed Loop: Generate a decision support report containing health score, diagnostic conclusion, verification results and maintenance suggestions, and execute the corresponding preset closed loop actions.
2. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 1, characterized in that, In step S5, the system simulation model is a digital twin model constructed based on the principles of fluid dynamics and thermodynamics. The specific process of performing fire extinguishing effectiveness simulation verification includes: S5.1 Initialize the simulation scenario: Use the measured state of all current bottle groups or the diagnosed equivalent abnormal state as the initial conditions, and set the environmental parameters and system startup logic. S5.2 Execute dynamic simulation solution: Based on the digital twin model, solve the dynamic physical equations of the entire process from the start-up of the cylinder group to the spraying of the extinguishing agent through the pipeline to the diffusion of the protected area, and calculate the pressure and flow distribution in the pipeline and the changes in the concentration of the extinguishing agent at key locations in the protected area over time in real time. S5.3 Comparison of key performance indicators: Compare the key performance indicators obtained from simulation calculations with the design specifications. The key performance indicators include at least one or more of the following: extinguishing agent injection time, time to reach design concentration, and design concentration immersion time. S5.4 Generate system-level impact assessment conclusions: Based on the comparison results, output an assessment conclusion on whether the system can still meet the design fire extinguishing performance under the current state.
3. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 2, characterized in that, In step S5.1, the method for setting the diagnosed equivalent abnormal state is as follows: For bottle groups with abnormal pressure, their current measured pressure value is used as the initial pressure for simulation. For bottle groups that are inferred to have leaks, an equivalent flow area corresponding to the inferred leakage rate is set for their valves in the model.
4. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 2, characterized in that, In step S5.2, the dynamic simulation solution process includes the following within each calculation time step: a. Calculate the outlet mass flow rate of each bottle group based on the instantaneous internal pressure and valve status; b. Calculate the dynamic distribution of extinguishing agent in each node and branch of the pipeline network based on fluid network theory; c. Calculate the jet flow rate based on the real-time pressure at each nozzle; d. Update the remaining amount and pressure status of the extinguishing agent inside the cylinder group based on the total mass of the outflow.
5. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 2, characterized in that, In step S5.3, the simulated fire extinguishing agent concentration-time curve within the protected area is compared with preset concentration and time thresholds to determine whether both conditions are met simultaneously. The time required to reach the fire extinguishing design concentration does not exceed the limit, the immersion time for maintaining the concentration above the fire extinguishing design concentration meets the standard, and the total spraying time complies with the specifications.
6. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 1, characterized in that, In step S1, the dynamic threshold includes a pressure standard threshold range that is corrected in real time based on ambient temperature; the association defines the protection zone, pipeline loop, and logic drive group to which the bottle group belongs.
7. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 1, characterized in that, In step S3, the multi-source bottle group status data includes direct status signals from sensors at the bottle head valve, as well as valve sealing status indirectly inferred by analyzing the slight pressure changes in the outlet pipe section of the target bottle group during the system quiescent period.
8. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 1, characterized in that, In step S4, the bottle group health assessment model is a machine learning-based model, whose training features include historical pressure sequences, ambient temperature sequences, and historical fault labels, and the health score is a continuous numerical value.
9. The automatic inspection method for a gas fire extinguishing system based on pressure and cylinder status monitoring according to claim 1, characterized in that, In step S6, the preset closed-loop action includes: updating the state of the digital twin model, triggering a backup bottle group logic switching instruction, or automatically generating and dispatching a maintenance work order to a designated terminal.