Industrial building fire dynamic risk assessment method and system

Abstract

The application discloses an industrial building fire-fighting dynamic risk assessment method and system, and relates to the technical field of fire safety risk assessment. The method comprises the following steps: acquiring fire protection reference boundaries, fire protection demand domain data and fire protection capability domain data of a fire risk unit; determining a fire protection demand expansion comprehensive quantity and a fire protection capability reduction comprehensive quantity respectively; determining a fire protection mismatch boundary field according to the two, generating a state machine boundary parameter set, and using the state machine boundary parameter set to modulate the evolution boundary of a bottom risk state machine, and outputting a risk level, a rectification priority, a review strategy and a release observation condition. The application can identify the approaching, occupying and blocking states of dynamic fire load to the fire protection space, and identify the situation that the fire protection facilities appear normal but the actual protection capability decreases, thereby improving the assessment accuracy, stability and rectification closed-loop traceability.

Description

Technical Field

[0001] This invention relates to the field of fire safety risk assessment technology, and more specifically, to a method and system for dynamic fire risk assessment of industrial buildings. Background Technology

[0002] Dynamic fire risk assessment for industrial buildings is primarily used in scenarios such as factories, warehouses, production workshops, power distribution areas, temporary storage areas for packaging materials, finished product turnover areas, and precision manufacturing workshops. The assessment objects include the status of fire protection facilities, material stacking status, production operation status, evacuation accessibility, and local fire compartmentation status within the building space. Because the dynamic fire loads of pallets, turnover boxes, packaging materials, semi-finished products, finished product stacks, and mobile shelves in industrial buildings continuously change with production cycles and warehousing turnover, the fire risk assessment needs to reflect the real-time status of the site.

[0003] In the prior art, CN110555617B discloses a real-time dynamic quantitative assessment method for building fire risk based on the Internet of Things (IoT). It is applicable to different places such as shopping malls, industrial enterprises, public entertainment venues, high-rise buildings, construction sites, and subways. It collects and detects static and dynamic parameters of a specified place through the IoT, and outputs the fire risk index of the place in a timely and quantitative manner through an evaluation model algorithm, thereby realizing the identification and online dynamic assessment of fire risks in different places.

[0004] The core process of the existing solution includes: standardizing and classifying different site types and establishing corresponding site models and indicator systems; identifying risk areas and classifying and grading risks on the floor plan; dividing assessment indicator parameters into static and dynamic parameters; dynamic parameters include the output signal status values ​​of functional equipment such as fire hydrants, roller shutters, fire pumps, fire water tanks, wet alarm valves, end-of-line test devices, fire alarm controllers, smoke detectors, and electrical fire monitoring detectors; and then assigning values ​​to relevant parameters and outputting a fire risk index based on the assessment model. This process already possesses the common characteristics of data collection, risk area identification, dynamic parameter acquisition, and risk index output in dynamic fire risk assessment of industrial buildings.

[0005] However, in the actual operation of industrial buildings, fire risks do not only stem from abnormal conditions of fire-fighting equipment itself, but also often from the temporary intrusion of dynamic fire loads into the fire protection space. For example, fire hydrant water pressure may be normal, but the opening path of the cabinet door may be blocked by a tray; fireproof roller shutters may be online and normal, but the descent clearance area may be occupied by turnover boxes; sprinkler terminal pressure may be normal, but the effective sprinkler protection area may be eroded by high stacking; and electrical distribution cabinets may be in normal condition, but combustible packaging materials may be piled up in the prohibited stacking area. Although the existing solutions can incorporate the output signals of fire-fighting facilities as dynamic parameters into risk assessment, they do not further establish fire protection baseline boundaries, nor do they conduct a linkage analysis between the expansion of fire protection demand caused by dynamic fire loads and the reduction of fire protection capacity. As a result, the assessment results are difficult to accurately reflect the real risks after the local fire protection capacity is weakened by on-site materials, passageways, or process disturbances.

[0006] Therefore, it is still necessary to provide a method and system for dynamic fire risk assessment of industrial buildings. Based on fire risk units and fire protection benchmark boundaries, this system generates in parallel the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capacity reduction. By modulating the evolution boundary of the underlying risk state machine through the fire protection mismatch boundary field, the system can improve the accuracy, stability, interpretability and rectification closed-loop capability of dynamic fire risk assessment of industrial buildings. Summary of the Invention

[0007] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide a method and system for dynamic risk assessment of fire protection in industrial buildings. This system is used to assess the changes in fire protection requirements and fire protection capabilities caused by dynamic fire loads in a linked manner, and outputs risk levels, review strategies, and rectification closed-loop results through state machine boundary modulation, thereby solving the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A method for dynamic fire risk assessment of industrial buildings includes: acquiring fire protection baseline boundary, fire protection demand domain data, and fire protection capability domain data of fire risk units; determining a comprehensive quantity of fire protection demand expansion based on the fire protection demand domain data and the fire protection baseline boundary, and determining a comprehensive quantity of fire protection capability reduction based on the fire protection capability domain data and the fire protection baseline boundary; determining a fire protection mismatch boundary field based on the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capability reduction; generating a state machine boundary parameter set based on the fire protection mismatch boundary field, modulating the evolution boundary of the underlying risk state machine using the state machine boundary parameter set, and outputting risk level, rectification priority, review strategy, and observation release conditions based on the evolution state of the underlying risk state machine.

[0009] In a preferred embodiment, obtaining the fire protection reference boundary of the fire risk unit includes: dividing the industrial building into multiple fire risk units; generating a corresponding fire protection reference boundary for each fire risk unit; the fire protection reference boundary includes at least one of the following: effective area for fire hydrant access, clear area for fireproof roller shutter descent, effective area for sprinkler protection, clear width area for evacuation routes, smoke exhaust area, prohibited area for power distribution equipment, fire protection facility maintenance and operation area, prohibited water area for precision equipment, local ventilation protection area for chemicals, and special gas pipeline monitoring area.

[0010] In a preferred embodiment, the fire protection demand domain data includes at least one of the following: dynamic fire load object category, combustible properties, spatial outline, stacking height, stacking area, spatial coordinates, dwell time, direction of movement, distance from heat source, distance from electrical equipment, distance from special gas pipeline, distance from precision equipment, and occupancy status of fire protection reference boundary; the fire protection capability domain data includes at least one of the following: detection chain status, alarm chain status, fire extinguishing chain status, separation chain status, smoke exhaust chain status, evacuation chain status, local ventilation chain status, and special gas monitoring chain status.

[0011] In a preferred embodiment, determining the comprehensive amount of fire protection demand expansion based on the fire protection demand domain data and the fire protection reference boundary includes: determining the topology state level based on the topology state migration relationship between the dynamic fire load and the fire protection reference boundary; and determining the comprehensive amount of fire protection demand expansion based on the topology state level, the dynamic fire load continuous occupancy correction amount, the dynamic fire load combustible property correction amount, and the correction amount for proximity to heat sources, electrical equipment, special gas pipelines, or precision equipment.

[0012] In a preferred embodiment, the topology status levels include disjoint state, adjacent state, edge-cutting state, partially occupied state, continuously occupied state, and functional blocking state; the functional blocking state is that the dynamic fire load obstructs the opening path of the fire hydrant box door, the descent path of the fireproof roller shutter, the minimum clear width of the evacuation passage, the sprinkler water coverage space, the prohibited storage space for power distribution equipment, the local exhaust ventilation protection space, the special gas pipeline monitoring space, or the water-restricted protection space for precision equipment.

[0013] In a preferred embodiment, determining the comprehensive reduction amount of fire protection capability based on the fire protection capability domain data and the fire protection reference boundary includes: determining the comprehensive reduction amount of fire protection capability based on the reduction level of detection chain, alarm chain, fire extinguishing chain, separation chain, smoke exhaust chain, evacuation chain, local exhaust ventilation chain, special gas monitoring chain, cross-source consistency anomaly level, and apparent normal and effective reduction level; wherein, the apparent normal and effective reduction level is determined based on the status of fire protection facilities and the occupancy status of the fire protection reference boundary.

[0014] In a preferred embodiment, determining the fire protection mismatch boundary field based on the comprehensive expansion of fire protection demand and the comprehensive reduction of fire protection capacity includes: determining the positive expansion of fire protection demand based on the positive change of the comprehensive expansion of fire protection demand within a continuous analysis window; determining the positive reduction of fire protection capacity based on the positive change of the comprehensive reduction of fire protection capacity within a continuous analysis window; determining the growth trend of protection mismatch based on the positive expansion of fire protection demand and the positive reduction of fire protection capacity; and determining the fire protection mismatch boundary field based on the comprehensive expansion of fire protection demand, the comprehensive reduction of fire protection capacity, and the growth trend of protection mismatch. The boundary field is configured; the state machine boundary parameter set includes at least one of risk escalation boundary, risk hysteresis boundary, review trigger boundary, rectification lock boundary, release observation boundary, and disposal resource allocation boundary; wherein, the fire protection mismatch boundary field generates a normal boundary package, an observation boundary package, a tightening boundary package, a review boundary package, or a lock boundary package based on the level of the comprehensive quantity of fire protection demand expansion and the level of the comprehensive quantity of fire protection capability reduction; when the protection mismatch growth trend is greater than the protection mismatch growth trend threshold, the boundary package is upgraded by one level based on the currently generated boundary package, and the upgraded boundary package does not exceed the lock boundary package; the state machine boundary parameter set is generated based on the upgraded and modulated boundary package.

[0015] A dynamic fire safety risk assessment system for industrial buildings includes: a fire protection baseline boundary generation module for generating fire protection baseline boundaries for fire risk units; a dual-domain data acquisition module for acquiring fire protection demand domain data and fire protection capability domain data; a fire protection demand expansion analysis module for determining a comprehensive fire protection demand expansion quantity based on the fire protection demand domain data and the fire protection baseline boundary; a fire protection capability reduction analysis module for determining a comprehensive fire protection capability reduction quantity based on the fire protection capability domain data and the fire protection baseline boundary; a fire protection mismatch boundary field module for determining a fire protection mismatch boundary field based on the comprehensive fire protection demand expansion quantity and the comprehensive fire protection capability reduction quantity, and generating a state machine boundary parameter set based on the fire protection mismatch boundary field; and a bottom-level risk state machine module for modulating the evolution boundary of the bottom-level risk state machine using the state machine boundary parameter set, and outputting risk level, rectification priority, review strategy, and removal observation conditions based on the evolution state of the bottom-level risk state machine.

[0016] In a preferred embodiment, the system further includes: a fire risk unit modeling module for dividing industrial buildings into multiple fire risk units; a data credibility reconstruction module for performing consistency verification on video recognition data, warehouse management system data, production execution system data, electronic tag data, forklift positioning data, fire IoT data, inspection terminal data, special gas monitoring data, and local exhaust ventilation status data; a protection mismatch growth trend analysis module for generating a protection mismatch growth trend quantity based on the positive changes of the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capacity reduction within a continuous analysis window; and a risk output rectification closed-loop module for receiving video review, sensor retest, inspection terminal confirmation, or manual review results, and outputting risk level, rectification priority, review strategy, and conditions for lifting observation based on the evolution state of the underlying risk state machine.

[0017] In a preferred embodiment, the system is deployed on an edge computing industrial gateway, a local control server, or an industrial computer; the edge computing industrial gateway, local control server, or industrial computer includes a processor, a memory, and a communication interface; the communication interface is used to connect to a video surveillance system, a fire alarm controller, a fire protection IoT gateway, a warehouse management system, a production execution system, a forklift positioning system, an electronic tag system, an inspection terminal, a local exhaust ventilation system status acquisition terminal, a special gas monitoring system status acquisition terminal, and an industrial equipment status acquisition terminal; the memory is used to store fire risk unit data, fire protection benchmark boundary data, dynamic fire load data, fire protection status data, cross-source consistency verification records, state machine boundary parameter sets, state machine evolution records, rectification task records, review records, and de-observation records.

[0018] The technical effects and advantages of the present invention, a method and system for dynamic fire risk assessment of industrial buildings, are as follows: This invention establishes fire risk units and fire protection benchmark boundaries, incorporating spatial protection objects in industrial buildings, such as effective fire hydrant access areas, fireproof roller shutter descent clearance areas, effective sprinkler protection areas, clear width areas of evacuation routes, and prohibited stacking areas of electrical equipment, into the dynamic assessment scope. It can identify the approach, occupation, and obstruction status of dynamic fire loads such as pallets, turnover boxes, packaging materials, and cardboard box stacks on fire protection spaces, thereby improving the accuracy of locating fire hazards on-site in industrial buildings.

[0019] This invention generates a comprehensive quantity of fire protection demand expansion and a comprehensive quantity of fire protection capacity reduction in parallel, and constructs a fire protection mismatch boundary field by combining the quantity of protection mismatch growth trend. It can identify apparent normal and effective reduction states such as fire hydrant water pressure is normal but the door opening path is blocked, fireproof roller shutter is online and normal but the descent clearance area is occupied, and sprinkler terminal pressure is normal but the effective sprinkler protection area is eroded by high stacking, thereby improving the accuracy and interpretability of dynamic risk assessment results.

[0020] This invention modulates the evolution boundary of the underlying risk state machine by using the state machine boundary parameter set, so that risk escalation, review triggering, rectification locking and observation release have stable boundary conditions. This can reduce misjudgments caused by short-term occlusion, single-point sensor anomalies and process disturbances, and can output risk level, rectification priority, review strategy and observation release conditions, which facilitates the formation of a traceable closed loop for fire safety rectification of industrial buildings. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall architecture of a dynamic risk assessment method and system for fire protection in industrial buildings according to the present invention.

[0022] Figure 2 This is a schematic diagram of the fire risk unit and fire protection benchmark boundary modeling of the present invention.

[0023] Figure 3 This is a flowchart of the dual-domain data acquisition and credibility reconstruction process of the present invention.

[0024] Figure 4 This is a schematic diagram illustrating the generation of the boundary field and state machine boundary parameter set for fire protection mismatch in this invention.

[0025] Figure 5 This is a schematic diagram illustrating the evolution of the underlying risk state machine of this invention.

[0026] Figure 6 This is a schematic diagram illustrating the dynamic fire load intrusion and protection capacity reduction of the present invention.

[0027] Figure 7 This is a schematic diagram of the spatial relationship for risk review in the precision manufacturing equipment area of ​​the cleanroom according to the present invention.

[0028] Figure 8 This is a schematic diagram illustrating the edge computing deployment and data interaction of the present invention.

[0029] 100. Industrial building outline; 110. Production operation unit; 111. Production equipment; 120. Storage and stacking unit; 130. Precision equipment protection unit; 140. Power distribution unit; 150. Evacuation route unit; 151. Clean passage; 160. Personnel door; 210. Fire hydrant box; 211. Effective fire hydrant access area; 220. Fireproof roller shutter; 221. Clear space for fireproof roller shutter descent; 230. Sprinkler head; 231. Effective sprinkler protection area; 240. Distribution cabinet; 241. Power distribution equipment 250. Restricted storage area; 251. Precision equipment; 260. Precision equipment water-restricted area; 261. Local exhaust hood; 262. Chemical local exhaust protection zone; 270. Local exhaust outlet; 271. Special gas pipeline; 272. Special gas pipeline monitoring area; 311. Special gas monitoring point; 312. Pallet; 313. Turnover box; 320. Cardboard box stack; 330. Material turnover path; 341. Personnel evacuation path; 342. Chemical process equipment; 343. Local temperature rise zone; 344. Characteristic gas fluctuation zone. Detailed Implementation

[0030] 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.

[0031] Example 1, as Figure 1 As shown in the figure, this embodiment provides a dynamic risk assessment system for fire protection in industrial buildings. The system includes a fire risk unit modeling module, a fire protection baseline boundary generation module, a dual-domain data acquisition module, a data credibility reconstruction module, a fire protection demand expansion analysis module, a fire protection capacity reduction analysis module, a protection mismatch growth trend analysis module, a fire protection mismatch boundary field module, a bottom-level risk state machine module, and a risk output rectification closed-loop module.

[0032] The fire risk unit modeling module is used to divide industrial buildings into multiple fire risk units. The fire risk unit includes at least one of the following: production operation unit 110, storage and stacking unit 120, power distribution unit 140, equipment heat source unit, fire separation unit, fire protection facility protection unit, evacuation route unit 150, precision equipment protection unit 130, and chemical process unit.

[0033] The fire protection baseline boundary generation module is used to generate a corresponding fire protection baseline boundary for each fire risk unit. The fire protection baseline boundary includes at least one of the following: effective fire hydrant access area 211, fireproof roller shutter descent clearance area 221, sprinkler protection effective area 231, evacuation passage clearance area, smoke exhaust area, electrical equipment no-stacking area 241, fire protection facility maintenance and operation area, precision equipment water-free area 251, chemical local exhaust protection area 261, and special gas pipeline monitoring area 271.

[0034] The dual-domain data acquisition module is used to collect data in the fire protection demand domain and the fire protection capability domain. The fire protection demand domain data is used to characterize the approach, penetration, occupation, obstruction, and continuous change of dynamic fire loads on the fire protection baseline boundary. The fire protection capability domain data is used to characterize the effectiveness of detection chains, alarm chains, fire extinguishing chains, separation chains, smoke exhaust chains, evacuation chains, local ventilation chains, and special gas monitoring chains under the current operating conditions of the industrial building.

[0035] The data credibility reconstruction module is used to perform consistency verification on video recognition data, warehouse management system data, production execution system data, electronic tag data, forklift positioning data, fire protection IoT data, inspection terminal data, special gas monitoring data, and local exhaust status data, so as to reduce misjudgments caused by dust, water vapor, obstruction, reflection, sensor drift, network latency, and process heat source disturbances.

[0036] The fire protection demand expansion analysis module is used to generate a comprehensive quantity of fire protection demand expansion based on the topological state migration relationship between the dynamic fire load and the fire protection benchmark boundary.

[0037] The fire protection capability reduction analysis module is used to generate a comprehensive fire protection capability reduction based on the shortcomings of the detection chain, alarm chain, fire extinguishing chain, separation chain, smoke exhaust chain, evacuation chain, local ventilation chain and special gas monitoring chain, as well as the data reliability reconstruction results.

[0038] The protection mismatch growth trend analysis module is used to determine whether the increase in fire protection demand and the decrease in fire protection capability have formed a synchronous deterioration based on the positive changes of the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capability reduction within the continuous analysis window, and to generate the protection mismatch growth trend quantity.

[0039] The fire protection mismatch boundary field module is used to generate a state machine boundary parameter set based on the fire protection demand expansion comprehensive quantity, the fire protection capability reduction comprehensive quantity, and the protection mismatch growth trend quantity.

[0040] The underlying risk state machine module is used to modulate the evolution boundary of the underlying risk state machine using the state machine boundary parameter set, and output the risk level, rectification priority, review strategy and observation release conditions according to the evolution state of the underlying risk state machine.

[0041] The risk output rectification closed-loop module is used to receive video review, sensor retest, inspection terminal confirmation or manual review results, and output the fire risk unit number, current state machine status, risk level, rectification priority, affected fire protection benchmark boundary, review strategy, rectification object, observation release conditions and historical risk evolution records according to the evolution status of the underlying risk state machine.

[0042] Example 2, as Figure 2 As shown, this embodiment further illustrates the establishment method of fire risk units and fire protection benchmark boundaries based on embodiment one.

[0043] The system acquires the building floor plan, fire protection facility location map, fire compartmentation map, evacuation plan, storage location map, production equipment layout map, power distribution map, material turnover path map, precision equipment layout map, chemical pipeline layout map, and local exhaust system layout map of the industrial building. Based on the building floor plan, the system determines the outer contour 100 of the industrial building. This outer contour 100 is used to define the division of fire risk units, establish fire protection baseline boundaries, arrange evacuation passage units 150, and determine the spatial relative positions between each fire risk unit. Based on the building floor plan and the evacuation plan, the system determines personnel doors 160 and personnel evacuation paths 330. The personnel doors 160 indicate the entrances for personnel from each fire risk unit to the evacuation passage unit 150, and the personnel evacuation paths 330 indicate the evacuation direction and path for personnel from the production operation unit 110, storage and stacking unit 120, precision equipment protection unit 130, or power distribution unit 140 to the evacuation passage unit 150.

[0044] The system divides industrial buildings into multiple fire risk units based on fire compartment boundaries, production area boundaries, warehouse storage location boundaries, fire protection facility service areas, camera coverage areas, material turnover paths, precision equipment protection areas, and chemical process boundaries. Figure 2 A corresponding fire protection baseline boundary is established within each fire risk unit shown. The production equipment 111 is set within the production operation unit 110 and is used to characterize the fixed equipment in the industrial building that participates in production, processing, assembly, heat treatment, packaging, or transportation.

[0045] A basic data table is established for each fire risk unit. The basic data table includes the fire risk unit number, spatial coordinate range, fire compartment, adjacent risk units, associated fire protection facilities, associated fire separation components, associated evacuation routes, permitted stacking range, prohibited stacking range, location of heat source equipment, location of electrical equipment, location of precision equipment (250), location of special gas pipelines (270), location of local exhaust vents (262), and camera coverage relationship.

[0046] The system generates a fire protection baseline boundary for each fire risk unit. The fire protection baseline boundary is represented by a two-dimensional polygonal region, a two-dimensional grid region, or a three-dimensional spatial envelope. For areas involving sprinkler protection, stacking height, high-bay racking, cleanroom local exhaust ventilation, special gas pipelines 270, and water-free protection of precision equipment 250, a three-dimensional spatial envelope representation is preferred.

[0047] The effective fire hydrant access area 211 is generated based on the opening direction and angle of the fire hydrant box 210 door, the direction of hose access, and the personnel operating space. The fireproof roller shutter descent clearance area 221 is generated based on the roller shutter width, descent trajectory, bottom closure line, and the positions of the side guide rails. The sprinkler protection effective area 231 is generated based on the position of the sprinkler head 230, the protection radius of the sprinkler head 230, the beam bottom obstruction relationship, the shelf height, and the allowable stacking height. The evacuation passage clear width area is generated based on the passage design width, the location of the safety exit, and the evacuation direction. The smoke exhaust area is generated based on the smoke exhaust outlet location, the smoke exhaust fan's operating range, and the make-up air path. The electrical equipment no-stacking area 241 is generated based on the safety distances around the power distribution cabinet 240, cable trays, control cabinets, and charging equipment. The fire protection facility maintenance and operation area is generated based on the space required for fire protection facility inspection, maintenance, and emergency operations.

[0048] In a precision manufacturing scenario, the system determines the clean passage 151 based on the cleanroom layout, personnel access routes, equipment maintenance routes, and clean area isolation boundaries. The clean passage 151 is used to limit the passage space for personnel inspection, equipment maintenance, and anomaly verification within the precision equipment protection unit 130, and serves as one of the spatial benchmarks for judging whether dynamic fire loads affect the cleanroom's access and verification capabilities.

[0049] In a precision manufacturing scenario, the chemical local exhaust protection zone 261 is generated based on the location of the chemical process equipment 340, the location of the local exhaust hood 260, the location of the local exhaust outlet 262, the flow direction of the exhaust pipeline, the local exhaust coverage area, and the leakage monitoring range. The special gas pipeline monitoring zone 271 is generated based on the direction of the special gas pipeline 270, the location of the valve box, the location of the special gas monitoring point 272, and the emergency shut-off area. The special gas monitoring point 272 is used to collect information on leaks, concentration fluctuations, or alarm status near the special gas pipeline 270.

[0050] The output of this embodiment is a set of fire risk units and a set of fire protection baseline boundaries. This output serves as the basis for subsequent topology state identification, demand expansion analysis, capacity reduction analysis, and state machine boundary modulation.

[0051] Example 3, as Figure 3 As shown, this embodiment further illustrates the dual-domain data acquisition and credibility reconstruction method based on embodiment two.

[0052] The system collects fire protection demand domain data and fire protection capability domain data in parallel within each analysis window. The analysis window represents the data collection and calculation cycle for one dynamic assessment by the system, preferably ranging from 30 to 300 seconds. For areas with frequent forklift traffic, the analysis window is preferably 30 to 60 seconds; for storage and stacking areas, it is preferably 60 to 180 seconds; for fixed areas with production equipment, it is preferably 120 to 300 seconds; and for high-level cleanrooms and precision manufacturing equipment areas, it is preferably 30 to 120 seconds.

[0053] The fire protection requirement domain data includes dynamic fire load object categories, combustible properties, spatial outline, stacking height, stacking area, spatial coordinates, dwell time, direction of movement, distance from heat sources, distance from electrical equipment, 270° distance from special gas pipelines, 250° distance from precision equipment, and the occupancy status of fire protection baseline boundaries. This fire protection requirement domain data can come from video surveillance, warehouse management systems, production execution systems, forklift positioning systems, electronic tags, warehouse location sensors, inspection terminals, manual verification records, and cleanroom equipment management systems.

[0054] The fire protection capability domain data includes the status of the detection chain, alarm chain, fire extinguishing chain, separation chain, smoke exhaust chain, evacuation chain, local exhaust ventilation chain, and special gas monitoring chain. This fire protection capability domain data can originate from fire alarm controllers, fire IoT gateways, water pressure sensors, sprinkler terminal pressure sensors, fire pump control cabinets, fireproof roller shutter controllers, fire door status collectors, smoke exhaust fan status collectors, evacuation route video monitoring equipment, local exhaust ventilation system status collectors, special gas monitoring system status collectors, production execution systems, and inspection terminals.

[0055] During the data credibility reconstruction process, the system performs consistency verification on multi-source data within the same fire risk unit. For anomalies triggered by a single data source, the system prioritizes observation or verification and does not directly enter the rectification lockout state; for at least two types of data sources supporting the same anomaly within the same analysis window or adjacent analysis windows, the system increases the credibility level of the anomaly.

[0056] The cross-source consistency ratio is calculated using the following preferred model: ; in, This represents the cross-source consistency ratio of the i-th fire risk unit within the k-th analysis window; i represents the fire risk unit number; k represents the analysis window number. This indicates the number of data sources that give consistent conclusions on the same anomaly event; This indicates the total number of valid data sources participating in the verification.

[0057] when When the value is 0, the system does not calculate. The system will mark the corresponding abnormal events as data missing events. If the data missing event involves the effective area for fire hydrant access 211, the clear area for fireproof roller shutter descent 221, the clear width area for evacuation passage, the prohibited area for power distribution equipment 241, the monitoring area for special gas pipelines 271, or the prohibited water area for precision equipment 251, the system will enter the verification and confirmation state or trigger the data recovery process.

[0058] In a preferred embodiment As a verification gating parameter of the underlying risk state machine. When When the value is not less than 0.75, the system determines that the abnormal event has high confidence; if the fire protection mismatch boundary field output is at this time... or If a boundary packet is detected, the system will review the boundary trigger. The waiting window is set to 0 to 1 analysis window, enabling the underlying risk state machine to enter either the review and confirmation state or the rectification lock state. When When the value is greater than or equal to 0.50 and less than 0.75, the system marks the abnormal event as an anomaly pending review, suspends automatic escalation triggered by a single data source, and outputs video review, sensor retest, inspection terminal confirmation, or manual review tasks, keeping the underlying risk state machine in the demand expansion observation state, protection mismatch tightening state, or review confirmation state. When the value is less than 0.50, the system marks the abnormal event as a low-confidence anomaly and freezes the risk escalation boundary of the abnormal event. If the abnormal event involves a functional blockage or a reduction in the core protection chain, the system will not lower the current state machine state and will trigger a review and confirmation.

[0059] In high-end cleanrooms and precision manufacturing equipment areas, if an abnormal rise in ambient temperature, fluctuation of characteristic gas, or abnormal local exhaust ventilation is triggered by a single sensor, and the production execution system shows that the corresponding equipment is in the process-permitted heating, drying, or cleaning stage, the system will mark the abnormal event as a process-related event to be observed and restrict it from directly entering the rectification and locking state.

[0060] This embodiment can reduce misjudgments caused by dust, water vapor, strong glare, forklift obstruction, camera shake, sensor drift, network latency, and process heat source disturbances in industrial buildings.

[0061] Example 4, based on Example 3, further illustrates the generation method of the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capacity reduction.

[0062] The system generates a comprehensive quantity for expanding fire protection requirements based on the topological state migration relationship between dynamic fire loads and fire protection baseline boundaries.

[0063] The topological state is preferably divided into six levels. Level 0 is the separated state, indicating that the dynamic fire load has not approached the fire protection reference boundary. Level 1 is the adjacent state, indicating that the dynamic fire load is close to the fire protection reference boundary but has not entered it. Level 2 is the edge-entry state, indicating that the dynamic fire load has entered the edge of the fire protection reference boundary. Level 3 is the partial occupancy state, indicating that the dynamic fire load occupies a part of the fire protection reference boundary. Level 4 is the continuous occupancy state, indicating that the dynamic fire load occupies the fire protection reference boundary in multiple consecutive analysis windows. Level 5 is the functional interruption state, indicating that the dynamic fire load causes fire hydrants to be unusable, fire-resistant roller shutters 220 to be unable to descend, sprinkler protection space to be damaged, evacuation passages to be insufficiently wide, the prohibition of stacking electrical equipment to be violated, fire protection facilities to be unable to be maintained or operated, local exhaust protection boundary to be ineffective, or water-free protection conditions for precision equipment to be violated.

[0064] In a preferred embodiment, the system performs superposition calculations on the spatial contour of the dynamic fire load object and the fire protection reference boundary, and determines the topology status level based on distance, overlap ratio, functional path blocking status and number of continuous windows.

[0065] To improve the reproducibility of topology state level determination, the system defines the spatial envelope of dynamic fire load objects as follows: The spatial envelope of the fire protection reference boundary is , for and The shortest spatial Euclidean distance between them Spatial overlap rate, ;in, express and The overlapping area or overlapping volume, The area or volume of the fire protection reference boundary is represented; when the fire protection reference boundary is represented as a two-dimensional region, the area is used for calculation, and when the fire protection reference boundary is represented as a three-dimensional spatial envelope, the volume is used for calculation.

[0066] The system according to , The number of consecutively occupied windows and the status of blocked functional paths determine the topology status level. : when Greater than 1.0 meter, and When equal to 0, Setting it to 0 corresponds to a disjoint state.

[0067] when Greater than 0 and not greater than 1.0 meter, and When equal to 0, Set it to 1, which corresponds to the adjacent state.

[0068] when It equals 0, and When it is greater than 0 and not greater than 10%, Take 2, which corresponds to the edge cutting state.

[0069] when When the percentage is greater than 10% but not greater than 50%, and no complete functional blockade has been achieved, A value of 3 corresponds to a partially occupied state.

[0070] when When the percentage is greater than 10% but not greater than 50%, and remains in a localized occupied state for at least three consecutive analysis windows, without forming a complete functional blockade, A value of 4 corresponds to a continuously occupied state.

[0071] when When the dynamic fire load exceeds 50%, or obstructs the opening path of the fire hydrant box 210, the descent path of the fireproof roller shutter 220, the minimum clear width of the evacuation route, the sprinkler coverage space, the prohibited storage area 241 for power distribution equipment, the local exhaust ventilation protection space, the monitoring space corresponding to the special gas pipeline monitoring area 271, or the water-restricted area 251 for precision equipment. A value of 5 corresponds to the function blocking state. When the same dynamic fire load simultaneously satisfies multiple topology state levels, the system determines the state according to the higher level. .

[0072] In a preferred embodiment, when the dynamic fire load covers more than 50% of the preset opening fan-shaped surface of the fire hydrant box 210 door, cuts off the descent path of the fireproof roller shutter 220, makes the remaining effective net width of the evacuation passage less than 0.9 meters, makes the stacking height within the effective sprinkler protection zone 231 exceed the preset allowable stacking height, or obstructs the special gas sampling path, the system directly... Take 5.

[0073] The following optimal calculation model is used to calculate the comprehensive expansion of fire protection demand: ; in: This represents the comprehensive expansion of fire protection demand for the i-th fire risk unit within the k-th analysis window; This indicates the topological state level between dynamic fire loads and fire protection reference boundaries; This indicates the correction amount for continuous occupancy of dynamic fire loads; This indicates the correction amount for the flammability properties under dynamic fire load. This indicates a correction amount of 270 for proximity to heat sources, electrical equipment, special gas pipelines, or precision equipment; or 250 for proximity to such equipment. If the dynamic fire load appears only within one analysis window Set to 0; if the dynamic fire load appears consecutively in 2 to 5 analysis windows, Take 1; if the dynamic fire load persists for more than 5 analysis windows, Take 2.

[0074] If the dynamic fire load has low flammability properties... Take 0; if the dynamic fire load is cardboard boxes, plastic pallets, wooden pallets, ordinary packaging materials, or ordinary combustible semi-finished products, Take 1; if the dynamic fire load is flammable liquid, combustible dust, oil-containing materials, high-fire-load combustible materials, high-fire-load materials, or chemical containers, Take 2.

[0075] If the dynamic fire load is not near a heat source, electrical equipment, special gas pipeline 270, or precision equipment 250 Set to 0; if the dynamic fire load is close to one type of object, Take 1; if the dynamic fire load is simultaneously close to two or more types of objects, Take 2.

[0076] The system also generates a comprehensive reduction in fire protection capability based on the reduction status of the detection chain, alarm chain, fire extinguishing chain, separation chain, smoke exhaust chain, evacuation chain, local ventilation chain, and special gas monitoring chain.

[0077] The following optimal calculation model is used for the comprehensive reduction in fire protection capacity: ; in: This represents the comprehensive reduction in fire protection capacity for the i-th fire risk unit within the k-th analysis window; Indicates the detection chain reduction level; Indicates the alarm chain reduction level; Indicates the reduction level of the fire extinguishing chain; Indicates the reduction level of the separator chain; Indicates the reduction level of the smoke exhaust chain; Indicates the reduction level of the evacuation chain; Indicates the reduction level of the local exhaust chain; Indicates the reduction level of the special gas monitoring chain; Indicates the level of cross-source consistency anomaly; This indicates a normal appearance and an effective reduction grade. Cross-source consistency anomaly level Based on cross-source consistency ratio and the total number of valid data sources Confirmed. When When it is not less than 0.75, Take 0; when When the value is greater than or equal to 0.50 and less than 0.75, Take 1; when When less than 0.50, Take 2; when When the value is 0 and the corresponding abnormal event does not involve critical fire protection baseline boundaries, Take 3; when When the value is 0 and the corresponding abnormal event involves the effective fire hydrant access area 211, the fireproof roller shutter descent clearance area 221, the evacuation passage clearance area, the power distribution equipment prohibited storage area 241, the special gas pipeline monitoring area 271, or the precision equipment water prohibited area 251. Take 4; when When the state of 0 persists for at least three consecutive analysis windows, and the corresponding abnormal event involves the aforementioned key fire protection baseline boundary, Take 5.

[0078] The reduction levels for detection chains, alarm chains, fire extinguishing chains, separation chains, smoke exhaust chains, evacuation chains, local ventilation chains, and special gas monitoring chains are all assigned values ​​based on the corresponding protection chain's equipment offline status, operational failure status, abnormal status feedback, occupied status of associated fire protection baseline boundaries, number of consecutive abnormal windows, and verification confirmation results. The system prioritizes determining the equipment-side reduction level based on the equipment's offline status, operational failure status, and abnormal status feedback, and then adjusts the equipment-side reduction level upwards based on the occupied status of associated fire protection baseline boundaries, number of consecutive abnormal windows, and verification confirmation results. When the same protection chain simultaneously meets multiple reduction level conditions, the system determines the corresponding protection chain reduction level according to the highest level.

[0079] The reduction levels of each protection chain are preferably divided into levels 0 to 5, where: Level 0 indicates that the corresponding protection chain is in normal condition and the associated fire protection reference boundary is not occupied; Level 1 indicates that the corresponding protection chain has a slight delay, slight fluctuation, or a single low-confidence anomaly, but does not affect effective protection; Level 2 indicates that the corresponding protection chain has a single-point anomaly, or the equipment status needs to be retested, but has not yet affected the core protection function; Level 3 indicates that the corresponding protection chain has a core node anomaly, or the equipment status is normal but the corresponding fire protection reference boundary is partially occupied by dynamic fire loads, forming an apparent normal and effective reduction state; Level 4 indicates that the corresponding protection chain has multiple node anomalies, or the core protection space is continuously occupied, resulting in a significant decrease in protection capability; Level 5 indicates that the corresponding protection chain is ineffective, functionally blocked, or confirmed after review to be unable to provide effective protection in the early stages of a fire.

[0080] To quantify the apparent normality and effective reduction level The system is based on the status codes of the fire protection facilities. Spatial overlap rate Perform joint determination; A value of 1 indicates that the corresponding fire protection facilities, local ventilation equipment, or special gas monitoring equipment are online and in a normal self-test state; A value of 0 indicates that the corresponding device is in a faulty, offline, or self-test abnormal state.

[0081] when Take 1 and When equal to 0, Take 0. When Take 1 and When it is greater than 0 and not greater than 15%, Take 2. When take 1 and When it is greater than 15% and not greater than 40%, Take 3. When Take 1 and When it is greater than 40% and not greater than 75%, Take 4. When Take 1 and When the fire load exceeds 75%, or when the dynamic fire load directly cuts off the only physical operating surface, only sampling path, only descent path, or only evacuation clearance of the fire protection facilities, the fire load is insufficient. Take 5.

[0082] when When the value is 0, the system determines the comprehensive reduction in fire protection capacity based on the corresponding protection chain reduction level; when... When the fire protection baseline boundary is occupied by dynamic fire loads, the system uses a method of selecting 1. Identify a state that appears normal and has been effectively reduced.

[0083] In this embodiment, and These two calculations are generated in parallel within the same analysis window, with no sequential dependency and neither directly determining the final risk level. The above calculation model is used to determine the level range and subsequent boundary packets, not to directly form the final risk level through weighted summation.

[0084] Example 5, such as Figure 4 and Figure 5 As shown, this embodiment, based on embodiment four, further illustrates the implementation methods of the protection mismatch growth trend quantity, the fire protection mismatch boundary field, and the underlying risk state machine.

[0085] The system analyzes the data within the continuous analysis window. and The positive change generates a growth trend of protection mismatch. First, we calculate the positive expansion of fire protection demand: ; in: This represents the positive expansion of fire protection demand for the i-th fire risk unit within the k-th analysis window; This represents the comprehensive expansion of fire protection demand for the i-th fire risk unit within the (k-1)-th analysis window.

[0086] Then calculate the positive reduction in fire protection capacity: ; in: This represents the positive reduction in fire protection capability of the i-th fire risk unit within the k-th analysis window; This represents the comprehensive reduction in fire protection capability for the i-th fire risk unit within the (k-1)-th analysis window.

[0087] The following optimal calculation model is used to protect against the growth trend of mismatch: ; in: This represents the growth trend of protection mismatch in the i-th fire risk unit within the k-th analysis window; Therefore, the growth trend of protection mismatch increases only when the fire protection demand increases and the fire protection capability decreases simultaneously within the same fire risk unit; when both the fire protection demand and the fire protection capability are in a recovery trend, the growth trend of protection mismatch is not amplified. The system will , and The input is a fire protection mismatch boundary field, which includes a demand expansion axis, a capacity reduction axis, and a synchronous deterioration modulation amount. The output is a set of state machine boundary parameters.

[0088] ; in: This represents the state machine boundary parameter set of the i-th fire risk unit within the k-th analysis window; Indicates the risk escalation boundary; Indicates the hysteresis boundary of downshifting; Indicates the boundary for verification; This indicates that the rectification measures have been implemented and the boundaries have been locked. Indicates the lifting of the observation boundary; This indicates the boundary for allocating resources for disposal.

[0089] The risk escalation boundary is used to limit the number of consecutive abnormal windows, boundary package level, and confidence level required for the underlying risk state machine to enter the demand expansion observation state from the normal monitoring state, or from the demand expansion observation state to the protection mismatch tightening state; the risk hysteresis boundary is used to limit the number of consecutive stable windows, review pass conditions, and affected fire protection benchmark boundary recovery conditions required for the underlying risk state machine to return to the normal monitoring state from the rectification lock state or the recovery observation state; the review trigger boundary is used to limit whether to trigger video review, sensor retest, inspection terminal confirmation, or manual review; the rectification lock boundary is used to limit whether to generate a continuously effective rectification task and keep the corresponding fire risk unit in the rectification lock state before the review passes; the release observation boundary is used to limit whether to allow the release of risk observation after the rectification is completed; the disposal resource allocation boundary is used to determine the priority triggering object among video review, manual inspection, fire protection facility retest, material handling, local exhaust capacity review, and special gas sensor retest.

[0090] To avoid abrupt changes at the state machine boundaries, the system... Generate state machine boundary offsets: ; in: This represents the state machine boundary offset of the i-th fire risk unit within the k-th analysis window; Indicates the maximum allowable boundary offset; e represents the natural constant; This represents the boundary offset sensitivity coefficient; The threshold representing the growth trend of protection mismatch at the start boundary offset; when As the risk increases, the risk escalation boundary decreases, the review trigger boundary decreases, the rectification lock boundary decreases, and the downgrade hysteresis boundary increases. As a result, the system can quickly tighten the state machine boundary when protection needs and protection capabilities deteriorate simultaneously, and prevent frequent jumps when the risk recovers by using the downgrade hysteresis boundary.

[0091] In a preferred embodiment The value range is from 0.1 to 0.8. The value ranges from 1.0 to 3.0 analysis window units. The value ranges from 4.0 to 10.0. For a typical warehouse stacking unit of 120, The preferred value is 0.2. The preferred value is 1.0. The preferred value is 8.0. For the precision equipment protection unit 130, power distribution unit 140, special gas pipeline monitoring area 271, fireproof roller shutter descent clearance area 221, and evacuation passage clearance area, The preferred value is 0.5 to 0.7. The preferred value is 2.0 to 3.0. The preferred values ​​are 4.0 to 6.0.

[0092] During the dynamic correction phase, the system uses a 30-day parameter correction cycle to calculate the false escalation rate and the missed escalation rate within the same fire risk unit. The false escalation rate is the ratio of the number of escalations that, after review and confirmation, did not result in fire protection baseline boundary occupancy or protective chain reduction to the total number of escalations. The missed escalation rate is the ratio of the number of escalations that, after inspection, alarm, or post-event rectification records, confirmed fire protection baseline boundary occupancy or protective chain reduction but not yet entered the review and confirmation state by the system to the total number of anomalies. When the false escalation rate exceeds 15%, the system will... Increase by 10% to 20%, and Reduce by 5% to 10%; when the rate of missed upgrades exceeds 5%, the system will... Reduce by 10% to 20%, and Increased by 5% to 10%. (Revised) , and It is still limited to the above-mentioned range of values.

[0093] In a preferred embodiment, the fire protection mismatch boundary field will and They are divided into levels 0 to 5, and five types of boundary packets are generated according to priority rules; among them, For normal boundary packets, To observe the boundary packet, To tighten the boundary package, To verify the boundary packet, To lock the boundary packet; the to Boundary packets are used to change the state transition threshold of the underlying risk state machine. The final risk level is determined by the evolution state of the underlying risk state machine after modulation by the corresponding boundary packet. The system first judges... Then judge Then judge Then judge Final judgment .

[0094] when Level 5, or Level 5, or and When the sum is not less than 8, the system generates Lock the boundary packet.

[0095] When not generated ,and and When the sum is greater than or equal to 6 and not greater than 7, the system generates Verify boundary packets; when not generated ,and Level 4 or When it is level 4, the system generates Verify the boundary package.

[0096] When not generated or ,and and When the sum is greater than or equal to 4 and not greater than 5, the system generates Tighten the boundary packet; when not generated or ,and Level 3 or When it is level 3, the system generates Tighten the boundary package.

[0097] When not generated to ,and Level 2 or When it is level 2, the system generates Observe the boundary packet; when it is not generated to ,and and When at least one of them is level 1, the system generates Observe the boundary packet.

[0098] when equal to 0 and When equal to 0, the system generates Normal boundary package.

[0099] when Greater than At that time, the system will upgrade the current boundary packet by one level, with a maximum upgrade of one level per analysis window, and a maximum of no more than [missing information]. .

[0100] The underlying risk state machine includes the following states: normal monitoring state, demand expansion observation state, protection mismatch tightening state, review and confirmation state, rectification and locking state, and recovery observation state.

[0101] When the state machine is in normal monitoring state, and the fire protection mismatch boundary field outputs continuously for 1 to 2 analysis windows. At this time, the state machine enters the demand expansion observation state. When the state machine is in the demand expansion observation state, and the fire protection mismatch boundary field output... or When the state machine enters the protection mismatch tightening state, and the fire protection mismatch boundary field output is in the protection mismatch tightening state... ,or When the value is greater than or equal to 0.50 and less than 0.75, and the abnormal event involves the effective fire hydrant access area 211, the fireproof roller shutter descent clearance area 221, the evacuation passage clearance area, the electrical equipment prohibited storage area 241, the special gas pipeline monitoring area 271, or the precision equipment water prohibited area 251, the state machine enters the verification and confirmation state. When the state machine is in the verification and confirmation state, and the fire protection mismatch boundary field output... If the verification results confirm a functional blockage, the state machine enters a rectification lock state. When the state machine is in the rectification lock state, and the rectification task is completed, and the fire protection mismatch boundary field output decreases to a certain level for 3 to 10 consecutive analysis windows... or When the state machine enters the recovery observation state, and the observation release boundary is met, the state machine returns to the normal monitoring state; if the same fire risk unit reappears in the recovery observation state... or If the boundary packet is cleared, the system will immediately return to the verification and confirmation state or the rectification and locking state.

[0102] In a preferred embodiment, the system outputs a risk level based on the current state of the underlying risk state machine; the normal monitoring state corresponds to a low risk level, the demand expansion observation state corresponds to a risk level of concern, the protection mismatch tightening state corresponds to a higher risk level, the review and confirmation state corresponds to a higher or higher risk level, the rectification and locking state corresponds to a high risk level, and the recovery observation state corresponds to a risk level of concern; if the same fire risk unit reappears in the recovery observation state... or If the boundary packet is cleared, the system will immediately return to the verification and confirmation state or the rectification and locking state.

[0103] In this embodiment, the preferred requirement for removing the observation boundary is that no expansion of fire protection demand or reduction of fire protection capability occurs within 3 to 10 consecutive analysis windows, and the review record shows that the affected fire protection baseline boundary has been restored; In this embodiment, , as well as It can be obtained in any of the following ways.

[0104] Offline calibration is performed based on historical records of fire hazards in industrial buildings, fire inspection records, alarm records, rectification closed-loop records, and production operation records. The system can statistically analyze the correlation between dynamic fire load occupancy, fire protection chain reduction, and actual rectification results in different fire risk units, obtaining initial parameter values ​​applicable to the corresponding workshop type.

[0105] Configuration is based on experience and the spatial attributes of industrial buildings. This includes power distribution unit 140, prohibited storage area for electrical equipment 241, flammable material storage area, special gas pipeline monitoring area 271, fireproof roller shutter descent clearance area 221, evacuation passage unit 150, personnel evacuation route 330, and effective fire hydrant access area 211. and It is preferable to configure the state machine within the upper limit of the aforementioned corresponding value range, making the state machine boundary more sensitive to the growth of protection mismatch. For ordinary turnover areas, low fire load areas, and short-term logistics passage areas, and It is preferable to configure it within the lower limit of the above-mentioned corresponding value range to reduce false upshifts caused by short-term passage.

[0106] The system is dynamically revised based on the review results during the trial operation phase. During the trial operation period, the system records the actual results of risk escalation, review confirmation, rectification locking, and lifting of observation. When a certain type of fire safety risk unit frequently experiences erroneous escalation, the corresponding risk level is appropriately lowered. or improve When a certain type of fire risk unit has been missed in reporting or its review is delayed, the corresponding level should be appropriately increased. or reduce .

[0107] Based on the manual calibration of fire load density, floor height, equipment tolerance limit, material flammability, fire protection facility coverage and personnel evacuation conditions in a specific workshop by those skilled in the art; the manual calibration values ​​can be used as initial parameters, which will be subsequently corrected by the system based on historical evaluation records.

[0108] By using the above-described parameter acquisition method, even if different industrial buildings have differences in production processes, material types, spatial layouts, and fire protection facility configurations, those skilled in the art can still determine the parameters according to the rules disclosed in this embodiment. , and The range of feasible values.

[0109] Example 6, as Figure 6 and Figure 7 As shown, this embodiment further presents a typical industrial building scenario based on embodiments one through five.

[0110] Scenario 1: The effective access area 211 for fire hydrants is obstructed by pallet 311. In a certain storage unit 120, the system has already established the effective access area 211 for fire hydrants. Video surveillance identifies that the items stacked on pallet 311 have entered the effective access area 211 for fire hydrants, and there are four consecutive analysis windows. The warehouse management system shows that there are temporary turnover goods in this area, and forklift positioning shows that pallet 311 has not been moved after being placed. The system identifies this as a continuous occupancy state and generates a high comprehensive fire protection demand expansion. The fire hydrant water pressure sensor shows that the water pressure is normal, but video recognition shows that the opening path of the fire hydrant box 210 is obstructed. The system identifies the fire extinguishing chain as appearing normal and in an effective reduced state. Fire protection mismatch boundary field generation. or When the boundary package is reached, the state machine enters the verification and confirmation state or the rectification lock state. The system outputs the pallet 311 moving task, the fire hydrant box 210 door opening verification requirements and rectification completion confirmation conditions.

[0111] Scenario 2: The clearance area 221 for the descent of the fireproof roller shutter is occupied by a turnover box 312; a fireproof roller shutter 220 is installed between a production workshop and a storage area. The system has already established the clearance area 221 for the descent of the fireproof roller shutter. Video surveillance identifies the turnover box 312 as crossing the descent path of the fireproof roller shutter 220, with three consecutive analysis windows present. The fireproof roller shutter controller reports that the equipment is online and functioning normally. The system identifies this as a functional interruption state and identifies the separation chain as an action reduction state. A fire protection mismatch boundary field is generated. The boundary package and the underlying risk state machine enter the rectification lock state. The system outputs a task to clear obstacles in the fireproof roller shutter descent area, and requires video or manual verification to confirm that the turnover box 312 has been removed before allowing entry into the recovery observation state.

[0112] Scenario 3: Sprinkler protection zone 231 is eroded by high stacking. In a packaging material warehouse, the system uses video recognition and the warehouse location system to obtain the stacking height and determines that some cardboard box stacks 313 have encroached on sprinkler protection zone 231, with multiple analysis windows continuously present. The pressure sensor at the sprinkler end shows normal pressure. The system identifies this as a continuous occupancy state and increases the fire protection demand expansion quantity because cardboard boxes are considered high fire load materials. At the same time, the system determines that the sprinkler coverage space is damaged by the stacking height and identifies the fire extinguishing chain as a coverage reduction state. A fire protection mismatch boundary field is generated. At the boundary packet, the state machine enters the protection mismatch tightening state or the verification confirmation state, and outputs tasks such as reducing stacking height, adjusting stacking position, or inspection verification.

[0113] Scenario 4 involves flammable packaging materials appearing in the restricted storage area 241 for electrical equipment. In a certain electrical equipment area, the system has established a restricted storage area 241 corresponding to the distribution cabinet 240. The video identifies flammable packaging cartons entering the restricted storage area 241, and the electrical fire monitoring device detects residual current fluctuations. The warehouse management system has not recorded any materials allowed for temporary storage in this area. The system identifies a dynamic fire load entering the restricted storage area 241, and the fire protection demand is increased due to its proximity to electrical equipment. The detection chain enters a reduced state due to the abnormal electrical fire monitoring. Video recognition and electrical monitoring jointly support the same abnormal event, increasing the cross-source consistency ratio. The state machine enters a verification confirmation state or a rectification lock state, and the system outputs packaging material relocation tasks, electrical equipment retesting tasks, and manual inspection requirements.

[0114] Scenario 5 involves chemical anomalies and environmental disturbances in the precision manufacturing equipment area of ​​a cleanroom. In a high-end hardware manufacturing workshop or chip manufacturing cleanroom, the system has established a precision equipment water-restricted zone 251, a smoke-free zone, a chemical local exhaust ventilation protection zone 261, a special gas pipeline monitoring zone 271, and a clean passage protection zone corresponding to the clean passage 151. The precision equipment water-restricted zone 251 is used to limit the area where automatic sprinklers, fire-fighting water mist, or external water sources may cause accidental damage to high-value precision equipment 250. The chemical local exhaust ventilation protection zone 261 is used to characterize the boundary of local exhaust capacity required for chemical growth, cleaning, etching, encapsulation, or surface treatment processes. The local exhaust hood 260 is set above or to the side of the chemical process equipment 340 to collect volatile gases or local hot air flows generated during the operation of the chemical process equipment 340. The local exhaust outlet 262 is located at the exhaust inlet of the local exhaust hood 260 or the exhaust passage inlet to define the effective exhaust position of the chemical local exhaust ventilation protection zone 261. The special gas pipeline monitoring zone 271 is used to characterize the safety monitoring boundary required for the transportation, emission, and detection of special gases. In this scenario, the observation, verification, and local handling strategies output by the system are used to assist in fire management judgment and are not used to shield, delay, or replace the triggering logic of existing statutory fire alarms and fire linkage systems.

[0115] Within a certain analysis window, the sensor array detects a localized temperature rise and trace fluctuations in characteristic gases, but the video monitoring does not identify any open flame, smoke plume diffusion, or visible combustion features. The system determines a localized temperature rise zone 341 based on temperature sensor data, infrared thermography data, or equipment process status data, and determines a characteristic gas fluctuation zone 342 based on the detection results of the special gas monitoring point 272, localized exhaust ventilation status, and gas concentration change trends. The localized temperature rise zone 341 represents a localized area near the precision equipment 250 or chemical process equipment 340 where a temperature increase occurs, and the characteristic gas fluctuation zone 342 represents an area within the chemical localized exhaust ventilation protection zone 261 or the special gas pipeline monitoring zone 271 where characteristic gas concentration fluctuations occur. The system calls upon production execution system data, equipment process status data, localized exhaust ventilation status data, and video recognition data for cross-source consistency verification. If the production execution system reports that the equipment in this area is undergoing high-load chemical growth, heat treatment, cleaning and drying, or process heating, and the equipment process formula allows for a corresponding temperature rise range during this stage, the system identifies this heat source as a legitimate process heat source. If video surveillance does not detect any open flame characteristics, the characteristic gas fluctuation does not continue to rise, and the local exhaust ventilation chain remains effective, the system will maintain the overall expansion of fire protection demand in an adjacent or edge-cut-in state, and maintain the overall reduction of fire protection capability at a low reduction level.

[0116] Under the above circumstances, a fire protection mismatch boundary field is generated. When observing the boundary package, the underlying risk state machine enters the demand expansion observation state. The system outputs suggestions for local exhaust capacity review, manual fixed-point inspection tasks, special gas sensor retest tasks, and continuous analysis window observation conditions. It does not directly output the global spray trigger result, nor does it directly enter the rectification lock state.

[0117] If any of the following situations occur in the subsequent analysis window, the system will increase the comprehensive quantity of fire protection demand expansion or the comprehensive quantity of fire protection capacity reduction: local temperature rise continues to increase and exceeds the process allowable range; characteristic gas concentration continues to rise; local exhaust chain is reduced; video identifies smoke plume diffusion or flame characteristics; equipment process status and environment are abnormally inconsistent. At this time, the protection mismatch growth trend increases, and the fire protection mismatch boundary field changes from... Upgraded to or The underlying risk state machine enters either the protection mismatch tightening state or the verification confirmation state.

[0118] If further confirmation indicates a special gas leak, ventilation failure, open flame signs, or a reduction in the core protection chain, the system will generate [a response / instruction]. Once the boundary package is locked, the underlying risk state machine enters a rectification lockout state and outputs response strategies such as local isolation, enhanced fixed-point ventilation, personnel evacuation, equipment area verification, and confirmation by fire management personnel.

[0119] This scenario enables the invention to be applied to industrial building scenarios that are highly sensitive to false alarms and malfunctions, such as high-precision hardware manufacturing, chip manufacturing, semiconductor packaging, and precision electronic assembly. The system avoids misclassifying legitimate heat sources in the process as fire risks through cross-source consistency verification of process status, environmental status, video status, and fire protection chain status. Simultaneously, it can tighten the state machine boundaries when real anomalies continue to worsen.

[0120] Example 7: Based on Examples 1 to 6, this example further illustrates the abnormal data processing and fallback mechanism.

[0121] When the video recognition confidence level is below 0.75, the system does not directly enter the rectification and locking state, but instead calls the warehouse management system, electronic tags, forklift positioning data, production execution system data, or inspection terminal records for cross-source verification.

[0122] When a single fire sensor experiences a momentary anomaly, the system will first enter the demand expansion observation state or the protection mismatch tightening state, and require confirmation from multiple consecutive analysis windows before entering the verification and confirmation state.

[0123] When fire protection IoT data is missing, the system will mark the corresponding protection chain as having decreased data reliability and increase the sensitivity of the review trigger. If the data cannot be recovered after multiple consecutive analysis windows, the corresponding protection chain will enter a reduction state.

[0124] When a dynamic fire load briefly enters the fire protection reference boundary and then immediately leaves, the system allows entry into observation mode. If the same object or area repeatedly enters multiple consecutive analysis windows, the system increases the persistent occupancy correction.

[0125] Once the rectification is completed, the system enters a recovery observation state. Only when there are no further expansions in fire protection needs or reductions in fire protection capabilities within multiple consecutive analysis windows, and the review results are satisfactory, will the state machine be allowed to return to the normal monitoring state.

[0126] For conflict events affecting fire hydrant boxes 210 or fire hydrant access effective areas 211, fireproof roller shutters 220 or fireproof roller shutter descent clearance areas 221, evacuation passage units 150 or personnel evacuation routes 330, power distribution equipment prohibited areas 241, special gas pipeline monitoring areas 271 or precision equipment water prohibited areas 251, the system improves the sensitivity of review trigger boundaries and rectification locking boundaries.

[0127] When localized temperature rises, characteristic gas fluctuations, or abnormal particle counts occur in cleanrooms or precision manufacturing equipment areas, such as Figure 7As shown, the system prioritizes calling the production execution system, equipment process formula, local exhaust ventilation status, and video recognition results to verify the legality of the process. If the anomaly falls within the permissible disturbance range of the current process stage, the system enters the demand expansion observation state and outputs a local review task. If the anomaly exceeds the permissible range of the process, or if cross-source data points to a special gas leak, exhaust failure, plume diffusion, or a reduction in the core protection chain, the system enters the review confirmation state or the rectification lock state.

[0128] The risk levels, review strategies, rectification tasks, and state machine states output by this system are not intended to shield, disconnect, or bypass existing legally mandated fire alarm and fire linkage systems in industrial buildings. When existing fire alarm and linkage systems meet the legally mandated triggering conditions, the original fire protection system will still execute according to the preset linkage logic. This system is used to provide a basis for dynamic risk assessment, review tasks, rectification closure, and on-site management.

[0129] This embodiment can reduce the frequent changes in risk levels caused by single-point false alarms, short-term occlusion, sensor drift, network latency, process heat source disturbances, and unstable video recognition.

[0130] Example 8, as Figure 8 As shown, this embodiment provides an electronic device, preferably an edge computing industrial gateway, local control server, or industrial computer deployed in an industrial production site. The electronic device includes a processor, a memory, a communication interface, and a data bus. The memory stores a computer program, and when the processor executes the computer program, it implements the dynamic risk assessment method for fire protection of industrial buildings as described in any one of Embodiments 2 to 7.

[0131] The electronic device is preferably deployed in the industrial building site network to complete local fire risk unit data management, fire protection benchmark boundary retrieval, dual-domain data acquisition, data credibility reconstruction, dual comprehensive quantity calculation, protection mismatch growth trend analysis, fire protection mismatch boundary field generation, underlying risk state machine evolution, and risk output rectification closed-loop recording.

[0132] The communication interface is used to connect to video surveillance systems, fire alarm controllers, fire IoT gateways, warehouse management systems, production execution systems, forklift positioning systems, electronic tag systems, inspection terminals, local exhaust ventilation system status acquisition terminals, special gas monitoring system status acquisition terminals, and industrial equipment status acquisition terminals.

[0133] The industrial standard communication protocol includes at least one of the following: industrial fieldbus protocol, industrial Ethernet communication protocol, building automation standard protocol, and Internet of Things message transmission protocol. Through this industrial standard communication protocol, the electronic device can collect data on the status of underlying equipment, fire protection facilities, production processes, and environmental perception. It can also send risk assessment results, review tasks, rectification tasks, state machine status, and state machine boundary parameter sets to a local management platform, fire safety management terminal, or an authorized fire linkage prompt interface.

[0134] In a preferred embodiment, the electronic device only performs secure isolation reading of the underlying device status, without directly rewriting the native control program of the industrial equipment or directly replacing the legally mandated linkage judgment of the existing fire alarm controller. For scenarios requiring linkage, the electronic device prioritizes outputting risk warnings, review tasks, rectification tasks, suggestions for reviewing local ventilation capacity, or manual confirmation requests. Only when permitted by the existing fire linkage system of the industrial building will the electronic device output linkage suggestions or status warning signals to the corresponding linkage interface.

[0135] The edge computing industrial gateway can generate state machine boundary parameter sets and evolve the underlying risk state machine locally, reducing the impact of cloud transmission delay on dynamic risk assessment. In the event of a network interruption, the edge computing industrial gateway retains the most recent valid fire risk unit data, fire protection benchmark boundary data, and state machine status, and continues to perform observation, review, and rectification locking judgments according to local fallback rules.

[0136] The memory is used to store fire risk unit data, fire protection benchmark boundary data, dynamic fire load data, fire protection status data, cross-source consistency verification records, state machine boundary parameter sets, state machine evolution records, rectification task records, review records, and de-observation records.

[0137] This embodiment also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the dynamic risk assessment method for fire protection of industrial buildings as described in any one of Embodiments 2 to 7.

[0138] Through the above embodiments, this invention can transform dynamic fire load changes in industrial buildings into a comprehensive quantity of fire protection demand expansion, and transform the actual usability of fire protection facilities and protection chains into a comprehensive quantity of fire protection capacity reduction. Furthermore, it identifies the simultaneous deterioration of rising fire protection demand and declining fire protection capacity through the protection mismatch growth trend. The fire protection mismatch boundary field modulates the underlying risk state machine through a state machine boundary parameter set, ensuring that risk escalation, review triggering, rectification locking, and de-observation have interpretable boundary conditions. This improves the accuracy, stability, practicality, and traceability of rectification closed-loop in dynamic fire risk assessment of industrial buildings.

[0139] It should be noted that, for the sake of brevity, the foregoing method embodiments are described as a series of actions, but this does not mean that the application limits the order of the steps. Based on the ideas of this application, some steps can be executed in different orders or in parallel without affecting the functional implementation. Secondly, those skilled in the art should also understand that the specific embodiments described in the specification are preferred embodiments of the technical solutions of this application, and not limitations on the scope of protection of this application. All equivalent improvements or substitutions made within the spirit and principles of this application should be covered within the scope of protection of this application.

[0140] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dynamic risk assessment of industrial building fire protection, characterized in that, include: Acquire the fire protection baseline boundary, fire protection requirement domain data, and fire protection capability domain data of the fire risk unit; The fire protection demand expansion quantity is determined based on the fire protection demand domain data and the fire protection benchmark boundary, and the fire protection capacity reduction quantity is determined based on the fire protection capacity domain data and the fire protection benchmark boundary. The fire protection mismatch boundary field is determined based on the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capacity reduction. The state machine boundary parameter set is generated based on the fire protection mismatch boundary field. The evolution boundary of the underlying risk state machine is modulated using the state machine boundary parameter set. The risk level, rectification priority, review strategy and observation release conditions are output based on the evolution state of the underlying risk state machine.

2. The method for dynamic fire safety risk assessment of industrial buildings according to claim 1, characterized in that, The method for obtaining the fire protection baseline boundary of the fire risk unit includes: The industrial building is divided into multiple fire risk units as described above; Generate a corresponding fire protection baseline boundary for each fire risk unit; The fire protection baseline boundary includes at least one of the following: effective area for fire hydrant access, clear area for fireproof roller shutter descent, effective area for sprinkler protection, clear width area for evacuation routes, smoke exhaust area, prohibited area for power distribution equipment, fire protection facility maintenance and operation area, prohibited water area for precision equipment, local exhaust protection area for chemicals, and special gas pipeline monitoring area.

3. The method for dynamic fire safety risk assessment of industrial buildings according to claim 2, characterized in that, The fire protection demand domain data includes at least one of the following: dynamic fire load object category, combustible properties, spatial outline, stacking height, stacking area, spatial coordinates, dwell time, movement direction, distance from heat source, distance from electrical equipment, distance from special gas pipeline, distance from precision equipment, and fire protection benchmark boundary occupancy status. The fire protection capability domain data includes at least one of the following: detection chain status, alarm chain status, fire extinguishing chain status, separation chain status, smoke exhaust chain status, evacuation chain status, local ventilation chain status, and special gas monitoring chain status.

4. The method for dynamic fire safety risk assessment of industrial buildings according to claim 3, characterized in that, The determination of the comprehensive quantity of fire protection demand expansion based on the fire protection demand domain data and the fire protection baseline boundary includes: The topology state level is determined based on the topology state transition relationship between the dynamic fire load and the fire protection reference boundary. The comprehensive expansion of fire protection requirements is determined based on the topology status level, the dynamic fire load continuous occupancy correction, the dynamic fire load combustible property correction, and the correction for proximity to heat sources, electrical equipment, special gas pipelines, or precision equipment.

5. The method for dynamic fire safety risk assessment of industrial buildings according to claim 4, characterized in that, The topology state levels include disjoint state, adjacent state, edge-cutting state, local occupancy state, continuous occupancy state, and function blocking state. The functional blocking state refers to the dynamic fire load obstructing the opening path of the fire hydrant box, the descent path of the fireproof roller shutter, the minimum clear width of the evacuation passage, the sprinkler water coverage space, the prohibited storage space for power distribution equipment, the local exhaust ventilation protection space, the special gas pipeline monitoring space, or the water-restricted protection space for precision equipment.

6. The method for dynamic fire safety risk assessment of industrial buildings according to claim 3, characterized in that, The determination of the comprehensive reduction in fire protection capacity based on the fire protection capacity domain data and the fire protection benchmark boundary includes: The comprehensive reduction in fire protection capability is determined based on the reduction levels of detection chain, alarm chain, fire extinguishing chain, separation chain, smoke exhaust chain, evacuation chain, local ventilation chain, special gas monitoring chain, cross-source consistency anomaly level, and apparent normal and effective reduction level. The apparent normal and effective reduction level is determined based on the status of the fire protection facilities and the occupancy status of the fire protection reference boundary.

7. The method for dynamic fire safety risk assessment of industrial buildings according to claim 6, characterized in that, The determination of the fire protection mismatch boundary field based on the comprehensive expansion of fire protection demand and the comprehensive reduction of fire protection capacity includes: Based on the positive change of the comprehensive quantity of fire protection demand expansion within the continuous analysis window, determine the positive expansion quantity of fire protection demand; Based on the positive change of the comprehensive reduction in fire protection capacity within the continuous analysis window, determine the positive reduction in fire protection capacity; Based on the positive expansion of fire protection demand and the positive reduction of fire protection capacity, the growth trend of protection mismatch is determined; The fire protection mismatch boundary field is determined based on the comprehensive quantity of fire protection demand expansion, the comprehensive quantity of fire protection capacity reduction, and the quantity of protection mismatch growth trend. The state machine boundary parameter set includes at least one of the following: risk escalation boundary, risk downgrade hysteresis boundary, review trigger boundary, rectification lock boundary, observation release boundary, and disposal resource allocation boundary. The fire protection mismatch boundary field generates a normal boundary package, an observation boundary package, a tightened boundary package, a verification boundary package, or a locked boundary package based on the level of the fire protection demand expansion comprehensive quantity and the level of the fire protection capability reduction comprehensive quantity. When the protection mismatch growth rate exceeds the protection mismatch growth rate threshold, the boundary packet is upgraded by one level based on the currently generated boundary packet, and the upgraded boundary packet does not exceed the locked boundary packet. The state machine boundary parameter set is generated based on the modulated boundary packet.

8. A dynamic fire safety risk assessment system for industrial buildings, characterized in that, include: The fire protection baseline boundary generation module is used to generate the fire protection baseline boundary of the fire risk unit; The dual-domain data acquisition module is used to acquire data in the fire protection demand domain and the fire protection capability domain. The fire protection demand expansion analysis module is used to determine the comprehensive quantity of fire protection demand expansion based on the fire protection demand domain data and the fire protection baseline boundary. The fire protection capacity reduction analysis module is used to determine the comprehensive amount of fire protection capacity reduction based on the fire protection capacity domain data and the fire protection benchmark boundary. The fire protection mismatch boundary field module is used to determine the fire protection mismatch boundary field based on the fire protection demand expansion comprehensive quantity and the fire protection capability reduction comprehensive quantity, and to generate a state machine boundary parameter set based on the fire protection mismatch boundary field; The underlying risk state machine module is used to modulate the evolution boundary of the underlying risk state machine using the state machine boundary parameter set, and output the risk level, rectification priority, review strategy and observation release conditions according to the evolution state of the underlying risk state machine.

9. The dynamic risk assessment system for fire protection of industrial buildings according to claim 8, characterized in that, The system also includes: The fire risk unit modeling module is used to divide industrial buildings into multiple fire risk units. The data credibility reconstruction module is used to perform consistency verification on video recognition data, warehouse management system data, production execution system data, electronic tag data, forklift positioning data, fire protection IoT data, inspection terminal data, special gas monitoring data, and local exhaust ventilation status data. The protection mismatch growth trend analysis module is used to generate the protection mismatch growth trend quantity based on the positive changes of the comprehensive quantity of fire protection demand expansion and the comprehensive quantity of fire protection capacity reduction within the continuous analysis window; The risk output rectification closed-loop module is used to receive video review, sensor retest, inspection terminal confirmation or manual review results, and output risk level, rectification priority, review strategy and observation release conditions according to the evolution state of the underlying risk state machine.

10. The dynamic risk assessment system for fire protection of industrial buildings according to claim 9, characterized in that, The system is deployed on an edge computing industrial gateway, a local control server, or an industrial computer. The edge computing industrial gateway, local control server, or industrial computer includes a processor, memory, and communication interface. The communication interface is used to connect to video surveillance systems, fire alarm controllers, fire IoT gateways, warehouse management systems, production execution systems, forklift positioning systems, electronic tag systems, inspection terminals, local exhaust ventilation system status acquisition terminals, special gas monitoring system status acquisition terminals, and industrial equipment status acquisition terminals. The memory is used to store fire risk unit data, fire protection benchmark boundary data, dynamic fire load data, fire protection status data, cross-source consistency verification records, state machine boundary parameter sets, state machine evolution records, rectification task records, review records, and de-observation records.

Images