Distributed smoke sensing-temperature sensing composite fire early warning system for underground garage

Abstract

The invention relates to a distributed smoke sensing-temperature sensing composite fire early warning system for an underground garage, which comprises a distributed composite acquisition module, an area data processing unit, a central early warning platform and a linkage control module, and each module realizes data interaction and instruction transmission with an RS485 bus through an LoRa / WiFi dual-mode communication protocol. According to the method, smoke sensing-temperature sensing-object temperature multi-parameter composite monitoring is adopted, D-S evidence theory data fusion is combined, the limitation of traditional single-parameter monitoring is broken through, the false alarm rate is greatly reduced, and based on distributed deployment of the wireless Mesh network, full-coverage non-blind-area monitoring of the underground garage is achieved; meanwhile, the problems that a traditional wired system is complex in wiring and high in maintenance cost are solved, a triangulation positioning algorithm and a dynamic threshold adjustment technology are introduced, fire source 1m precision positioning and self-adaptive early warning in different environments are achieved, the early warning precision is improved, the fault self-diagnosis and standby power supply functions are achieved, and the system is suitable for popularization and application. And normal operation of the system can be ensured when the commercial power is interrupted or the module fails.

Description

Technical Field

[0001] This invention relates to the field of fire early warning system technology, specifically a distributed smoke-temperature composite fire early warning system for underground parking garages. Background Technology

[0002] With the acceleration of urbanization, underground parking garages have become an essential infrastructure for residential communities and commercial complexes. However, underground parking garages are enclosed spaces with poor ventilation, high vehicle density, and often store flammable and explosive materials (such as gasoline and tires). In the event of a fire, the fire can spread rapidly, and smoke can quickly disperse, easily leading to explosions, asphyxiation, and other serious accidents. Furthermore, traditional fire warning systems have many shortcomings. 1. Limitations of single-point monitoring: Existing smoke or heat detectors are mostly deployed independently at single points, which can only monitor local areas and cannot achieve large-scale coverage. The dense parking of vehicles in underground garages leads to many blind spots and makes it easy to miss early fires. 2. High false alarm rate for single parameter: Traditional systems rely on a single parameter such as smoke concentration or temperature to trigger warnings. Interference factors such as vehicle exhaust, humid environment, and mechanical heat in underground parking garages can easily lead to false alarms. Frequent false alarms not only reduce the reliability of the system, but also waste emergency rescue resources. 3. Delayed early warning response: Existing systems mostly use wired connections for data transmission, which are complex and costly to maintain. When detectors malfunction or lines are damaged, data cannot be fed back in a timely manner, resulting in delayed fire warnings and missing the best opportunity to extinguish fires. 4. Insufficient positioning accuracy: After a fire breaks out, traditional systems can only roughly locate the area, not the specific parking space or fire source, which is not conducive to rescuers quickly locating and extinguishing the fire, and prolongs the fire response time.

[0003] Therefore, there is an urgent need to design an underground parking garage fire early warning system with wide coverage, low false alarm rate, fast response speed and accurate positioning, in order to solve the shortcomings of existing technology and improve the fire prevention and control capabilities of underground parking garages. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a distributed smoke-temperature composite fire early warning system for underground parking garages. By deploying composite acquisition modules in a distributed manner, integrating multi-dimensional data fusion analysis, and utilizing wireless communication technology, it achieves early and accurate monitoring of fires in underground parking garages, low false alarm warnings, and precise location of fire sources, thereby improving the safety and reliability of fire prevention and control in underground parking garages.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a distributed smoke-heat composite fire early warning system for underground parking garages, characterized in that it includes a distributed composite acquisition module, a regional data processing unit, a central early warning platform, and a linkage control module. Each module achieves data interaction and command transmission through a LoRa / WiFi dual-mode communication protocol and an RS485 bus, forming a complete closed loop of "monitoring-analysis-early warning-linkage". The distributed composite acquisition module uses a single parking space in the underground garage as the basic monitoring unit. One module is fixed at a pre-set installation point on the ceiling or column above each parking space. Additional modules are added in key risk areas such as passage intersections, ventilation openings, power distribution boxes, and charging piles to ensure that there are no blind spots in the monitoring coverage. Each module self-organizes through a wireless Mesh network and supports dynamic routing and AES-128 data encryption. The regional data processing unit is divided into jurisdictional areas with 50-100 acquisition modules, and has a built-in industrial-grade PLC and a storage module of more than 16GB. It is responsible for data reception, multi-source fusion and fault diagnosis. The central early warning platform is built on a dual-redundant industrial server, equipped with 64GB of memory and a hybrid storage array, supports stable operation 24 / 7, and has functions such as data visualization, fire classification and precise fire source location. The linkage control module is connected to the existing fire protection and security system of the garage through a relay / analog interface, and can automatically trigger directional fire control and evacuation actions according to the fire level.

[0006] Furthermore, the distributed composite acquisition module is encapsulated in an IP65 waterproof and dustproof shell, internally integrating a smoke detection unit, a temperature detection unit, a data preprocessing unit, a wireless communication unit, and a backup power supply unit. The smoke detection unit is a laser scattering sensor with a detection range of 0.01-20%obs / m and a resolution of 0.01%obs / m, and incorporates a built-in exhaust gas / dust interference recognition algorithm. The temperature detection unit integrates a thermocouple (-40℃-125℃, accuracy ±0.5℃) and an infrared sensor (-20℃-300℃, accuracy...). ±1℃, respectively monitoring the temperature of the environment and the target object; the data preprocessing unit uses an STM32L476 microcontroller (32-bit ARM Cortex-M4 core, 80MHz main frequency), which reduces noise and standardizes data through Kalman filtering algorithm; the wireless communication unit supports dual-mode switching between LoRa (433 / 868MHz band, maximum transmission 1km) and WiFi (150Mbps rate); the backup power unit is a 5000mAh lithium battery pack, with a battery life of ≥48h when the mains power is interrupted, and has a float charging protection function.

[0007] Furthermore, the regional data processing unit uses a Siemens S7-1200 series PLC, which has three core functions: First, data fusion function, which uses the DS evidence theory algorithm to construct a trust assignment function from three-dimensional data of smoke concentration, ambient temperature, and target temperature, and then synthesizes it to determine fire risk, improving the accuracy of identification by more than 85% compared to single parameter identification; Second, threshold dynamic adjustment function, which has a built-in multi-scenario environmental database (including data from morning and evening peak hours, rainy / dry weather, etc.), and analyzes historical data through random forest algorithm to automatically adjust the warning threshold, such as raising the smoke threshold by 20% during morning and evening peak hours to avoid false alarms from exhaust gas; Third, fault self-diagnosis function, which uses heartbeat detection (three consecutive no feedbacks determine communication failure), power supply voltage monitoring (mains power 180-240V / backup power 3.6-4.2V), and data consistency verification (comparing with data from surrounding modules) to identify module faults in real time, generate fault information with number, type, and location, and upload it.

[0008] Furthermore, the central early warning platform adopts a B / S architecture and an Android / iOS mobile app. The hardware is a Dell PowerEdge R750 server (dual-socket Xeon Gold processor), and the software is developed based on Windows Server 2019. It features data visualization capabilities, displaying module status through electronic floor plan color coding (green / yellow / red representing normal / early warning / alarm respectively). Historical data storage is ≥1 year and supports multi-dimensional query and export. The fire classification function divides risks into three levels: Level 1 warning (slight smoke / temperature anomalies) triggers audible and visual alerts and SMS notifications; Level 2 alarm (smoke ≥5%obs / m and temperature ≥60℃ or target temperature ≥150℃) activates the garage's audible and visual alarms and some linkages; Level 3 alarm (smoke ≥10%obs / m and temperature ≥80℃ or target temperature ≥250℃) triggers full linkage and automatically dials 119. Fire source location uses a triangulation algorithm, selecting at least three near-source modules and calculating coordinates based on the deviation between smoke (weight 0.6) and temperature (weight 0.4), achieving a positioning accuracy ≤1m, while also predicting the fire spread trend.

[0009] Furthermore, the linkage control module uses a Schneider M262 PLC, which is connected to the central early warning platform via an RS485 bus (9600bps baud rate). The linkage logic is divided into two levels: when a level two alarm occurs, the ventilation / smoke exhaust fan in the fire source area is activated, the ventilation vents in non-fire areas are closed, emergency lighting and access control at entrances and exits are turned on, and the garage entrance gate is closed; when a level three alarm occurs, the sprinkler system in the fire source area is activated, and the locations of the three nearest fire hydrants are marked on the platform. The module has a manual control panel, which allows manual start and stop of associated equipment in case of system failure. After the linkage command is executed, a log is generated and stored in the central platform for easy traceability.

[0010] Furthermore, the wireless mesh network supports self-organizing network expansion and dynamic routing. Newly added acquisition modules automatically join the network after power-on without the need for parameter reconfiguration. Communication data is encrypted using AES-128, and each module is assigned a unique device number to ensure a one-to-one correspondence between data and modules. The network also has a bandwidth adaptive function, automatically switching to LoRa low-power mode when the amount of monitored data is small, and switching to WiFi high-speed mode when transmitting historical data or images, reducing power consumption by more than 30% and extending module battery life.

[0011] Furthermore, the specific steps of the DS evidence theory data fusion are as follows: Step 1, determine the identification framework of "no fire / suspected fire / early fire / fire spread"; Step 2, take smoke, ambient temperature, and target temperature as independent evidence bodies, and calculate the trust level of each evidence body for different states according to parameter deviation; Step 3, use the DS synthesis rules to synthesize evidence bodies pairwise to obtain a comprehensive trust allocation; Step 4, if the difference between the highest and second highest trust levels is ≥0.2, output the judgment result; otherwise, mark it as "to be confirmed" and combine it with the surrounding module data for secondary analysis to avoid misjudgment.

[0012] Furthermore, the triangulation algorithm calculation process is as follows: Step 1, retrieve the coordinates of the three most recent modules. (Establish a coordinate system with the fixed corner of the garage wall as the origin); Step 2, calculate the smoke deviation ΔS and temperature deviation ΔT for each module, according to... Determine the weights; Step 3, through Calculate the coordinates of the fire source; Step 4: If the error is >1m, increase the number of modules (4-5) and recalculate until the accuracy meets the standard.

[0013] Furthermore, it also includes a regular self-check function, with the self-check cycle set through the central platform (default 2:00-2:30 daily). The self-test process includes: the central platform sending instructions to the regional units; the regional units detecting the sensor response speed (delay ≤ 500ms), data transmission delay, and backup power supply voltage (≥ 3.6V) of the acquisition module; the regional units self-testing the communication interface, storage module, and data processing capabilities; all self-test data being aggregated to the central platform, with maintenance notifications pushed when abnormalities occur, and a qualified report generated and stored when normal conditions are met, ensuring system reliability.

[0014] Furthermore, it also includes a remote linkage unit for the fire command center. This unit connects to the central early warning platform via the 4G / 5G public network, supporting the automatic uploading of three-level alarm information (including garage address, fire source coordinates, fire level, and real-time monitoring data) to the local fire command center system. At the same time, it receives remote control commands issued by the fire command center, such as adjusting the sprinkler range and opening fire exit access control. After the commands are verified by the central platform, they are forwarded to the linkage control module for execution. The unit has a built-in data caching function, which stores alarm information when the network is interrupted and automatically retransmits it after the network is restored, ensuring that the fire department can obtain the fire dynamics in real time and shorten the rescue response time.

[0015] Compared with the prior art, the technical solution of this application has the following beneficial effects: 1. This invention employs multi-parameter composite monitoring of smoke, temperature, and object temperature, combined with DS evidence theory data fusion, to overcome the limitations of traditional single-parameter monitoring, significantly reducing the false alarm rate. Based on the distributed deployment of a wireless mesh network, it achieves full coverage monitoring of underground parking garages without blind spots. At the same time, it solves the problems of complex wiring and high maintenance costs of traditional wired systems. It introduces triangulation algorithm and dynamic threshold adjustment technology to achieve 1m accuracy in locating fire sources and adaptive early warning in different environments, improving the accuracy of early warning.

[0016] 2. The system of this invention is flexible in deployment. The number of composite acquisition modules and regional data processing units can be flexibly adjusted according to the area and number of parking spaces of the underground garage, adapting to garages of different sizes. It has fault self-diagnosis and backup power functions to ensure that the system can still operate normally when the mains power is interrupted or the module fails, thus improving reliability. The linkage control module can be seamlessly connected with the existing garage fire protection system without large-scale modification of existing facilities, reducing application costs. Attached Figure Description

[0017] Figure 1 This is a framework diagram of the distributed smoke-temperature composite fire early warning system for underground parking garages according to the present invention; Figure 2 This is a data processing flowchart of the distributed smoke-temperature composite fire early warning system for underground parking garages according to the present invention; Figure 3 This is a flowchart illustrating the decision-making and early warning process of the distributed smoke-temperature composite fire early warning system for underground parking garages according to the present invention. Figure 4 This is a data loop flowchart for the distributed smoke-temperature composite fire early warning system for underground parking garages according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1-4 The distributed smoke-heat composite fire early warning system for underground parking garages in this embodiment includes a distributed composite acquisition module, a regional data processing unit, a central early warning platform, and a linkage control module. The specific structure and connection relationship of each part are as follows: (a) Distributed composite acquisition module Deployment method: A grid-based distributed deployment is adopted, with each parking space in the underground garage as the basic unit. A composite acquisition module is installed on the ceiling or column above the parking space. At the same time, acquisition modules are added in key areas such as garage passage intersections, ventilation openings, and power distribution boxes to ensure no blind spots in monitoring. The modules are connected to each other through a wireless mesh network to form a full-coverage monitoring network. Hardware structure: Each composite acquisition module includes a smoke detection unit, a temperature detection unit, a data preprocessing unit, a wireless communication unit, and a backup power supply unit; Smoke detection unit: It adopts a laser scattering smoke sensor with a detection range of 0.01-20%obs / m and a resolution of 0.01%obs / m. It can identify tiny smoke particles and avoid interference from vehicle exhaust and other sources. Temperature sensing unit: integrates thermocouple and infrared temperature sensor. The thermocouple detects the ambient temperature (-40℃-125℃, accuracy ±0.5℃), and the infrared sensor detects the temperature of the vehicle surface and surrounding objects (-20℃-300℃, accuracy ±1℃), realizing dual temperature monitoring of the environment and the target. Data preprocessing unit: Built-in STM32L476 microcontroller, which filters (Kalman filter algorithm) and reduces noise in smoke and temperature sensor data, and removes abnormal interference data; Wireless communication unit: Supports LoRa and WiFi dual-mode communication. LoRa is used for long-distance (maximum 1km) low-power data transmission between modules, and WiFi is used for high-speed communication with the regional data processing unit. Backup power unit: Powered by lithium battery, normally powered by the garage mains power, automatically switches to backup power when the mains power is interrupted, with a battery life of ≥48 hours. (ii) Regional Data Processing Unit Number of modules: Divide the underground garage into areas, with 50-100 composite acquisition modules corresponding to 1 area data processing unit, which is installed in the garage area duty room or equipment room; Functionality implementation: Data reception and fusion: Receives smoke and temperature data from all composite acquisition modules within the jurisdiction, and uses a multi-dimensional data fusion algorithm (DS evidence theory) to fuse and analyze three parameters: smoke concentration, ambient temperature, and object temperature to determine whether there is a fire risk. Dynamic threshold adjustment: The built-in garage environment database stores normal smoke and temperature data under different time periods (such as morning and evening rush hours, night) and different weather conditions (humid and dry). The warning threshold is dynamically adjusted through machine learning algorithms (random forest) to reduce the false alarm rate. For example, when vehicle exhaust increases during morning and evening rush hours, the smoke warning threshold is automatically increased to avoid false alarms. Fault self-diagnosis: Real-time monitoring of the communication status, power supply status, and sensor accuracy of the composite acquisition module. If the module experiences communication interruption or sensor failure, fault alarm information is immediately generated and uploaded to the central early warning platform. (III) Central Early Warning Platform Hardware architecture: Built on industrial-grade servers, equipped with redundant power supplies and storage arrays to ensure stable system operation; Software features: Data visualization and storage: The system adopts a B / S architecture and displays the underground parking garage floor plan in real time through a web interface. It marks the location of each composite acquisition module and the monitoring data (smoke concentration and temperature values ​​are distinguished by color, with green for normal, yellow for warning, and red for alarm). It also stores historical data (storage time ≥ 1 year) and supports data query and export. Fire risk assessment: Based on the fused data results, fire risk is divided into three levels: Level 1 Warning (Suspected Fire): Smoke concentration exceeds the normal range but does not reach the alarm threshold, or temperature rises abnormally (ambient temperature is 5℃-10℃ higher than the normal average). The platform issues an audible and visual alert and pushes the warning information to the mobile phones of management personnel. Level 2 alarm (early fire): When the smoke concentration reaches the alarm threshold (≥5%obs / m) and the ambient temperature is ≥60℃, or the object temperature is ≥150℃, the platform will activate the garage's audible and visual alarm and simultaneously push alarm information to the property management and fire control room. Level 3 Emergency Alarm (Fire Spread): When the smoke concentration is ≥10%obs / m and the ambient temperature is ≥80℃, or the object temperature is ≥250℃, the platform will automatically trigger the linkage control module and dial the 119 fire alarm number (automatically broadcasting the garage location and fire source area). Precise fire source location: By using a multi-module data triangulation algorithm, combined with the position coordinates of each composite acquisition module and the degree of data anomaly, the coordinates of the fire source are calculated with a positioning accuracy of ≤1m (i.e., accurate to a specific parking space), and the location of the fire source and its spread trend are marked on the platform map. (iv) Linkage Control Module Connections: Connected to the central early warning platform via RS485 bus, and also connected to the ventilation system, sprinkler system, smoke exhaust system, emergency lighting system and access control system of the underground parking garage; Linkage Functions: When a level 2 alarm is triggered, the ventilation and smoke exhaust fans in the fire area will be automatically activated, and the ventilation openings in other areas will be closed to prevent the spread of smoke. At the same time, the emergency lighting system will be turned on, the garage exit access control will be opened, and personnel will be guided to evacuate. When a Level 3 emergency alarm is triggered, in addition to the actions mentioned above, the sprinkler system in the fire area will be automatically activated to suppress the spread of the fire; if there is a fire hydrant in the garage, the platform will simultaneously display the location of the nearest fire hydrant for the convenience of rescue personnel. III. Work Process Real-time monitoring: The distributed composite acquisition module collects real-time data on smoke concentration, ambient temperature, and object temperature in the underground parking garage. After preprocessing, the data is transmitted to the regional data processing unit via a wireless mesh network. Data fusion analysis: After receiving the data, the regional data processing unit uses the DS evidence theory to perform multi-parameter fusion and combines it with dynamically adjusted early warning thresholds to determine the fire risk level; Early warning and location: If the warning is determined to be a Level 1 warning or a Level 2 / 3 alarm, the regional data processing unit will upload the results to the central early warning platform. The platform will generate visual early warning information and determine the location of the fire source through a triangulation algorithm. Linkage control: Based on the fire level, the central early warning platform sends instructions to the linkage control module to trigger the corresponding ventilation, sprinkler, lighting and other system actions, and pushes alarm information to relevant personnel; Fault Handling: If the regional data processing unit detects a fault in the composite acquisition module, it immediately uploads the fault information to the central early warning platform, which then prompts the management personnel to carry out repair or replacement.

[0020] Example I. System Deployment The underground parking garage of a commercial complex (10,000 m²) 2 Taking a parking lot with 200 spaces as an example, the system deployment is as follows: Distributed composite acquisition module: One module is installed above each of the 200 parking spaces, and additional modules are installed at 10 intersections, 5 ventilation openings, and 3 power distribution boxes, for a total of 218 modules. The modules form a self-organizing network through a wireless Mesh network, with the LoRa communication frequency set to 433MHz and the WiFi communication rate set to 150Mbps. Regional data processing unit: Divided into 3 regions (approximately 70 modules in each region), with 1 processing unit installed between the devices in each region, using an industrial-grade PLC (model S7-1200) and equipped with a 16GB storage module. Central Early Warning Platform: It uses a Dell PowerEdge R750 server with dual redundant power supplies and runs Windows Server 2019. The early warning platform software is developed in Java and supports access via web (Chrome, Edge browsers) and mobile (Android / iOS APP).

[0021] Linkage control module: adopts Schneider M262PLC, which is connected to the existing ventilation system (10 fans), sprinkler system (50 sprinkler heads), smoke exhaust system (8 smoke exhaust fans), emergency lighting (200 emergency lights) and 3 exit access control in the garage. II. System Testing False alarm rate test: During morning and evening rush hours (frequent vehicle entry and exit, high exhaust gas concentration) and in humid weather (relative humidity ≥85%), the system operated continuously for 72 hours, and the number of false alarms was 0, which meets the design requirements. Early warning response test: When simulating a vehicle spontaneous combustion (using a heating device to raise the temperature of the object to 200°C and generate smoke at the same time), the system detects the abnormality within 15 seconds, triggers a level 2 alarm within 30 seconds, and starts the ventilation and smoke exhaust system within 1 minute, demonstrating a fast early warning response speed. Positioning accuracy test: In simulated fires at different parking spaces, the system positioning error was ≤0.8m, which is accurate to the specific parking space, and the positioning accuracy meets the design requirements.

[0022] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0023] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A distributed smoke-heat composite fire early warning system for underground parking garages, characterized in that, It includes a distributed composite acquisition module, a regional data processing unit, a central early warning platform, and a linkage control module. Each module achieves data interaction and command transmission through the LoRa / WiFi dual-mode communication protocol and RS485 bus, forming a complete closed loop of "monitoring-analysis-early warning-linkage". The distributed composite acquisition module uses a single parking space in the underground garage as the basic monitoring unit. One module is fixed at a pre-set installation point on the ceiling or column above each parking space. Additional modules are added at key risk areas such as passageway intersections, ventilation openings, power distribution boxes, and charging piles to ensure monitoring coverage without blind spots. Each module self-organizes through a wireless Mesh network and supports dynamic routing and AES-128 data encryption. The regional data processing unit is divided into jurisdictional areas with 50-100 acquisition modules, and has a built-in industrial-grade PLC and a storage module of more than 16GB. It is responsible for data reception, multi-source fusion and fault diagnosis. The central early warning platform is built on a dual-redundant industrial server, equipped with 64GB of memory and a hybrid storage array, supports stable operation 24 / 7, and has functions such as data visualization, fire classification and precise fire source location. The linkage control module is connected to the existing fire protection and security system of the garage through a relay / analog interface, and can automatically trigger directional fire control and evacuation actions according to the fire level.

2. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, The distributed composite acquisition module is encapsulated in an IP65 waterproof and dustproof shell and integrates a smoke detection unit, a temperature detection unit, a data preprocessing unit, a wireless communication unit, and a backup power supply unit. The smoke detection unit is a laser scattering sensor with a detection range of 0.01-20%obs / m and a resolution of 0.01%obs / m, and it has a built-in exhaust gas / dust interference recognition algorithm; the temperature detection unit integrates a thermocouple and an infrared sensor to monitor the ambient temperature and the temperature of the target object, respectively. The data preprocessing unit uses an STM32L476 microcontroller to reduce noise and standardize data through a Kalman filter algorithm; the wireless communication unit supports LoRa and WiFi dual-mode switching; the backup power unit is a 5000mAh lithium battery pack, which provides ≥48 hours of battery life during mains power outages and has a float charging protection function.

3. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, The regional data processing unit uses a Siemens S7-1200 series PLC, which has three core functions: 1) Data fusion function: Through the DS evidence theory algorithm, a trust assignment function is constructed from three-dimensional data of smoke concentration, ambient temperature and target temperature. After synthesis, the fire risk is determined, which improves the accuracy of identification by more than 85% compared with single parameter identification. 2) Dynamic threshold adjustment function: It has a built-in multi-scenario environment database and analyzes historical data through the random forest algorithm to automatically adjust the warning threshold; 3) Fault self-diagnosis function: Through heartbeat detection, power supply voltage monitoring, and data consistency verification, it identifies module faults in real time, generates fault information containing number, type, and location, and uploads it.

4. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, The central early warning platform adopts a B / S architecture and an Android / iOS mobile APP. The hardware is a Dell PowerEdge R750 server, and the software is developed based on Windows Server 2019. It has data visualization capabilities, which use color coding on an electronic floor plan to display the status of modules. Historical data is stored for ≥1 year and supports multi-dimensional query and export. The fire classification function divides risks into three levels: Level 1 warning (slight smoke / temperature abnormality) triggers audio-visual alerts and SMS notifications; Level 2 alarm (smoke ≥ 5%obs / m and temperature ≥ 60℃ or target temperature ≥ 150℃) activates the garage's audible and visual alarm and some related systems. A Level 3 alarm (smoke ≥10%obs / m and temperature ≥80℃ or target temperature ≥250℃) triggers full linkage and automatically dials 119; The fire source location uses a triangulation algorithm, selecting more than three near-source modules, and calculating the coordinates based on the deviation of smoke (weight 0.6) and temperature (weight 0.4). The location accuracy is ≤1m, and the fire spread trend is predicted at the same time.

5. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, The linkage control module uses a Schneider M262 PLC, which is connected to the central early warning platform via an RS485 bus. The linkage logic is divided into two levels: In the event of a Level 2 alarm, activate the ventilation / smoke exhaust fan in the fire area, close the ventilation openings in non-fire areas, turn on emergency lighting and access control at entrances and exits, and close the garage entrance gate. When a level 3 alarm is triggered, the sprinkler system in the fire source area will be activated, and the locations of the three nearest fire hydrants will be marked on the platform. The module has a manual control panel, which allows manual start and stop of associated equipment in case of system failure. Logs are generated after the linkage commands are executed and stored on the central platform for easy traceability.

6. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, The wireless Mesh network supports self-organizing network expansion and dynamic routing. Newly added acquisition modules automatically join the network after being powered on, without the need to reconfigure parameters. Communication data is encrypted using AES-128, and each module is assigned a unique device number to ensure that data corresponds one-to-one with module. The network also features bandwidth adaptive functionality, automatically switching to LoRa low-power mode when the amount of monitored data is small, and switching to WiFi high-speed mode when transmitting historical data or images, reducing power consumption by more than 30% and extending the module's battery life.

7. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 3, characterized in that, The specific steps of the DS evidence theory algorithm data fusion are as follows: Step 1: Establish the identification framework of "no fire / suspected fire / early fire / fire spread"; Step 2: Using smoke, ambient temperature, and target temperature as independent evidence, calculate the confidence level of each evidence for different states based on parameter deviations. Step 3: Use the DS synthesis rule to synthesize evidence pairs to obtain a comprehensive trust allocation; Step 4: If the difference between the highest and second highest trust levels is ≥0.2, output the judgment result; otherwise, mark it as "to be confirmed" and combine it with the data from surrounding modules for secondary analysis to avoid misjudgment.

8. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, The triangulation algorithm calculation process is as follows: Step 1: Retrieve the coordinates of the 3 most recent modules (Establish a coordinate system with the fixed corner of the garage wall as the origin); Step 2: Calculate the smoke deviation for each module. Temperature deviation ,according to Determine the weights; Step 3, through Calculate the coordinates of the fire source; Step 4: If the error is >1m, increase the number of modules (4-5) and recalculate until the accuracy meets the standard.

9. The distributed smoke-heat composite fire early warning system for underground parking garages according to claim 1, characterized in that, It also includes a regular self-check function, and the self-check cycle can be set through the central platform (default is 2:00-2:30 every day). The self-test process includes: the central platform sending instructions to the regional units, and the regional units detecting the sensor response speed (delay ≤ 500ms), data transmission delay, and backup power supply voltage (≥ 3.6V) of the acquisition module. Area unit self-test communication interface, storage module and data processing capabilities; All self-test data is aggregated to the central platform. When an anomaly occurs, a maintenance notification is pushed out. When the system is functioning normally, a qualified report is generated and stored to ensure system reliability.

10. The distributed smoke-temperature composite fire early warning system for underground parking garages according to claim 1, characterized in that, It also includes a remote linkage unit for the fire command center, which connects to the central early warning platform via the 4G / 5G public network and supports the automatic uploading of level 3 alarm information to the local fire command center system; at the same time, it receives remote control commands issued by the fire command center, and the commands are forwarded to the linkage control module for execution after being verified by the central platform; The unit has a built-in data caching function that stores alarm information when the network is interrupted and automatically retransmits it after the network is restored, ensuring that the fire department can obtain real-time fire dynamics and shorten the rescue response time.

Images