1. Introduction
Fires not only inflict significant damage on the immediate area but also trigger environmental hazards, such as widespread smoke and the release of chemical gases. These factors can lead to extensive destruction and irreversible consequences, posing threats to both human populations and the environment. On April 29, 2020, a fire broke out at a logistics warehouse center[1] in Icheon, Gyeonggi Province, South Korea. Investigations revealed that a safety oversight likely led to the release of flammable gases into the air; these gases came into contact with sparks from other construction units, triggering multiple explosions. The incident resulted in at least 38 deaths and 10 injuries, with ten consecutive explosions, intense flames, thick smoke, and the spread of toxic gases. On June 17, 2021, a fire occurred at the Coupang Logistics Center[2] in South Korea, burning for five days and covering an area of over 127,000 square meters. Tragically, a firefighter lost his life during this blaze. On August 19, 2021, a fire broke out at the RAM Enterprise warehouse[3,4] in Coventry, UK, where household goods were stored over an area of 2,500 square meters; the extensive smoke severely affected the lives of nearby residents.
Improving fire management strategies necessitates controlling heat sources prior to ignition, thereby reducing the likelihood of a fire. To address this challenge, existing fire-fighting technologies will be analyzed, and an ultra- early fire prevention system designed to "nip fires in the bud" will be proposed. This system will augment current fire-fighting procedures and detail advancements in related technologies. The approach is particularly vital in large facilities such as libraries, logistics centers, and chemical plants, where insufficient manpower can lead to poor management and an increased risk of fire. In unattended automated mechanical fields, fires may occur unexpectedly, underscoring the need for proactive measures. The proposed preventive fire safety technology monitors the environment before a fire occurs, promptly identifies heat sources upon detecting an abnormal event, and notifies relevant units for appropriate action. This proactive system helps prevent the spread of fire, thereby reducing potential property damage and casualties.
The contributions of this research are as follows:
1. A comprehensive collection and analysis of existing fire protection systems and technologies, with a focus on their applications in fire scenarios.
2. The development of IFPPS, which integrates data collection, intelligent real-time computing, and information communication across multiple platforms. This integration enhances fire warning accuracy and timeliness, thereby reducing fire losses.
3. Preliminary experiments demonstrate that IFPPS detects anomalies more rapidly than traditional systems.
This paper is organized as follows: Section 1 presents the background and research contributions of this study. Section 2 discusses the current state of research techniques. Section 3 elaborates on the system architecture employed in this study. Section 4 presents a proof of the concept of the proposed system. Finally, Section 5 concludes the paper.
2. Related Work
The timeline of a fire can be divided into multiple stages[5], with the active fire phase generally consisting of four stages: Incipient, Growth, Fully Developed, and Decay[6]. As shown in Figure 1, the graph is based on the heat release and time. There is also a momentary state between the Growth and Fully Developed stages known as Flashover. This state can be influenced by the environment and does not necessarily occur, but most fires follow the pattern presented in Figure 1. Distinct stages of a fire call for tailored rescue strategies and specialized equipment. This section offers a comprehensive analysis of the firefighting systems endorsed by international organizations across various fire stages, as well as the diverse technologies employed in fire safety operations.
2.1 Overview of international fire safety management
The standards established by the National Fire Protection Association (NFPA) and the International Fire Code (IFC) are crucial for ensuring robust fire safety measures. These internationally recognized guidelines provide comprehensive technical definitions and systematic approaches for preventing, detecting, and mitigating fire hazards. They cover a range of systems-from automatic detection and alarm verification to dynamic fire extinguishing solutions and hazardous materials compartmentation-designed to protect lives and property. Implementing NFPA and IFC standards not only harmonizes fire safety practices across diverse environments, but also enhances community resilience by reducing risks and ensuring prompt, effective emergency responses, thereby improving overall safety outcomes.
The National Fire Protection Association (NFPA) is dedicated to addressing complex safety challenges[7], including fire prevention, wildfire preparedness, electrical safety, and hazardous materials management, all aimed at reducing community risks and enhancing public safety. The association has proposed a series of technical definitions and system development specifications related to fire safety. NFPA 72[8] outlines two principal systems for fire notifications and related communications: the Fire Alarm System, which monitors and reports fire alarms or the status of signal-initiating devices and responds accordingly, and the Emergency Communications System, which disseminates information regarding emergencies such as fires, human-caused incidents, natural disasters, and accidents. In its efforts to curtail fire occurrences and facilitate early detection, NFPA 76[9] introduces two concepts: Early Warning Fire Detection (EWFD) Systems, which employ smoke, heat, or flame detectors to identify fires before high heat endangers human life or disrupts telecommunications services, and Very Early Warning Fire Detection (VEWFD) Systems, which detect low-heat fires before they can threaten telecommunications services. Additionally, NFPA 101[6] describes the temporary measure of Fire Watch[10], where one or more individuals are assigned to a designated area to ensure timely emergency notifications, prevent fires, or extinguish small fires promptly, thereby protecting the public from fire-related hazards and life safety risks.
The International Fire Code (IFC) encompasses a broad spectrum of fire safety measures. It addresses not only firefighting systems but also the compartmentation of hazardous materials, pre-disaster prevention measures, and the development of comprehensive emergency plans and preparations during fire incidents. The code incorporates several systems featuring automatic fire extinguishing and alarm functionalities. The first system, the Alarm Verification Feature[11], integrates automatic fire detection with alarm signaling. Smoke detectors promptly report alarm conditions to ensure the swift reception of valid alerts. The second system, the Automatic Fire-extinguishing System[11], is designed to automatically detect fires and discharge extinguishing agents directly into the affected area. The third system, the Automatic Water Mist System[11], employs a water supply, a pressure source, and a distribution piping network equipped with supplementary nozzles to deliver water as fine droplets, thereby controlling, suppressing, or even extinguishing fires. Compared to conventional sprinkler systems[12], water mist systems can address a broader range of fire types while using less water. Furthermore, fire protection systems, including sprinkler systems, standpipe systems, and fire alarm systems[13,14], are designed to detect, control, and extinguish fires or smoke, while simultaneously alerting building occupants.
The names of various fire protection systems will be organized and concisely described based on their functions, areas of application, and the specific fire stages they address, as outlined in Table 1. Static fire protection systems are permanently installed in designated locations to collect environmental data, including carbon monoxide, smoke, and temperature readings. In contrast, dynamic fire protection systems are flexible and adaptable to changing environments and varying scales. Among these systems, Fire Watch is primarily used to monitor previously extinguished fire sources to prevent rekindling, providing a clear distinction between this and other fire protection systems.
2.2 Integration of intelligent technologies in fire safety management
In today's rapidly advancing technological landscape, numerous methods exist to enhance fire safety management. This necessitates a review of recent literature on related technologies to promote awareness in this field. Lin et al. [15] have investigated Lithium-Ion Batteries (LIBs) and proposed several fire-fighting strategies based on their characteristics. LIBs are prone to fire and explosion under conditions such as overheating, overcharging, crushing, puncturing, and short-circuiting. The combustion of LIBs exhibits the following characteristics: intense exothermic reactions within the battery can cause a rapid temperature increase; when battery pressure reaches a threshold, the expelled flammable gases and electrolytes can form explosive flames; the heat release rate of the flames can reach up to 100 kW; closely packed batteries can facilitate rapid flame propagation; and the chemical reactions during combustion produce large amounts of toxic and flammable gases, posing risks of reignition and explosion. Based on the characteristics of LIBs, the following fire-fighting strategies have been proposed. The first strategy, Fire Detection Tube Technology, employs a device that releases a fire extinguishing agent when a preset temperature is reached, thereby enabling point-to-point suppression of the fire source. The second strategy, Collaborative Fire-Extinguishing Method, combines the advantages of multiple fire extinguishers to enhance fire-fighting effectiveness. The third strategy, Intermittent Spray, involves the periodic application of fire extinguishing agents to prolong suppression duration and improve agent efficiency. The fourth strategy, Suppression Microcapsule, entails affixing microcapsules to the battery surface; when the battery temperature rises abnormally, the microcapsules release a fire extinguishing agent to curb the temperature increase. The fifth strategy, Ventilation and Explosion Suppression, addresses the significant amounts of toxic and flammable gases that may remain in a space after flames are extinguished. When oxygen enters the previously enclosed area, it can trigger explosions and fires; therefore, increased ventilation or the release of inert gases is recommended to reduce the likelihood of explosions.
A. Z. Erkinjonovich et al.[16] designed an automatic shut-off valve to prevent fires and explosions resulting from natural gas leaks. This valve automatically terminates the gas supply within a short period, thereby preventing leakage and reducing fire risk. J. Balen et al.[17] developed the FireBot autonomous monitoring robot system for early fire detection and autonomous suppression. FireBot is equipped with capabilities for autonomous navigation, obstacle avoidance, video surveillance, fire prevention and detection, and fire suppression. By detecting fires early and extinguishing them autonomously, it mitigates the risk of casualties and property damage. FireBot is programmed with various risk-level protocols, ranging from notifying the maintenance department for inspection in low-risk scenarios to activating fire suppression procedures and alerting the fire department in high-risk situations. R. Pohle et al.[18] developed an indoor fire prevention system based on micro-sensor drones and fixed sensor nodes. This system employs thermal imaging along with sensors for temperature, humidity, and related gas components to conduct preventive analyses of indoor environments and identify potential fire incidents. M. M. Rahman et al.[19] conducted a qualitative failure analysis of an IoT-supported industrial fire detection and prevention system. The system comprises a perception layer, data processing layer, and transmission layer, all managed by a Smart Hub System (SHS). The perception layer collects environmental information and transmits it to the SHS; upon fire detection, the SHS activates sprinklers, sounds alarms, and sends text messages to users.
Deep learning-based object recognition technology can further enhance fire safety management. A. Biswas et al.[20] developed an early fire hazard monitoring system based on deep learning, employing an improved Inception-v3 model for image detection of fires and smoke. F. M. Talaat et al.[21] proposed a smart fire detection algorithm utilizing YOLO-v8 for object detection, which identifies fire-related categories such as "flame," "smoke," or "embers" and alerts the relevant authorities while outputting the detected images. L. Zheng et al.[22] developed a real-time analysis system for fire detection and early warning in power systems. This system uses monitoring equipment for object detection, leveraging smoke and flame detection to provide early fire warnings. Additionally, a web-based user interface has been developed for real-time fire monitoring.
The reviewed literature demonstrates that integrating building-related information with existing technologies can lead to the proposal of a real-time fire protection system structure aimed at reducing fire risk. This system emphasizes real-time connectivity and monitoring between building management systems and local fire department control consoles. C.-J. Hsiao et al.[23] employ Building Information Modeling (BIM) and simulation technologies for disaster prevention, utilizing Virtual Reality (VR), Augmented Reality (AR), and Simulated Reality (SR) to simulate and validate disaster prevention, evacuation, and rescue operations. C. Zhang[24] examines the risks, challenges, and prevention strategies associated with structural fires. He identifies human activity as the primary cause, with key fire-causing actions including cooking, heating, smoking, electrical applications, and arson. Fire prevention measures, such as fire doors, fire curtains, fire and smoke dampers, and fire-resistant building materials, can help prevent ignition and limit the early spread of fire. Additionally, Zhang notes that intelligent fire detection systems contribute to reducing fire risk. Z. Liu et al.[25] analyze fire hazards and preventive measures from the perspectives of villages, streets, and buildings, taking into account factors such as local building materials, living habits, and wind direction. They propose that fire simulations, adjustments in building locations, improvements in road design, and enhancements in fire-resistant building materials can collectively mitigate fire risk. Table 2 compiles the various technologies applied in the field of firefighting based on the above.
3. Intelligent Fire Prevention Patrol System
Based on the analysis of existing fire protection systems and related literature, it is apparent that neither practical applications nor the literature offer a clear definition of an intelligent autonomous patrol system. This paper proposes an Intelligent Fire Prevention Patrol System (IFPPS)[26] composed of three key components: data collection, intelligent real- time computing and recognition, and information communication. As shown in Figure 2, the system can be implemented on various patrol equipment for monitoring public environments, and it is capable of detecting fire sources at a very early stage to achieve pre-ignition prevention, thereby avoiding casualties and property damage and enhancing public safety.
At the data collection stage, as shown in Figure 2(a), environmental data is comprehensively gathered using various heterogeneous sensing devices, such as sensors, cameras, and thermal imagers. This data includes temperature, humidity, various gas concentrations, as well as thermal imaging and RGB images. These sensors precisely detect various environmental parameters, ensuring comprehensive and accurate data acquisition. These sensing devices can be installed on a variety of carriers, ranging from large machinery to handheld devices. For large transportation vehicles, sensing devices can be mounted on heavy machinery such as forklifts. These forklifts move around factories or warehouses and are equipped with high-resolution cameras and thermal imagers. This equipment provides extensive and detailed environmental data, assisting management personnel in monitoring conditions over large areas. These devices not only capture and record real-time on-site imagery but also automatically analyze anomalies, offering early warning and reporting functions that significantly enhance management efficiency and safety. For small carriers, sensing devices are installed on handheld tools or small platforms such as safety helmets used by inspection personnel. For example, embedding sensing devices in flashlights enables inspectors to conduct effective environmental monitoring during routine checks. These handheld devices can be equipped with night vision, thermal imaging technology, and high-sensitivity sensors, ensuring clear observation in dark or low-visibility environments. Additionally, handheld sensing devices can connect to mobile devices for real-time data uploading and remote monitoring, allowing backend management personnel to stay immediately informed about on-site conditions.
Devices equipped with various sensors feed the collected heterogeneous environmental data into the intelligent real-time computation and recognition module of the IFPPS, as shown in Figure 2(b). Within the IFPPS concept, this module can leverage advanced algorithms, such as those based on expert system, artificial intelligence machine learning, deep learning algorithms, and generative AI applications, enabled by continuously improving hardware and software computing capabilities. This allows for real-time data analysis and interpretation, rapidly identifying any abnormal events. Once an abnormal environmental state or event is detected, the IFPPS can accurately pinpoint its precise location, thereby facilitating a timely response.
Finally, the information communication function plays a crucial role in this system. The environmental monitoring data collected by the sensing devices is analyzed in real time by the intelligent real-time computation and recognition module and then transmitted to the backend control center for monitoring and management via the following communication protocols including but not limited to, such as Ethernet’s TCP/IPv4 or TCP/IPv6, wireless network WiFi (standard: IEEE 802.11)[27], Bluetooth (standard: IEEE 802.15.1)[28], ZigBee (standard: IEEE 802.15.4)[29], MQTT (standard: ISO/IEC 20922)[30], and CoAP (standard: RFC 7252)[31] and so on. Once an abnormal event occurs, the system can promptly notify the relevant units, enabling them to take appropriate measures based on real-time information and thereby enhancing the efficiency and effectiveness of emergency response.
The Intelligent Fire Prevention Patrol System (IFPPS) delivers significant benefits by integrating real-time monitoring, rapid detection, and immediate response to potential fire hazards. Utilizing advanced sensors, intelligent computing, and efficient communication, IFPPS identifies early signs of fire, enabling prompt intervention before escalation occurs. This proactive approach minimizes property damage, reduces casualties, and enhances overall public safety. Its versatile design allows deployment across various platforms and environments, including industrial, commercial, and public spaces. By offering continuous surveillance and swift action, IFPPS effectively improves risk management and fire safety, ultimately safeguarding communities and preserving valuable assets while continuously adapting to emerging challenges.
By comparing the IFPPS with existing systems, their differences can be clearly identified and distinguished, as shown in Table 3. One characteristic of the IFPPS system is that various sensors and computing units can be installed on edge devices mounted on track vehicles or robots for mobile detection. This approach overcomes the limitation of fixed-point systems, which cannot accurately report the location of the fire source. Thus, in the earliest stage of a fire, the primary difference among the IFPPS, VEWFD, and EWFD systems lies in their detection range; the IFPPS can perform both static and dynamic detection, whereas the VEWFD and EWFD systems rely on fixed-point detection, are limited to specific areas, and require additional equipment to cover new zones. In contrast, the IFPPS can flexibly adjust its detection range using mobile devices. In terms of functionality, the EWFD system is designed to detect high-heat fires, while both the IFPPS and VEWFD systems can achieve extremely early detection of low-heat fires to prevent ignition. Furthermore, the IFPPS can accurately locate the fire source, ensuring that the fire department can precisely identify the ignition point.
Based on the four stages of fire development and the sequential order of heat release rates and application times, various existing fire protection systems are classified, as shown in Figure 3. Specifically, during the fire growth stage, the system issues alarms and notifies relevant personnel; during the full development stage, various firefighting equipment is employed to control or extinguish the fire; and finally, during the decay stage, it is necessary to verify whether there is a risk of rekindling in the environment. Moreover, the IFPPS system is capable of detecting subtle heat generation in real time and continuously monitoring the environment prior to the early stage of a fire, thereby extinguishing potential fires in their incipient state.
4. A Proof of Concept for the IFPPS
The prototype of the IFPPS, as shown in Figure 4, was validated for feasibility in a real-world environment by the research team[32]. A dual-row, double-layer shelving system was set up to simulate a large-scale warehouse environment in a logistics center. A patrol robot served as the carrier, equipped with multiple sensing devices, including a thermal imager, temperature and humidity sensors, and a carbon dioxide sensor, to capture environmental data. By integrating deep learning-based object detection technology with thermal imaging, the system detects heat sources in the environment and transmits smoldering fire source information to the main system via a wireless network.
In this experiment, the heat source was generated by a smoldering mosquito coil inside a cardboard box rather than by heat released from an open flame. The thermal imager detected a heat source inside the box, as shown in Figure 5(a), while no obvious anomalies were visible from the outside shown in Figure 5(b). This confirms that the IFPPS system can detect extremely low-heat fire sources at the earliest stage of a fire.
In addition to the autonomous patrol robot based on IFPPS shown in Figure 6, a photoelectric smoke detector and a visual flame and smoke detector were used to detect environmental abnormalities. These devices are positioned at the top row and the upper left corner of Figure 6, respectively. In this experiment, the photoelectric smoke detector did not trigger any alarms, while the visual flame and smoke detector only detected smoke when new smoldering material was added. These results demonstrate that the proposed IFPPS-based prototype system can rapidly detect smoldering sources before flames develop.
5. Conclusions
Through the collection and analysis of existing fire protection systems and related literature, this study defines and proposes the Intelligent Fire Prevention Patrol System (IFPPS) as an advanced fire safety solution. The IFPPS integrates data acquisition, intelligent real-time computing, and information communication, ensuring cross-platform adaptability. A comparative analysis with conventional fire detection systems highlights IFPPS’s dynamic detection capability, overcoming the limitations of fixed-point systems. This system significantly improves fire warning accuracy and response efficiency, thereby reducing potential fire losses. Unlike traditional photoelectric smoke detectors and visual flame detectors, which require visible smoke or flames to trigger alarms, the proposed IFPPS prototype demonstrates superior early heat detection, capable of identifying smoldering heat sources before ignition. Preliminary experiments confirm that IFPPS can detect subtle heat anomalies in real time, allowing proactive notification to relevant units before a fire occurs, thereby preventing fires and minimizing the risk of casualties. Additionally, the system's versatility across both fixed and mobile platforms enhances its application potential, providing valuable insights and a foundation for future innovations in intelligent fire safety management.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
