Applied Sciences

Research

20 pages, 1978 KB

Open AccessArticle

UAV-Based Forest Fire Early Warning and Intervention Simulation System with High-Accuracy Hybrid AI Model

by Muhammet Sinan Başarslan and Hikmet Canlı

Appl. Sci. 2026, 16(3), 1201; https://doi.org/10.3390/app16031201 - 23 Jan 2026

Viewed by 155

In this study, a hybrid deep learning model that combines the VGG16 and ResNet101V2 architectures is proposed for image-based fire detection. In addition, a balanced drone guidance algorithm is developed to efficiently assign tasks to available UAVs. In the fire detection phase, the [...] Read more.

In this study, a hybrid deep learning model that combines the VGG16 and ResNet101V2 architectures is proposed for image-based fire detection. In addition, a balanced drone guidance algorithm is developed to efficiently assign tasks to available UAVs. In the fire detection phase, the hybrid model created by combining the VGG16 and ResNet101V2 architectures has been optimized with Global Average Pooling and layer merging techniques to increase classification success. The DeepFire dataset was used throughout the training process, achieving an extremely high accuracy rate of 99.72% and 100% precision. After fire detection, a task assignment algorithm was developed to assign existing drones to fire points at minimum cost and with balanced load distribution. This algorithm performs task assignments using the Hungarian (Kuhn–Munkres) method and cost optimization, and is adapted to direct approximately equal numbers of drones to each fire when the number of fires is less than the number of drones. The developed system was tested in a Python-based simulation environment and evaluated using performance metrics such as total intervention time, energy consumption, and task balance. The results demonstrate that the proposed hybrid model provides highly accurate fire detection and that the task assignment system creates balanced and efficient intervention scenarios. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

31 pages, 4778 KB

Open AccessArticle

ESCFM-YOLO: Lightweight Dual-Stream Architecture for Real-Time Small-Scale Fire Smoke Detection on Edge Devices

by Jong-Chan Park, Myeongjun Kim, Sang-Min Choi and Gun-Woo Kim

Appl. Sci. 2026, 16(2), 778; https://doi.org/10.3390/app16020778 - 12 Jan 2026

Viewed by 170

Abstract

Early detection of small-scale fires is crucial for minimizing damage and enabling rapid emergency response. While recent deep learning-based fire detection systems have achieved high accuracy, they still face three key challenges: (1) limited deployability in resource-constrained edge environments due to high computational [...] Read more.

Early detection of small-scale fires is crucial for minimizing damage and enabling rapid emergency response. While recent deep learning-based fire detection systems have achieved high accuracy, they still face three key challenges: (1) limited deployability in resource-constrained edge environments due to high computational costs, (2) performance degradation caused by feature interference when jointly learning flame and smoke features in a single backbone, and (3) low sensitivity to small flames and thin smoke in the initial stages. To address these issues, we propose a lightweight dual-stream fire detection architecture based on YOLOv5n, which learns flame and smoke features separately to improve both accuracy and efficiency under strict edge constraints. The proposed method integrates two specialized attention modules: ESCFM++, which enhances spatial and channel discrimination for sharp boundaries and local flame structures (flame), and ESCFM-RS, which captures low-contrast, diffuse smoke patterns through depthwise convolutions and residual scaling (smoke). On the D-Fire dataset, the flame detector achieved 74.5% mAP@50 with only 1.89 M parameters, while the smoke detector achieved 89.2% mAP@50. When deployed on an NVIDIA Jetson Xavier NX (NVIDIA Corporation, Santa Clara, CA, USA)., the system achieved 59.7 FPS (single-stream) and 28.3 FPS (dual-tream) with GPU utilization below 90% and power consumption under 17 W. Under identical on-device conditions, it outperforms YOLOv9t and YOLOv12n by 36–62% in FPS and 0.7–2.0% in detection accuracy. We further validate deployment via outdoor day/night long-range live-stream tests on Jetson using our flame detector, showing reliable capture of small, distant flames that appear as tiny cues on the screen, particularly in challenging daytime scenes. These results demonstrate overall that modality-specific stream specialization and ESCFM attention reduce feature interference while improving detection accuracy and computational efficiency for real-time edge-device fire monitoring. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

16 pages, 2444 KB

Open AccessArticle

The Decomposition Mechanism of C₄F₇N–Ag Gas Mixture Under High Temperature Arc

by Tan Liu, Yi Ding, Congrui Zhang and Xingjian Kang

Appl. Sci. 2026, 16(1), 356; https://doi.org/10.3390/app16010356 - 29 Dec 2025

Viewed by 190

Abstract

The global phase-out of sulfur hexafluoride (SF6), an insulating gas with high global warming potential (GWP), has driven the search for eco-friendly alternatives in high-voltage equipment. Perfluoroisobutyronitrile (C4F7N) emerges as a promising candidate due to its low GWP and high dielectric strength. However, [...] Read more.

The global phase-out of sulfur hexafluoride (SF6), an insulating gas with high global warming potential (GWP), has driven the search for eco-friendly alternatives in high-voltage equipment. Perfluoroisobutyronitrile (C4F7N) emerges as a promising candidate due to its low GWP and high dielectric strength. However, its chemical stability under circuit breaker conditions, especially when interacting with vaporized contact materials such as silver, remains a key concern. This study investigates the decomposition mechanisms of C4F7N in the presence of silver vapor using quantum chemical calculations at the B3LYP/LanL2DZ level. A reaction network comprising 35 pathways and 12 transition states were identified. All structures were confirmed as valid stationary points via frequency analysis and intrinsic reaction coordinate (IRC) calculations. Three primary reaction pathways between C4F7N and Ag were delineated, leading to secondary reactions that generate low-weight molecules and Ag-containing species such as AgF and AgCN. Key energy barriers and temperature-dependent equilibrium constants (K_eq) were determined to evaluate pathway feasibility. This work provides fundamental insights into the high-temperature interfacial chemistry of C4F7N with Ag, offering essential data for assessing its material compatibility and long-term reliability as a sustainable insulation medium in power systems. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

29 pages, 5198 KB

Open AccessArticle

Distributed Fire Classification and Localization Model Based on Federated Learning with Image Clustering

by Jiwon Lee, Jeongheun Kang, Chun-Su Park and Jongpil Jeong

Appl. Sci. 2024, 14(20), 9162; https://doi.org/10.3390/app14209162 - 10 Oct 2024

Cited by 4 | Viewed by 2398

Abstract

In this study, we propose a fire classification system using image clustering based on a federated learning (FL) structure. This system enables fire detection in various industries, including manufacturing. The accurate classification of fire, smoke, and normal conditions is an important element of [...] Read more.

In this study, we propose a fire classification system using image clustering based on a federated learning (FL) structure. This system enables fire detection in various industries, including manufacturing. The accurate classification of fire, smoke, and normal conditions is an important element of fire prevention and response systems in industrial sites. The server in the proposed system extracts data features using a pretrained vision transformer model and clusters the data using the bisecting K-means algorithm to obtain weights. The clients utilize these weights to cluster local data with the K-means algorithm and measure the difference in data distribution using the Kullback–Leibler divergence. Experimental results show that the proposed model achieves nearly 99% accuracy on the server, and the clustering accuracy on the clients remains high. In addition, the normalized mutual information value remains above 0.6 and the silhouette score reaches 0.9 as the rounds progress, indicating improved clustering quality. This study shows that the accuracy of fire classification is enhanced by using FL and clustering techniques and has a high potential for real-time detection. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

14 pages, 2254 KB

Open AccessArticle

Risk Assessment of Hydrogen Cyanide for Available Safe Egress Time in Fire Simulation

by Oh-Soo Kwon, Ho-Sik Han and Cheol-Hong Hwang

Appl. Sci. 2024, 14(16), 6890; https://doi.org/10.3390/app14166890 - 6 Aug 2024

Cited by 2 | Viewed by 2602

Abstract

The majority of fatalities in building fires are attributed to asphyxiation caused by toxic gases. Hydrogen cyanide (HCN) is one of the toxic gases that can be released during a fire, posing a lethal risk to humans even at low concentrations. However, analysis [...] Read more.

The majority of fatalities in building fires are attributed to asphyxiation caused by toxic gases. Hydrogen cyanide (HCN) is one of the toxic gases that can be released during a fire, posing a lethal risk to humans even at low concentrations. However, analysis of the risk posed by HCN in fire risk assessments using fire simulations is relatively rare. This study conducted fire simulations to examine the potential risks of HCN to occupants during a fire. The simulations considered various fire conditions in residential buildings by varying fuel types, fire growth rates, and HCN yields. The relative risk score (RRS) was derived based on the time to reach the threshold values of parameters considered critical for life safety. The results of the fire simulations indicated that the RRS for HCN was approximately 20–40 points higher than that of O₂, CO, and CO₂, reaching a maximum of 92 points. However, the risk posed by HCN was found to be limited in comparison to the risks associated with temperature and visibility. Nevertheless, considering that the primary cause of fatalities in fires is asphyxiation due to toxic gases, HCN must be regarded as a critical factor in fire risk assessments. Additionally, since HCN yield values can increase up to nine times depending on temperature and ventilation conditions, the risk posed by HCN could be significantly higher. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

27 pages, 3790 KB

Open AccessArticle

Identification and Discrimination of Petrol Sources by Nuclear Magnetic Resonance Spectroscopy and Machine Learning in Fire Debris Analysis

by Yanita Yankova, Silvia Cirstea, Michael Cole and John Warren

Appl. Sci. 2024, 14(12), 5177; https://doi.org/10.3390/app14125177 - 14 Jun 2024

Cited by 1 | Viewed by 1675

Abstract

Petrol is considered the most common fire accelerant. However, the identification and classification of petrol sources through the years has proven to be a challenging field in the investigation of fire debris analysis. This research explored the possibility of identifying petrol sources by [...] Read more.

Petrol is considered the most common fire accelerant. However, the identification and classification of petrol sources through the years has proven to be a challenging field in the investigation of fire debris analysis. This research explored the possibility of identifying petrol sources by high-field NMR methods accompanied by ML (machine learning). The automated identification and classification of petrol brands were achieved for first time based on the ML classification model developed in this research. A hierarchical classification model was constructed using local classifiers to categorize neat or weathered petrol into its sources. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

23 pages, 7796 KB

Open AccessArticle

The Effect of Inclined Conditions on the Consequences of Fires Caused by Spilled Flammable Liquids: Development of Inclined Spreading Extent Formulae

by Daeyu Baeg, Hyunho Lee, Seungyul Lee and Jung Kwan Seo

Appl. Sci. 2024, 14(2), 745; https://doi.org/10.3390/app14020745 - 15 Jan 2024

Cited by 1 | Viewed by 2151

Abstract

The accidental spillage of flammable liquids on in-service ships and offshore installations may lead to pool fires, which are likely to spread over a particularly large area in large compartments under ship motion, resulting in extensive damage. However, the effect of the spreading [...] Read more.

The accidental spillage of flammable liquids on in-service ships and offshore installations may lead to pool fires, which are likely to spread over a particularly large area in large compartments under ship motion, resulting in extensive damage. However, the effect of the spreading extent of liquid fuel due to inclined ship motion on pool fire consequences has not been considered in the existing literature. Thus, in this study, fuel discharge experiments were conducted to investigate the spreading behaviour under different substrate inclination angles and discharge rates. The experimental results were analysed to derive closed-form expressions to predict the spreading extent of liquid fuel in large compartments. Additionally, the effects of surface inclination on fire consequences were investigated using the Fire Dynamics Simulator in terms of the heat release rate. The findings can provide guidance for effective fire safety design and establishing a realistic fire modelling methodology for ships and offshore installations. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

13 pages, 5614 KB

Open AccessArticle

Configuration Approaches of CFAST for Prediction of Smoke and Heat Detector Activation Times in Corridor Fires

by Hyo-Yeon Jang and Cheol-Hong Hwang

Appl. Sci. 2023, 13(24), 13161; https://doi.org/10.3390/app132413161 - 11 Dec 2023

Cited by 1 | Viewed by 2676

Abstract

In performance-based design for domestic buildings, there is a growing need for real-time comparison between the Available Safe Egress Time and Required Safe Egress Time through the integration of fire and evacuation simulations. Therefore, the utilization of the Consolidated Model of Fire and [...] Read more.

In performance-based design for domestic buildings, there is a growing need for real-time comparison between the Available Safe Egress Time and Required Safe Egress Time through the integration of fire and evacuation simulations. Therefore, the utilization of the Consolidated Model of Fire and Smoke Transport (CFAST) has been discussed as an alternative to the Fire Dynamics Simulator (FDS), which has high computational costs; requires sufficient experience in the numerical calculation of fire dynamics, along with various input parameters; and has limitations in coupling with evacuation simulations. In this study, the prediction performance of CFAST for the activation times of smoke and heat detectors was evaluated. Specifically, it is essential to configure the mass movement between adjacent computational regions for smoke concentration. For achieving adequate predictive performance, the temperature should be determined according to the ceiling jet velocity generated by the fire source. Therefore, a method for setting a computational domain that can produce reasonable prediction results while considering the characteristics of CFAST for different types of smoke and heat detectors is proposed. Full article

(This article belongs to the Special Issue Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Advanced Analysis and Technology in Fire Science and Engineering - 2nd Edition

Share This Special Issue

Special Issue Editor

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Published Papers (8 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI