Traditional fire smoke detection methods mostly rely on manual algorithm extraction and sensor detection;however,these methods are slow and expensive to achieve discrimination.We proposed an improved convolutional neu...Traditional fire smoke detection methods mostly rely on manual algorithm extraction and sensor detection;however,these methods are slow and expensive to achieve discrimination.We proposed an improved convolutional neural network(CNN)to achieve fast analysis.The improved CNN can be used to liberate manpower.The network does not require complicated manual feature extraction to identify forest fire smoke.First,to alleviate the computational pressure and speed up the discrimination efficiency,kernel principal component analysis was performed on the experimental data set.To improve the robustness of the CNN and to avoid overfitting,optimization strategies were applied in multi-convolution kernels and batch normalization to improve loss functions.The experimental analysis shows that the CNN proposed in this study can learn the feature information automatically for smoke images in the early stages of fire automatically with a high recognition rate.As a result,the improved CNN enriches the theory of smoke discrimination in the early stages of a forest fire.展开更多
Existing almost deep learning methods rely on a large amount of annotated data, so they are inappropriate for forest fire smoke detection with limited data. In this paper, a novel hybrid attention-based few-shot learn...Existing almost deep learning methods rely on a large amount of annotated data, so they are inappropriate for forest fire smoke detection with limited data. In this paper, a novel hybrid attention-based few-shot learning method, named Attention-Based Prototypical Network, is proposed for forest fire smoke detection. Specifically, feature extraction network, which consists of convolutional block attention module, could extract high-level and discriminative features and further decrease the false alarm rate resulting from suspected smoke areas. Moreover, we design a metalearning module to alleviate the overfitting issue caused by limited smoke images, and the meta-learning network enables achieving effective detection via comparing the distance between the class prototype of support images and the features of query images. A series of experiments on forest fire smoke datasets and miniImageNet dataset testify that the proposed method is superior to state-of-the-art few-shot learning approaches.展开更多
To improve the performance of the forest fire smoke detection model and achieve a better balance between detection accuracy and speed, an improved YOLOv4 detection model (MoAm-YOLOv4) that combines a lightweight netwo...To improve the performance of the forest fire smoke detection model and achieve a better balance between detection accuracy and speed, an improved YOLOv4 detection model (MoAm-YOLOv4) that combines a lightweight network and attention mechanism was proposed. Based on the YOLOv4 algorithm, the backbone network CSPDarknet53 was replaced with a lightweight network MobilenetV1 to reduce the model’s size. An attention mechanism was added to the three channels before the output to increase its ability to extract forest fire smoke effectively. The algorithm used the K-means clustering algorithm to cluster the smoke dataset, and obtained candidate frames that were close to the smoke images;the dataset was expanded to 2000 images by the random flip expansion method to avoid overfitting in training. The experimental results show that the improved YOLOv4 algorithm has excellent detection effect. Its mAP can reach 93.45%, precision can get 93.28%, and the model size is only 45.58 MB. Compared with YOLOv4 algorithm, MoAm-YOLOv4 improves the accuracy by 1.3% and reduces the model size by 80% while sacrificing only 0.27% mAP, showing reasonable practicability.展开更多
The main goal of the fire management strategy is to avoid people interaction with fire hazards,therefore several studies are carried out for optimal escape route planning as well as smoke management and mechanical ven...The main goal of the fire management strategy is to avoid people interaction with fire hazards,therefore several studies are carried out for optimal escape route planning as well as smoke management and mechanical ventilation design.Underground spaces are a special case of study due to their technical characteristics that create different fire spread conditions compared to conventional buildings.In this paper,a comparative study on the fire smoke propagation is carried out taking into account the use and performance of smoke mitiga-tion techniques.The Underground Hazardous Waste Management Repository of Lavrion Technological and Cultural Park in Greece,is selected as a case study to achieve the paper objectives.A Computer Fluid Dynamic software is used to simulate various fire scenarios based on the specific characteristics of the underground space.Parameters that have significant impact on evacuation procedure and human health such as visibility and carbon monoxide concentration are monitored.Finally,the results are presented and an evaluation method is proposed based on evacuation simulation results,in order to predict the impact of the smoke control design on the conditions inside the underground space.The results reveal that the ventilation shaft(passive system)offers an effective smoke control without any other combined control system and that the installation of smoke curtains,which are activated on time,offers significant improvement on smoke spread control extending the available egress time(ASET)in case of fire.展开更多
Understanding the characteristics of smoke flow in tunnel fire is very important for tunnel safety. The characteristics of tunnel fire were analyzed. The smoke development in different situations of an engineering exa...Understanding the characteristics of smoke flow in tunnel fire is very important for tunnel safety. The characteristics of tunnel fire were analyzed. The smoke development in different situations of an engineering example was simulated using commercial CFD software PHOENICS 3.5 by field modeling method. The spreading rules and characteristics of concentration field and temperature field of smoke flow with different longitudinal ventilation speeds were studied, which may provide the theoretical background for evacuation design in tunnel fire. The effective measures of fire rescue and crowd evacuation were also described.展开更多
Fire smoke movement of multi-floor and multi-room (MFMR) fire was studied at the model test building in State Key Laboratory of Fire Science (SKLFS). The ingredient, temperature, air pressure difference and air veloci...Fire smoke movement of multi-floor and multi-room (MFMR) fire was studied at the model test building in State Key Laboratory of Fire Science (SKLFS). The ingredient, temperature, air pressure difference and air velocity of smoke were measured and analyzed. Meanwhile, the hazard of smoke ingredient to exposed occupants was analyzed based on the national standard, Occupational Exposure Limit for Hazardous Agents in the Workplace (GBZ2-2002). The experimental results showed that the maximum temperature difference in MFMR fire was located along the vertical height from the fire source. With the spreading and diffusion of smoke, the temperature of smoke layer would tend to be no difference. In the fire of woodpile and kerosene, the main smoke ingredients such as SO 2 , CO and CO 2 would first exceed human’s average physiological limit, while smoke ingredients such as NO and NO 2 would come behind. Because of the higher fluctuation range and frequency of air pressure difference of smoke in multi-layer building fire, the fire smoke would spread around everywhere of the passageway and made the human evacuation more difficult.展开更多
Fire Dynamics Simulator v3.0 was used to investigate and management in a realistic indoor sports center. An atrium fire test case assess fire smoke transport and ustrated the code＇s superiority over code-type empiri...Fire Dynamics Simulator v3.0 was used to investigate and management in a realistic indoor sports center. An atrium fire test case assess fire smoke transport and ustrated the code＇s superiority over code-type empirical models for both accuracy and capability. Four fire scenarios in the Tsinghua University Sports Center were then simulated. The smoke layer＇s descent speed was predicted for each case. The importance of the door effect was revealed and an additional mechanical ventilation system for the building was proved to be of no help. The door effect must be carefully considered in future fire safety designs.展开更多
基金National Natural Science Foundation of China(31670717)Natural Science Foundation of Heilongjiang Province(LH2020C051)。
文摘Traditional fire smoke detection methods mostly rely on manual algorithm extraction and sensor detection;however,these methods are slow and expensive to achieve discrimination.We proposed an improved convolutional neural network(CNN)to achieve fast analysis.The improved CNN can be used to liberate manpower.The network does not require complicated manual feature extraction to identify forest fire smoke.First,to alleviate the computational pressure and speed up the discrimination efficiency,kernel principal component analysis was performed on the experimental data set.To improve the robustness of the CNN and to avoid overfitting,optimization strategies were applied in multi-convolution kernels and batch normalization to improve loss functions.The experimental analysis shows that the CNN proposed in this study can learn the feature information automatically for smoke images in the early stages of fire automatically with a high recognition rate.As a result,the improved CNN enriches the theory of smoke discrimination in the early stages of a forest fire.
基金The work was supported by the National Key R&D Program of China(Grant No.2020YFC1511601)Fundamental Research Funds for the Central Universities(Grant No.2019SHFWLC01).
文摘Existing almost deep learning methods rely on a large amount of annotated data, so they are inappropriate for forest fire smoke detection with limited data. In this paper, a novel hybrid attention-based few-shot learning method, named Attention-Based Prototypical Network, is proposed for forest fire smoke detection. Specifically, feature extraction network, which consists of convolutional block attention module, could extract high-level and discriminative features and further decrease the false alarm rate resulting from suspected smoke areas. Moreover, we design a metalearning module to alleviate the overfitting issue caused by limited smoke images, and the meta-learning network enables achieving effective detection via comparing the distance between the class prototype of support images and the features of query images. A series of experiments on forest fire smoke datasets and miniImageNet dataset testify that the proposed method is superior to state-of-the-art few-shot learning approaches.
文摘To improve the performance of the forest fire smoke detection model and achieve a better balance between detection accuracy and speed, an improved YOLOv4 detection model (MoAm-YOLOv4) that combines a lightweight network and attention mechanism was proposed. Based on the YOLOv4 algorithm, the backbone network CSPDarknet53 was replaced with a lightweight network MobilenetV1 to reduce the model’s size. An attention mechanism was added to the three channels before the output to increase its ability to extract forest fire smoke effectively. The algorithm used the K-means clustering algorithm to cluster the smoke dataset, and obtained candidate frames that were close to the smoke images;the dataset was expanded to 2000 images by the random flip expansion method to avoid overfitting in training. The experimental results show that the improved YOLOv4 algorithm has excellent detection effect. Its mAP can reach 93.45%, precision can get 93.28%, and the model size is only 45.58 MB. Compared with YOLOv4 algorithm, MoAm-YOLOv4 improves the accuracy by 1.3% and reduces the model size by 80% while sacrificing only 0.27% mAP, showing reasonable practicability.
文摘The main goal of the fire management strategy is to avoid people interaction with fire hazards,therefore several studies are carried out for optimal escape route planning as well as smoke management and mechanical ventilation design.Underground spaces are a special case of study due to their technical characteristics that create different fire spread conditions compared to conventional buildings.In this paper,a comparative study on the fire smoke propagation is carried out taking into account the use and performance of smoke mitiga-tion techniques.The Underground Hazardous Waste Management Repository of Lavrion Technological and Cultural Park in Greece,is selected as a case study to achieve the paper objectives.A Computer Fluid Dynamic software is used to simulate various fire scenarios based on the specific characteristics of the underground space.Parameters that have significant impact on evacuation procedure and human health such as visibility and carbon monoxide concentration are monitored.Finally,the results are presented and an evaluation method is proposed based on evacuation simulation results,in order to predict the impact of the smoke control design on the conditions inside the underground space.The results reveal that the ventilation shaft(passive system)offers an effective smoke control without any other combined control system and that the installation of smoke curtains,which are activated on time,offers significant improvement on smoke spread control extending the available egress time(ASET)in case of fire.
基金Project(20033179802) supported by the Science and Technology Program of China Western Transportation Development
文摘Understanding the characteristics of smoke flow in tunnel fire is very important for tunnel safety. The characteristics of tunnel fire were analyzed. The smoke development in different situations of an engineering example was simulated using commercial CFD software PHOENICS 3.5 by field modeling method. The spreading rules and characteristics of concentration field and temperature field of smoke flow with different longitudinal ventilation speeds were studied, which may provide the theoretical background for evacuation design in tunnel fire. The effective measures of fire rescue and crowd evacuation were also described.
基金This work was supported by the National Natural Science Foundation of China(Grant No.50106017)China National Key Basic Research Special Funds(Grant No.2001CB409600)the 10th Five-year Tackle Key Plan of China Science and Technology(Grant No.2001BA803B01).
文摘Fire smoke movement of multi-floor and multi-room (MFMR) fire was studied at the model test building in State Key Laboratory of Fire Science (SKLFS). The ingredient, temperature, air pressure difference and air velocity of smoke were measured and analyzed. Meanwhile, the hazard of smoke ingredient to exposed occupants was analyzed based on the national standard, Occupational Exposure Limit for Hazardous Agents in the Workplace (GBZ2-2002). The experimental results showed that the maximum temperature difference in MFMR fire was located along the vertical height from the fire source. With the spreading and diffusion of smoke, the temperature of smoke layer would tend to be no difference. In the fire of woodpile and kerosene, the main smoke ingredients such as SO 2 , CO and CO 2 would first exceed human’s average physiological limit, while smoke ingredients such as NO and NO 2 would come behind. Because of the higher fluctuation range and frequency of air pressure difference of smoke in multi-layer building fire, the fire smoke would spread around everywhere of the passageway and made the human evacuation more difficult.
基金Supported by the National Key Basic Research and Development (973) Program of China (No. 2001CB409600)
文摘Fire Dynamics Simulator v3.0 was used to investigate and management in a realistic indoor sports center. An atrium fire test case assess fire smoke transport and ustrated the code＇s superiority over code-type empirical models for both accuracy and capability. Four fire scenarios in the Tsinghua University Sports Center were then simulated. The smoke layer＇s descent speed was predicted for each case. The importance of the door effect was revealed and an additional mechanical ventilation system for the building was proved to be of no help. The door effect must be carefully considered in future fire safety designs.
