Large quantities of water are often used to extinguish fires when accidental fuel leakage occurs during storage and transportation. This may lead to spill fires on water. Boilover and splash of heavy oil spill fires in particular can pose a serious thermal threat to surrounding facilities and personnel. In this work, a series of diesel spill fire experiments were conducted on the surface of water. The results showed that, for the non-ignition cases, the fuel spread velocity was fast at first, then maintained a long period of steady spread, which can be successfully predicted by a developed spread model. During the ignition process, the burning of diesel fuel is divided into four phases. Following a brief quasi-steady burning phase, we observed an expansion of the burning area during the intermittent boilover phase, which was primarily driven by boilover. During the quasi-steady burning phase, the burning rate was lower than that of pool fires, which is attributed to the heat loss between the diesel and water layers. This heat loss also results in a lower flame height than the pool fire, and a dimensionless equation was proposed to eliminate discrepancies. During the intermittent boilover phase, the increase in the burning area was used to characterize the boilover intensity, which was found to be negatively correlated with the number of boilovers. Furthermore, the emergence of the boilover also caused flame radiation to rise rapidly; it was about 19% to 30% higher than that in the quasi-steady burning phase.
Ground-penetrating radar (GPR) has been proven effective for detecting subsurface pipes in a nondestructive way, typically with manual processing and decision-making. However, existing automatic models for segmenting the target hyperbolas often lack generalization across different pipe radii, varying subsurface media, and complex field conditions. This is especially reflected in B-scans with diverse or small-scale hyperbolas, often accompanied by cluttered and irregular noise. In this paper, an automatic preprocessing model is proposed to enhance the interpretation of B-scans under challenging conditions. The model includes a ground reflection removal algorithm (GRRA), the data gravitational force enhancement (DGFE) method, and a global–local thresholding technique consisting of dilation-based local thresholding and segmentation (DLTS). First, a frequency-domain filter based on the fast Fourier transform and a spatial filter are applied to the raw B-scan to remove obstructive ground reflection strips. Owing to the minimal intensity differences among the target hyperbola, multiples, and background, the DGFE approach is introduced to amplify the main body of the hyperbola, distinguishing it from the noise. Finally, the target hyperbola is extracted from the grayscale image by an integrated thresholding approach. The approach initially employs global thresholding to eliminate all information except for part of the hyperbola, followed by DLTS, which uses a dilation operation with local thresholding to fully segment the hyperbola. The proposed model is evaluated on two distinct datasets and compared with several state-of-the-art methods. The results demonstrate its effectiveness, particularly in terms of cross-dataset generalization.
Assessing the frequency of fires and the resulting economic losses is crucial for allocating emergency resources and firefighting personnel across different areas. In this study, spatial regression methods were employed to analyze the correlation between fire risk and various influencing factors (economic vulnerability, population vulnerability, building vulnerability, etc.), and an optimal model determination procedure was proposed and validated. The results indicate that the spatial lag model with population density, per capita income, and the degree of personnel concentration around high-risk points of interest (POIs) as indicators was the best model for predicting fire frequency. Moreover, logarithmic transformation of the indicators effectively improved the prediction accuracy in low-population density areas. An ordinary least squares model with the illiteracy rate and average surface water resources as indicators was the best model for predicting direct economic losses from fires, and the distribution of high-risk POIs can qualitatively explain the differences between the predicted results and actual data. The present work not only enriches the research on city-scale fire risk assessment but also reveals that the optimal regression model determination process can provide technical support for the application of spatial regression methods in fire risk assessment.