| China is in the transition period of economy and society. In this period, fire hazards are greatly breeding, and the fire safety management is especially difficult. The pressure of fire safety is unprecedented. With the great attention of Chinese government and the collaborative effort of society, fire situation is steadily improving, but the situation of high-casualty fire (HCF) is serious. In order to deepen the understanding of China HCFs, the present work conducts studies form the perspective of frequency-sized power law distribution, fatality levels and influence factors, provincial fire characteristics, time-scaling property, fire risk and the suppressive effect of fire prevention campaigns. Given the difficulty of obtaining the number of injured people, HCFs were represented by fires with three or more fatalities in the present work. The HCF data was derived from the Fire Statistical Year Book of China and China Fire Services, which are both edited by the Fire Service Bureau and are the most official fire statistics that are available at the present time.The frequency-size power-law distribution and time-scaling property of HCFs were investigated firstly. The results show that HCFs exhibit a frequency-size power-law distribution regardless of whether the fire size is represented by fatalities, direct loss or burned area. The frequency-fatality power-law distribution is a common phenomenon in fire accidents, even in international data. Six factors (place, cause, time of day, season, year and region of the fire) were analyzed to assess their effects on the frequency-fatality distribution and compared using the scaling exponent. Factors, such as a non-residential place, an electrical cause, the winter season and regions with strong economies, cause fire frequency to decrease slowly with increases in fatalities. That is, these factors are associated with higher fatalities.Then we investigated the associations between fatality levels and influence factors that involve place, cause, time of day, month, year and province. The variable fatality in a fire has four levels:3,4-5,6-9and≥10. The results show that hotels, welfare houses and hospitals tend to be strongly associated with fatality leve110. The fires caused by work-related tasks tend to precipitate a relatively high number of fatalities and are strongly associated with fatality level6-9. Fires that occur in the daytime (8:00-19:59) are associated with higher fatalities than fires that occur at night (20:00-7:59). The months in the cold season, such as winter or the beginning of spring, tend to be associated with fatality levels4-5,6-9and≥10. CA dynamically portrayed the fatality tendency over the past eight years, and2007tended to be associated with fatality level≥10. Fatality c haracteristics of provinces are identified, and Beijing, Shandong and Jilin are strongly associated with fatality leve≥10. To explore whether associations between influence factors and fatality levels of high-casualty fires in China resemble corresponding associations of HCFs in the United States, data on fires with fatality level>5in the two countries were collected. The results of four sets of comparisons indicate that the associations between influences and fatality levels in the two countries present contrasting features. Some practical applications are briefly discussed.For the purpose of investigating provincial fire characteristics, principal component analysis (PCA) and clustering analysis were used. Comparing to previous work, the information of HCFs is taken into account in the quantitative analysis of fire characteristics. Four principal components are adopted to portray fire characteristics, and they are "comprehensive fire situation","the ratio of general fire to HCFs","comprehensive fatal fire situation" and "the ratio of high fatality fire to low fatality fire". In order to make the results clear and easy to understand, clustering analysis was conducted. According to results, the Zhejiang and Guangdong provinces have the most serious fire characteristics. The socio-economic information was acquired through the provincial region information. Finally, the effect of5influencing factors (GDP, per capita GDP, population, fire stations per million people and years) on the provincial fire characteristics were explored by Relative Risk.The time-scaling properties of HCFs were investigared. The time-scaling properties were detected by means of Fano Factor(FF), Allan Factor(AF) and detrended fluctuation analysis (DFA). The results of FF and AF show that the HCF sequences with death≥3and death≥4exhibit obvious time scaling behavior after the fractal onset times. The scaling exponents of FF and AF decrease significantly with increasing fatality, which reflects that HCF sequences with more fatalities tend to behave as Poisson process. The sequence of HCFs with death≥6can be considered as a Poisson process according to the comparison of FF (AF) curve and Poissonian95%confidence curve. The DFA scaling exponent of HCF sequence involving fatality≥3is approximate0.551±0.005, indicating that this sequence exhibits long-range correlations. With the increase of fatality threshold the DFA scaling exponent gradually decreases to about0.5, which reflects that the HCFs with high fatality levels are likely to be uncorrelated. Furthermore, the sequence of HCFs with death≥6can be regarded as uncorrelated because its DFA scaling exponent is0.496±0.003.We analyzed the risk of HCFs with six or more fatalities and extreme HCFs with very high fatalites. Risk was analyzed by employing a Bayesian hierarchical model yielding analytical functions for both frequency (Poisson) and severity distributions (Generalized Pareto). This approach addresses a key problem in risk estimation, namely the scarcity of data resulting in high uncertainties in particular for the risk of extreme events, where the risk is extrapolated beyond the historically most severe accidents. The risks of China HCFs, the United States HCFs, China non-residential HCFs and China residential HCFs are162.26,91.67,120.23and40.72person/year, respectively. The risk of extreme HCFs is assessment by the product of1%exceedance and frequency, and the value of China HCFs, the United States HCFs, China non-residential HCFs and China residential HCFs are511.51,292.64,391.48and142.50person/year, respectively.In China, the post-accident fire prevention campaigns are very popular. Many impressive safety campaigns are carried out after the occurrence of the deadliest fire accidents. The aim of following work is to explore the suppressive effect of6fire prevention campaigns initiated after disastrous fires that resulted in very high casualty and significant losses. Time series intervention models are employed to investigate the effects of fire prevention campaigns on the monthly time series of fires (1997-2010), fire deaths (1997-2010) and high-casualty fires (2000-2010). The suppressive effects produced by campaigns that followed the Luoyang Dongdu fire and the new CCTV building fire are determined. The effects of the campaigns on the fire series following both fires exhibit an exponential decay pattern with a sudden "pulse" decrement after the start of the intervention, followed by a gradual deceleration back to the original pre-intervention level with no permanent effect. The effect of the campaign precipitated by Luoyang Dongdu fire on fire series is greater than that from the campaign initiated after new CCTV building fire. The campaign precipitated by the Luoyang Dongdu fire also produces an approximate damped sine wave suppressive effect with a permanent decrement of43deaths. The high-casualty fire series is not influenced by fire prevention campaigns precipitated by disastrous fires, which is inconsistent with the campaigns’hypotheses that HCFs would decrease. |
PDF Full Text Request