我国煤矿安全事故统计分析及预测  

Statistical analysis and prediction of coal mine safety accidents in China

作　　者：黄继广 马汉鹏[1,2] 范春姣 姚绒绒 HUANG Ji-guang;MA Han-peng;FAN Chun-jiao;YAO Rong-rong(North China Institute of Science and Technology,Beijing 065201,China;National Safety Training Center of Coal Mines,Beijing 065201,China)

机构地区：[1]华北科技学院,北京065201 [2]中国煤矿安全技术培训中心,北京065201

出　　处：《陕西煤炭》2020年第3期34-39,6,共7页Shaanxi Coal

基　　金：中央高校基本科研业务费赞助(3142017066);中央高校基本科研业务费赞助(3142019051)。

摘　　要：为分析我国煤矿死亡事故发生的规律和准确判断未来我国煤矿的整体安全生产形势,根据2009—2018年我国煤矿死亡事故统计数据,从煤矿事故发生级别、事故类型2个维度对我国的煤矿生产安全事故进行了研究。通过SPSS 25.0软件建立Brown线性趋势模型,并对我国2009—2018年煤矿百万吨死亡率进行验证,最后预测2019—2021年我国煤矿百万吨死亡率。结果表明2009—2018年煤矿事故起数和死亡人数处于持续下降趋势,我国煤矿安全形势持续稳定向好;在事故发生级别上,一般事故发生的事故起数最多,累计造成的人员死亡也是最高,其次是较大事故和重大事故,特别重大事故发生起数在2016年出现了回升,造成了大量人员伤亡,但是2017年出现回落,没有发生特大事故;在事故类型中,顶板事故发生频率最高,瓦斯事故是一次伤害造成死亡人数最多的;Brown线性趋势模型对2009—2018年煤矿百万吨死亡率验证表明,相对误差基本上都在30%以内,R^2=0.931,接近于1,均方根误差(RMSE)和正态化的BIC均较小,表明Brown线性趋势模型模型的拟合效果比较好,预测精度较高。对我国煤矿事故发生规律的研究和煤矿百万吨死亡率的预测验证分析,使煤矿企业制定完备的安全管理措施和避免煤矿事故的发生具有重要意义。In order to analyze the law of coal mine death accidents in China and accurately judge the overall safety production situation of coal mines in the future,according to the statistics data of coal mine death accidents in China from 2009 to 2018,the coal mine production safety accidents in China were studied from the level and type of coal mine accidents.Based on the Brown linear trend model established by SPSS 25.0 software,the death rate per million-ton of coal mines in China in 2009—2018 was analyzed,and the death rate per million-ton of coal mines in China from 2019 to 2021 was predicted.The results show that from 2009 to 2018,the number of coal mine accidents and deaths continues to decline;in the accident level,the number of general accidents is the largest,and the cumulative death is the highest;in the accident type,the roof accident frequency is the highest,and the gas accidents cause the most deaths at a time.Using Brown linear trend model to analyze the death rate per million-ton of coal mines in 2009—2018 shows that the relative errors are basically within 30%,R^2=0.931,close to 1,and the root mean square error(RMSE)and the normalized BIC are smaller,which shows that Brown linear trend model has better fitting effect and higher prediction accuracy.The research on the occurrence law of coal mine accidents and the prediction of the death rate per million-ton of coal mines indicate that it is of great significance for coal mine enterprises to formulate complete safety management measures to avoid the occurrence of coal mine accidents.

关 键 词：煤矿事故 统计分析 SPSS软件 百万吨死亡率 Brown线性趋势模型 

分 类 号：TD77[矿业工程—矿井通风与安全]