摘要
为了分析煤与瓦斯突出预测输入变量的不确定性对于预测结果的影响,提出一种采用抽样盲数的估计法.在不确定变量的取值置信区间内进行分段,用每一段区间的均值及取值概率作为样本反映该区间的不确定信息,以不确定变量的各个样本的集合作为反映其不确定信息的抽样盲数.分别将各个输入变量的抽样盲数的样本组合作为预测器的输入进行预测,得到反映煤与瓦斯突出预测结果不确定性的抽样盲数,从而得出煤与瓦斯突出预测结果的概率分布及置信区间.结果表明:所建议的方法是可行的,它不仅可以处理任意概率分布的不确定性信息,还能避免卷积运算从而减少了计算量,且具有能够反映出煤与瓦斯突出各个级别的可能性的优点.3组检验样本的预测煤与瓦斯突出级别为A,D和A的概率分别为0.9045,0.7775和0.9893,与实际结果相符.
In order to evaluate the influence of the uncertainties on the forecasting results of coal and gas outburst, a novel estimation approach based on sampled-blind-number （SBN） was proposed. An uncertain variable was divided into several segments within its confidence interval. The uncertain behavior of an element in one segment was described by the averaged value and probability within the segment, which formed a sample of the element. The SBN of an element consists of the set of all samples of the element. Various combinations of the corresponding element samples were used as the input of the coal and gas outburst forecaster, respectively, and the SBN describing the uncertainty of the forecasting result was worked out. The probability distribution and the confidence interval of the forecasting results can also be obtained. The results show that the proposed approach is feasible. Not only the uncertainties with arbitrary probability distribution can be dealt with, the amount of calculation is also reduced due to the convolution operation being avoided. The probability of each level of the coal and gas outburst can be evaluated. The forecasting probabilities of the three testing situations with the coal and gas outburst levels of A, D and A are 0. 904 5, 0. 777 5 and 0. 989 3, respectively, which are in positive with the actual conditions.
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2010年第2期227-232,共6页
Journal of China University of Mining & Technology
基金
电子信息产业发展基金招标项目(XDJ2-0514-27)
国家高技术研究发展计划(863计划)(2005AA133070)
关键词
煤与瓦斯突出
不确定性
抽样盲数
概率分布
置信区间
预测
coal and gas outburst
uncertainty
sampled-blind-number
probability distribution
confidence interval
forecasting
