摘要
针对半定量与定量风险评价方法的优缺点,提出指标概率综合风险评价新方法(简称I-P法)。该方法不仅能够克服半定量评价法容易产生"避重就轻"现象的弊端,又能克服定量评价法中因缺失指标失效概率而无法应用的不足。I-P法下层指标采用基于专家经验的指标评价方法,而后利用BP神经网络方法建立指标相对分值d与失效概率Pf(i)的关系曲线,依据此曲线将指标分值转化为失效概率,并对上层指标采用基于失效概率的定量评价。采用该方法对某管道的河沟道水毁进行了风险评价,结果表明,该方法能够解决在某些特殊条件下基于指标评价法的结论与管道实际风险状态不一致的问题。
In view of the advantages and disadvantages of semi-quantitative and quantitative risk assessment methods, an index ＆ probability comprehensive risk assessment method （I-P method） is put forward. This method can not only overcome the disadvantage of semi-quantitative evaluation method ＂shying away the important and dwelling on the trivial ＂, but also overcome the disadvantage of the quantitative evaluation method that the quantitative evaluation method can not be applied due to lack of the failure probability of indicators. The underlying indexes in I-P method apply index assessment method based on expert＇ s experience, and then the relationship between the relative score of the indexes and failure probability is established using BP neural network. According to the relationship curve, the relative score of the indexes is transformed into failure probability, and the quantitative evaluation of top indexes is based on the failure probability. The water damage risk of some oil pipeline in river section is evaluated using I-P method, and the result shows that the method can overcome the problem that the evaluation conclusion of index assessment method is not consistent with the actual risk of pipeline under certain special conditions.
出处
《石油工业技术监督》
2015年第5期31-34,39,共5页
Technology Supervision in Petroleum Industry
关键词
风险评价
指标概率综合法
河沟道水毁
失效概率
BP神经网络
risk assessment
index ＆ probability comprehensive assessment method
water damage in river section
failure probability
BP neural network
