基于时间序列相似性分析的管道风险评估权重设置方法研究

A Weighting Method for Risk Assessment of Pipelines Based on Similarity Analysis of Time Series

提出一种权重设置方法,以改善穆氏管道风险评估模型。穆氏模型是一种系统化风险评估方法,但它把导致管道失效的4大类因素都同样赋予100分,因而不能反映不同类型因素对管道风险的不同作用。较好的方法是对不同类别设置不同权重,文章基于风险概念来设置权重,即对与事故后果关系密切的因素类别赋予较大的权重,管道失效因素类别与事故后果的关系通过历史事故统计数据形成的时间序列相似性程度来确定,时间序列的相似性用欧氏距离来衡量。并以美国管道安全办公室(OPS)统计的管道安全数据为例说明了该方法的使用。

This paper proposes a weight assigning method to improve Muhlbauer risk assessment model for pipelines. Muhlbauer model is a systemic pipeline risk analysis method,but it has a deficiency that the four hazard index categories contributing to pipeline failures are equally assigned 100 scores and can not reflect their different importances for pipeline risk. A proper way is to assign weights to the categories. The weights are determined based on the concept of risk in this paper,that is,the category which has a closer relationship with accident consequence is assigned with greater weight. The relationship is measured by the similarity be-tween their corresponding time series created from historical accident statistical data. The similarity is calcu-lated by Euclidean distance. To illustrate the usage of this method,an example of assigning weights to the four categories is proposed based on the data collected by the Office of Pipeline Safety (OPS) of the United States.