Research on the relationship between causal factors and consequences of incidents occurred in tank farm using ordinal logistic regression

Abstract

The analysis of incidents occurred is beneficial for the prevention of incidents in the future and the improvement of system design and management capability. This paper presents statistical analysis of the incidents related to tank farms of China from 1960 to 2018. Apart from major accidents which attract national or international media attention, a large number of near misses and ordinary incidents were also collected in order to more fully present and analyze problems in the design and management of tank farms. This study concentrates on exploring the relationship between causal factors and different levels of consequences. The causal factors were considered to cover five major categories which were equipment failure, environmental factors, human factors, organization/management factors and other factors specifically and subdivided into fifteen subcategories, and the consequences were divided into four levels according to financial losses and casualties. The cross-tabulation analysis of causal factors and consequence levels was performed to reveal the contribution of different factors to specific consequences and present each factor which tends to be classified what the level of consequence. Further analyses aim to determine whether there were significant correlations between the cause factor and the specific level of consequences based on the ordinal logistic regression. The results show that different causal factors tend to trigger consequences of various degrees of severity, and several organization/management factors have the significant effect on the consequences of higher level.