This paper presents statistical models for predicting volume of oil spills in pipeline failures. It is based on performance and failure data of cross-country oil pipelines.
The two-parameter Welbull distribution was used to predict the size (volume) of the oil spill. The goodness-of-fit for the models has been evaluated using Chi-square criterion. The two-parameter Welbull distribution proved to be an effective means of statistically predicting oil spill size.
The statistical models developed make it possible to assess both the probability of volume of oil spillage due to all causes of pipeline failures, and of individual causes as well.