An optimized SVM pipeline risk assessment method based on artificial bee colony algorithm

Pipeline risk assessment is one important part of pipeline risk management, and it is aimed at identifying the important factors that possibly lead to pipeline accidents by means of risk investigation and analysis to make the pipeline risk management more scientific. In this paper, an optimized SVM (Support Vector Machine) pipeline risk assessment method based on artificial bee colony algorithm was proposed in order to carry out accurate assessment on pipeline risks in the state of daily operation. In this method, the pipeline risk assessment model is established, the technological operation characteristics of products pipelines, forward/backward crude oil pipelines and mixed-transportation crude oil pipelines are collected from the viewpoint of technological operation, and the sample characteristic set is built up. Then, 4 types of pipeline characteristic sets were tested. It is indicated that in the case of small sample, the optimized SVM pipeline risk assessment method based on artificial bee colony algorithm has higher accuracy and good applicability, and it can provide the accurate risk assessment result according to the actual running state of pipelines.