Prediction of Residual Service Life of Pipelines Based on Unbiased Grey and Markov Chain Composite Model

The study of predicting the residual service life of pipelines is of great significance for preventing pipeline leakage and making a reasonable test strategy. Due to a limited amount and random fluctuations of actual pipeline corrosion data, an improved unbiased grey GM(1,1) model was built based on traditional grey GM(1,1) model to improve prediction accuracy. Then an unbiased grey and Markov chain composite model was established based on unbiased grey GM(1,1) model and used for predicting residual service life of pipeline based on pipeline corrosion data from an ocean oil field crude oil processing system. The results show that the pipeline reached the fifth leakage state after being used for 8 years and should be repaired or replaced. The prediction accuracy of the unbiased grey and Markov chain composite model was above 94%. The result met the requirements of engineering accuracy, indicating that the proposed model can be used to predict the residual service life of pipelines.