Digital twin for oil pipeline risk estimation using prognostic and machine learning techniques

Abstract Digital Twin technology is emerging as the digitization platform to enhance the industrial information processing and management in concern with virtual and physical entities. It paves the path for integrated industrial data analysis by combining IoT and Artificial Intelligence for better data interpretation. At present in oil industry, pipelines prevail to be feasible mode, the risk probability rate is getting increased and maintenance system becomes difficult with attention to the earlier prediction of accidents risks by undertaking entire pipeline. This paper aims to provide the frame structure of Digital Twin based on machine learning and prognostics algorithms model to analyze and predict the risk probability rate of oil pipeline system. Prognostics focuses on the detection of a failure precursor by estimating risk condition with respect to the pressure data towards the evaluation of remaining useful life (RUL). The abnormality of pressure attribute is taken in prognostic analysis for risk probability estimation followed by Dirichlet Process Clustering and Canopy clustering to segregate the abnormal pressure drop and rise. Using multiple oil substation data integration platform, the features are extracted using manifold learning methods and the best feature probability rates are evaluated using kernel based SVM algorithm to provide on-time control action on the entire oil pipeline system through efficient wireless data communication between server and the oil substations. As a result, the proposed work creates Virtual Intelligent Integrated Automated Control System to predict the risk rate in oil industry by integrating entire transmission lines through enhanced wireless information networks in remote locations.