Global Practice of AI and Big Data in Oil and Gas Industry

Abstract

Abstract This chapter introduces typical cases of artificial intelligence and big data application in oil and gas industry. In the upstream field, it introduces how to combine digital rock physics with big data and AI to optimize recovery efficiency. Digital core (also known as digital petrophysics—DRP) technology enables more reliable physical information about pore-scale multiphase flows to determine the cause of low recovery and provides new ways for different injection solutions to improve reservoir performance. Combined with digital rock technology and AI, we can integrate the characteristics of digital rock databases into logging and well data, and use a variety of advanced classification techniques to identify the remaining oil and gas potential. Through the multi-phase flow simulation, the multi-scale model can predict the best injection method for maximum recovery under different conditions and propose possible solutions to optimize crude oil production. In the downstream field, the application of AI and big data analysis in planning and scheduling systems, process unit optimization, preventive maintenance of equipment, and other aspects is introduced. Among them, the molecular-level advanced planning and scheduling system (MAPS) can realize the cost performance measurement under different production schemes for potential types of processable crude oil, which is conducive to more accurate selection of crude oil and prediction of crude oil properties. In addition, the whole process simulation can be used to understand the product quality changes under different crude oil blending schemes and different unit operating conditions, which is conducive to timely adjusting the product blending schemes according to economic benefits or ex-factory demands. The operation conditions of secondary units and even the properties of mixed crude oil can be deduced according to different product quality requirements. In the process of optimization, the Continuous Catalytic Reforming (CCR) unit in the refinery process, for example, introduces the application of large data analysis, including correlation analysis, single index detection, multidimensional data anomaly detection, and the parameters of the single objective optimization, a multi-objective parameter optimization analysis, unstructured data analysis, and forecast analysis based on material properties. Good practices in CCR units have also been extended to other oil refining and chemical units, such as Fluid Catalytic Cracking (FCC) and ethylene cracking. In terms of equipment preventive maintenance, it introduced how to integrated application of Internet of things, deep machine learning, knowledge map and other technology to build real-time and on-line distributed equipment health monitoring and early warning system, for early detection of equipment hidden danger, early warning, early treatment of providing effective means, guarantee equipment run healthy and stable for a long period of time, to reduce unplanned downtime losses. In particular, the establishment of equipment prediction model based on time series and AI can realize effective monitoring and early warning of equipment faults such as shaft displacement, shaft fracture, shell cracking, power overload, and prediction of equipment remaining life.