Abstract

Since about 2002, Industry 4.0, and Digital Twin (DT) research as we know it today, enjoyed increasing attention in academia and industry. The research on the DTs of control valves is currently not as mature as in other critical fields. It is however of utmost importance given the role control valves play in the process, nuclear, and petroleum industries. As a common and essential part of the pipeline system, valves are used to regulate fluid flow within pipelines to achieve a desired flow condition. However, the local flow characteristics within the valve domain are difficult to determine experimentally because of complex geometry and inaccessibility of the flow domain. The application of DT technology in pipelines allows real-time fluid flow data transmission both at global and local levels, and performance monitoring of the system as well as for predicting remaining useful life. Big data analytics and machine learning can be leveraged to improve valve system performance prediction as well as to improve safety. This paper describes initial work carried out on developing a control valve digital twin which incorporates tools for monitoring local flow conditions in the control valve. Computational Fluid Dynamics (CFD) is used to determine the internal characteristics of the fluid inside the valve, and the data are analysed and managed through the development of a DT of the valve system.