Value maximisation with process digital twins and data analytics The process digital twin uses real-time simulation and data analytics to optimise operations and generates invaluable insights
Jesse Mallhi, Paras N Shah, Sathiyanarayanan A, N C Chakrabarti, Narendar Mitta, and Vikas Deshmukh Reliance Industries Limited
R eliance Jamnagar, which includes two ultra-large refineries, a petcoke gasification complex, and a refinery off-gas cracker petrochemical complex, is at the forefront of adopting the latest digital technologies. A unified digital platform is currently being developed to integrate all functional domains, including operations, maintenance, technical services, safety, “ Process digital twins use first- principles, hybrid, and self-learning machine learning and artificial intelligence models to achieve key business objectives ” procurement, planning, and security. The goal is to enhance productivity at the Jamnagar complex by simplifying processes, automating tasks, eliminating digital breaks, and facilitating collaboration among multiple stakeholders. The digital modules, including digital twins, smart devices, virtual assistants, handheld devices, and smart CCTVs, will significantly
improve safety, reliability, and productivity, ultimately leading to increased profitability. Process digital twin The term ‘process digital twin’ is being used uniformly across all types of industries, although people from different fields have different understandings of what a digital twin is. As Markus Meissner aptly stated, “Sometimes good things come in pairs: The Digital Twin.” In continuous process manufacturing facilities, such as refining and petrochemical complexes, a process digital twin is a digital replica of all unit operations within a process plant. This includes chemical reactions, separation processes, heat transfer, and other operations, all modelled to mimic the actual operation and performance of the plant. Process digital twins use first-principles, hybrid, and self-learning machine learning (ML) and artificial intelligence (AI) models to achieve key business objectives, including operational excellence, reliability, availability, and safety. To mimic the actual operations, multiple models are required to be built and deployed (see Figure 1 ): • At the core of the process digital twin is a real-time process simulation model based on first principles and hybrid models. This provides detailed, real-time insights into process operations. • Next is the data analytics model, based on ML and AI. This model excels in early event detection, the development of soft sensors, and the identification of optimal operating zones. • The final key element is the interactive dashboard. This integrates high-quality data and value-added information derived from both
Data analytics models
Process models
Interactive dashboard
First principle/ hybrid based
Data driven
Visualisation
Physics, chemistry, engineering simulation
Statistical, ML & AI to generate insights
Integrate value- added information for e ective and timely decision
Figure 1 Key components of process digital twin
Refining India
60
Powered by FlippingBook