Hybrid model development • Construct a process-specific base model using chemical engineering know-how as a starting point. • Review the plant’s setup, available sensors, and measurements, and any operational constraints. • Identify the relevant process data to measure and optimise performance. This process data brings hybrid models to life. By carefully integrating unknowns using arbitrary functions derived from the data, hybrid models ensure a good fit for all data. • Once all relevant effects within the plant have been captured and considered, advanced “ The process digital twin facilitates a new way of working by providing near-real-time optimisation for improving operational excellence and bringing out hidden opportunities ” ML techniques are applied. This creates AI- enhanced, data-driven, hybrid process models that, in contrast to rigid and inflexible first principles models, can learn intuitively and adapt quickly to changing conditions. • Using intelligent algorithms to consider all data observed from a continuous stream, generates invaluable insights. This means it is possible to act quickly and with confidence on real-time recommendations for process control variables to reach optimisation goals and critical KPIs. Process digital twins – new way of working The process digital twin, which consists of real-time process simulation models and data analytic models, facilitates a new way of working by providing near-real- time optimisation for improving operational excellence and bringing out hidden opportunities. • It provides a precise view of operational constraints and capabilities. • Real-time monitoring of KPIs for process, equipment, energy, and APC enables us to identify and close the performance gaps immediately. • It provides insight into the best timing for catalyst replacement during its end-of-run
conditions, based on catalyst remaining life, increased utilities consumption, and yield patterns. • Real-time early event detection, along with predictive and prescriptive models, will help avoid process upsets and interruptions. • A real-time process reliability model helps to avoid corrosion and erosion issues. • Faithful ‘steady-state’ plant models will be available for offline ‘what-if’ analysis for various plant requirements, such as: Identifying alternate feedstocks that can be processed within the existing hardware to maximise profits. Changes in operating conditions for the respective changes in product/grades, changes in feed/product pricing, and other variables. Conducting debottlenecking and revamp studies to identify bottlenecks in the plant for capacity enhancement. Implementing ‘end-to-end optimisation’ that covers all the unit operations of the process units. collaboration among multiple departments across Reliance. There is a clearcut roadmap for implementing process digital twins for all process units in the coming years. The significance of the process digital twin emerges from its applications in monitoring and control, leading to continuous optimisation and full autonomy in the near future. VIEW REFERENCES Roadmap ahead The process digital twin is a result of
Paras N Shah paras.n.shah@ril.com Jesse Mallhi jesse.mallhi@ril.com
Vikas Deshmukh vikas.deshmukh@ril.com Narendar Mitta Narendar.Mitta@ril.com N C Chakrabarti Nc.Chakrabarti@ril.com Sathiyanarayanan A sathiyanarayanan.a@ril.com
Refining India
63
Powered by FlippingBook