rates are minimised without compromising safety, dryer regeneration is optimised to real moisture load, cooling demand is matched to operating state, and compression is stabilised to avoid energy peaks, transforming the BOP from a cost centre into a performance lever. PASS: intelligent safety integration The integration of smart optical analysers with the Process Alarm And Safety System (PASS) marks a critical evolution in plant safety. Traditionally, safety systems have been static and conservative, separated from optimisation systems by design. Their purpose was protection, not performance. Optical analysers change this relationship by providing continuous, high-integrity measurement of safety-critical variables. Oxygen partial pressure in hydrogen systems, moisture and oxygen in inerting networks, salt and water in crude feeds, and quality risk in blending are measured directly rather than inferred indirectly. This enables PASS to move from threshold- based protection to condition-based protection. Instead of reacting only when limits are exceeded, abnormal trends are detected earlier and corrected before hazardous conditions develop. When combined with AI and DRL, PASS becomes adaptive. Alarm thresholds remain absolute and inviolable, but warning zones and pre-alarm strategies can be optimised dynamically to reduce nuisance alarms while strengthening true hazard detection. Safety and optimisation cease to be competing objectives; they become complementary functions of the same measurement architecture. Hydrogen as natural extension of refinery optimisation Hydrogen electrolysers represent not a departure from refinery engineering, but its continuation. The same optimisation principles that govern CDU operation and blending economics apply directly to hydrogen production. Electrolysers must maintain strict purity, operate within narrow safety margins, minimise energy consumption, and protect long-term asset integrity. Oxygen partial pressure measurement becomes the central safety variable. DRL and AI learn how operating envelopes, purge strategies, drying systems, and transient behaviour influence efficiency and safety.
This makes hydrogen systems an ideal extension of refinery optimisation philosophy. Measurement-driven control, golden setpoints, and lifecycle optimisation are as applicable to hydrogen as they are to crude processing. PaaSS: optimisation across the lifecycle The final evolution is embedding analysers into a Process-as-a-Service System (PaaSS) lifecycle. In this model, analysers, models, AI, safety logic, and optimisation are not project deliverables but living services that evolve with the plant. Analysers are continuously monitored for signal health and drift. Models are retrained as crude slates, blend recipes, or hydrogen operating modes change. AI policies are updated as the plant learns. PASS logic remains aligned with real process behaviour to ensure optimisation does not degrade after commissioning. It is alive, adaptive, and economically relevant for the asset lifetime. Conclusion CDU optimisation, gasoline blending, and hydrogen electrolysis express the same challenge: operating complex systems as close as possible to their true optimum without compromising safety. Smart optical analysers, combined with chemometrics, AI, and DRL, transform measurement into the foundation of continuous learning and autonomous optimisation. They reduce Capex and Opex, strengthen safety, improve sustainability, and unlock higher performance. The future of process optimisation lies not in individual instruments, but in intelligent measurement ecosystems that continuously learn, adapt, and improve.
VIEW REFERENCES
Gregory Yakhnin greg.yakhnin@modcon.group Ariel Kigel arielk@modcon-systems.com Gadi Briskman gadib@modcon-systems.com Parul Varma parulv@modcon-systems.com Ravi Krishnamoorthy ravik@proccesanalytik.com
Refining India
21
Powered by FlippingBook