Transforming process safety and efficiency in modern refineries Smart optical analysers, along with chemometrics, DRL and AI, turn measurements into a foundation for continuous learning and optimisation
Gregory Yakhnin, Ariel Kigel, Gadi Briskman, Parul Varma, and Ravi Krishnamoorthy Modcon
L ong before artificial intelligence (AI) optimisation challenges. A crude distillation unit (CDU) is not simply a separation process; it is a continuous balancing act between energy efficiency, product quality, corrosion risk, operational stability, and economic yield. Changes in crude composition, salt or water content, and furnace duty or reflux ratio continually shift this equilibrium. In this sense, the modern refinery has always been an optimisation system. What has changed is not the problem’s complexity, but the ability to observe and control it. The difference between reactive operation and proactive optimisation is defined by measurement quality. When key became familiar, refineries were already tackling one of industry’s most complex properties are measured continuously and reliably in real time, optimisation becomes predictive rather than corrective. The same logic applies to gasoline blending. Although often perceived as a commercial activity, blending is one of the most technically demanding control problems in the refinery. Octane, vapour pressure, density, sulphur, and emissions specifications must be met simultaneously while minimising giveaway and maximising component utilisation. Without accurate, real-time quality measurements, blending systems operate conservatively, protecting compliance at the expense of margin. These examples reveal a fundamental principle: measurement is not instrumentation. Measurement is economics, safety, and sustainability translated into data. Today, hydrogen electrolysers are being added alongside traditional refining units, and the
same principle holds. Electrolysers must operate within strict safety limits, maintain purity targets, minimise energy consumption, and protect asset integrity. The molecules differ, but the optimisation challenge remains the same. This convergence forms the foundation of smart optical analysers, supporting CDU optimisation, gasoline blending, and hydrogen production through continuous, high-integrity data. Four pillars of advanced process optimisation Across refining and hydrogen applications, all effective optimisation strategies converge towards four inseparable pillars (see Figure 1 ): Safety: Safety remains the first and non- negotiable objective. It includes preventing corrosion and fouling in CDUs, controlling flammable or explosive atmospheres, ensuring hydrogen purity and oxygen exclusion, and reducing human exposure to hazardous environments. Measurement accuracy directly
Safety
Economics
Sustainability
Performance
Figure 1 Four pillars of advanced process optimisation
Refining India
17
Powered by FlippingBook