Kazakhstan journal for oil & gas industryKazakhstan journal for oil & gas industry2707-42262957-806XKMG Engineering10901510.54859/kjogi109015Design of a Combined Isomerization-Catalytic Reforming Unit for Increasing Gasoline Octane Number Using an AI-Based Quality Prediction ModelAbdigaliAruzhan<p>Master's student of OP 7M07101 - Chemical Engineering</p>abdigali_aa_1@enu.kzhttps://orcid.org/0009-0005-9636-622XAikenovaNuriya<p>Senior Lecturer, PhD</p>aikenova_nye@enu.kzhttps://orcid.org/0000-0002-1144-4008ZhakhanshaevaZhansaya<p>student of OP 6B05306 - Chemistry of organic substances and polymers</p>zhakhanshayeva_zhn@enu.kzhttps://orcid.org/0009-0009-1804-0549L.N. Gumilyov Eurasian National University831106202621092026Copyright © , Abdigali A., Aikenova N., Zhakhanshaeva Z.<p style="font-weight: 400;">This study presents the design of a combined octane-boosting system employing two complementary technologies: light naphtha isomerization with a deisopentanizer (DIP) recycle for producing AI-92 gasoline, and catalytic reforming for producing AI-95 and AI-98 grades. Three commercial gasoline grades were characterized using pycnometric density measurements, capillary viscometry, and thermogravimetric analysis (TGA) coupled with differential thermal analysis (DTA). The physicochemical data were used as input features for machine learning models (Linear Regression, k-Nearest Neighbors, Decision Tree) to predict the research octane number (RON).</p> <p style="font-weight: 400;">A comprehensive mathematical model was developed for both process units: (1) a pseudo-first-order kinetic model for n-pentane and n-hexane isomerization on Pt/Al₂O₃ chlorinated catalyst with material and heat balance calculations for the isomerization reactor and DIP column; and (2) a kinetic dehydrogenation model for the catalytic reforming reactor with material balance, heat balance, and reformate yield calculations. The isomerization unit achieves a product RON of 87–88 suitable for AI-92 blending, while the reforming unit produces reformate with RON 95–100 for AI-95 and AI-98 grade production. A techno-economic analysis confirmed the feasibility of the combined configuration.</p>gasoline octane numberartificial intelligencemachine learningisomerizationcatalytic reforminglight naphthakinetic modelmaterial balanceheat balancerefinery design.бензиннің октан саныжасанды интеллектмашиналық оқытуизомерленукаталитикалық риформингжеңіл нафтакинетикалық модельматериалдық балансжылу балансымұнай өңдеу зауытының дизайны.Октановое число бензинаискусственный интеллектмашинное обучениеизомеризациякаталитический риформинглегкая нафтакинетическая модельматериальный баланстепловой баланспроектирование нефтеперерабатывающего завода.<p style="font-weight: 400;">The production of high-octane gasoline meeting stringent environmental standards such as Euro-5 represents a major challenge for petroleum-refining industries worldwide, and particularly in Kazakhstan, where refinery modernization programs aim to eliminate low-octane grades (e.g., A-80) and produce fuels complying with Euro-4/Euro-5 requirements [1]. At the same time, the rapid development of artificial intelligence (AI) and digitalization has been declared a strategic national priority [2].</p> <p style="font-weight: 400;">Within the oil refining sector, two principal technologies are most widely employed to increase gasoline octane number: catalytic reforming and isomerization [3,4]. Catalytic reforming converts naphtha (C₇–C₁₀ fraction) into high-octane reformate by promoting dehydrogenation of naphthenes to aromatics, isomerization, and dehydrocyclization. The reformate possesses a high RON of 95–102, making it the primary blending component for high-octane gasoline grades AI-95 and AI-98 [4]. However, reformate contains significant amounts of aromatics and benzene, which are regulated under modern environmental legislation [3].</p> <p style="font-weight: 400;">Light naphtha isomerization, in contrast, rearranges straight-chain C₅/C₆ paraffins into their branched isomers, producing a sulfur-free, benzene-free, olefin-free isomerate with RON up to 92 depending on the process configuration [3,6,7]. This makes isomerate an ideal blending component for the AI-92 grade and a valuable addition to the gasoline pool for reducing overall aromatics content.</p> <p style="font-weight: 400;">The combination of isomerization and catalytic reforming in a single refinery scheme is widely recognized as the most effective strategy for meeting the full spectrum of gasoline quality requirements: isomerization handles the light naphtha fraction to produce clean AI-92 base stock, while catalytic reforming upgrades the heavier naphtha fraction to provide the high-octane reformate needed for AI-95 and AI-98 [3,4,6]. This integrated approach maximizes the utilization of the entire naphtha distillation range.</p> <p style="font-weight: 400;">Recent advances in machine learning have introduced powerful tools for predicting fuel properties. Lin et al. (2025) developed a sparse-autoencoder and ensemble learning framework for predicting octane loss during hydrotreating [8]. Bounaceur et al. (2024) proposed a deep-learning QSPR model for RON, MON, and cetane number estimation [9]. These studies demonstrate that AI can serve as a reliable surrogate model within the refinery design loop [8–10].</p> <p style="font-weight: 400;">The objective of this work is to design a combined isomerization–catalytic reforming system using an AI-based model to predict gasoline quality. The specific tasks include: (1) experimental characterization of three commercial gasoline grades; (2) development and comparison of ML models for RON prediction; (3) design of a light naphtha isomerization unit with DIP recycle for AI-92 production; (4) design of a catalytic reforming unit for AI-95 and AI-98 production; (5) comprehensive mathematical modeling of both units; and (6) techno-economic assessment.</p> <p style="font-weight: 400;">This study demonstrated the design of a combined isomerization–catalytic reforming system for producing the full spectrum of commercial gasoline grades, integrated with an AI-based quality prediction model. The main findings are:</p> <p style="font-weight: 400;">Physicochemical characterization of three commercial gasoline grades (AI-92, AI-95, AI-98) revealed strong correlations between density, viscosity, and octane number (|r| ≈ 0.99), confirming these parameters as reliable ML features. The Decision Tree model achieved exact RON predictions (MAE = 0.00, R² = 1.00) for all three grades.</p> <p style="font-weight: 400;">A light naphtha isomerization unit with DIP recycle was designed for AI-92 production. The mathematical model includes pseudo-first-order kinetics (Ea = 105.4 kJ/mol for n-C₅), material and heat balances confirming near-isothermal reactor operation (ΔT = 0.017 °C), and McCabe–Thiele DIP column design (70 trays, D = 5.0 m, R = 7.54). The isomerate RON = 87.4 serves as the primary blending component for AI-92 gasoline.</p> <p style="font-weight: 400;">A catalytic reforming unit with three semi-regenerative reactors was designed for AI-95 and AI-98 production. The kinetic model includes Langmuir–Hinshelwood dehydrogenation kinetics and associated material/heat balances. The reformate RON = 98.5 at 500 °C with 85.4% liquid yield provides the high-octane base for AI-95 (blended RON 95.3) and AI-98 (blended RON 98.2).</p> <p style="font-weight: 400;">Techno-economic analysis confirmed the viability of the combined system: isomerization ROI ≈22%/year (payback 3.5–4.0 years), reforming ROI ≈25%/year (payback 3.0–3.5 years). The combined CAPEX is $83–105M with a blended ROI of ~24%/year.</p> <p style="font-weight: 400;">The gasoline blending analysis demonstrated that the full range of commercial grades can be produced by adjusting the proportions of isomerate, low-severity reformate, and high-severity reformate in the blending pool.</p> <p style="font-weight: 400;">Future work should expand the ML training dataset, validate the mathematical models through Aspen HYSYS simulation, and optimize the blending strategy through multi-objective optimization.</p>1.International Energy Agency. Kazakhstan Energy Profile. 2022. https://www.iea.org/reports/kazakhstan.2.Tokayev, K.-J. Address of the President of the Republic of Kazakhstan, 2025.3.Osman, M.A.; Mujtaba, I.M. Experiment and Simulation of an Optimum Design of Naphtha Recycle Isomerization Unit. Processes 2023, 11, 3406.4.Jechura, J. Gasoline Upgrading. CBEN 409 Lecture Notes, Colorado School of Mines, 2019.5.Wikipedia. Alkylation unit. Available online: https://en.wikipedia.org/wiki/Alkylation_unit6.Axens. How to Achieve Gasoline Pool Octane Requirement. 2023.7.KBR. MAX-ISOM Isomerization Technology. 2023.8.Lin, Y.; et al. AI-Based Prediction Framework for Octane Loss during Gasoline Hydrotreating. Preprints 2025, 202504.2593.9.Bounaceur, R.; et al. A Multimodal Learning Model for RON, MON and CN estimation. Fuel 2024, 381, 133295.10.Design of novel high-performance fuels with artificial intelligence. Appl. Energy Combust. Sci. 2025, 21, 100338.11.Mohamed, M.F.; et al. Improving gasoline quality from MIDOR light naphtha isomerization unit. Egypt. J. Pet. 2017, 26, 111–124.12.Muhammed, N.I.; et al. Raising the RON using optimized SMB technology. Fuel 2023, 337, 127178.13.Benavides, A.; et al. Predicting octane number. Processes 2023, 11(5), 1437.14.Brudzewski, K.; et al. Gasoline quality prediction using electronic nose. Fuel 2006, 85(5–6), 553–559.15.Al Ibrahim, E.; Farooq, A. Octane prediction from infrared spectroscopy. Energy Fuels 2019, 33(11), 11027–11036.16.Brown, M.E. Introduction to Thermal Analysis, 2nd ed.; Springer: 2001.17.Kumari, A.; et al. Petroleum Product Analysis. Energy Sources, Part A 2025.18.Pedregosa, F.; et al. Scikit-learn: Machine Learning in Python. JMLR 2011, 12, 2825–2830.19.Fix, E.; Hodges, J.L. Discriminatory Analysis. Int. Stat. Rev. 1989, 57, 238–247.20.Breiman, L.; et al. Classification and Regression Trees; CRC Press: 1984.21.Trost, T.; et al. Rheological Properties of Fuels. Fuels 2021, 2, 289–303.22.Kashyap, D.; et al. Gasoline Composition and Engine Performance. Energies 2024, 17, 4567.23.Volkova, G.G.; et al. Kinetics of n-Pentane Isomerization over Pt/Al₂O₃. React. Kinet. Catal. Lett. 2007, 91, 63–70.24.Chuzlov, V.A.; et al. Mathematical Modeling of Light Naphtha Isomerization. Chem. Eng. J. 2019, 378, 122178.25.Seader, J.D.; Henley, E.J. Separation Process Principles, 3rd ed.; Wiley: 2011.26.Ancheyta-Juárez, J.; Villafuerte-Macías, E. Kinetic Modeling of Naphtha Catalytic Reforming Reactions. Energy Fuels 2000, 14(5), 1032–1037.27.Rahimpour, M.R.; et al. A novel dynamic modeling and simulation of industrial naphtha reformer. Chem. Eng. Process. 2003, 42(1), 67–82.28.Smith, R.B. Kinetic Analysis of Naphtha Reforming with Platinum Catalyst. Chem. Eng. Prog. 1959, 55(6), 76–80.