Refining India March 2026 Issue

March 2026 REFINING INDIA Technologies driving a sustainable future

ALL FOR PAPERS NOW OPEN! ndia Technology Conference returns on 21-22 September 2026 at Le Méridien, New PTQ Magazine and IDS, it gathers senior refinery leaders and global tech firms to ndustry innovations. With India’s refining capacity set to grow to 450 million tonnes year by 2030, this is the perfect platform to showcase your solutions. e looking for, but not limited to, the following key conference themes: ROCHEMICAL AND CHEMICAL INTEGRATION / EXPANSION • NEW PROCESS / NOLOGIES • PROCESS OPTIMISATION / ENERGY EFFICIENCY • DIGITALISATION, ANALYTICS AND AI • ENERGY TRANSITION TECHNOLOGIES come submissions that go beyond these themes. All submissions will enter a review process and are not automatically guaranteed for inclusion.

SAVE THE DATE! Refining India 2026 will be held on 21-22 September 2026 in New Delhi Brought to you by PTQ magazine and with support from IDS, Refining India is back bigger and better! SUBMISSION DEADLINE: 31 MARCH 2026 21-22 SEPTEMBER 2026 NEW DELHI

nce to share new perspectives at the leading event in the refining sector. Please submit presentation along with a 200-word abstract to presentations@refiningindia.com.

CALL FOR PAPERS NOW OPEN! The 13th Refining India Technology Conference returns on 21-22 September 2026 at Le Méridien, New Delhi . Hosted by PTQ Magazine and IDS, it gathers senior refinery leaders and global tech firms to explore the latest industry innovations. With India’s refining capacity set to grow to 450 million tonnes per year by 2030, this is the perfect platform to showcase your solutions.

refiningindia.com

Maximise FCC unit flexibility with JM Additives and INTERCAT addition systems Johnson Matthey’s FCC Additives, along with INTERCAT TM addition systems give you the flexibility to handle opportunity feeds and capture product value in your FCC unit.

VISIT MATTHEY.COM OR SCAN THE QR CODE TO LEARN MORE

Contents

March 2026

2 Artificial intelligence: a frontier technology powering India’s oil refining revolution Manoj Sharma

5 Transitioning refinery hydrogen systems from grey to green Ishita Bansal Expert on Energy Transition & Digital Transformation

AI as a strategic catalyst Sujoy Choudhury The Chatterjee Group (TCG)

Transforming process safety and efficiency in modern refineries Gregory Yakhnin, Ariel Kigel, Gadi Briskman, Parul Varma, and Ravi Krishnamoorthy Modcon Hydrogen blistering in rich amine flash drum of VGO HDS unit Avinash Sankpal, Arun Kumar, Aman Kumar Sahoo, Prashant Nandanwar, and Abdul Quiyoom Kochi Refinery, BPCL

Utilising process simulation to reduce SOx emissions Ganank Srivastava Bryan Research & Engineering, LLC

Understanding heat integration losses in refineries Tania Guha Engineers India Limited

Accelerating energy transition through sustainable process Arun Kuniyil, Bhanu Prasad S G, Pramod Kumar, Sriram S, and V.K. Maheshwari HP Green R&D Centre, Hindustan Petroleum Corporation Limited

Emerging technologies for low-carbon intensity syngas: Part 1 Ajay Misra Sr. Consultant (Fertilizers & Petrochemicals)

Benefits of high-performance Pt-based isomerisation catalyst Vimal K Upadhyay, Pushkar Varshney, Satyen K Das, Atul Ranjan, RK Kaushik Singha,

and Alok Sharma Indian Oil Corporation Ltd Chaitanya Sampara Viridis Chemicals Pvt Ltd

End-to-end digital enablement of petcoke gasification complex Saswat Panda and Vijay Sastri Reliance Industries Limited

© 2026. The entire content of this publication is protected by copyright. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means – electronic, mechanical, photocopying, recording or otherwise – without the prior permission of the copyright owner. The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while every care has been taken in the preparation of all material included the publisher cannot be held responsible for any statements, opinions or views or for any inaccuracies.

Refining India

Artificial intelligence: a frontier technology powering India’s oil refining revolution

Editor Manoj Sharma editor@refiningindia.com +91 989 9077 595 Managing Editor Rachel Storry

rachel.storry@emap.com tel +44 (0)7786 136440 Editorial Assistant Lisa Harrison lisa.harrison@emap.com Graphics Peter Harper Business Development Director Paul Mason info@decarbonisationtechnology.com tel +44 844 5888 771 Managing Director Richard Watts richard.watts@emap.com

For decades, oil refining was a game of steady-state engineering and rigid schedules. However, as we move through 2026, the industry is shedding its ‘legacy’ skin. The integration of artificial intelligence (AI) is fundamentally rewriting the refinery’s DNA. The

journey from manual, analogue operations to the AI- driven refineries of 2026 has been nothing short of a revolution. This transition was not an overnight switch but a strategic evolution. The trajectory of AI adoption in Indian refining exemplifies a broader industrial transition from mechanical systems to intelligent, data-driven operations. At the heart of refining operations lies complexity, involving hundreds of interconnected units, massive data streams from sensors, and narrow margins for error. Traditional control systems, while reliable, often struggle to adapt to dynamic conditions. AI bridges this gap by learning from historical and real-time data, enabling predictive, adaptive, and autonomous decision-making. This shift marks a move from reactive operations to proactive intelligence. AI implementation in Indian refineries has expanded significantly between 2024 and early 2026, with a clear distinction between the large-scale infrastructure focus of bigger players and a process-driven improvement focus by others. Indian refiners are not only integrating AI into massive hardware and digital infrastructure projects to drive global competitiveness but also focusing on indigenising AI technologies for enhancing operational safety, efficiency, and sustainability. Operational excellence and core refining improvements One of the most significant impacts of AI in refining is in process optimisation. Machine learning models continuously analyse variables such as temperature, pressure, flow rates, and feedstock composition to optimise reactors, distillation columns, and heat exchangers. Refineries use digital twins to simulate ‘what if’ scenarios for complex units like the crude distillation unit (CDU) and fluid catalytic cracking (FCC). The result is higher throughput, improved product quality, and reduced energy consumption – critical factors in an industry where

EMAP, 10th Floor Southern House Wellesley Grove, Croydon CR0 1XG

Cover Story AI-driven refineries will become more adaptive and efficient

Refining India

even marginal efficiency gains translate into substantial economic benefits. The adoption of advanced process control, integrated automation systems, and real-time data analytics across pipeline and refining networks further complements AI strategies. These implementations, which leverage Internet of Things (IoT) sensors, Supervisory Control And Data Acquisition (SCADA) systems, and enterprise resource planning tools, provide the foundational digital layer necessary for AI applications to function effectively. AI is also redefining maintenance strategies. Instead of relying on scheduled or reactive maintenance, refineries now deploy predictive maintenance systems that forecast equipment failures before they occur. By analysing vibration data, acoustic signals, and operational trends, AI helps prevent unplanned shutdowns, extend equipment life, and enhance plant reliability. In the realm of safety and environmental compliance, AI-powered computer vision and analytics are proving invaluable. Automated monitoring systems detect gas leaks, flaring inefficiencies, and safety violations in real time. As global pressure mounts to reduce emissions and improve sustainability, AI offers refiners a practical pathway to meet environmental targets while maintaining profitability. Beyond the plant gates, AI in Indian oil refining is expanding into end-to-end value chain optimisation. Advanced AI models analyse global price trends, demand signals, and logistics data to improve crude procurement decisions, optimise inventory levels, and streamline distribution. AI-assisted market forecasting, including real-time scenario testing, enables refiners to mitigate price volatility risks and align output with anticipated demand patterns. Challenges on the frontier Despite these advancements, several hurdles remain. Legacy systems and fragmented data infrastructures complicate real-time data integration – a prerequisite for advanced AI efficacy. There is also a talent shortage in AI engineering and data science within

the petroleum sector. Strategic investments in workforce upskilling and collaborations between industry, academic institutions, and technology partners will be essential to scale frontier AI deployments. Conclusion: a strategic imperative for the future AI in India’s oil refining sector is transitioning from operational support to a strategic driver of competitiveness and sustainability. As the industry embraces frontier technologies – from generative models to autonomous AI agents – refineries will become more adaptive, resilient, and efficient. The successful integration of these capabilities will not only improve refinery performance but also support India’s broader energy security and climate objectives in an era of global transformation. However, projections of the AI sector’s energy needs over the coming years are massive in scale. The International Energy Agency expects AI’s energy demand to double between now and 2030, presenting a serious challenge to energy security in many nations and regions where large data centre developments are planned. models to autonomous AI agents – refineries will become more adaptive, resilient, and efficient ” Manoj Sharma is an executive leader with more than 35 years of experience in petroleum refining, petrochemical operations, and strategic management. He has proven expertise in refinery optimisation, green initiatives (CCUS, green H₂, biofuels), crude oil trading, risk management, and digital transformation. He has a strong background in international business, process engineering, and corporate governance as a board director. He holds an International MBA from the University of Ljubljana, Slovenia, and a BE in chemical engineering from Punjab University, Chandigarh. “ As the industry embraces frontier technologies – from generative

Refining India

UNPARALLELED INNOVATORS. PROBLEM- SOLVING PARTNERS.

At Ketjen, we’re dedicated to nonstop innovation, delivering industry leading catalysts that provide timely solutions, enhance performance,

increase yields, and maximize returns.

Learn more at KETJEN.COM

Transitioning refinery hydrogen systems from grey to green The engineering considerations for transitioning refinery hydrogen systems from reformer-based grey hydrogen to electrolyser-based green hydrogen

Ishita Bansal Expert on Energy Transition & Digital Transformation

H ydrogen is a critical utility in modern refineries, underpinning key conversion and upgrading processes, such as hydrotreating, hydrocracking, delayed coking, and residue upgrading. Hydrogen availability, purity, and pressure stability directly influence refinery throughput, product quality, catalyst activity, hydrogen partial pressure, and overall unit severity. In hydrogen-intensive refineries, even short-duration disruptions in hydrogen supply can rapidly translate into throughput loss, off-spec products, accelerated catalyst deactivation, increased make-up hydrogen demand, or forced unit rate reductions. As a result, hydrogen system reliability is treated as a core operational requirement rather than a secondary utility consideration. In most Indian refineries, hydrogen demand is predominantly met through captive steam methane reforming (SMR) units. These systems are designed for continuous, steady-state operation, delivering high-purity hydrogen at relatively stable pressures and flow rates aligned with base-load refinery operation. Over decades of operation, refinery hydrogen networks, headers, compressor configurations, purification schemes, and control philosophies have been optimised around the predictable operating characteristics of reformers, shift reactors, CO₂ removal systems, and downstream purification units. Hydrogen system design has therefore prioritised stability, availability, and predictability over operational flexibility. While SMR-based hydrogen systems are well proven and operationally robust, they also represent a significant source of direct carbon

dioxide (CO2) emissions and expose refineries to volatility in natural gas pricing, long-term fuel supply security, and geopolitical risk. With tightening emissions norms, increasing scrutiny on refinery carbon intensity, and long- term uncertainty around fossil fuel economics, refinery hydrogen has emerged as one of the most impactful and technically complex levers for decarbonisation. Green hydrogen, produced through water electrolysis using low-carbon electricity, offers a potential pathway to reduce the carbon intensity of refinery hydrogen supply. However, transitioning from grey to green hydrogen in operating refineries is not a simple substitution exercise. Unlike greenfield projects, “ Green hydrogen, produced through water electrolysis using low-carbon electricity, offers a potential pathway to reduce the carbon intensity of refinery hydrogen supply ” operating refineries are constrained by legacy infrastructure, tightly integrated process units, limited operating margins, brownfield space constraints, and stringent reliability requirements. Introducing electrolyser-based hydrogen therefore brings a new set of engineering, operational, safety, and system- integration challenges that must be addressed in a structured and risk-managed manner. This article examines the key engineering considerations involved in transitioning refinery hydrogen systems from reformer-based grey hydrogen to electrolyser-based green

Refining India

minimal and designed primarily for short-term buffering. Hydrogen networks rely heavily on real-time balancing between production and consumption. Control philosophies are built around predictable reformer behaviour, slow dynamic response, and high mechanical availability. Any disturbance in hydrogen supply can quickly

hydrogen. The focus is on practical integration within existing refinery infrastructure, rather than greenfield concepts or policy-driven targets, with emphasis on maintaining operational stability, safety, flexibility, and long- term resilience during the transition. Existing refinery hydrogen systems: design and operating philosophy Most refinery hydrogen networks have evolved around SMR-based production with a strong emphasis on continuous, steady-state operation. Hydrogen is typically produced at relatively high pressure and purity, then distributed through a refinery-wide header system, supplying multiple consumers with differing pressure, flow, and purity requirements. These consumers often include hydrotreaters, hydrocrackers, delayed coking units, residue upgrading units, and hydrogen recovery systems, each with different sensitivities to hydrogen pressure fluctuations, purity deviations, and availability. To manage these diverse requirements, refineries deploy a hierarchy of hydrogen headers, make-up compressors, recycle compressors, purge gas compressors, and purification systems. Pressure swing adsorption (PSA) units play a critical role in upgrading hydrogen purity and recovering hydrogen from off-gases generated in process units, such as hydrotreaters and catalytic reformers. Recovered hydrogen is reintegrated into the network, improving overall hydrogen utilisation efficiency and reducing dependence on fresh hydrogen production. Intermediate hydrogen storage is typically

cascade into process unit instability, making hydrogen system reliability a dominant design and operating constraint. These established design principles strongly influence how green hydrogen can be introduced. Electrolyser-based hydrogen must integrate into a system optimised for stability rather than variability, requiring deliberate engineering adaptation rather than simple capacity addition. Characteristics of green hydrogen production Electrolyser-based hydrogen production differs fundamentally from SMR systems in both operating behaviour and system interfaces. Electrolysers are electrically driven and capable of rapid load changes, enabling flexible operation across a wide turndown range. While this flexibility allows alignment with variable power availability and electricity pricing, it introduces short-term variability that is unfamiliar to refinery hydrogen systems designed for steady-state operation. Electrolyser outlet pressure and purity depend on technology selection and balance-of-plant configuration. Alkaline and proton exchange membrane electrolysers exhibit different dynamic characteristics, efficiency profiles, degradation mechanisms, and maintenance requirements. Frequent start-stop operation, power interruptions, voltage fluctuations, and load cycling can influence membrane life, stack degradation rates, and hydrogen output stability. Electrolyser availability also becomes closely coupled to power system reliability and

Refining India

electricity market conditions, particularly when linked to renewable energy sources. From an engineering perspective, these characteristics necessitate careful assessment of how electrolytic hydrogen interfaces with existing headers, compressors, purification systems, and downstream consumers. Variability must be absorbed without compromising unit stability, safety, or product quality. Transition pathways: incremental integration strategies For most operating refineries, immediate and complete replacement of grey hydrogen with green hydrogen is neither practical nor desirable. Incremental transition strategies therefore represent the most viable and lowest- risk approach. Common pathways include partial substitution of SMR output with electrolytic hydrogen, hybrid grey-green hydrogen systems and phased capacity expansion aligned with refinery revamp or turnaround cycles. These approaches allow refineries to gain operational experience with green hydrogen while limiting exposure to technical and commercial risk. Engineering studies must determine the maximum green hydrogen fraction that can be absorbed without destabilising the hydrogen network. Key considerations include SMR minimum turndown limits, hydrogen header pressure control margins, PSA performance envelopes, compressor turndown constraints, and consumer sensitivity to transient supply variations. In most configurations, SMRs continue to provide base-load hydrogen, while electrolysers operate as supplementary or load- following sources. Hybrid configurations provide flexibility, enabling progressive emissions reduction while preserving reliability and operational confidence. Hydrogen header integration and pressure management The refinery hydrogen header is the backbone of hydrogen distribution and often represents the most critical integration challenge. Electrolyser output must be matched to header pressure requirements through compression, pressure control, or intermediate buffering arrangements.

Introducing variable hydrogen sources into a header designed for a steady supply requires revisiting control philosophies. Pressure control strategies must prioritise delivery to critical consumers under all operating conditions. Flow prioritisation, constraint handling, and automated load-shedding logic may require enhancement to manage transient events, such as electrolyser trips, power disturbances, or renewable intermittency. In some cases, segregated headers or dedicated green hydrogen injection points may be justified to limit propagation of variability. These design decisions must be evaluated against capital cost, operational complexity, and reliability objectives. Storage, buffering and demand-side management Hydrogen storage plays a critical role in mitigating variability introduced by green hydrogen production. Intermediate storage enables partial decoupling of hydrogen production and consumption, improving system resilience and operational flexibility. Engineering considerations include storage integration with hydrogen network control systems. While large-scale hydrogen storage remains capital-intensive, even modest buffer volumes can significantly smooth short-term fluctuations. Demand-side management becomes capacity sizing, pressure rating, safety classification, line-pack utilisation, and increasingly important in hybrid systems. Prioritisation of hydrogen consumers, optimisation of unit operating severity, catalyst life management, and coordination with turnaround planning can help manage hydrogen availability during transient conditions. Safety and materials compatibility are central to any modification of hydrogen systems. Changes in operating pressure, cycling frequency, temperature, or impurity profile influence risks related to hydrogen embrittlement, fatigue, leakage, and accelerated degradation. Introducing electrolytic hydrogen may require reassessment of metallurgy in pipelines, valves, Materials compatibility and safety considerations

Refining India

Economic and implementation considerations

Transitioning from grey to green hydrogen involves trade-offs between capital expenditure, operating cost, emissions reduction, and reliability. Electrolyser capital costs, electricity pricing, utilisation factors, infrastructure modification

costs and policy incentives all influence project economics. Phased implementation allows refineries to align investments with asset life and technology maturity. In the near to medium term, hybrid systems are expected to dominate, balancing decarbonisation goals with operational and financial constraints. Operational readiness and commissioning considerations Operational readiness is a critical yet frequently underestimated aspect of transitioning refinery hydrogen systems from grey to green. While electrolyser installation and mechanical completion represent visible project milestones, successful integration depends heavily on commissioning philosophy, sequencing logic, control system readiness, and operator preparedness. Unlike SMR units, electrolysers introduce new operating dynamics, including frequent start- stop cycles, rapid load-following behaviour, and dependency on power availability. Commissioning plans must therefore explicitly address transient scenarios, such as grid disturbances, electrolyser trips, black-start recovery, and coordinated restart with hydrogen consumers operating at different severity levels. Start-up sequencing logic must ensure that hydrogen header pressure and purity are stabilised before admitting green hydrogen into sensitive units such as hydrocrackers. Temporary operating envelopes, conservative ramp rates, and staged integration are often required during early commissioning phases. Control system integration is central to operational readiness. Distributed control systems must accommodate variable hydrogen

compressors, and storage vessels. Higher cycling frequencies and different impurity profiles may necessitate revised inspection and maintenance strategies. Leak detection systems, ventilation provisions, hazardous area classification, operating procedures, and emergency response plans must be reviewed and updated to maintain compliance with refinery safety management systems. Electrical and utility system implications Electrolysers introduce large, dynamic electrical loads into refinery utility systems. Their operation must be coordinated with captive generation, grid supply, backup systems, and load-shedding philosophies. Electrical system studies are required to assess impacts on load profiles, short-circuit levels, voltage stability, harmonics, and power quality. Power reliability becomes a direct determinant of hydrogen availability, linking electrical system performance with process reliability. Role of digitalisation and advanced control Digitalisation is a key enabler for managing hybrid hydrogen systems. Advanced control systems, energy management platforms, and predictive analytics support optimal electrolyser dispatch, hydrogen balancing, and power cost optimisation. Digital twins and scenario analysis tools allow refineries to test operating strategies under different power and hydrogen availability scenarios, reducing operational risk and supporting long-term optimisation.

Refining India

Refinery-specific integration considerations in the Indian context Indian refineries face additional constraints that influence green hydrogen integration. Grid reliability varies significantly by region, making power quality and backup strategies critical design considerations. Water availability and quality for electrolysis must also be assessed carefully, particularly in water-stressed regions. Brownfield space constraints often limit optimal electrolyser layout, influencing piping complexity, pressure drops, and safety zoning. Regulatory approvals, permitting timelines, and coordination with multiple agencies can further affect project execution. Successful integration therefore requires site-specific engineering solutions rather than generic templates, with close coordination between refinery, utility, and grid stakeholders. Conclusions and way forward Transitioning refinery hydrogen systems from grey to green is fundamentally an engineering challenge rather than a simple technology “ Incremental, hybrid approaches provide a pragmatic pathway for Indian refineries to begin decarbonising hydrogen while preserving reliability and flexibility ” substitution. Success depends on disciplined integration with existing infrastructure, robust safety management and strong cross- disciplinary coordination. Incremental, hybrid approaches provide a pragmatic pathway for Indian refineries to begin decarbonising hydrogen while preserving reliability and flexibility. As technology costs decline and operational experience grows, green hydrogen can progressively assume a larger role in refinery hydrogen networks, supporting long- term competitiveness in a decarbonising energy landscape.

input while maintaining stable header pressure and prioritised supply to critical units. Alarm management, override logic, and interlocks must be validated under dynamic conditions to avoid nuisance alarms, operator overload, or unintended unit trips. Operator training is equally important. Refinery personnel accustomed to reformer- based hydrogen production may have limited familiarity with electrolyser behaviour. Structured training programmes, dynamic simulations, and supervised trial runs significantly reduce operational risk during initial integration. Common pitfalls observed in early green hydrogen integration projects Early refinery projects integrating green hydrogen have highlighted several recurring pitfalls. A common technical issue is the underestimation of hydrogen production variability and its interaction with existing headers, leading to pressure oscillations, compressor hunting, or PSA instability. Insufficient buffering and storage are another frequent challenge. Projects relying solely on real-time balancing often experience hydrogen shortfalls during power disturbances or renewable intermittency, forcing curtailment of hydrogen consumers and eroding operator confidence. Electrical system integration issues are also prevalent where harmonics, voltage flicker, or short-circuit impacts were inadequately assessed during design. In some cases, electrolyser availability has been constrained by grid events unrelated to refinery operations. Contracting and organisational issues also emerge. Poor definition of battery limits between electrolyser vendors and refinery systems, unclear O&M responsibilities, and delayed handover of control philosophies can undermine otherwise sound engineering designs. Materials and safety risks are sometimes underestimated, particularly where increased cycling accelerates fatigue or embrittlement in legacy equipment. Projects that fail to update inspection and maintenance regimes often encounter premature equipment degradation.

Ishita Bansal ishitabansal.me@gmail.com

Refining India

AI as a strategic catalyst

Enterprise-wide AI integration combined with deep process expertise enables step-change improvements in operational performance

Sujoy Choudhury The Chatterjee Group (TCG)

I ndia has emerged as one of the fastest- growing major economies, consistently outpacing both advanced and emerging peers. Despite ongoing geopolitical challenges and energy market volatility in 2025, the country continues to demonstrate robust growth, maintaining its position as a global growth leader. According to the World Economic Outlook (April 2025), India’s GDP is projected to grow by 6.2% in 2025, 6.3% in 2026, and 6.5% in 2030, well ahead of other major economies, such as China, the US, Japan, and the UK. In 2025, India made global headlines by overtaking Japan to become the fourth-largest economy in the world. With a GDP of $4.187 trillion and a per capita GDP of $2,934, India combines strong economic growth with a large and diverse population. The oil and gas sector is a cornerstone of India’s economic engine, contributing around 15% to the country’s GDP and meeting more than 30% of its total energy demand (see Figure 1 ). Notable government initiatives Sustaining and strengthening India’s oil and gas

infrastructure is vital for national growth, energy independence, and maintaining geopolitical stability amid evolving global energy dynamics. In this light, the government initiatives (see Figure 2 ) are critical as the industry faces mounting pressures and challenges (outlined in the next section). Industry status quo: limitations of traditional control systems The challenges outlined above also help set the context for understanding the oil and gas sector deeply, which has traditionally been very conservative in its approach, predominantly emphasising reliability and safety in its operations. Historically, the sector has relied heavily on established, classical advanced process controls (APCs) to manage plant processes. These systems, while foundational, often lack the dynamic adaptability needed to fully optimise complex industrial operations. The industry’s cautious stance has limited the adoption of AI, focusing mainly on proven methods to ensure steady and safe operations rather than embracing innovative AI-driven

Upstream

Downstream

Midstream

2nd ranked in ethanol - blended petrol 4th largest LNG importer

3rd largest oil consumer 3rd largest oil importer

4th largest rening capacity 7th largest exporter of rened petroleum products India currently operates 19 Public Sector Undertaking (PSU) reneries, 3 Private Sector reneries and 1 Joint Venture renery . The country’s refining capacity increased from 215.1 million m etric t ons per annum (MMTPA) in April 2014 to 256.8 MMTPA in April 2024 .

The country’s total oil consumption is approximately 5.7 million barrels per day , with global demand exceeding 100 million barrels per day. India depends heavily on imports, sourcing 90% of its crude oil - over 243 million tonnes annually - at a cost surpassing $143 billion , highlighting strategic vulnerabilities and the sector’s economic weight.

Natural gas demand is expected to reach nearly 103 billion cubic metres (bcm) annually by 2030 from approximately 64 bcm in 2024. India aims to double its natural gas share in the energy mix from 6 - 7% to 15% by 2030 .

Figure 1 India’s energy position – a snapshot in numbers.

Source: PIB, PNGRB studies and Energy Statistics India 2024

Refining India

Refining Capacity Expansion : India aims to nearly double its refining capacity to 450 - 500 million tonnes per annum (MMTPA) by 2030, with several new refinery projects under way (e . g. West Coast Refinery, HPCL Rajasthan Refinery, Numaligarh Refinery expansion ) Production-Linked Incentive (PLI) Scheme for Petroleum Rening: Incentivises capacity expansion and technological upgrades in reneries to boost domestic manufacturing and exports. The Pradham Mantri Ujjwala Yojana (PMUY) , including its expanded Ujjwala 2.0 phase, aims to provide clean cooking fuel by oering subsidised LPG connections to economically weaker households across India.

Hydrocarbon Exploration and Licensing Policy (HELP) and its Open Acreage Licensing Policy (OALP) : Policies to boost domestic production with revenue sharing and year-round bidding, including 25 new blocks in OALP’s 10th round. Oilelds (Regulation and Development) Amendment Act, 2025 : Modernises regulations to attract investment with investor-friendly clauses and streamlined lease renewals. Draft Petroleum & Natural Gas Rules, 2025 : Supports interation of renewables, mandates emissions monitoring and establishes carbon capture frameworks. National Seismic Programme Expansion : Enhances exploration using advanced technologies and AI-driven seismic data analysis.

Figure 2 Driving growth: India’s energy policy priorities for 2025.

Source: PIB, DGH, MoPNG

solutions that could unlock greater efficiency and productivity. The reasoning above also partly explains why continuous process industrial players, including those in oil and gas, have been very slow to implement AI. Complex processes depend on physical relationships across a wide range of variables, making modelling and improvement through analytics challenging. Additionally, data is often scarce, siloed, or of low quality because it resides in disparate IT-OT systems related to individual equipment rather than being integrated across the entire process. Hence, many organisations must rely on experienced operators’ intuition to manage changing conditions. This convergence of technical, operational, and organisational barriers has historically locked the refinery and petrochemical operations into incremental improvements rather than a transformative change. AI-driven transformation of refinery and petrochemical operations These historical constraints are no longer insurmountable. Where traditional approaches have struggled with fragmented data and manual optimisation, modern AI platforms – combining hybrid intelligence, advanced analytics, and deep process expertise – can potentially overcome these long-standing constraints, enabling continuous, intelligent improvement across refinery operations. Additionally, the continuous nature of refinery operations makes them prime candidates for

improvement through analytics. Scientific relationships between process inputs and outcomes guarantee that targeted interventions can be quantified and tracked over time. Unlike batch processes, continuous processes enable ongoing modelling, rapid testing, and refinement. AI can reshape refinery operations by replacing rigid, rule- based control systems with continuous, intelligent, and self-optimising frameworks that adapt dynamically to changing process conditions. Through the integration of “ Through the integration of advanced ML, optimisation algorithms, and real-time sensor data, refining and petrochemical operations can now drive improvements in throughput, yield, and energy efficiency ” advanced machine learning (ML), optimisation algorithms, and real-time sensor data, refining and petrochemical operations can now drive improvements in throughput, yield, and energy efficiency while strengthening asset reliability and safety performance. The rise of Industry 4.0 and embedded sensors generates vast streams of data, enabling real- time optimisation with the power of cloud computing. In this regard, AI can effectively handle a multitude of variables and their interdependencies, and arrive at optimal models in much less time. Agentic AI – where individual

Refining India

Downstream use cases

Customer retail

Asset management use cases Look, listen, feel (LLF) automation Inspection data management system (IDMS) Smart meters and sensors Smart defence (Cyber OT security) AV/VR, APC-based training for O&M E -permits Unified asset performance and mechanical integrity management Digital twin for APM Digital enablement of paper/manual processes HSE & sustainability use cases Dynamic risk assessments (DRA) Fuel compliance management (FCM) GHG accounting and reporting Renewable energy certicate (REC) accounting Renewable energy production Carbon capture utilisation and storage (CCUS)

Supply chain optimisation Customer 360 – Omnichannel CS and support Customer loyalty Cashow analytics City gas distribution – customer portal

AI for production optimisation Pricing optimisation dashboard

Gas processing optimisation Plant performance monitoring Integrated supply chain optimisation Real time optimiser (RTO) Yield and throughput optimisation Grade market and sequence optimisation

Upstream use cases

AI for subsurface data analysis Cloud and OSDU for subsurface data Digitisation of historical data ESG compliance for upstream Mobility-enabled field operations Well - developed applications Augmented w ell development Advanced r ig scheduling Autononomous well operations

Cracker health monitoring Energy demand forecasting Automated fuel re-ordering Digital product passport Digital terminals AI/ML-based yield and quality

Electronic proof of delivery (ePOD) End-to-end lube inventory visibility Loyalty program me s for customers Hyper-localisation of customer data Remote fuel retail monitoring Commodity trading and risk management automation

Midstream use cases

Pipeline health analysis Intelligent pipeline management Pipeline corrosion detection Turnaround optimisation with drones Mobility - based field services

Figure 3 An illustrative compendium of AI-driven use cases

agents manage specific components of the whole model, exercise autonomous decision- making, and interact with other agents to arrive at a global optimal solution – holds significant potential to transform the oil and gas industry. AI-powered supply chain optimisation synchronises procurement, production, and dispatch decisions across the refinery network. Using predictive analytics and reinforcement learning, AI continuously recalibrates schedules, tank allocations, and logistics plans based on live market signals, feedstock variability, and demand fluctuations. This enables smarter crude selection, reduced demurrage costs, and tighter inventory control, thereby minimising working capital and enhancing responsiveness to market volatility. In production operations, AI-driven process optimisation eliminates manual trial-and-error by using digital twin simulations to test and implement optimal parameters in real time. It dynamically fine-tunes blending ratios, energy loads, and unit constraints to maximise product quality and minimise resource waste. Predictive maintenance models simultaneously forecast equipment degradation patterns from vibration, pressure, and flow data, enabling early intervention and avoiding costly shutdowns. By creating a connected, data-driven operations layer across the refinery’s value chain – from crude input to product dispatch

– AI enables real-time decision-making, higher margins, and improved sustainability, marking the transition toward the fully autonomous ‘digital refinery’ of the future. In this regard, multiple readily implementable use cases (see Figure 3 ) can be developed across streams using AI, enabling integration that drives significant gains such as controlling costs, maximising yields, optimising plant operations to handle various feeds, manufacturing high-yield catalysts, and streamlining supply chains. Against this backdrop, there is a pressing need to accelerate the adoption of advanced AI to deliver real, quantifiable monetary benefits across every operating asset – where even marginal improvements in yields, conversion rates, or reductions in energy and feed consumption can substantially enhance margins and justify investments in AI/ML. Uncovering a live AI-driven industry example of RTO These transformative capabilities are not purely theoretical; they are real-world implementations across the industry. They demonstrate the tangible value AI-driven optimisation delivers. Figure 4 shows a live industry example illustrating the impact of AI-enabled real-time optimisation. The real-time optimisation (RTO) system integrates plant data, including feedstock rates, composition, and operating parameters,

Refining India

Input feed

Constraints Operating parameters – temperature, pressure etc. Equipment duties

Optimiser

Individual feedstock rate Individual feed composition Any recycle rate Recycle composition

Iteration i th

Objective function

Plant model output of products/ inputs to AI objective function

Operating parameters Operating parameters – temperature, pressure etc. No. of furnaces online at any point

Prices and costs

User inputs

Optimal output

Market prices of products Cost of inputs

Operating data, LIMS data

Optimal operating conditions and optimised objective Desired product mix Operating parameters – temperature,

APC/MPC

Predictive model

pressure, CV % open etc. Input feed mix volumes Fuel gas consumption

Oine until optimiser

Minimisation of error

Figure 4 A case of live RTO example

validated with real-time and laboratory inputs. It uses predictive plant models and an AI-based optimiser to run iterative scenarios, searching for optimal conditions under varying constraints such as temperature, pressure, and equipment limits. • Input handling: Operational data from distributed control systems (DCS), laboratory information management systems (LIMS), and manual user entries are validated and fed into the plant model. • Hybrid modelling: The plant model uses both first-principles physics and AI-driven data science to simulate product outputs under different operating scenarios. • AI optimisation loop: The AI optimiser evaluates the plant model’s outputs and seeks configurations that maximise profits or yields, considering current market prices and costs, while satisfying constraints. • Automatic setpoint recommendation: After finding optimal targets (for example, product mix, temperatures, pressures, feed rates), the system sends these recommendations to advanced process control (APC/MPC) systems for real-time implementation, continuously minimising error between target and plant operations. AI-driven RTO delivers key business outcomes:

• Profit maximisation: Fine-tuned process control unlocks more value from existing assets without major capital expenditure. • Yield improvement: Optimised product mix and process parameters enable consistently higher yields from varied feedstock. • Energy savings: Reduced waste and smarter energy usage lower the operational cost and carbon footprint. • Bottleneck identification: Continuous optimisation highlights process bottlenecks, supporting debottlenecking strategies and cost- effective expansion decisions. Figure 5 reflects not just the value realised in terms of economic benefits for a $3 billion petrochemical major in India. This use case is also directly relevant to the value of AI adoption in the petrochemical context, demonstrating how advanced optimisation blends engineering expertise and AI to drive measurable operational and financial benefits. Current state of AI adoption in global oil and gas industry While this case of RTO implementation

exemplifies what is possible, such comprehensive, enterprise-scale AI

deployments remain the exception rather than the norm across the global oil and gas industry.

Refining India

$58 m illion in economic benefit for a $3 b illion petrochemical major

$0.1M Sales analytics & forecasting

$2.5M Feedstock cost analysis

$9.2M Grade market & sequence optimisation

$8M Remote monitoring

$0.2M Lab analytics

$5M Asset performance

$0.1M Smart contracts

$5.5M Logistics optimisation

$2M Optimising economic benet from spread analysis

$10M Hydrocarbon loss analysis

$5M Real-time optimiser (RTO)

$6.7M Inventory optimisation

$2M Smart turnaround

$2M Pricing analytics

AI and data platform Monitor and predict

Optimise and improve

Procurement

Buisness planning

Production

Maintenance

Distribution

Marketing and retail

Figure 5 AI-powered RTO: $58 million of benefits realised

Strategic advantage of mature AI platforms Bridging this gap requires a fundamental shift: moving from fragmented, point-solution AI deployments to integrated, enterprise-grade platforms engineered for scaled transformation. For oil and gas companies seeking to do a technology transformation and implement AI on a large scale, the choice of an AI platform becomes critical. Established, enterprise-grade AI platforms offer significant advantages over nascent solutions through proven scalability, robust security frameworks, and comprehensive integration capabilities. The advanced AI platforms distinguish themselves through critical capabilities: • Modular analytics engine: Low-code and no- code-based developments combining traditional machine learning, generative AI/large language models (LLM), and automated data analysis for enterprise-scale agentic applications. • Semantic data framework : Ontology and knowledge graph-based integration that contextualises data and grounds AI insights in operational realities. • Unified data architecture : Real-time and batch data ingestion and transformation across relational and NoSQL formats, with comprehensive lakehouse storage for all data types. • Self-service intelligence : Dashboarding and operational reporting with on-demand and scheduled distribution, enabling real-time visibility into critical process variables and AI- driven recommendations. • Enterprise governance, and compliance : Advanced workbench with user administration,

Global oil and gas companies have launched numerous AI pilot projects, yet comprehensive, enterprise-wide AI adoption remains limited. Most AI applications are concentrated in exploration and production, particularly in seismic analytics, drilling optimisation, predictive maintenance, and field-level emissions monitoring. In contrast, midstream and downstream sectors, such as refining, petrochemicals, logistics, and customer-facing “ Despite significant national investments, AI integration across the entire value chain is generally at advanced pilot or early deployment phases rather than full operational maturity ” operations, have seen relatively few large- scale AI deployments. Despite significant national investments, AI integration across the entire value chain is generally at advanced pilot or early deployment phases rather than full operational maturity. This highlights both the promise and the challenges of scaling AI to transform the oil and gas industry comprehensively. This gap stems from challenges such as fragmented data, siloed systems, and few examples of AI enabling true operational autonomy or end-to-end transformation. As a matter of fact, most companies have still tested AI in pockets, and the major value depends on overcoming these barriers and moving from pilots to scaled impact.

Refining India

Enterprise-wide AI integration and future outlook As AI maturity deepens, the shift toward agentic AI frameworks highlights enhanced autonomy but does not diminish the critical importance of human-in-the-loop governance. This integrated and cautious approach could power enterprises to move from experimental deployments to organisation- wide AI operations, unlocking transformative value across the oil and gas value chain. Enterprise-wide AI integration combined with deep process expertise enables step-change improvements in operational performance through data-driven, predictive workflows. The convergence further drives accelerated digital transformation: from isolated pilots to full-scale deployments across exploration, production, and asset optimisation. To sum it up, in this changing scenario, the future of AI in oil and gas depends solely on pairing advanced AI platforms with domain expertise. The strategic integration of the two would drive lasting operational transformation. The Indian government’s visionary policies, significant infrastructure investments, and technology-driven initiatives are set to accelerate the sector’s modernisation and sustainability. By advancing AI adoption through mature, proven platforms alongside sustainability objectives, India aims to build a resilient, self-reliant oil and gas industry that balances economic growth with environmental responsibility, supporting its energy security and net-zero ambitions on the global stage. Velocity to value Powered by ontologies, knowledge graphs, and agentic AI, TCG Digital’s AI analytics platform tcgmcube unifies and contextualises diverse data landscapes, enabling systems to sense variability, reason intelligently, and act autonomously. By combining advanced AI-ML models with semantic understanding, it transforms complex data into actionable intelligence, helping businesses accelerate innovation and achieve measurable impact.

row/column-level authorisation, role- based configuration, audit trails, and model performance tracking for regulated oil and gas environments. • Cloud-agnostic architecture : Containerised deployment across on-premises, AWS, Azure, and Google Cloud without vendor lock-in, ensuring operational flexibility and resilience. • Pre-built industrial workflows : Ready-to- deploy applications for predictive maintenance, asset optimisation, and process control with native integration to in-house systems and APIs. By orchestrating all these critical AI functions from data to insights to deployment – all within an integrated ecosystem, enterprises can potentially accelerate their technology transformation from months to weeks while maintaining enterprise-grade governance. Agentic AI framework, problem of hallucination, and need for guardrails Building on these integrated platform capabilities, the emerging agentic AI framework introduces new levels of autonomy, where intelligent agents independently plan, decide, and act across interconnected refinery operations. However, this autonomy also comes with its own set of unique challenges of AI hallucination, where AI generates plausible but inaccurate outputs that could jeopardise safety and efficiency. The issue of hallucination mostly stems from the basic fact that most LLMs typically lack domain-specific training or face-siloed enterprise data, which complicates reliable decision-making. Hence, to ensure trust and reliability, robust guardrails are critical. These include rigorous validation, continuous human-in- the-loop oversight, real-time monitoring, and flagging mechanisms to detect and correct hallucinations. Manual supervision remains vital to maintain alignment with operational realities and safety protocols. This combination of autonomous intelligence reinforced by human governance forms the cornerstone of safe, scalable AI adoption in complex industrial environments. Building on this foundation, it is essential to consider the evolving role of agentic AI and the governance frameworks that ensure responsible adoption.

Sujoy Choudhury sujoy@tcgind.com

Refining India

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20 Page 21 Page 22 Page 23 Page 24 Page 25 Page 26 Page 27 Page 28 Page 29 Page 30 Page 31 Page 32 Page 33 Page 34 Page 35 Page 36 Page 37 Page 38 Page 39 Page 40 Page 41 Page 42 Page 43 Page 44 Page 45 Page 46 Page 47 Page 48 Page 49 Page 50 Page 51 Page 52 Page 53 Page 54 Page 55 Page 56 Page 57 Page 58 Page 59 Page 60 Page 61 Page 62 Page 63 Page 64