Global Lighthouse Network 2026

Tüpraş and Qatar Shell have prepared their sites for a volatile world FIGURE 8 In-house developmentAI-driven crude supply chain synchr onization Manual pr ocesses and siloed planning caused delays, excess inventory and higher costs. An AI-driven system now optimizes crude pr ocur ement, pr oduction scheduling and logistics in r eal time. The crude slate layer evaluates cargo impacts months ahead using simulations, the pr oduct slate aligns production with demand for ecasts, while Aspen PIMS optimization and the logistics slate synchr onize transport with r eal-time tank levels and disruptions.-5% Inventory levels+6 weeks Planning time horizon visibilityTüpraș İzmit, Türkiye In-house developmentAutomated condensate optimization informed by live data feed fr om of fshor e wells With gas feed capped, daily manual choke adjustments for of fshor e wells limited condensate pr oduction. Now , an AI system optimizes valve settings for each well every 30 minutes. A compositional integrated network (CIN) and physics-based models simulate hydrocarbon flow and calculate optimal settings to inform an advanced pr ocess contr ol (APC) system’ s real-time adjustments, eliminating manual intervention.-90% Engineering cost-98% Response time to system upsetQatar Shell Ras Laffan, Qatar Joint-developmentIntelligent crude blending and r efinery optimization To ensur e optimal crude blend quality and maximize yield, the site built a 3-layer optimization system. Petr omix dynamically adjusts blending strategies based on live data such as crude quality and vessel schedules, while a r efinery-wide optimization layer leverages digital twins and soft sensors to recommend hourly setpoints. Integrated with Aspen PIMS and IoT analytics, the system runs 10,000+ simulations to align operations with market conditions.+10p.p. Capacity usage ratio-16% Demurrage (due to tank & berth allocation)Tüpraș İzmit, Türkiye Joint-developmentReal-time, physics-based AI for asset integrity and performance Pressurized steel integrity challenges can lead to downtime, safety risks and reduced operational efficiency. By combining structural mechanics with real-time operational data, physics-based AI delivers ultra-fast simulations of stress and fatigue for critical assets, far surpassing traditional methods. These insights empower proactive maintenance and operational decisions, extending asset life and enhancing reliability while significantly reducing costs and risks.+6 years Critical equipment lifetime-64% Annualized CapExQatar Shell Ras Laffan, Qatar In-house developmentIntegrated platform for pr oduct distribution and logistics excellence Manual dock scheduling was causing long delays, high demurrage and lower on-time delivery . A self-developed driver hub platform integrates ERP , terminal automation system (TAS), IoT sensors and RFID with an AI orchestration engine. It unlocks r eal-time appointment-to-exit planning and dynamic slot optimization, r educing manual work and enabling 30 FTEs to be r edeployed to analytics r oles.-75% Average truck loading time+10p.p. On time deliveryTüpraș İzmit, Türkiye Notes: ERP = enterprise resource planning, FTE = full-time equivalent (worker), PIMS = pipeline integrity management system, RFID = radio frequency identification. Demurrage is a charge payable to the owner of a chartered ship on failure to load or discharge the ship within the time agreed. Source: Global Lighthouse Network.