Akademik Veri Yönetim Sistemi
Journal of Engineering Research (Kuwait), 2025 (SCI-Expanded, Scopus)
Natural gas network inefficiencies, infrastructure constraints and supply risks are the biggest obstacles to Turkey's ambition to become an efficient regional natural gas hub. This study aims to address these issues and enhance the performance of Turkey's natural gas infrastructure by integrating network flow modeling with an AI-based Network Flow Optimization System (NFOS). This study uses a Python-based simulation framework that represents network flow modeling as a directed graph. The gas flow optimization problem is defined as a cost-minimization function considering network operational constraints. The AI-based NFOS uses machine learning and reinforcement learning techniques to address real-time supply changes and effectively predict network inefficiencies. The application of AI-based NFOS system resulted in improved operational flexibility, resulting in a 15–20 % improvement in flow efficiency compared to traditional static optimization methods, and achieved 5–10 % cost savings through its ability to make dynamic gas allocation decisions and successfully mitigate the impact of supply disruptions. Strategic infrastructure development, combined with AI-based management, enhances energy efficiency and enables Turkey to establish itself as an efficient regional gas hub. Furthermore, incorporating renewable gas sources, such as biomethane and hydrogen, into the optimization model would contribute to defining a sustainable gas infrastructure in the long term.