YAPAY ZEKÂ VE MAKİNE ÖĞRENMESİ İLE ARITMA TESİSLERİNDE ENERJİ VERİMLİLİĞİ VE SÜREÇ İYİLEŞTİRME POTANSİYELİNİN DEĞERLENDİRİLMESİ

This study evaluates the potential gains of artificial intelligence (AI) and machine-learning (ML) models for improving process efficiency and reducing energy consumption in wastewater treatment plants. A comprehensive literature review shows that models such as SVR, ANN, CNN, PLS-R, RF/GBT, Isolation Forest and deep reinforcement learning can reduce aeration energy by 25–35 % and overall plant energy use by 10–25 % worldwide. For Türkiye, the sector’s annual electrical load is estimated at roughly 1.6 TWh; AI-driven optimisation could save 15–25 % of this demand, translating into 0.9–1.5 billion TRY in operating costs and 110 000–180 000 t CO₂ emissions each year. Model success is found to depend strongly on high-quality data streams, process-specific algorithm selection and operator training. The findings indicate that national incentives and standardised data infrastructures would enable large-scale AI/ML deployment, thereby advancing Türkiye’s sustainable water-management objectives through lower energy use, improved process stability and reduced environmental impact.