Thermal efficiency impacts of structural and environmental variables in combined cycle plants: A machine learning approach to relocation scenario

Combined Cycle Power Plants (CCPPs) are essential in modern power systems due to their high thermal efficiency and relatively low greenhouse gas emissions. However, their performance is highly influenced by structural and environmental factors. Key variables such as turbine configuration, cooling method, cooling water temperature (CWT), ambient temperature (AT), atmospheric pressure (AP), and relative humidity (RH) directly affect thermodynamic efficiency. This study examines operational data from two CCPPs located only 2.5 km apart but operating under different conditions. CCPP A, at 145 m altitude, uses dry cooling and a single-shaft (1 + 1) setup. CCPP B, at 35 m, utilizes seawater cooling and a multi-shaft (2 + 1) configuration. A hypothetical relocation scenario was developed to estimate how CCPP A would perform under CCPP B's conditions. Three machine learning models—Multiple Linear Regression (MLR), Bagged Trees, and Artificial Neural Networks (ANN)—were trained using real operational data. The results reveal that relocating CCPP A to the region of CCPP B could lead to a net power increase of 15.68 MW and an approximate thermal efficiency improvement of 1.65 %. These outcomes highlight the value of predictive modeling and location-specific design for optimizing CCPP performance in varying geographical contexts. Single shaft (1 + 1) configuration power plant diagram (CCPP A).