Simulation and 4E analysis of absortion refrigeration, M-CYCLE, and hybrid M-CYCLE cooling systems for gas turbine performance by machine learning

Abstract

High compressor intake-air temperatures not only degrade gas turbine (GT) performance but also lead to increased NOx emissions. Absorption (AB) refrigeration, which utilizes heat as its energy source, presents an efficient solution for cooling GT intake air. Complementing this, M-cycle technology—an indirect evaporative cooler (IEC)—achieves cooling below the wet-bulb temperature. In a hybrid pre-cooling system, intake air is cooled near the dew-point temperature by the M-cycle pre-cooler first and then further cooled by the AB system. This configuration requires minimal electricity, utilizing waste heat from the GT to generate power of AB refrigeration system, generate steam for used in factories that require it, beside the condensed water from the AB system to supply the M-cycle cooler, resulting in an almost zero-energy pre-cooling solution. In this study, performance comparisons between AB, M-Cycle, and Hybrid system are investigated. Also, different Siemens and General Electric gas turbines have been considered and simulated by machine learning. In addition, Energy, Exergy, Economic, and Environmental impact analysis have been performed for the considered cooling systems. Machine learning has been utilized for simulating and analyzing performance, as well as assessing the economic and environmental impacts, of the AB, M-Cycle, and hybrid cycle. Machine learning modeling accurately predicts performance across three gas turbine cooling systems, with comparative analysis revealing the Hybrid system delivers superior overall performance (30.12% power improvement, 27.31% cost reduction, 34.58% emission reduction), while the Absorption system excels in thermodynamic efficiency and the M-Cycle offers balanced performance, indicating system selection should align with specific operational priorities.