Hybrid approach for improving heat transfer in copper square ducts with ribs for gas turbine blade cooling systems

Abstract

The increasing global demand for efficient and high-performance newlinepower generation systems has driven the evolution of gas turbine (GT) newlinetechnologies. Industrial gas turbines are widely used for electricity generation newlinedue to their high power-to-weight ratio and rapid startup capabilities. newlineHowever, a major limiting factor in improving turbine performance is the newlinehigh temperature at the turbine inlet, which can lead to thermal stress, newlinematerial degradation, and eventual failure of turbine blades. Maintaining newlineblade surface temperatures within a safe operating range is therefore crucial. newlineThis has necessitated the development of advanced blade cooling techniques newlinethat can sustain higher turbine inlet temperatures without compromising newlinecomponent durability. Among various cooling methods, internal cooling using newlinerib turbulators has emerged as a highly effective approach to enhance heat newlinetransfer within turbine blade cooling passages. Rib turbulators induce newlineturbulence in the cooling fluid, thereby increasing convective heat transfer. newlineThis thesis focuses on optimizing internal cooling in copper square ducts newlinerepresentative of turbine blade passages by employing a hybrid technique newlineknown as the Gannet Optimization Algorithm combined with Spiking Deep newlineResidual Network (GOA-SDRN). This hybrid strategy leverages the global newlineoptimization capabilities of the Gannet Optimization Algorithm and the newlinepredictive strength of the Spiking Deep Residual Network to maximize heat newlinetransfer performance. The study involves simulating turbulent air flow newlinethrough ribbed square ducts under ambient hot air conditions at various newlineReynolds numbers (Re), while maintaining a constant hydraulic diameter. newlineSeveral rib configurations were analyzed, with special attention to their newlineinfluence on the Nusselt number (Nu), which quantifies heat transfer newlineenhancement. newline