Analysis of Supercritical CO2 Cycle Using Zigzag Channel Pre-Cooler: A Design Optimization Study Based on Deep Neural Network

The role of a pre-cooler is critical to the sCO2-BC as it not only acts sink but also controls conditions at main compressor’s inlet that are vital cycle’s overall performance. Despite their prime importance, studies on pre-cooler’s design hard find in literature. This partly due unavailability data around complex thermohydraulic characteristics linked with operation close point. Henceforth, current work deals designing and optimizing by utilizing machine learning (ML), an in-house recuperator design, analysis code (RPDAC), cycle point (CDPC). Initially, computed using 3D Reynolds averaged Navier-Stokes (RANS) equation used train (ML) model based deep neural network (DNN) predict Nusselt number (Nu) friction factor (f). trained ML then optimization (RPDAC) generate various designs pre-cooler. Later, RPDAC was (CDPC) understand impact Finally, multi-objective genetic algorithm optimize geometry environment power cycle. Results suggest can approximate 99% 90% certainty operating regime. Cycle simulation results performance calculation be misleading without considering pumping power. Moreover, study indicates temperature ranging from 307.5 308.5 channel’s 28,000 30,000 would good compromise between efficiency size.