Prediction of inner pinch for supercritical CO2 heat exchanger using Artificial Neural Network and evaluation of its impact on cycle design

Abstract

Supercritical CO2 (S-CO2) power cycles have received much attention due to their desired advantages in various applications. Due to the substantial difference in specific heat of S-CO2 at different pressure, unique characteristics inside heat exchangers, especially in the recuperator, are observed and affect the cycle design. In this research, the problem of inner pinch occurring inside the S-CO2 recuperator is identified and resolved by suggesting a new framework for the definition of heat exchanger effectiveness using the point of zero inner pinch. The design methodology of S-CO2 cycles is improved using this definition by developing a module to calculate the zero inner pinch using artificial neural network (ANN) to undergo a learning process. As a result, the computation time required for cycle analysis is reduced by an order of 10(3) in the case of simple recuperated Brayton cycle, under the accuracy of 10(-7) error bound. As the readers gain the access to the developed module for calculating the zero inner pinch, the procedure for S-CO2 cycle optimization will become far more manageable for researchers, and as a result, this result can allow the conceptualization of even further complex layouts in S-CO2 cycle research.