Exergy Optimization of Organic Rankine Cycle System Using Genetic Algorithm
The application of exergy analysis in thermodynamic systems is rapidly expanding alongside energy analysis, providing valuable insights into processes causing exergy destruction and losses. Environmental concerns have driven increased investigation on heat recovery, with the Organic Rankine Cycle (ORC) emerging as an effective solution for converting low-temperature waste heat into useful work. This research performs a comparative and optimization assessment of various carbon dioxide Rankine cycles—namely simple, cascade, and split configurations—specifically for recovering waste heat from gas turbines. This research employs a multi-objective optimization strategy, validated by simulation outcomes, that integrates a Genetic Algorithm with various machine learning techniques—such as Random Forest, XGBoost, Artificial Neural Networks, Ridge Regression, and K-Nearest Neighbors—to forecast the performance of the cycle. Results highlight the split cycle's superior power generation, achieving optimized performance metrics of 7.99 MW net power output, 76.17% heat recovery, 26.38% system efficiency, and 57.96% exergy efficiency, positioning it as a promising solution for waste heat recovery applications.