Enhancing working fluid selection for novel cogeneration systems by integrating predictive modeling: From molecular simulation to process evaluation.

Wang, Lili; Zong, Fang; Liu, Zhengguang; Yang, Jiawen; Xia, Li; Zhang, Xuxue; Zhao, Wenying; Sun, Xiaoyan; Xiang, Shuguang

In this work, an efficient predictive model for screening working fluids of the Organic Rankine Cycle (ORC) was proposed by using 120 °C of low-temperature waste heat. The proposed model was implemented by combining the structural parameters of working fluids and the optimal performance of the ORC system. Four quantum chemical descriptors (molecular volume, HOMO orbital energy, mulliken electronegativity, molecular polarity index) were screened to characterize the Quantitative Structure-Property Relationship (QSPR) model. The model was constructed using Support Vector Regression. The results demonstrate that the Leave-One-Out determination coefficient of the model achieves a commendable value of 0.8997, thereby underscoring its robustness. The model's practical application reveals its proficiency in predicting the optimal efficiency of the working fluids within the ORC, surpassing the results obtained with the commonly working fluids, R245fa. A novel ORC process was proposed based on the QSPR. The total annual costs (TAC) for the proposed process obtained a reduction of 2.01%, and exergy destruction experienced a decrease of 3.13%. The reduction of the annual operational cost and equipment costs were 30.63% and 5.37%, respectively. To summarize, the new model facilitates efficient screening of working fluids and offers a theoretical foundation for the industrial application of such fluids. [Display omitted]

PREDICTION models; WASTE heat; WORKING fluids; MOLECULAR volume; OPERATING costs; RANKINE cycle

Process Safety & Environmental Protection: Transactions of the Institution of Chemical Engineers Part B, 2024, Vol 183, p587

0957-5820

Academic Journal

10.1016/j.psep.2024.01.026