Akademik Veri Yönetim Sistemi
Engineering Applications of Artificial Intelligence, cilt.160, 2025 (SCI-Expanded, Scopus)
The heat exchanger network (HEN) is an essential component in the petroleum refinery, particularly within the crude distillation unit. It plays a critical role in minimizing furnace heat loads by harnessing the residual heat of product streams from the Crude Oil Distillation Unit (CDU). While this process helps in energy recovery, there is an ongoing effort to further optimize its efficiency. In this study, the quantification of exergy efficiency and exergy destruction within the HEN of the crude distillation unit under uncertain process conditions is conducted using an advanced exergy analysis technique in combination with artificial neural networks (ANN). Initially, an advanced exergy analysis of the HEN in the crude distillation unit under steady-state conditions was performed. Subsequently, an uncertainty of ±10% was incorporated into the process variables, converting the process model into the dynamic mode. From that dynamic model, a data set of 600 samples was generated. ANN models were developed using datasets to predict exergy efficiency, exergy destruction, unavoidable and avoidable exergy destruction, and modified exergy efficiency. The resulting performance indicators, with correlation coefficients (R), are 0.9923, 0.99266, 0.9829, 0.99243, and 0.99135 for overall exergy efficiency, overall exergy destruction, avoidable exergy destruction, unavoidable exergy destruction, and modified exergy destruction respectively. This study lays the groundwork for advancing Refinery 4.0, holding promising potential for future initiatives to enhance energy efficiency in the industry.