Exergy Analysis and Estimation of Optimum Cut Point Temperature of a Crude Distillation Unit under Uncertainty

Abstract

Energy-efficient design and operation have been the focus of research in process industries to mitigate global warming and realize a circular economy. The crude distillation unit (CDU) is a critical component in the refining process, but it also consumes a significant amount of energy. It is estimated that the CDU is responsible for 35-40% of the total energy consumption in a refinery. It highlights the need for efficient operation and process optimization to reduce energy consumption and costs. Improved operation and technology advancements can lead to significant energy savings in the CDU process. Optimum values of tray temperature, also known as the cut-point temperature, have been a challenge considering the uncertainty around crude composition and process conditions. Apart from cut-point temperature optimization, an analysis of energy and exergy is conducted to assess the energy efficiency of the CDU and identify potential areas for improvement. Compared to conventional energy analysis, exergy analysis is a more comprehensive method for evaluating the performance of the CDU, as it incorporates the second law of thermodynamics and traditional energy analysis techniques. In this study, we integrate the exergy analysis aspect in our previous study based on the hybrid framework of the Taguchi method and genetic algorithm (GA). A crude distillation unit (CDU) simulation was created using Aspen HYSYS to evaluate crude oil assays from Pakistan's Kunnar and Zamzama regions to improve performance. Multiple variations of the crude assay were created by introducing artificial uncertainty in the actual crude composition and operating conditions, resulting in hundreds of scenarios being examined to evaluate the effect of uncertainty. The hybrid model combining the Taguchi and genetic algorithms was created in MATLAB and integrated with Aspen HYSYS simulation to determine the optimal cut points. Minimizing exergy destruction in a column per kilo barrel of diesel production was set as an objective function. Three hundred and ten data samples comprised of a variant in process conditions and optimized cut points from the hybrid network were generated. Based on the results, an artificial neural network model was developed to predict optimal cut points for increased diesel production. The results produced by the artificial neural network (ANN) were then used directly in the Aspen HYSYS model, bypassing the hybrid structure. The results of the Hybrid optimization and ANN models were similar, indicating that the ANN model could accurately predict the optimal cut points for optimized diesel production. For the Kunnar crude, a 27% iv increase in diesel production and a 26% decrease in exergy destruction within the column per kilo barrel of diesel were observed compared to straight-run results. For the Zamzama crude, there was a 12% increase in diesel production and a 13.22% decrease in exergy destruction within the column per kilo barrel of diesel.