Development of a Self-Optimization Framework of Reactive Units for Energy Efficient Operation of a Petroleum Refinery

Abstract

In this work, integrated frameworks of the artificial neural networks (ANN) with genetic algorithm (GA) and particle swarm optimization (PSO) were developed to realize higher exergy efficiency of reactive units of a refinery under uncertainty in process conditions. Initially, a steady-state Aspen model was used to perform exergy analysis for quantifying exergy efficiency, irreversibility and improvement potential of the plant. The process model was then transformed to a dynamic mode by inserting ±5% uncertainty in process conditions, i.e., temperature, pressure, and mass flow rate, to generate a dataset of 216 samples for integrated naphtha and isomerization process and 200 for delayed coking process. An ANN model was developed using the dataset to predict exergy efficiency. The ANN model was used as a surrogate in GA and PSO environments to achieve higher exergy efficiency under uncertainty. The optimized process condition derived through GA and PSO based approach were fed to Aspen model for cross-validation. The integrated naphtha and isomerization process had an overall exergy efficiency, irreversibility, and improvement potential of 50.57%, 34955.55 kW, and 17276.98 kW, respectively. Whereas the delayed cocking process had an overall exergy efficiency, irreversibility, and improvement potential of 77.61%, 29204.035 kW, and 6539.51 kW, respectively. The correlation coefficient of ANN model was 0.97432 for integrated naphtha and isomerization process and 0.99051 for delayed coking process. Performance of the GA and the PSO based approaches were comparable, and they significantly enhanced the exergy efficiency of the plant when compared to standalone Aspen model of the process.