Exergetic Optimisation of Atmospheric and Vacuum Distillation System Based on Bootstrap Aggregated Neural Network Models

This paper presents a bootstrap aggregated neural network-based strategy for the modelling and optimisation of crude distillation unit incorporating the second law of thermodynamics. Exergy analysis pinpoints the location and magnitude of the losses and is a tool for determining how efficient a process is. Exergy analysis of processes gives insights into the overall energy use evaluation of the process, potentials for efficient energy use of such processes can then be identified, and energy-saving measures of the processes can be suggested. The focus is to improve the exergy efficiency of the crude distillation and hence reduce the energy consumption. To overcome the difficulties in developing detailed mechanistic models, data-driven models such as artificial neural network (ANN) models can be utilised. Real-time optimisation of distillation columns is made feasible by using ANN models which can be quickly developed from process operation data. To enhance the reliability of ANN models, bootstrap aggregated neural network (BANN) is used in this study. A further advantage of BANN is that model prediction confidence bounds can be obtained. BANN models for exergy efficiency and product qualities are developed from simulated process operation data and are used to maximise exergy efficiency while satisfying product quality constraints. The standard error of the individual neural network predictions is taken as the indication of model prediction reliability and is incorporated in the optimisation objective function. Application to a crude distillation system (comprising of ADU and VDU) shows good improvement in the exergy efficiency of the unit and no additional costs of equipments. A further analysis was to investigate the effects of preflash units on the exergy efficiency of the ADU and VDU. The analysis gives realistic and promising results. The method could be applicable in determining feasible and energy-efficient operating and design conditions for the crude distillation unit.