(373av) Real-Time Optimization of an Oil Refinery Heat Exchanger Network
The steady-state detection step was implemented in MATLABÂ® and data reconciliation, parameter estimation, and economical optimization were carried out in GAMSÂ®. The F-modified method of steady-state detection (Cao and Rhinehart) featured a reliable use in an online application, provided that the frequency of collected data would be adequate and the tuning of the constants Ê being critically done.
The data reconciliation and parameter estimation are coupled, as a DRPE â Data Reconciliation and Parameter Estimation, estimating global heat transfer coefficients and hot fluid volume flow of each heat exchanger. Finally, the optimization step were executed in two parts: with an objective function maximizing HEN total heat transfer and another objective function for minimizing the temperature difference between each HEN branch. According to the results of both optimization strategies, the potential increase of HEN energy recovery could be between 5,1% to 5,7%.