Optimal Experimental Design and Parameter Estimation of a Stacked Bed Hydrotreating Process
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
This master thesis deals with a project conducted at IFP Énergies nouvelles during the period from April to August 2014. It describes a methodology for estimating the kinetic parameters of stacked catalysts for the hydrotreating process. This methodology is built on extending the kinetic model for one catalyst and knowing that the feed has to pass all the catalysts in the order they are put in the reactor. The parameters of this new model are estimated by an adaptive version of the differential evolution algorithm. Another part of the work investigates how the hydrotreating experiments have to be conducted for obtaining the best precision of the estimates.This is done using a Bayesian optimal design methodology and the uncertainty for some of the best designs is further quantified in a simulation study. The results show that the estimation procedure is successful in estimating the kinetic parameters in a precise and effective way. Further it shows that a necessary requirement for obtaining reasonable precision is to use a scheme of sampling more temperatures than catalyst ratios and to obtain experimental points with as large differences in the temperatures and as large differences in the catalyst ratios as possible.
Place, publisher, year, edition, pages
2015. , 41 p.
Hydrotreating, Stacking of catalysts, Nonlinear modeling, Adaptive differential evolution, Bayesian optimal design
Probability Theory and Statistics
IdentifiersURN: urn:nbn:se:liu:diva-119903ISRN: LIU-IDA/STAT-A--15/006—SEOAI: oai:DiVA.org:liu-119903DiVA: diva2:829666
IFP Énergies nouvelles
Subject / course
Nordgaard, Anders, Senior lecturer
Villani, Mattias, Professor