TY - JOUR
T1 - Retrofit Design of Hydrogen Network in Refineries
T2 - Mathematical Model and Global Optimization
AU - Jagannath, Anoop
AU - Madhuranthakam, Chandra Mouli R.
AU - ElKamel, Ali
AU - Karimi, Iftekhar A.
AU - Almansoori, Ali
N1 - Funding Information:
The authors thank the Petroleum Institute, Abu Dhabi, for the providing the resources and financial support for this work. The authors would also like to thank the Department of Foreign Affairs, Trade and Development, Canada, for their financial support through the Canadian Commonwealth Scholarship Program (2010).
Publisher Copyright:
© 2018 American Chemical Society.
PY - 2018/4/11
Y1 - 2018/4/11
N2 - The problem of retrofit design of refinery hydrogen networks is addressed in this work, using the mathematical superstructure optimization. The superstructure of retrofit hydrogen network design contains hydrogen using, producing, and purifying units; along with compressors to facilitate hydrogen distribution. The developed mathematical model is formulated as a mixed integer nonlinear programming model (MINLP), with the objective being minimum total annual cost. The nonlinearity in the model is because of the bilinear, posynomia, and linear fractional terms. A new heuristic method is presented which helps in assigning suction and discharge pressures for the newly retrofitted compressor. With such an assignment, the nonlinearity in the model is now only confined to bilinear terms. This bilinear MINLP model is solved to global optimality using the proposed global optimization algorithm. Tests on some literature examples show that the proposed algorithm can reach global solutions faster than some commercial MINLP global solvers.
AB - The problem of retrofit design of refinery hydrogen networks is addressed in this work, using the mathematical superstructure optimization. The superstructure of retrofit hydrogen network design contains hydrogen using, producing, and purifying units; along with compressors to facilitate hydrogen distribution. The developed mathematical model is formulated as a mixed integer nonlinear programming model (MINLP), with the objective being minimum total annual cost. The nonlinearity in the model is because of the bilinear, posynomia, and linear fractional terms. A new heuristic method is presented which helps in assigning suction and discharge pressures for the newly retrofitted compressor. With such an assignment, the nonlinearity in the model is now only confined to bilinear terms. This bilinear MINLP model is solved to global optimality using the proposed global optimization algorithm. Tests on some literature examples show that the proposed algorithm can reach global solutions faster than some commercial MINLP global solvers.
UR - http://www.scopus.com/inward/record.url?scp=85045331945&partnerID=8YFLogxK
U2 - 10.1021/acs.iecr.7b04400
DO - 10.1021/acs.iecr.7b04400
M3 - Article
AN - SCOPUS:85045331945
SN - 0888-5885
VL - 57
SP - 4996
EP - 5023
JO - Industrial and Engineering Chemistry Research
JF - Industrial and Engineering Chemistry Research
IS - 14
ER -