Wallerand, Anna SophiaKermani, MaziarKantor, Ivan DanielMaréchal, François2017-12-012017-12-012017-12-01201710.1016/B978-0-444-63965-3.50195-1https://infoscience.epfl.ch/handle/20.500.14299/142400WOS:000417380000195Industrial waste heat is abundant and represents significant energy inefficiency for many processes. With increasing emphasis on improving industrial energy efficiency, heat pump systems (including refrigeration) offer a solution by valorizing low-temperature waste heat. Optimization of industrial heat pump systems attempts to reach the cost-optimal configuration of equipment (compressors, evaporators, etc.), the sizes, operating conditions (pressures levels, temperatures), and working fluids which can be expressed as a mixed integer nonlinear programming (MINLP) problem. This work presents a general MINLP heat pump superstructure which incorporates enhanced features such as fluid after-cooling (after compression) and inter-cooling (during multi-stage expansion) while considering pressure levels and fluid selection. The MINLP is solved using a bi-level mathematical approach to explore a large solution space. The superstructure was applied to a set of MILP literature cases and it is shown that the MILP sub-problem performs well; furthermore, the full MINLP superstructure achieves up to 10% improvement compared to the literature optimal scenario with respect to the total annualized cost.industryheat pumpgenetic algorithmheat pump superstructureindustrial refrigerationmathematical programmingSFOE_IHPSCCER_EIPprocess_integrationtext::book/monograph::book part or chapter