Exploring optimal working fluids and cycle architectures for organic Rankine cycle systems using advanced computer-aided molecular design methodologies

The combination of computer-aided molecular design (CAMD) with an organic Rankine cycle (ORC) power-system model presents a powerful methodology that facilitates an in- tegrated approach to simultaneous working-fluid design and power-system thermodynamic or thermoeconomic optimisation. Existing CAMD-ORC models have been focussed on simple subcritical, non-recuperated ORC systems. The current work introduces partially evaporated or trilateral cycles, recuperated cycles and working-fluid mixtures into the ORC power-system model, which to the best knowledge of the authors has not been previously attempted. A necessary feature of a CAMD-ORC model is the use of a mixed-integer non-linear programming (MINLP) optimiser to simultaneously optimise integer working- fluid variables and continuous thermodynamic cycle and eco- nomic variables. In this paper, this feature is exploited by in- troducing binary optimisation variables to describe the cycle lay- out, thus enabling the cycle architecture to be optimised along- side the working fluid and system conditions. After describing the models for the alternative cycles, the optimisation problem is completed for a defined heat source, considering hydrocar- bon working fluids. Two specific case studies are considered, in which the power output from the ORC system is maximised. These differ in the treatment of the minimum heat-source outlet temperature, which is unconstrained in the first case study, but constrained in the second. This is done to replicate scenarios such as a combined heat and power (CHP) plant, or applications where condensation of the waste-heat stream must be avoided. In both cases it is found that a working-fluid mixture can per- form better than a pure working fluid. Furthermore, it is found that partially-evaporated and recuperated cycles are optimal for the unconstrained and constrained case studies respectively