This work presents a Computer-Aided Molecular Design (CAMD) method for the synthesis and selection of binary working fluid mixtures used in Organic Rankine Cycles (ORC). The method consists of two stages, initially seeking optimum mixture performance targets by designing molecules acting as the first component of the binaries. The identified targets are subsequently approached by designing the required matching molecules and selecting the optimum mixture concentration. A multiobjective formulation of the CAMD-optimization problem enables the identification of numerous mixture candidates, evaluated using an ORC process model in the course of molecular mixture design. A nonlinear sensitivity analysis method is employed to address model-related uncertainties in the mixture selection procedure. The proposed approach remains generic and independent of the considered mixture design application. Mixtures of high performance are identified simultaneously with their sensitivity characteristics regardless of the employed property prediction method.
Available at: http://works.bepress.com/sakis_papadopoulos/16/