Safe and Sustainable by Design Assessment of Organic Rankine Cycle Working Fluids for Binary-Cycle Geothermal Power Plants: A Combined Entropy-TOPSIS Model with Global Sensitivity Analysis

Abstract

Working-fluid selection is decisive for the safety, environmental footprint, and thermodynamic performance of the organic Rankine cycle (ORC) units that convert medium-enthalpy geothermal brine into electricity, yet fluid-selection studies rarely embed all three dimensions inside a single, auditable decision model. This study operationalises the European Commission Joint Research Centre (JRC) Safe and Sustainable by Design (SSbD) framework as a quantitative multi-criteria case study for a representative 150 degC binary-cycle geothermal plant. Six candidate fluids (R245fa, R1233zd(E), isobutane, isopentane, propane and ammonia) are scored on nine indicators nested within the three SSbD pillars — Safety, Environmental Sustainability, and Functionality — the last populated by a first-principles, Carnot-referenced thermodynamic sub-model rather than by literature look-up alone. Indicator weights are derived by combining objective Shannon-entropy weighting with a neutral, regulation-consistent subjective baseline, and fluids are ranked with the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). Under the neutral baseline, R245fa attains the highest closeness coefficient (CC = 0.810), ahead of the ultra-low-global-warming-potential (GWP) alternative R1233zd(E) (CC = 0.763). A four-tier sensitivity analysis — one-factor-at-a-time tornado analysis, a manually implemented Sobol–Saltelli variance decomposition, an 8,000-fold Dirichlet Monte-Carlo rank-stability test, and one-way thermodynamic/life-cycle-assessment parameter sweeps — shows that this ranking is not robust: the Environmental Sustainability dimension alone explains 56–62% of the output variance, and a weight shift of roughly 30% toward that pillar is sufficient to make R1233zd(E) the preferred fluid, which occurs in 32% of Dirichlet-sampled weight vectors. The results demonstrate that SSbD assessments of energy-conversion fluids are only decision-useful when reported jointly with an explicit, quantitative sensitivity envelope, and they provide a transparent, fully reproducible template for extending the SSbD framework from chemicals and materials into renewable-energy hardware design.