Parametric multi-response optimisation of various organic Rankine cycle configurations for geothermal heat source application using Taguchi – grey relational analysis

Abstract

In this study, statistical methods (Taguchi, analysis of variance (ANOVA), and grey relational analysis (GRA)) are used to evaluate the impact, contribution ratios, and order of importance of parameters on the energy and exergy efficiencies of the simple organic Rankine cycle (SORC) and dual pressure organic Rankine cycle (DORC). The parameters being investigated are the working fluid (A), pinch point temperature difference of the evaporator (B) and condenser (C), degree of superheating (D), evaporator temperature (E), condenser temperature (F), turbine isentropic efficiency (G), pump isentropic efficiency (H), and low-pressure evaporator temperature (J, for DPORC only). Whereas the Taguchi method determines the optimum parameter combination for maximum system performance, ANOVA weighs the influence of individual parameters on the performance of the target function, and GRA optimizes the multi-response characteristic function. The condenser and evap-orator temperatures, pinch point temperature difference of the condenser and turbine isentropic efficiency are revealed as the major process parameters for multi-response performance characteristics of SORC, with an influence factor of 44.79%, 20.96%, 14.81% and 10.69%, respectively. While considering three different working fluids: HFE7000 (1), R245fa (2), and R141b (3), the combination A1B1C1D3E2F1G3H3 is determined as the optimum operating condition for multi-response per-formance characteristic of SORC with first- (energy) and second- (exergy) law efficiencies calculated as 18.64% and 51.69%, respectively. For DPORC, the turbine isentropic efficiency, condenser temperature, and pinch point temperature difference of the condenser and evaporator are the main process parameters for multi-response performance with 41.90%, 17.80%, 14.75%, and 10.47% impact factors, respectively. The best operating condition is obtained as A1B1C1D3E2F1G3H3J2 with first- and second-law efficiencies computed as 13.17% and 57.33%, respectively.