Uncertainty Assessment of Equations of State with Application to an Organic Rankine Cycle

Evaluations of equations of state (EoS) with application to process systems should include uncertainty analysis. A generic method is presented for determining such uncertainties from both the mathematical formand the data for obtaining EoS parameter values. The method is implemented for the Soave–Redlich–Kwong (SRK), the Peng-Robinson (PR) cubic EoS, and the perturbed-chain statistical associating fluid theory (PCSAFT) EoS, as applied to an organic Rankine cycle (ORC) power system to recover heat from the exhaust gas of a marine diesel engine with cyclopentane as the working fluid. Uncertainties of the EoS input parameters, including their corresponding correlation structure, are quantified from the data using a bootstrap method. A Monte Carlo procedure propagates parameter input uncertainties onto the process output. Regressions have been made of the three cubic EoS parameters from both critical point matching and vapor pressure and density data, as used for the three PC SAFT parameters. ORC power uncertainties of 2-5 % are found for all models from the larger data sets. Mean power values for the cubic EoS are similarfor both parameter regressions. The mean power from the PC-SAFT EoS is less than for the cubic EoS, withno overlap of the uncertainty distributions.