Based on experience from wet gas compressor testing at NTNU (low-pressure air-water fluid) and K-Lab full-scale testing (normal operating conditions, high pressure and hydrocarbon fluids), this paper documents important aspects relating to the uncertainty evaluation of wet gas compressor performance test results.
The Monte Carlo method for evaluation of uncertainty on a wet gas compressor system is outlined, and the resulting uncertainties of key compressor performance parameters are presented.
Furthermore, a sensitivity analysis has been performed to evaluate how uncertainties in the output of the model can be appointed to different sources of uncertainty in the inputs, thus identifying main contributors to the uncertainties.
The importance of accurately determining the fluid composition, the properties of the fluid components and how these affect the fluid characterization are discussed. Together with the choice of equation of state, the characterization directly affects the simulated fluid properties, and great care is required to obtain reliable compressor performance results.
Uncertainties of physical properties originating from the thermodynamic simulation show that compressor power and gas density ideally should be determined from direct measurements and not from thermodynamic simulations.