NICFD and the PIV technique

Abstract

The growing interest in organic Rankine Cycle (ORC) based power systems has encouraged ample amount of literature on the design methodologies for unconventional turbo-machinery. These machines generally operate in the so-called Non-Ideal Compressible Fluid Dynamic (NICFD) region of the working fluid where the thermophysical properties and transport properties models, and optical properties are experimentally unexplored. Therefore, these design methods need to be validated using state-of-the-art measurement techniques like Particle Image Velocimetry (PIV). PIV has not been implemented to study the non-ideal behaviour of such fluids, and therefore a feasibility study of PIV in these unconventional media is required. This work deals with exploring the possible challenges that could occur while applying the PIV technique to measure the flow comprising of non-ideal fluids in low speed and high-speed regime.

The fluids for which the feasibility is studied are Octamethylcyclotetrasiloxane (D4) and Hexamethyldisiloxane (MM) which are frequently used working fluids for ORC power systems. The equation of state used to calculate thermo-physical properties of these fluids is briefly discussed. The viscosity of these fluids is calculated to assess the tracer particle response characteristics and check for large variations of viscosity with the thermodynamic variables. To be able perform optical diagnostics, one also has to explore the optical properties of the working fluid --- especially the refractive index. Therefore, a theoretical study of refractive index and influence of thermodynamic properties on the refractive index is studied. Conventional seeding techniques are reviewed and its feasibility for the fluids of interest is discussed.

A test facility called the Non-Intrusive Vapour Analyser (NIVA) was designed to conduct PIV in low speed vapour flows induced by a rotating disk. A suspension of D4 and 170 nm titania particles was evaporated to obtain a seeded volume of D4 vapour, on which PIV can be performed. The signal-to-noise ratio (SNR) was calculated to verify sufficient light scattering property of the titania particles. The seeding technique of evaporating the suspension of D4 + titania yields sufficiently homogeneous seeding distribution. Mean velocity fields of the vapour flow in the NIVA at different disk rotation speeds could be measured with acceptable uncertainties. Considering a vast difference in flow conditions at high-speeds, a theoretical study of high-speed MM flow in a de-Laval nozzle is done to explore challenges that could occur in application of PIV. Large gradients in density are typical of dense gas expansions. This subsequently results in large gradients in optical properties like refractive index. Challenges to particle imaging due to inhomogeneous refraction of light are investigated by preliminary estimation of position error and velocity error along the nozzle axis. A conceptual design of the seeding system is proposed that can operate at high-pressures and does not risk contamination of the working fluid.

It was concluded from the experimental results in NIVA that PIV is feasible in low-speed vapour flows and can measure velocity fields with an average uncertainty of less than 1%. Also, refractive index gradients in high-speed vapour flows could cause unacceptable errors of greater than 1% in PIV measurements. These errors depend on the complexity of the fluid and the distance between the measurement plane and nozzle wall.