On the bias error due to obscured vectors in particle image velocimetry for concentrated solid–liquid flows

Abstract

Particle image velocimety (PIV) is a valuable tool for single-phase and multiphase flow velocity measurements. Many previous works have used PIV in solid–liquid (and other multiphase) flow to obtain velocity fields for the fluid phase. Early works noted the potential for bias error in multiphase PIV measurements, noting the need to constrain experiments to low solid concentrations. Refractive index matching (RIM) techniques have enabled acceptable PIV images to be obtained at much higher solid concentrations, and the technique is now used extensively. In the single-phase flow literature, the errors induced by measurements close to solid boundaries are well known, and similar errors occur in multiphase measurements except that they affect all velocity locations rather than just near-wall locations. By extending the near-wall analysis that is well understood for single-phase flow to a multiphase flow scenario, this paper demonstrates that even when RIM is perfect, multiphase PIV still suffers from a considerable statistical bias error when obtaining mean velocity measurements. Considering this error is important for planning and conducting accurate PIV measurements.