Development of two-frequency planar doppler velocimetry instrumentation
Date published
Free to read from
Authors
Supervisor/s
Journal Title
Journal ISSN
Volume Title
Publisher
Department
Type
ISSN
Format
Citation
Abstract
This thesis describes the development of the two-frequency Planar Doppler Velocimetry (2n-PDV) flow measurement technique. This is modification of the Planar Doppler Velocimetry (PDV) technique that allows the measurement of up to three components of the flow velocity across a plane defined by a laser light sheet. The 2n-PDV technique reduces the number of components required to a single CCD camera and iodine cell from the two CCDs in conventional PDV. This removes the error sources associated with the misalignment of the two camera images and polarisation effects due to the beam splitters used in conventional PDV. The construction of a single velocity component 2n-PDV system is described and measurements made on the velocity field of a rotating disc and an axisymmetric air jet. The system was then modified to make 3D velocity measurements using coherent imaging fibre bundles to port multiple views to a single detector head. A method of approximately doubling the sensitivity of the technique was demonstrated using the measurements made on the velocity field of the rotating disc and was shown to reduce the error level in the final orthogonal velocity components by ~40 to 50%. Error levels of between 1.5ms-1 and 3.1ms-1 depending upon observation direction are demonstrated for a velocity field of ±34ms-1. The factors that will influence the selection of a viewing configuration when making 3D PDV measurements are then investigated with the aid of a computer model. The influence of the observation direction, the magnitude of the flow velocity, and the transformation to orthogonal velocity components are discussed. A new method using additional data in this transformation is presented and experimental results calculated using four-measured velocity components are compared to those found conventionally, using only three components. The inclusion of additional data is shown to reduce the final error levels by up to 25%.