This thesis investigates two methods of optical analysis of multiphase fluids. These two methods are
nephelometry and the statistical analysis of scattered light intensity fluctuations.
Nephelometry is an established technique for investigating particulate suspensions. In this work the
basic technique is combined with neural network processing to develop a system capable of
automatically distinguishing and quantifYing different suspensions, in particular suspensions of oil.
Evidence obtained in this study suggests that neural networks can distinguish the light scattering from
suspensions of different size distributions and produce a more accurate estimate of volume fraction than
commonly used turbidity measurements.
Non-Gaussian, fluctuating light intensities arise from the scattering of light from a varying population
of suspended particles. Successful measurement of these intensity fluctuations makes feasible new
instrumentation based on the statistical behaviour of the detected signal. Analyses that could prove
possible include particle number, size, type and flow characteristics. Photon counting methods have
traditionally been used to measure fluctuations from random media but the lower cost of analogue pin
diodes makes them the preferred choice of detector if they can be applied usefully. A method of
quantifYing the effect of noise from the diode detectors and removing it from the statistics of the
fluctuations is developed from a model of the pin diode detectors. Experimental results show that
detector noise can be compensated for in the analysis of scattered light fluctuations. Results also
indicate that the model used to describe the scattering process is valid and that further work can lead to
a practical instrument for the study of suspensions.
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.