
Improved Riffle Washer Design for the Australian Sultana Industry.
The aim of this work is to investigate and optimise the design of
riffle washers for the Australian sultana industry for improved heavy
contaminant removal in the processing of dried vine fruit. The industry
produces an annual average of 75,000 tonnes and estimates suggest that
approximately 1 stone per tonne of packed fruit is leaving the
processing sheds. The sustained production of high quality
contaminant-free fruit is essential to the continued demand for
Australian product. Prior knowledge on the subject was non-existent and
previous specific riffle washer performance knowledge was limited to
anecdotal information.
A riffle washer model was constructed using a 200mm wide clear acrylic
flume which highlighted the extremely turbulent and chaotic nature of
the flow and separation processes that were occurring. Subsequent
testing showed no difference in the sizing of particles throughout the
length of the machine, and no effect for the first riffle for varying
duration of flow of only fluid after fruit flow ceased. However, there
was a difference in the extraction efficiency for the first riffle for
equivalent factory fruit flowrates increasing from 0.5 to 8 t/h. The
model riffle washer also drew attention to the complex phenomena of
surging in the riffle washer, which in combination with the varied flow
across the machine width and uneven factory fruit flowrates, produced a
non-uniform two-phase flow.
The size, shape, density and constitution of sultanas and their
contaminants were assessed. Tests in a drop tube showed a difference in
the mean terminal velocities between sultanas and the majority of heavy
contaminants. Analysis of the physical properties of contaminants that
had previously escaped from a riffle washer and those that had been
extracted, showed no differences in their size, shape of makeup
distributions.
Assessment of the physical characteristics of the different industry
designs was conducted and the results showed diverse configurations.
Test work was carried out to evaluate the heavy contaminant extraction
efficiency of the different designs through the introduction into the
riffle washers of tagged contaminants that had been returned by
customers, after previously passing through riffle washers, and
contaminants that were recently extracted by a riffle washer. A range of
extraction efficiencies from 49 to 99.7% was determined. No difference
was found in the extraction efficiency of customer returned or riffle
washer extracted particles, or between ten minute and two hour
contaminant residence times in the riffle washers.
An improved riffle washer design was developed which is theoretically
capable of reducing the present contaminant levels of approximately 1
stone per tonne of packed fruit to 1 stone for every three years of
Australian production. This is achieved by using three consecutive sets
of highly efficient riffles incorporated in two parallel machines and
the removal of random elements in the complex flow.