The paper describes molasses exhaustion experiments to allow prediction of the
maximum practical exhaustion that could be obtained in factory operation for a given
molasses. The results emphasise the importance of high impurity to water (l/W) ratios in
obtaining low molasses purities. However, different samples showed widely varying
exhaustion characteristics, with exhausted molasses purities varying by up to three units
among the samples at the same I/W ratio. For some factories, molasses sampled at
different times during the crushing season also exhibited markedly different exhaustion
potentials. Exhausted molasses purities at a set consistency value were interpolated from the exhaustion curves and used to develop a revised expected purity formula. The new regression gives similar purity predictions to the SRI expected purity formula around
RS/Ash levels of unity, but the influence of the RS/Ash ratio is significantly reduced.
There is still considerable scatter of the data around the prediction line and extra
variables are needed to reduce the prediction error further. It would be preferable if
these variables were available by factory laboratory analysis or through rapid
turnaround contract analysis.