Title:
|
Anaerobic digestion of liquid wastewaters from food industry using continuously stirred tank reactors
|
This thesis describes an investigation on the application of anaerobic contact process treating
a variety of wastewaters from food industries (raw molasses, purge wastes, ice-cream and
diary wastewaters). The raw molasses is one of the wastes produced from the sugar industries,
which contains high concentrations of organic matters and high concentrations of sulphate
and was firstly investigated in this work. A series of batch experiments were carried out to
establish a suitable feedstock for pilot scale digesters. A 1% mass molasses medium with a
COD: N: P ratio of 200: 5: 1 in a buffer system has been shown to give a successful digestion of
molasses with a methane yield of 0.36 l/g CODrem and a COD removal of 89 % from the batch
experiment 3.
Two pilot scale digesters (14.13 1) were fed with the molasses medium that was established
from the batch experiment 3 as suitable for this work. Several organic loading rates were
investigated. After feeding with the molasses medium for about 30 days, both pilot digesters
have reached steady-state conditions with respect to a COD removal in excess of 90 % and
methane yield in excess of 0.321/g CODrem.
The process kinetics of anaerobic contact reactor for treating four different wastewaters was
investigated. The Monod equation and the Contois equation were used to develop two basic
steady-state models. The kinetic parameters required for the application of the steady-state
models were determined by using the results obtained from four individual experiments for
treating four different wastewaters. Both models were evaluated with four experimental
results obtained in this work and a set of routine analysis data obtained from a full scale
anaerobic contact reactor treating the ice-cream wastewater for Birds Eye Walls Ltd.,
(Gloucester). It is found that the Contois model would be more suitable than the Monod
model for express the process kinetics of the anaerobic contact process in full scale due to the
reason that the Contois model considers the effect of variable influent substrate concentration
into the prediction. In addition, the Contois model was found to be better than the Monod
model to express the process kinetic when treating the wastewater containing the organic
matters, which has a slow rate of hydrolysis (i. e. lipids/proteins).
|