THE TREATMENT OF DAIRY WASTEWATER BY THE ANAEROBIC
UP-FLOW PACKED BED REACTOR
Robert C. Backman, Graduate Student
Frederic C. Blanc, Professor
James C. O'Shaughnessy, Associate Professor
Department of Civil Engineering
Northeastern University
Boston, Massachusetts 02115
INTRODUCTION
Many industries discharging wastewater into municipal sewerage systems are required to utilize
pretreatment. Conventional aerobic processes have been proven to achieve suitable pretreatment
quality for many food-processing wastewaters. The high operation expenses often make them unattractive.
The anaerobic filter process has been shown to overcome the aerobic disadvantages because: 1)
no energy is needed for oxygen transfer; 2) high COD removal efficiencies can be achieved when
operating at high organic loading rates (in excess of 10 kg COD/mVday); 3) anaerobic filters can be
operated intermittently without losing the ability to regain efficient treatment; 4) shock loads can be
accepted as long as sufficient buffering capacity is available to maintain a pH level suitable for meth-
anogenic bacteria; and 5) low net production of biological solids in an anaerobic filter yields less
residual sludge. This occurs because of a long SRTs.
The purpose of this study was to evaluate the ability of an up-flow packed bed reactor (UPBR)
operating in the mesophilic temperature range (35 C) to treat a dairy wastewater containing greater
than 95% biodegradable solids. Operational parameters varied were hydraulic retention time and
substrate concentration. Waste characteristics measured throughout the study were chemical oxygen
demand (COD), biochemical oxygen demand (BOD), pH, alkalinity, biogas composition, biogas production rate, volatile acids, and total and volatile suspended solids. Also flow regimes were characterized for conditions before inoculation and when the reactors were considered at a biological
equilibrium condition.
MATERIALS AND METHODS
Reactor Design
The six bench-scale anaerobic filter units illustrated in Figure 1 were constructed from 15.46 cm
(6 in.) PVC (Schedule 80) tubing. The dimensions for the filters were: 101.6 cm (40 in.) overall length,
14.63 cm (5.761 in.) inside diameter, with a wall thickness of 1.1 cm (0.432 in.). The total volume
of the empty column was 17.09 liters (0.604 ct ft). At the top and bottom of each column, 5.08 cm
(2 in.) from the edge of the columns was located a dispersion plate. The plates had twenty-five evenly
spaced holes. These were used to develop a uniform flow as the waste entered the filters and to retain
effluent solids.
The reactors contained 71 cylindrical pall rings, 50.8 mm (2 in.) tall by 50.8 mm (2 in.) diameter
with a unit surface area equaled to 118 mVm3 (35.47 ftVft5). The rings were randomly placed between
the two dispersion plates. The liquid volume of the reactors with the pall rings was 16.41 liters (0.58
cu ft), which resulted in a porosity of 0.96. Since bacterial growth was expected to only develop
between the dispersion plates, the effective volume was considered to be 15.172 liters (0.536 cu ft).
Four sampling ports were located at 20.32 cm (8 in.) intervals throughout the column height. A
20.32 cm (8 in.) plexiglas plate was placed at the bottom and top to seal the columns. The influent
361