Introduction

In a world
driven towards sustainability, energy savings are nowadays a priority. In conventional
activated sludge wastewater treatment plants and membrane bioreactors,
the energy expenditure due to aeration, mixing, pumping or filtration is
affected by the viscosity of activate sludge (AS). This is a key parameter
governing the bioreactors hydrodynamics and oxygen transfer efficiency is such
aerobic processes (Duran et al., accepted).

Activated
sludge is a complex system including colloids that could be considered as a
non-Newtonian fluid (Seyssiecq et al., 2003; Ratkovich et al., 2013). Its rheological behaviour is
mainly controlled by the suspended solid (MLSS) concentration (Duran et al.,
accepted) and the applied shear rate.

Empirical and
semi-empirical correlations have been proposed to estimate average shear rate
in ideal reactors such as bubble column (Sanchez Perez et al 2006). However, in
activated sludge reactors, it is difficult to predict the average shear rate
using these correlations due to complex geometries of full scale reactors and several
operating conditions (different air flow rates, mixers shape and
localisation...). However the knowledge of the average shear rate in real
wastewater treatments reactors is foreseen to greatly help understanding the
impact of the rheological behaviour of activated sludge on process efficiency.

The aim of this
study was therefore to propose an innovative methodology to estimate the
average shear rate in full scale reactors based on the shear thinning
dependence of biological flocs. This methodology is
applied on a full scale wastewater treatment plant (WWTP) and the spatial and
time dependencies of the shear share is investigated.

Materials and
methods

Wastewater treatment plant

The
experimental set-up was installed in a full scale MBR WWTP located in Briis-sous-Forges, France. The
facility is designed to treat the urban wastewater of 17,000 PE, and is operated
under extended aeration (F/M ratio < 0.1 kg BOD₅ (kg VSS)^-1
d^-1). The biological plant is composed of an anoxic tank (equipped
with mixers) followed by an aerobic tank (equipped with one mixer and intermittent
aeration through a fine bubble diffusion system) and three filtration zones
(equipped with plane membranes of filtration and coarse bubble aeration systems).

Floc size measurements

The distributions
of the floc size were determined using a particle
size analyser Malvern Mastersizer 3000. It measures
particle size in a liquid phase in the range 0.02 ? 2000 μm.
The optical index of the dispersant was set to 1.33 which corresponded to the
value for water (Wan et al., 2009). As biological aggregates are mainly
composed of organic matter, the optical index was set to 1.596 (Wan et al.,
2009).

In order to
determine the floc size distribution just after
sampling, the analyser was transported to the WWTP closed to biological tanks.

Results

Influence of hydrodynamic sequencing on
aggregate strength

To analyse the
influence of shear forces on floc size, a first
experiment was done to study dynamically the cohesion-break mechanisms. The aim
was to subject the flocs to successive hydrodynamic
stresses and to determine if the resulting aggregate structure and the floc size distributions were modified by their history.
This experiment was done in a wet sample
dispersion unit (Hydro-EV, Malvern) at a MLSS concentration of
0.1 g.L^-1to prevent
multiple diffractions during particle size determination.

The
hydrodynamic sequencing is made up of twelve successive stages with rapid and
slow mixing periods (Figure 1a).

 (a)

(b)

Figure 1. (a) Hydrodynamic sequencing and (b)
variation of the average diameter with time during hydrodynamic sequencing

As depicted in
Figure 1b, the kinetics of aggregation are
significantly slower than the break-up ones and seems to depend on the duration
of the previous rapid mixing phases. Once the flocs
have been subjected to the highest shear rate, their steady state size remains
the same for rapid mixing stages.

A standard
vessel to study the influence of shear rate on average floc
size from an aerated tank

To further understandthe mechanisms involved and to control the
average shear rate, a standard vessel has been designed. Each sample is
placed in a standard cylindrical vessel equipped with a Rushton
turbine or a marine propeller in order to cover a large range of known shear
rate (Figure 2). To characterize the influence of shear
rate on the average size of flocs, two measurements
have been done for each applied shear rate: (i) one
after a 30?minute- stabilization period, (ii) the second one after 1h. It has
been observed experimentally that after subjecting flocs
to a constant shear rate during 30 minutes, the average size of flocs stayed constant.

Rheological
measurements performed during this study highlight that the activated sludge
could be considered as a Newtonian fluid only for a MLSS concentration lower or
equal to 0.5 g.L^-1. To estimate the average
shear rate in the vessel whatever the agitation speed applied, activated
sludge samples are thus diluted with WWTP effluent to a MLSS concentration of 0.5 g.L^-1 (constant viscosity at this concentration).

Figure 2. Standard vessel

In turbulent flow for Newtonian fluids, the shear rate
level corresponding to the operating conditions induced during this experiment
can be calculated using equations [1] and [2]:

With: N_p the
Power number (-), ρ the density of fluid (kg m^-3), d_i the diameter of the impeller (m), µ the
dynamic viscosity (Pa.s) and N is the agitation speed (s^-1).

(a)

(b)

Figure 3. (a) Mean
steady state floc size versus mean shear rate in the standard
vessel and (b) Mean steady state floc size versus Kolmogorov microscale [

experiments /

 d₅₀ = η /

d₅₀ = 0.5 x η /

 d₅₀ = 2 x η]

Figure 3a
highlights the floc size dependence to the mean shear
rate. The average floc size decreases with the shear
rates applied. For this sample, the average floc size
is higher than 100 µm for shear rates around 30 s^-1and lower than 60 µm for shear rates higher than 250 s^-1.

In order to highlight
the mechanisms controlling the size of flocs, the
average size of flocs are
compared to the Kolmogorov microscale
(

) corresponding to the size of the smallest
eddies during mixing (linked to the shear level in the vessel). It could be
determined using equation [3].

With: η
the Kolmogorov length scale (m), ν the kinematic viscosity of the fluid (m²
s^-1), ε the average rate of dissipation of turbulence kinetic energy (m² s^-3)and the shear
rate (s^-1)

Whatever the
shear rate applied, the average floc size is close to
the Kolmogorov microscale
(ranged between 0.5 η and η, see Figure 3b). Flocs are thus essentially submitted to viscous and shear effects.

Toward an indirect measure of the shear
rate in aerated tank and MBR

For each
samples, a specific database of floc size for
different shear rate is build using the previous protocol. This database allows
to evaluate the breakage mechanisms and to highlight differences in the
cohesion of flocs between samples. Such differences
may be linked to differences in the microstructure of flocs
resulting from different feeding and hydrodynamics conditions.

In order to quantify the average shear rate in each biological tank and
for each operating condition (mixing, aeration), several measurements of size for
samples directly taken from the AS tank were done (less than 1 minute between
sample and measurement). The first results highlight that the mean floc size was 83 µm during aerated phases (with a minimal
value of 70 µm and a maximal value of 87 µm) corresponding to an average shear
rate of 90 s^-1 (see Figure 3a). More data are being acquired and will
be presented in the full paper. The impact of operating conditions and tank
characteristics on mean floc size and associated
shear rate will be investigated. These results will be compared with literature
data obtained on bubble columns.

References

Duran, C.,
Fayolle, Y., Pechaud, Y., Cockx, A., Gillot, S. Impact of the activated sludge
suspended solids on its non-Newtonian behavior and oxygen transfer in a bubble
column. Submitted.

Ratkovich, N., Horn,
W., Helmus, F.P.,  Rosenberger, S.,  Naessens, W., Nopens, I.,Bentzen, T.R. 2013.
Activated sludge rheology: A critical review on data collection and modeling.
Water Research 47, 463-482.

Sanchez Perez,
J.A., Rodriguez Porcel, E.M., Casas
Lopez, J.L., Fernandez Sevilla, J.M., Chisti, Y. 2006. Shear rate in stirred tank and bubble
column bioreactors. Chemical Engineering Journal 124, 1-5.

Seyssiecq, I., Ferrasse, J.H., Roche, N. 2003. State-of-the-art:
rheological characterisation of wastewater treatment
sludge. Biochemical Engineering Journal 16 (1), 41-56.

Wan, J., Mozo, I., Filali,
A., Liné, A., Bessière, Y.,
Spérandio, M. 2011. Evolution of bioaggregate strength during aerobic granular sludge
formation. Biochemical Engineering Journal 58-59,
69-78.