120 EUROPEAN PRACTICES

AERATION SYSTEMS

Modem designs of plants depend on many factors, but particularly the

effluent quality required and the size of population served. However, all
types of conventional activated sludge plants involve use of aeration sys­
tems and final clarifiers, and it is thus appropriate to review the characteris­
tics of these units before proceeding to consider the influence of the effluent
quality required and size of population served on plant design.
While examples of many types of aeration systems are to be found, in
the majority of plants aeration is achieved either by diffusing air through
the mixed liquor or by agitating it mechanically in open tanks. A valuable
summary is presented in a Manual o f Practice issued by the Institution
of Water and Environmental Management in the U.K. [13].

DIFFUSED AIR PLANTS

It is well established that the mass of oxygen that can be transferred

into mixed liquor from a given volume of air under given conditions
increases with decreasing bubble size. On the other hand, the difficulties
and associated costs of producing bubbles tend to increase with decrease
in the size sought. The practical compromise most frequently adopted in
Europe is to provide fine-bubble aeration systems delivering bubbles usu­
ally of around 2 to 2.5 mm in equivalent diameter.
This is most commonly achieved by delivering filtered air through
porous sintered materials, such as the Alundum domes popular in the U.K.
(Figure 3.4), or porous plastic tiles or discs. In certain proprietary designs
the plastic discs are sufficiently flexible to expand as back-pressure rises,
thus increasing orifice size and providing opportunity for clogging particles
to escape.
It is well established that in vertically sided tanks with diffusers located
at the bottom on a flat floor, the percentage of oxygen absorbed from the air
under given conditions increases linearly with the depth of mixed liquor.
However, construction and other costs tend to increase often more rapidly
with depth, especially if the water table is not far below the surface, and thus
there is again a practical compromise that in Europe normally results in the
preferred depth being about 3.5 to 4 m. It is recognized that aeration effi­
ciency is greatest when the air diffusers required to achieve the desired rate
of aeration are distributed as uniformly as possible over the floor.
However, because the oxygen demand to be satisfied declines from
inlet to outlet in a “ plug-flow” plant and the a-factor increases as the
liquid phase becomes purified, the number of diffusers provided in each
pocket is reduced (tapered) from inlet to outlet. A typical arrangement is
shown in Figure 3.5.
Figure 3.4 Dome diffuser. (Figure reproduced with the kind permission of Water Engineering
Ltd., England.)

Figure 3.5 Diffused air activated sludge plant equipped with dome diffusers. (Figure reproduced
with the kind permission of Water Engineering Ltd., England.)

121
122 EUROPEAN PRACTICES

One of the more notable variants that has been used in Europe, though
mainly in Scandinavia, is the Inka System in which air is injected from
a perforated grill, mounted about 0.6 m below the mixed liquor surface
on one side only of a longitudinal baffle running the full length of the
aeration tanks. A spiral flow is induced, which carries the bubbles for
some distance around the tank and shears them to smaller sizes than would
be produced from the relatively large orifices in still water.
Because of the small hydrostatic head to be overcome, fans are used
rather than compressors and this offers some advantage in reducing unpro­
ductive dissipation of energy. However the oxygen input per unit energy
under the most favourable conditions, probably around 2.2 kg/kWh is
somewhat lower than attainable with fine-bubble porous diffusers, and the
difference may well be even more unfavourable toward the outlet ends of
plug-flow plants where the required oxygenation capacities are rela­
tively low.

M ECHANICAL AERATORS

In some European countries a substantial proportion of works have

been equipped with mechanical aerators. In the U.K., for example, this
proportion is about two-thirds of the roughly 150 plants servicing popula­
tions greater than 10,000. However the proportion of the total flow treated
in these mechanically aerated plants is somewhat lower than that treated
in diffused air plants.
Mechanical aeration systems include several types of both vertical shaft
and horizontal shaft aerators. Three types of vertical shaft aerators are
shown in Figure 3.6. The single most popular type in the U.K. has been
the Simplex aeration cone, which is mounted at the top of a draft tube as
indicated in Figure 3.7. In compartmentalized plants, each pocket (usually
square in the plan) is equipped with a cone at its center and often fillets
are added to the base of the tank to assist vertical circulation; also, vertical
baffles are sometimes used to prevent horizontal rotation of the tank
contents. Depths are usually about 5.5 m and a freeboard of 1.5 m is
provided to retain spray. With other types of vertical spindle aerators not
employing a draft tube, the tanks are usually somewhat shallower.
On the European continent the proportion of mechanically aerated plants
employing horizontal-spindle aerators is greater than in the U.K. A popular
type is the “ Mammoth” rotor developed in Germany (Figure 3.8). This
type of rotor is manufactured in lengths up to 9 m and diameters up to 1
m, and typically operates at immersion depths up to 300 mm and speeds
of rotation around 70-80 rev/min. Oxygen inputs per unit energy between
2 and 2.5 kg/kWh are achievable. The aerators are usually individually
 Aeration Systems 123

powered by weatherproof, fan-cooled motors (up to about 120 kW in

power) acting through reduction gear boxes.
Adjustable weirs are often incorporated at the outlet ends of mechani­
cally aerated units to enable the depth of immersion of the rotating impellers
and thus the rate of aeration to be varied in response to demand. In addition,
the aerators may be covered by acoustic shrouds to reduce noise. Normally
one can expect mechanical aeration systems to deliver oxygen inputs per
unit energy of around 2 kg/kWh when driving force is near the maximum.

Figure 3.6 Three types of vertical shaft aerators (photograph courtesy of Biwater Treatment
Limited).
 Figure 3.7 Simplex cone aeration unit (photograph courtesy of Biwater Treatment Limited).

Figure 3.8 Mammoth aeration rotor (photograph courtesy o f Biwater Treatment Limited).

124
 Alternatives to the Main Types of Aeration Units 125

“ HYBRID” SYSTEMS

In passing it is noteworthy that in aeration zones in which the organic

load is relatively high (for example at the inlet ends of “ plug-flow” plant
or in uniformly mixed plants required to achieve only partial purification)
fine-bubble diffusers are reported to undergo relatively rapid clogging
requiring more intensive maintenance or replacement than those in more
lightly loaded zones [14]. Also α -factors for the heavily organically loaded
zones tend to be quite low.
In contrast, mechanical aerators do not suffer from these drawbacks. It
has thus been suggested that the optimum design for a plug-flow compart­
mentalized plant could involve the use of mechanical aerators in the first
few compartments and a fine-bubble diffused air system in the remainder.
A plant in which this configuration has been adopted is expected to be
completed by about the end of 1991 at Blackburn Meadows in Shef­
field, U.K.

ALTERNATIVES TO THE MAIN TYPES OF AERATION UNITS

THE ICI DEEP SHAFT

The ICI Deep Shaft process was developed in the early 1970s [15].
In the process, screened and degritted crude sewage is fed with returned
activated sludge into a vertical shaft up to 150 m deep though usually
between 50 and 100 m deep and divided into downcomer and riser
sections. Air is injected into the riser initially to set up a recirculating
flow, after which air is released into the downcomer at a point low
enough to ensure that the density of the air-water mixture in the riser
is lower than that in the downcomer, thus both maintaining circulation
and to achieving aeration. A feature of this method of aeration is that
high rates of oxygen transfer can be achieved, thus making it feasible
to match the oxygen demand of particularly strong sewages. Experience
has shown that because gas bubbles coming out of solution in the riser
attach themselves to sludge particles, tending to make them difficult
to settle, it is necessary either to render the sludge settleable before
transferring mixed liquor to secondary clarifiers or to use an alternative
means of separation such as flotation. Various means of rendering the
sludge settleable have been used, including vacuum degasification, but
that which now appears to be most commonly preferred is to aerate
the mixed liquor using conventional porous diffusers usually for between
1 and 2 hours. The flowsheet for a major plant in the U.K., performance
of which is described later, is given in Figure 3.9.
126 EUROPEAN PRACTICES

Figure 3.9 Process diagram of Tilbury deep shaft sewage treatment works (Irwin, R. A., W. J.
Brignal and M. A. Biss 1989. “ Deep Shaft Process at Tilbury,’’ J. Inst. Wat. Envir. Mangt., 3:281).

Despite the commitment of capacity and thus of land area to degassing

of sludge, the absence of primary tanks and the quite small area occupied
by the shaft itself make it substantially less demanding on space than
conventional AS plants.

OXYGEN AS PLANTS (OASP)

Although a few notable successes have been reported in the use of

oxygen for aeration rather than air in industrial effluent treatment, OASP
has not often been preferred in European sewage treatment plants. There
are a few plants of the “ closed” UNOX type, developed in the U.S.A.
by the Union Carbide Corporation, and some others using the “ open”
type such as the FMC Marox System and the British Oxygen Company’s
(BOC) Vitox System [16]. The main features of the Vitox System are
shown in Figure 3.10.
In closed systems the oxygen is dispersed into covered tanks, the con­
tents of which are mixed by vertical spindle vane impellers. In open
systems the oxygen is injected into the throat of a Venturi tube in a
sidestream through which mixed liquor is continuously recycled into other­
wise conventional open aeration tanks.
Advantages claimed for the closed systems include lower sludge produc­
tion, higher sludge densities, and greater reactivity per unit biomass than
attained in conventional plants. However, probably the main reason for
the limited use has been that any such advantages appeared to be insuffi­
ciently clear cut to overcome the drawbacks of greater complexity, a lower
rate of stripping of C 0 2, and thus the necessity usually to incorporate a
 Secondary Settlement 127

second conventional stage to achieve nitrification when limits on ammonia

or N in the effluent have to be met. Moreover, they would not appear to
integrate very conveniently into modem nitrification and denitrification
schemes for nutrient removal, control of sludge settleability, and energy
recovery.
The open systems have been used mainly to uprate the oxygenation
capacities of older conventional plants. Just such an application has recently
been reported for one of the main plants at Dusseldorf in Germany where
the oxygen injection system is of a proprietary design known as Lindesol-
vax-B [17]. Applications of this type do not appear to have been many,
but they may increase if the need for uprating outpaces construction of
new works.

SECONDARY SETTLEMENT

As with aeration units, there are only a few features that distinguish
European practice significantly from that in the U.S.A. and other developed
countries outside Europe. Thus, as noted earlier, detention times are usually
chosen to be about 2 h at peak and 6 h at average flow, with the correspond­
ing surface overflow rates being about 1.5 and 0.5 m/h.
Circular radical flow tanks with mechanical scrapers are the most popu­
lar design for larger works. For these it is usually considered advisable for:
• side walls to be not less than 2 m deep, to avoid carryover of

Figure 3.10 Vitox high pressure sidestream dissolver.

128 EUROPEAN PRACTICES

solids picked up from the settled sludge in currents travelling to

the overflow weirs
• the slopes of the conical floors to be not less than 5° to the
horizontal, to facilitate transfer to the withdrawal sump, and not
more than 30°, to limit depth and construction costs
• the maximum diameter of tanks to be not more than about 50 m
• weir loadings to be between 100 and 250 m3/md

Pyramidal tanks with walls in the lower section sloping at about 60° to
the horizontal are often used in small works.
Perhaps the most notable development in Europe within the last decade
has been the much wider recognition of the necessity to ensure that this
mass-flux must not exceed the capacity of the units to remove and recycle
them. This requirement was first brought into prominence by work in the
U.S.A. by Dick and his co-workers [18,19]. Both Dick and Ford and
Eckenfelder have suggested that a better indication of the settling velocity
of sludge under the dynamic conditions of full-scale tanks than is afforded
by the traditional sludge volume index (SVI), or its reciprocal the sludge
density index (SDI), would be obtained by measuring velocity while gently
stirring the sludge [20]. The local element of modem European practice
has evolved from the development by White at the former WRc Stevenage
Laboratory in the U.K. of such a new form of test to measure the stirred
settled volume index (SSVI) [21].
In this test, mixed liquor is settled in a standard apparatus in which it
is very gently stirred by a vertical ring impeller rotating at 1 rev/min as
indicated in Figure 3.11. The impeller eliminates wall effects and induces
a consistent degree of flocculation similar to that occurring under the
dynamic conditions of full-scale tanks.
The SSVI is calculated in the same way as SVI:

αοΛ/τ _ % °f total volume occupied by settled sludge after 30 min

Ij u V I —
concentration of MLSS (as %)

By analogy with SDI, the stirred sludge density index (SSD) is the
reciprocal of SSVI. Because MLSS concentration influences settling veloc­
ity it is usual to measure SSVI at a concentration of 3.5% if possible so
as to facilitate comparison with performances in other plants or in the
same plant under different operating conditions.
White then deduced a relationship between the maximum mass-flow
of suspended solids R (kg/h) recycled from the base of a settlement tank
to the downward velocity of the liquor, U (m/h), and the average SSVI as:

(3.5)
 Process Sequences 129

Figure 3.11 WRc standard settling apparatus (White, M. J. D. 1976. “ Design and Control of
Secondary Settlement Tanks,” J. Inst. Wat. Pollut Control, p. 461).

in which A (m2) is the cross-sectional area of the tank. This formed the
basis of the convenient nomogram shown in Figure 3.12, which enables
the applied solids loading to be compared with the predicted maximum
acceptable loading for a given SSVI.

PROCESS SEQUENCES

As indicated at the beginning of the preceding section, process selections

are very dependent on effluent quality required and size of population to
be served. The influence of effluent quality can adequately be illustrated
by reference to five types of requirement, namely to reduce:
(1) Carbonaceous BOD and SS partially (e.g., by 60-85 percent) to meet
limits in respect of SS and BOD of 60:40 mg/1 as average values
or better
(2) SS and BOD by 90-95 percent to meet limits of 30:20 as average
values
(3) SS, BOD, and ammoniacal-nitrogen to meet limits of 10:10:5 (or
lower in the last case sometimes down to 3 mg/1)
130 EUROPEAN PRACTICES

Figure 3.12 Nomograph for the calculation of predicted and applied solids loading (White, M.
J. D 1976. “ Design and Control of Secondary Settlement Tanks,” J. Inst. Wat. Pollut. Control,
p. 464).

(4) SS, BOD, and ammoniacal-N as in (3) but in addition total nitrogen
(Kjeldahl nitrogen plus oxidized nitrogen) to 10 mg/1 (or lower down
to around 2 mg/1)
(5) SS, BOD, ammoniacal nitrogen, total phosphorus (P) to 1 or 2 mg/
1 and often in addition total or total nitrogen as in (3)

These last two requirements [(4) and (5)] can conveniently be considered
together under the heading of nutrient removal.
The sequences chosen to fulfil these requirements are considered below
in relation to the size of population served. There is no sharp demarcation
between this last feature and the types of plants installed, though broadly
as populations served fall below 2 0 ,0 0 0 there would be a rapidly increasing
proportion of plants either of relatively shallow depth and simple construc­
tion occupying relatively large areas of land per unit of load treated (such
as the oxidation ditch) or of special simplified design (such as the Putox
process used in Austria— see later section). For simplicity it is convenient to
describe such plants in the context of facilities for “ smaller populations,”
though recognizing that a few of them serve populations up to about
2 0 0 ,0 0 0 and there is some overlapping with more intensive elaborately
 Process Sequences 131

engineered plants, with deeper reinforced concrete treatment units provided

for “ larger populations.”

LARGER POPULATIONS

Partial Purification

The use of the activated sludge process solely for partial purification
followed by release to a watercourse nowadays tends to be limited to
discharges near the mouth of large estuaries or to coastal waters. The
practice is not very widespread, but where it is employed the plants
are usually of the uniformly mixed or contact stabilisation types with
sludge loadings in the range from about 0.5 to 1 g/g day. Basic
flowsheets are usually as in Figure 3.1(a) and (c). In principle, such
plants could be used as the first stage of a two-stage process for
producing final effluent of much higher quality, but it is more common
to use a combination of high-rate biofilter followed by low-rate activated
sludge plant for this purpose.
High-rate AS plants were particularly popular in Germany prior to
1970, sludge loadings usually being in the range of 1 to 3 g/g day. Thus,
for example, the Kohlbrandhoft plant at Hamburg was designed to remove
about two-thirds of the incoming BOD at a sludge loading of 3 g/g day [20].
An example of a large modem inland high-rate plant discharging to a
large river (the Danube) is afforded by the main works for Vienna-Sim-
mering, which treats the mainly domestic flow from more than 2 million
people. Process design was by Von der Emde and his colleagues [22,23].
Settled sewage is treated in plug-flow mechanically aerated (vertical spin­
dle) aeration units having a total residence time of 1.5 h and operating
with a sludge loading of 1.1 g/g day and sludge age of 1.2 days (Figure
3.13). The aerators were powered with DC motors so that aeration intensity
could be varied and matched to demand. Because of the high water table
the treatment units were all rather shallow, the aeration basins for example
being not more than 2.6 m deep. An unusual feature of the plant is that
the final clarifiers are rectangular. The plant was required to achieve a 70
percent reduction in BOD but in practice reductions of 8 8 percent were
obtained, except during periods when high loads of centrate from a separate
sludge dewatering facility were returned to the works. Under more normal
conditions, effluent BOD and SS averaged 31 and 21 mg/1, respectively,
oxygen uptake per unit mass of BOD was 0.5 g/g, and sludge production
0.78 g/g BOD removed. These figures are reported to agree well with the
theory of Marais and Ekama [24]. The sludge productions were also in
reasonable agreement with an empirical equation derived by Hopwood
and Downing [25].
132 EUROPEAN PRACTICES

Figure 3.13 Layout of main treatment plant Vienna-Simmering.

On occasion ferrous sulfate is added, mainly to control bulking. Because

additionally the high organic loadings applied induce anaerobic zones
favouring mobilization of phosphorus at the inlet ends of the aeration
units, substantial removal of P can be achieved in the plant (see later
section).

30:20 Effluent

Whereas activated sludge plants for producing effluent of this quality

were quite common at inland sites, indeed almost the norm about two
decades ago, the increasing tendency nowadays is to design for limitation
of release of ammonia.
Many types of ASP have been used to produce 30:20 effluents as
average values though probably the most common type comprises a com­
partmentalised aeration unit aerated either by a fine-bubble diffused air
system or mechanical aeration. Loadings would normally be around 0.2-
0.4 g/g day.
Among the exceptions is the world’s largest deep shaft plant at Tilbury
in the U.K. [26]. This treats a flow of some 35,000 m3/day of a strong
sewage (BOD 600) containing a considerable proportion of industrial
wastewater before discharge to the Thames Estuary. It operates at a sludge
 Process Sequences 133

loading of around 1 g/g day and although performance figures for a whole
year do not yet appear to have been reported, during four months of the
warmer part of the year the effluent complied with a 95-percentile 30:20
standard. It seems possible that such a standard might not be met at the
above loadings in winter, in which case one might have to regard operation
in this mode as rather more in the category of partial purification.

Fully Nitrified Effluent (5 mg/l Limit for Ammonia or Better)

There are many fully nitrifying European plants. Up until the last
decade the majority of these were simply lightly loaded conventional
compartmentalized plants, with probably a slight preponderance in
numbers of diffused air over mechanically aerated plants, and almost
certainly a margin in favour of diffused air in terms of numbers of
people served. In some cases the rate of aeration was automatically
controlled according to the response of membrane electrode DO sensors
and the oxygenation capacity was tapered. However, in many cases,
particularly in diffused air plants, the rate of aeration toward the outlet
end of the aeration units was often excessive, thus involving wastage
of energy. Within the last decade it has become the practice to include
an anoxic zone, usually as the first compartment in a chain, because
of experience indicating that passage of mixed liquor through such a zone
improves settleability of sludge and also reduces energy requirements by
virtue of the fact that some of the oxygen injected into the aerobic
zones and converted into nitrate is recovered. However, this is at the
expense of having to provide somewhat more capacity (that of the
anoxic zone) than would be necessary simply to produce a fully nitrified
effluent meeting the same limit on ammonia. A bonus is that the
nitrogen content of the effluent is significantly reduced; and the tendency
of sludge to rise in final clarifiers in warm weather is probably also
somewhat lessened. Another feature of modem practice stems from
the recognition that reduction in longitudinal mixing improves sludge
settleability. This effect is shown in Figure 3.14, in which SSVI for
a large number of plants is plotted against dispersion number as reported
for example by Chambers and Jones [8 ]. These authors have shown
that dispersion number, a hydraulic parameter representative of the
degree of longitudinal mixing, is related approximately but with adequate
accuracy to the number of compartments in the aeration units by the
expression:

(3.6)
134 EUROPEAN PRACTICES

Figure 3.14 Effect of longitudinal mixing on sludge settleability. (Reprinted from Water Science &
Technology, Vol 20, figure 2, p. 129, with permission from the publishers, Pergamon Press, and
the copyright holders, IAWPRC.)

where the group of terms on the left hand side is the dispersion number
D = longitudinal dispersion coefficient
u = average horizontal velocity of mixed liquor
L = the length of the aeration unit
N = the number of compartments into which the aeration units are divided

The implication is that if the compartments were perfectly mixed and

in cascade (an arrangement adopted in at least one large European plant)
then SSVI would approach a minimum if N were 5 or more. In practice,
to allow for longitudinal dispersion not being fully eliminated by baffles
and to the desire in the interests of economy to match rate of aeration as
closely as possible to oxygen demand, the number of compartments
adopted in modem plants is usually between 5 and 16.
Using an early version of the WRc model, a design study of a typical
situation—in which settled sewage having average SS, BOD, and ammonia
of, respectively, 95, 120, and 30 mg/1 and fluctuating diumally in flow
and composition between peak and minimum values around, respectively,
1.5 and 0.5 times the mean was treated in a diffused air plant—indicated
 Process Sequences 135

that in an aeration unit having eight compartments, of which the first

would be an anoxic zone, the average magnitudes of kLa (per hour) required
just to maintain about 1-2 mg/1 DO in the aerobic units under steady
average conditions were 6.5, 5.0, 5.0, 2.5, 2.5, 1.0, and 1.0. For this
simulation, total sewage retention time was just under 10 hours, MLSS
3700 mg/1, temperature 12°C, and sludge age 9.3 days (i.e., sufficient for
full nitrification).
In practice, the aeration units would have to be designed to cater for
peak load conditions, and for these the corresponding values of kLa were
14, 12, 10, 8 , 6 , 4, 2. However, the degree of tapering in the air flow and
number of diffusers required to deliver the air would not be as great as
that of the values of kLa because of the influence of the α -factor, which
was expected to vary along the length in the sequence 0.4, 0.4, 0.5, 0.6,
0.7, 0.8 and 0.9. As a result, the maximum air flow and the appropriate
number of diffusers required in the first aerobic compartment, was calcu­
lated to be only about 1.7 times that for the last compartment, whereas
the corresponding ratio for kLa was 4 (Figure 3.15).
The incorporation of an anoxic zone in the first compartment of
plug-flow nitrifying plants has now become almost a standard feature
of modem design, the size of such zones usually being such as to
provide retention times of between 0.5 and 1 hour and usually bringing
about a 40-50 percent removal of N [27,28]. The average rate of
denitrification declines with increasing retention and becomes too slow
to make larger units economic. A recent U.K. design is described in
a subsequent section concerned primarily with automation and control.

Figure 3.15 Diffuser layout and air flow rates for nitrifying activated sludge plant as determined
by WRc model (Chambers, B. and G. L. Jones. 1985. “ Energy Saving by Fine-Bubble Aeration.’’
Wat. Pollut. Control, p. 82).
136 EUROPEAN PRACTICES

Another example is the plant recently installed at Valeton, southwest

of Paris, in which the aerobic zones are concentric “ plug-flow” circular
channels surrounding a central anoxic zone [29]. This anoxic zone is
covered and the contents are kept mixed by circulation of the liberated
gas containing mainly N2 and C 0 2. DO in the aerated zones is
automatically controlled to vary from about 1.5 mg/1 at the inlet to
3.0 mg/1 at the outlet.

Nutrient Removal

The extent to which nutrient removal is practised in Europe still varies

considerably from country to country but the EU Urban Waste Water
Treatment Directive (1991) [1] has given the technology a large impetus.
Examples of biological phosphorus removal can now be found in most of
the EU countries certainly in Germany, Sweden, Denmark, France, UK,
Holland and Italy. Research into relevant technology was first instigated
in the early 1960s, particularly in countries such as Switzerland and those
of Scandinavia, which had either many inland lakes or in the case of
Norway semi-enclosed water bodies such as fjords that were already suffer­
ing or were thought likely to suffer from increasing eutrophication. In
other countries already with a high degree of pollution control such as
France, Germany, and the U.K. nutrient removal was not at the time
considered necessary for various reasons including:
• Many lakes serving as sources for public supply were in upland
areas not subject to significant pollution.
• Algal growth in rivers and coastal waters was not usually at a
level then considered unacceptable.
• The technology of nutrient removal was still in its infancy and the
most appropriate method in which to invest the substantial sums
needed were far from clear.
• Nutrients were often derived from several sources and it was not
obvious that eliminating them from sewage effluent alone would
necessarily be beneficial in many cases.

It is fair to say that many of these reasons have now been challenged
and in the last 10 years many more rivers, estuaries and the coast have
been deemed “ sensitive” under the EU’s UWWTD and hence nitrogen
and phosphorus removal processes are being widely installed as retrofits
or new plants to meet the directive.
For the less wealthy countries, in which pollution control was much
less advanced, remediation of other even more serious effects of pollution
had greater priority.
While some of these features are still evident today, nutrient removal
 Process Sequences 137

programs have been considerably extended in recent years and now em­
brace several more countries. Thus for example it has been reported that
the number of nutrient removal plants in Scandinavia is to be increased
from 50 to 500. In Germany new regulations have been introduced placing
limits on the concentration of N in municipal plant effluent, and these
limits will require many plants to be appropriately modified. Similar limits
have also been imposed on individual plants in Austria, France, and
Holland.
So far as the authors are aware, there are no European plants achieving
N removal by other than biological processes, and very few make use of
external sources of carbon to energise the denitrification processes. On
the other hand, while biological processes for P removal are included in
or planned for some plants, the majority depend at least partly on chemical
precipitation.
When a large proportion of N must be removed to meet modem stan­
dards limiting inorganic, kjeldahl, or total nitrogen to 10 mg/1 or less, then
in larger works recirculation or spatial alternation processes are often
employed with basic configurations as shown in Figure 3.16(a) and (b).
An alternative approach, popular in Austria, is to design for simultane­
ous nitrification-denitrification. Experiences in oxidation ditches (see later
section), showing that in a nitrifying plant—if the oxygen introduced in

Figure 3.16 Most common configurations used for biological denitrification in large plants.
138 EUROPEAN PRACTICES

MLSS as it passed through the region around aerators was not enough to
prevent anoxic conditions developing before the MLSS once again passed
into a zone of high aeration—both anoxic and aerobic zones could coexist
in the same aeration unit. This principle has been made use of in compara­
tively large AS plants in Europe, notably that at Blumental, Vienna, where
unsettled sewage from a population numbering more than 2 0 0 ,0 0 0 is
treated in long baffled aeration units aerated by a series of Mammoth
rotors (Figure 3.17) [30,31,32]. After some experimentation to decide the
most favourable method of operation, the practice now adopted is to vary
the aeration intensity according to the oxygen uptake rate of the sludge.
This is determined by pumping mixed liquor to a small tank in which
aeration is constant and recording DO continuously. Under these constant
conditions, DO is inversely proportional to oxygen uptake rate (OUR)
and so aeration intensity can be varied according to the DO recorded.

Figure 3.17 Vienna Blumental treatment plant (Reprinted from Progress in Water Technology,
Vol. 8, 1977, p. 629, with permission from Pergamon Press and IAWPRC, the copyright holders.)
 Process Sequences 139

Figure 3.18 Dissolved oxygen profiles in aeration tanks 1 and 2 of Vienna Blumental plant.
(Reprinted from Progress in Water Technology, Vol. 8, 1977, p. 629, with permission from
Pergamon Press and IAWPRC, the copyright holders.)

Performance depends on loadings, as expected, but at low loadings—

around 0.12 g/g day—BOD removals of about 95 percent and TN of about
8 6 percent are achieved. The average proportion of the volume occupied
by anoxic zones was not fully defined, but appears from published diagrams
to be often between 30 and 40 percent, thus explaining the much greater
percentage removal of nitrogen than is obtained in designs— such as those
that have become popular in the U.K.—in which the anoxic zone occupies
only 10-15 percent of total aeration tank volume (Figure 3.18).
In the case of the new plants to be built in Denmark and Scandinavia,
the design has evolved from the joint researchers of six organisations
collaborating in the socalled HYPRO project [33]. The main features of
this design, shown in Figure 3.19, involve chemical precipitation of P in
primary sedimentation tanks, hydrolysis of the sludge to release biodegrad­
able organics, and feeding of these organics into a denitrification-nitrifica­
tion reactor with recycle to provide “ fuel” for denitrification.
The design of plants for the biological removal of phosphorus has
tended to follow the practice pioneered in South Africa in which the
sewage is first passed through an anaerobic zone to mobilise phosphorus
so that it can be subsequently removed in the following aerobic stages
by so-called “ luxury uptake.” A recent European version introduced in
Denmark is the Bio-Denipho process shown in Figure 3.20.
An alternative approach has been adopted in several Austrian plants in
the catchment areas of lakes, where new limits of 1 mg P/l in effluents
are now mandatory. Although not originally designed to remove P. these
plants have been modified to do so by addition of ferrous sulfate at the
140

Figure 3.19 HYPRO process for removal of N and P. (Reprinted with the permission of Springer-Verlag, Heidelberg.)
 Process Sequences 141

Figure 3.20 Bio-Denipho process for N and P removal.

inlet end of the aeration units. This P removal can be maximised by

adjusting aeration intensity to produce nearanaerobic conditions in the first
of the two aeration tanks in series, thus fostering mobilisation of P in a form
that can be both assimilated in luxury uptake and chemically precipitated
[26,27,28]. It has been shown that in a plant at Wulkatal—when no iron
is added if organic loadings are low, or when aeration intensity is too high
so that nitrate appears in the effluent—concentration of P increases with
an increase in concentration of nitrate. With nitrate-N up to more than 1
mg/1, 80 to 90 percent biological removal of P is obtained; however, with
N 0 3—N at 3 mg/1 the biological removal reduces to 30 percent. When
iron was added in dosages of 1-2.8 g/g P, the target limit of 1 mg P/l
could be met [34].
In the main treatment plant of Vienna (Figure 3.13), in which the
organic loading is too high to achieve nitrification, biological phosphorus
removal in the absence of added iron was no more than about 50 percent,
but it was substantially increased by addition of iron, reducing final effluent
content to around 2-3 mg P/l. In this case no mandatory limit had to be met.
In several cases in which alteration of existing structures seemed likely
to be uneconomic, addition of units simply to precipitate phosphorus from
the secondary treated effluent has been preferred. An interesting recent
design is that of the Crystalactor developed in Holland (Figure 3.21). In
142 EUROPEAN PRACTICES

Figure 3.21 Process flow scheme for Crystalactor phosphate removal plant. (Reprinted from
Water Science & Technology, Vol. 23, figure 1, p. 820, with permission from the publishers,
Pergamon Press, and the copyright holders, IAWPRC.)

this device calcium phosphate is precipitated in granular form in a fluidized

bed reactor [35]. Reported experience for the first year of operation indi­
cates that effluent concentrations can be restricted to 0.5 mg/1 or less.
In the UK there are a small number of biological phosphorus removal
plants [36,37,38]. These are in the minority because most of the plants
needing P removal are small to medium-sized plants (<50,000 pe) where
it is more cost effective to use iron salts [39]. The cheapest route is to
dose crystalline ferrous sulphate but it is not always convenient to fit the
crystal dissolver on site and so in many small plants ferric sulphate dosing
is applied.
 Process Sequences 143

The UK Biological phosphorus removal plants have also followed the

practice established first in South Africa and later in the USA and Canada
in the Bardenpho process (Figure 3.22), the UCT (University of Cape
Town) processes (Figure 3.23) and the JHB Johannesburg processes (Fig­
ure 3.24). All of these processes have been called ‘‘main stream’’ biological
phosphorus removal processes because the phosphorus is removed in the
excess waste sludge under aerobic conditions as part of the normal process.
The key to the processes is the release of the phosphorus from the sludge
phase into the liquid phase in the anaerobic stage. The phosphorus in the
liquid phase is then taken up again by the sludge during the aerobic stage
but more is taken up in this stage than is usually the case, i.e. a higher
concentration of phosphorus is achieved in the bacteria of the sludge of
a biological phosphorus removal process. Essentially the bacteria capable
of higher (excess or “ luxury uptake” ) of phosphorus are selected by
having an anaerobic selector stage. The keys to the process lie in (a)
providing volatile fatty acids (VFAs) which are taken up in the anaerobic
stage and trigger the release of phosphorus, and (b) maintaining a truly
anaerobic condition in the first stage. The difference between the three
main ‘three tank’ systems (Figures 3.22 to 3.24) relate to protection of
the anaerobic stage from dissolved oxygen or nitrate in the return sludge.
The Bardenpho system is the simplest and original system whereas the
UCT and JHB systems pay more attention to removing recycled nitrate.
These processes have been applied throughout Europe in the past 5 years.
In the UK the Johannesburg process has been successfully tested on part
of the Thames Water Beckton WwTP in East London [37,38] a range of
Biological P removal processes have been tested at Severn Trent Water’s
Milcote plant at Stratford-on-Avon [36], and Anglian Water have also
tested the Biological P removal options at their plants at Cambridge and
Great Billing (Wellingborough).
The main plants in Berlin now use Biological P removal processes
(with some chemical P removal back-up [40]. Biological P removal will
also be used at the new treatment works at Rostock in Germany [41].
Denmark, Sweden, Norway and Switzerland are the countries with the
highest percentage of wastewater to which phosphorus removal is applied
[42]. It reaches up to 90% in Sweden and Switzerland. Both chemical and
biological techniques of phosphorus removal are used but there is a trend
towards biological phosphorus removal because of the lower sludge pro­
duction and because the sludge has a fertiliser value. In Denmark all the
mainstream biological phosphorus removal methods are used as well as
Bio-denipho (Figure 3.20) [42].
In addition to the mainstream processes there is one major ‘‘side stream’’
process (or “ sludge stream” process), this is the Phostrip process patented
by the Biospherics Corporation in the USA. This process is shown in
144

Figure 3.22 Bardenpho arrangement.

 Figure 3.23 Layout of the University of Cape Town/Virginia Initiative Process (UCT/VIP).
145
146

Figure 3.24 The Johannesburg (JHB) process.

 Process Sequences 147

Figure 3.25. In the process the same biological principles are followed as
with the main stream processes. However, the anaerobic stage is created
in a separate side stream which takes a proportion of the recycled sludge.
When the phosphorus is released the stripped sludge is recycled back into
the main plant to allow it to pick up more phosphorus. The phosphorus-
rich liquid in the side stream is precipitated with lime and the phosphorus-
free liquid then returned to the main plant. Since the main plant is not
changed in size and operation the Phostrip process may result in smaller
plant. The Phostrip process is licensed to Severn Trent Water for Europe,
with the exception of Italy.
Both the main stream and side stream processes have shown it is possible
to get down to less than 2 mg total P/l and if carefully operated to less
than 1 mg total P/l.

SMALLER WORKS

Although many types of ASP are used for small populations, including
various proprietary package plants, in terms of population served the most

Figure 3.25 Phostrip® side-stream P removal.

148 EUROPEAN PRACTICES

popular single type is the oxidation ditch. Possibly the most numerous
though serving a much smaller total population is the Putox process, of
which there are nearly one thousand in Austria alone.

The Oxidation Ditch

Oxidation ditches were introduced into Europe by Pasveer (in Holland)

in 1953 and though now widely used elsewhere they are still to be found
in greater numbers and wider variety in Europe than in other parts of the
world. Essentially, oxidation ditches are lightly loaded extended aeration
AS plants in which the aeration units are endless channels around which
mixed liquor is circulated and aerated, most commonly by mechanical
agitation.
The first channel-type AS plants in Europe were in fact built in Sheffield
in the U.K. in 1920, but the aerators used—rotating flat paddles—were
inefficient. The ditches pioneered by Pasveer were equipped with much
more efficient “ brush” aerators in which the brushes—composed of metal
combs attached to a horizontal rotating axis— were of a type originally
devised by Kessener and later developed by Baars at the Dutch Research
Institute TNO [43,44].
The original plants in Holland were designed to provide an inexpensive
simple form of secondary treatment for small communities often located
on land with a high water table. Accordingly, the ditches were quite
shallow and often constructed by consolidating the excavation with simple
linings rather than by use of reinforced concrete, which was employed
only for the piers supporting the brushes.
Crude rather than settled sewage was treated, and in many of the early
plants there was no separate secondary clarifier. The plants were either
operated in a fill-anddraw mode or were provided with a simple baffled
settlement zone within the channel.
The low sludge loadings adopted, commonly about 0.05 g/g day, con­
ferred considerable performance stability, enabled high-quality nitrified
effluents to be produced, minimised sludge production and rendered the
sludge well-mineralized, thus facilitating its disposal on land.
Since those early days, the design and operating practice for ditches
has undergone a series of developments. In the first phase of such develop­
ments the brush aerators were replaced by cage rotors developed at TNO.
Then aeration rotors of much larger diameter (e.g., the so-called Mammoth
rotors, see Figure 3.8) were introduced, providing greater oxygenation
capacities and better vertical mixing, thus allowing channel depths to be
increased to about 3 m. Then, in the next phase, vertical shaft cone aerators
were introduced to create the “ Carrousel system” enabling depths to be
increased still further.
 Process Sequences 149

Such increases in OC and depth enabled plants to be designed for much

larger populations, and with this extension in the range of application it
was sometimes more appropriate to construct in reinforced concrete and
to build formal secondary clarifiers. In some cases it was appropriate to
include primary clarifiers also. Such developments enabled ditches to be
built to cater for populations of 150,000 or more (and some built for
industrial effluent treatment have catered for more than 300,000 PE).
Generally, however, ditches have been employed for smaller populations—
not often above 50,000, and usually considerably fewer.
It appears to be generally agreed that for this type of application, given
that adequate land is available, the process is cheaper than other forms of
ASP of the type used for large populations [45]. This economic advantage
stems mainly from the cheaper form of construction and the ability to
dispose of the well-stabilized sludge on rural land.
Figure 3.26 shows, in plan, four of the most common configurations
of ditch used in the traditional applications.
More recently, further modifications adopted have included use of dif­
fused air for aeration to induce circulation; the incorporation of in-channel
settling units (floating in one proprietary design); and the operation of a
combination of ditches sequentially in an automatically controlled fill-
and-draw mode. Also in departure from the original channel configuration,
plants have been built in which mixed liquor circulates through a block
of four aeration compartments of rectangular cross section in plan and
about 4 m deep (the Rotanox system).
Several of these recent designs of intermittent sequentially operating
plants have emanated from Denmark [46]. Thus, for example, the so-

Figure 3.26 Oxidation ditch configurations. (Reprinted by courtesy of Effluent and Water Treat­
ment Journal.)
150 EUROPEAN PRACTICES

Figure 3.27 Type D intermittent oxidation ditch system. (First published in the Proceedings of
the International Conference on Oxidation Ditch Technology, Amsterdam, 1982.)

called Type D ditch operates in four phases as indicated in Figure 3.27.

There is a continuous discharge, and both water level and aeration rotor
immersion are constant. A disadvantage is that the degree of utilisation
of the installed power and capacity of the rotors is rather low. To overcome
this, two other designs have been evolved in Denmark, the Type VR and
the Type T ditches.
The main features of the type VR, which is meant for PEs in the range
2000-6000, are shown in Figure 3.28. At each end of the partition wall
there is a hinged flapgate that can open only towards the rotor. Depending
on the direction of rotation of the reversible rotor and thus the flow of
mixed liquor, one gate will open and the other shut. This enables one part
of the ditch to be used for aeration while the other serves as a settlement
zone. Treated effluent is drawn off continuously either from one or the
other of the two side weirs, which are operated alternately automatically.
In the type T plant there are three ditches, two outer ones interlinked with
a middle one by large-diameter pipes. There are six phases of automatically
controlled operation involving the two outer ditches serving alternately

Figure 3.28 Type VR oxidation ditch system. (First published in the Proceedings of the Interna­
tional Conference on Oxidation Ditch Technology, Amsterdam, 1982.)
 Process Sequences 151

and intermittently as clarifiers as indicated in Figure 3.29. By this mode

of operation the utilization of the capacity of the aeration rotors and their
motors is improved, relative to that in the type D plants, by 50 percent.
Broadly, the relationships between loading and performance obtained
in ditches designed to meet conventional standards for SS, BOD, and
ammonia are consistent with those for other forms of ASP when account
is taken of the oxygenation capacity provided.
Figure 3.30 shows compilations of performance data consistent with
this view published in 1982. The data imply that either by original intent
or because populations have increased some ditches were receiving sludge
loads approaching 0.1 g/g day but the majority were around half this figure
or even lower. In principle it would no doubt be possible to operate at
loadings above 0.1 g/g day, but with increasing loading an increasing risk
of prejudicing the various favourable features of the process (stability,
good effluent quality, low sludge production).
In the most recent phase of development of the process, attention has
been turned to its operation in a nitrification-denitrification mode by creat-

PHASE F. 7.00-8.00a.m.

Figure 3.29 Type T oxidation ditch system. (First published in the Proceedings of the International
Conference on Oxidation Ditch Technology, Amsterdam, 1982.)
152 EUROPEAN PRACTICES

Figure 3.30 Average yearly effluent BOD as a function of sludge loading (First published in the
Proceedings o f the International Conference on Oxidation Ditch Technology, Amsterdam, 1982.)

ing one or more anoxic zones and thus enabling sludge settleability to be
improved, energy employed for nitrification to be recovered, and release
of nitrogen to the environment to be reduced.
It can be shown that—if under winter conditions with minimum sewage
temperatures of say 7°C (as in say France, Germany, Holland, and the U.K.)
sludge were produced at the rate of 0.3 g/g BOD removed—then for a plant
loaded at 0.05 g/g day containing 3000 mg/1 MLSS and treating settled sew­
age having a BOD of 200 mg/1 the retention period would be 1.33 days but
the time required to nitrify all but a small residual of ammonia entering at
 Process Sequences 153

40 mg/1 would be no more than about 0.25 days. Thus, even at peak flow,
by reducing residence time to about 0.5 days a large proportion of the ditch
could be made anoxic without risk of significant release of ammonia.
Studies of rates of denitrification of mixtures of sewage and activated
sludge indicate that about 7 mg/1 N 0 3—N might be removed in the first
10 minutes of contact but the subsequent rate would fall to about 4.5 mg/
1 h [47]. Under these circumstances, to denitrify 40 mg/1 N 0 3—N would
demand retention in the anoxic zone for about 7.5 h. At steady flow the
residence time provided would probably be sufficient to ensure almost
complete denitrification, but some leakage of nitrate could occur during
periods of peak flow [48].
Dynamic models specific to oxidation ditches do not appear to have
been produced to take account of the influence of diurnal fluctuations in
flow and strength of sewage. In their absence, the preferred approach to
design of ditches of the older style—in which all the contents were circu­
lated continuously—has been to allocate a substantial proportion of the
length of the ditch, up to about 40 percent, as potentially available for
denitrification, to feed sewage directly to this section, to monitor DO
immediately above this zone and control intensity of aeration so as to
ensure that anoxic conditions are maintained.
In plants in which part of the capacity functions intermittently as a clari­
fier, the timing of the operation sequence can be modified to include a phase
or phases in which power to the aeration rotors is reduced so as merely to
provide mixing rather than aeration. This is the case in the type T plant shown
in Figure 3.31, the method being known in Denmark as the Bio-denitro pro­
cess. The main difference between the two sequences is that in the Bio-
denitro process ditch 1 is operated anoxically during phase A and ditch 3 is
operated anoxically in phase D. The duration of the phases can be varied to
some extent depending on the amount of nitrogen to be removed, within
the overall constraint of having to provide a long enough period of aerobic
conditions at low temperatures to ensure that full nitrification is achieved if
there is a low limit on ammonia in the effluent.

The Putox Process

In the last 25 years more than 1000 small activated sludge plants for
populations ranging from 10 to 500 have been built in Austria, mostly of
the Putox design, basic features of which are shown in Figure 3.32 [49].
The treatment units comprising a two-compartment septic tank, a
uniformly mixed aeration chamber, final clarifier, and pump well, are
accommodated in excavations so that little is visible aboveground. The
average residence time in the septic tanks is about 1 day, in the aeration
unit about 0.5 day, and in the final clarifier around 7 h. Residence
154 EUROPEAN PRACTICES

Figure 3.31 Bio-denitro operation of Type T oxidation ditch. (First published in the Proceedings
o f the International Conference on Oxidation Ditch Technology, Amsterdam, 1982.)

time in the clarifier at peak flow is about 5.2 h. Oxygenation capacity

provided is 1.5 g/g BOD load and aeration is by coarse bubbles over
a 2-month period if observation in one plant during the winter of 1971
BOD of the effluent averaged 6.5 mg/1 and ammonia 2 mg/1. Subse­
quently it was shown that by aerating intermittently, substantial removal
of nitrogen could be achieved by denitrification during the periods
when the aerators were switched off.

INSTRUMENTATION, CONTROL, AND AUTOMATION (ICA)

The extent to which some degree of automatic control of AS plants

is practised has increased considerably in the last decade, and at the
 Instrumentation, Control, and Automation (ICA) 155

Figure 3.32 Main features o f Putox process.

very least control of aeration intensity according to DO concentration

of the mixed liquor would now normally be provided in new plants.
In many modem works the application of ICA is much more extensive,
embracing most elements of the process sequence, though only about
2 percent of works in Europe, mainly in Germany, are fully computer
controlled.
This feature could become more widely applied, particularly in plants
designed for nutrient removal by biological means since experience in
other parts of the world, notably South Africa, has shown that optimisation
of performance is difficult to achieve in the face of substantial diurnal
variations in flow and composition.
A good example of modem design is afforded by the contact-stabiliza­
tion diffused air AS plant at Holdenhurst near Bournemouth in the U.K.
(Figure 3.33) [50]. This works treats a DW flow of some 55 ml/d to
produce effluent with an average BOD of 8 mg/1 and ammonia content
of 0.6 mg N/l. A feature of the works is that flow is partially balanced
by retaining sewage in the main interceptor sewer feeding the plant by
operation of an automatically activated penstock. Also controlled are the
operation of inlet works, primary sedimentation tanks, the AS plant includ­
ing final clarifiers and sludge return, effluent recirculation pumps, and
sludge treatment. This is achieved by the action of six “ intelligent outsta-
tions” serially linked to a host computer and a further outstation linked
via a modem as shown schematically in Figure 3.34.
156 EUROPEAN PRACTICES

Figure 3.33 Layout o f Holdenhurst sewage treatment works (Robinson, M. S. 1990. “ Operating
Experiences o f Instrumentation, Control and Automation at Holdenhurst STW, Bournemouth.’’
J. Instn. Wat. & Envir. Mangt., 4:560, 564, 565)

In addition to the supervision of outstations the functions of the host

computer include:
• data logging
• alarm monitoring
• data display, including printing and graph plotting
The complete system includes about 400 instruments and some 25 km of
connecting cables. Analogue and digital inputs are monitored respectively
every 10 and 6 seconds.
The aeration units are divided into two trains, each subdivided into
stabilisation and anoxic contact zones, the anoxic zone being incorporated
in accordance with modem U.K. practice referred to earlier for control of
sludge settleability and for energy recovery. In the aerated contact and
stabilization zones, DO is controlled to a target value according to the
output from a DO sensor. A cascade PID control loop generates a difference
signal that is used to generate movements of the values controlling the
air flow to the porous disc diffusers. Operation of these valves alters the
 Instrumentation, Control, and Automation (ICA) 157

pressure in the air main. The changes in pressure are used to alter the
vane angle on the blowers and thus the blower output. The reference
pressure is automatically adjusted if the opening of the most open of the
valves exceeds a dead band of 65 percent of fully open. A second or third
blower is started when the vane angle has resulted in a blower delivering
maximum output for 1 h. Blower selection is based on accumulated running
hours so as to even out usage according to the arrangements shown in
Figures 3.35 and 3.36. Each air valve has a mechanical stop that serves
to maintain a minimum air flow to prevent external fouling of the discs.
MLSS concentrations are monitored continuously according to concen­
tration but not controlled. Instead, surplus activated sludge is drawn off
at a rate preset by the works management.
The ASP effluent and final works effluent are monitored for SS, ammo­
nia, and temperature. If the effluent quality contravenes pre-set limits, the
ASP effluent is returned to the storm tanks by screw pump.

Figure 3 3 4 Arrangement o f computer, outstation and peripherals (Robinson, M. S. 1990. “ Op­

erating Experiences o f Instrumentation. Control and Automation at Holdenhurst STW, Bournem­
outh.’’ J. Instn. Wat. & Envir. Mangt., 4:560, 564, 565)
158 EUROPEAN PRACTICES

Figure 3.35 Outstation 0— air pressure control system (Robinson, M. S. 1990. “ Operating
Experiences of Instrumentation, Control and Automation at Holdenhurst STW, Bournemouth.’’
J. Instn. Wat & Envir. Mangt., 4:560, 564, 565).

SOME OTHER PROCESS VARIANTS

SEQUENCING BATCH REACTORS

Sequencing Batch Reactors (SBRs) have attracted a good deal of atten­

tion in Europe in the past 5 years. The original activated sludge processes
in Manchester and Salford, UK in 1914, were “ fill and draw plants”
[51,52,53] but they were gradually replaced by continuous flow systems
which possibly needed less manual attention. The interest in the SBR
processes started in the mid 1970’s by Irvine and colleagues in the USA
[54,55] and was taken up enthusiastically by Goronszy in Australia, USA

Figure 3.36 Outstation 3— aeration control system (Robinson, M. S. 1990. “ Operating Experi­
ences of Instrumentation, Control and Automation at Holdenhurst STW, Bournemouth.’’ J. Instn.
Wat. & Envir. Mangt., 4:560, 564, 565).
 Some Other Process Variants 159

and more lately in Europe [56]. A conference reviewing the developments

has recently been held [57,58]. Over the last four years they have started
to be applied in Europe. The first new SBR type system in the UK has
been built by Yorkshire Water at Wath-upon-Deame and Dwyr Cymru
Welsh Water are intending to put SBRs in at sites in Swansea (South
Wales) and Ganol (North Wales). In Germany they are being applied to
the treatment of flows from small communities [59]. They are also starting
to be more readily applied outside the municipal sewage treatment field.
In France they have been applied to treating winery wastewaters [60] and
in the Netherlands they are being applied for the treatment of the wastewater
from cleaning up road and railway vehicles [61].
The recent interest in SBR processes probably derives from the fact
that the size of the SBR system can be smaller (and hence cheaper), than
the overall size of the conventional systems (aeration tanks plus settlement
tanks). The control of the fill and draw (drain) cycles can be achieved
by the more sophisticated electrical and computer control packages now
available. This allows very precise control of a whole range of cycle
options. It has been shown all the processes possible in continuous-flow
system BOD removal, nitrification, denitrification and biological phospho­
rus removal can be achieved in SBR type systems [56,57].
Another benefit of the SBR system is that they do not suffer from
sludge building and hence have less problems with solid liquid separation.
They can still produce stable foam (often called “ chocolate mousse” ) but
this need not cause a problem because the withdrawal of the liquid effluent
may be done from below the liquid foam interface.
There are now also a number of systems which combine fixed-film
processes with SBRs and some with sequenced fixed-film processes [57].
It seems likely that the potential for SBR systems will continue to be
investigated and developed in the next few years. Originally SBRs were
only considered for small systems but they are now being exploited for
larger populations.

FLUIDIZED BEDS

Although not strictly AS plants, fluidized beds are sufficiently similar

in principle to justify inclusion in this chapter. They may in fact be
described as suspended fixed-film processes in that the particles on
which the bacteria grow (usually sand) are in suspension in the up-
flowing liquid. It has been shown that the rate of BOD removal and
nitrification is very similar in mass loading rate terms (kg BOD/kg
MLSS-d) for Fluidized Beds as for Activated Sludge processes [62].
Where they differ is the volumetric loading rate (kg BOD/m3 of reactor-d)
for the same removal. This is because the Fluidized Bed Reactors can
160 EUROPEAN PRACTICES

achieve biomass concentrations of 5 to 10 times (10 —> 20 gMLSS/1)

higher than the Activated Sludge Process which is limited to about 5
g MLSS/1 by the solid liquid separation stage. This inclusion is merely
for the purpose of recording that although fluidised beds have been
the object of substantial investigation in Europe, there appear to be no
full-scale European plants for sewage treatment. Probably the reason
is that, as research studies have shown, the advantages of savings in
space do not necessarily outweigh the disadvantages of greater complex­
ity, including a requirement to use oxygen rather than air, and uncertainty,
owing to lack of extensive “ track records,” about the possibility of
operational difficulties—for example in sludge-sand separation. More­
over, the evolution of the aerated filter (which uses air rather than
oxygen) has probably appeared to offer a more attractive means of
reducing land requirements and rendering plants less obtrusive.

AERATED FILTERS

Though aerated filters can hardly be described as variants of ASP

they depend on aeration of biomass which, though not fully fluidized is
nevertheless not wholly static in the manner that it would be in, say, a
trickling filter. It thus has some of the characteristics of conventional ASP,
and so for completeness can appropriately be included in this section.
Such aerated filters are becoming increasingly popular, particularly in
France where the most widely used version, the Biocarbone process, has
been pioneered by Compagnie Generate des Eaux [63], In this process
sewage is passed through a drowned bed of small shale particles into
which air is injected about 1 m above the bottom of the bed. This rising
flow of air though insufficient to fluidise the shale bed, helps to maintain
hydraulic permeability and supplies the necessary oxygen to meet the
respiratory demand of the biomass that attaches to the particles and metabo­
lises the biodegradable components of the sewage. The section below the
point of air injection acts as a purely physical filter, and penetration of
SS is so small that secondary clarifiers are not required (Figure 3.37).
Biomasses per unit volume are typically equivalent to about 15,000-
20,000 mg/1 so that, for a given effluent standard, loadings per unit volume
can be applied that are several times those treatable in conventional AS
plants.
Depending on choice of loadings, aerated filters can be operated simply
to meet limits on BOD and SS or to produce nitrified effluents in a single
stage or to function as second-stage high-rate nitrifying units. Excess
sludge, which otherwise would accumulate in the bed, is removed by
backwashing intermittently.
 Some Other Process Variants 161

Figure 3.37 Biological aerated filter system (Biocarbone).

ROPE-TYPE BIOFILM REACTORS

A novel form of reactor has recently been introduced at Geiselbullach in

Germany as a means of uprating a conventional plant serving a population
numbering about 0.25 million to meet new standards including a limit of
10 mg/1 ammonia-N from May to October. The method adopted consisted
of suspending media in the aeration units on which biofilms could develop,
thus in effect creating a hybrid of the AS and aerated filter processes [64].
Various types of media were examined in preliminary trials, that selected
being a so-called Ring-Lace material consisting of flexible ropes of modi­
fied PVC incorporating a large number of woven rings. The ropes were
suspended vertically in the aeration units and although attached in cages
were able to move in response to liquid turbulence. Although the results
so far published cover a period of less than a year, it appears that the
biomass in the aeration units (expressed as equivalent MLSS) could be
increased substantially by at least 1700 mg/1 when no chemicals were
added to precipitate P, and by 5000-8000 mg/1 when P-precipitation was
in progress. These increases had a consequential benefit on nitrification.
FUTURE TRENDS

For various reasons, ASP can be expected to continue. These reasons

are that:
• AS plants can be expected to last at least 50 years.
• There is a large investment in such plants.
162 EUROPEAN PRACTICES

• A great many plants in Europe are less than 20 years old.

• Most such plants can be uprated without major difficulty to meet
such new standards as can be foreseen.
• There is no reason to suppose that biological treatment per se will
be superseded by better and cheaper methods in the foreseeable
future.
It can be expected that ASP will continue to be the dominant method
of sewage treatment in Europe well into the next century. At the same
time it seems clear that demands for reducing odours and more generally
rendering sewage treatment works less obtrusive are likely to result in a
trend in new plants for larger populations towards use of processes that
are less demanding on space and that can if desired be totally enclosed
above or below ground. Although there may be ways of reducing the size
of AS plants, involving for example the use of membrane processes to
separate and recycle biomass and thus to increase MLSS, it seems more
probable that designers will prefer to specify aerated filters. For small- to
moderate-sized towns, these are likely to be small enough to enclose; and
when used in conjunction with cross-flow membrane filters (a combination
already shown to be feasible for removal of carbonaceous BOD to produce
effluent that would comply with, say, a hypothetical 3:3 standard) can be
made very small indeed, for example with residence times of less than 10
minutes. Almost certainly the other major factor that will influence design
in the next two decades will be requirements to limit release of the nutrients
N and P in a very much wider range of circumstances than hitherto
considered necessary. This seems most likely to be accomplished by inter­
nal modification of existing plants to facilitate biological removal of N
and P; and by addition of supplementary units, including aerated filters
and anoxic filters for N removal by nitrification-denitrification, and of
chemical precipitation processes for P removal.
For smaller works, the oxidation ditch seems sure to remain the most
popular AS process. Conceivably, at some sites where ditches might other­
wise be used, reed beds will be preferred if the promising early perform­
ances of existing units are continued in the longer term.

REFERENCES

1 Council o f European Com m unities. 1991. Directive concerning urban waste water
treatment. (91/271/E E C ) O fficial Jo u rn a l o f the E uropean C om m unities. L I 35/40,
30 M ay 1991.
2 U .S.E P A , 1975. A G uide to the S election o f C ost-E ffective W a stew ater T reatm ent
System s, E P A -430-9-75/002.
3 Spearing, B. W., ed. 1987. “ S ew age Treatment Optim ization M odel— User Manual