IOA IUVA World Congress & Exhibition, Paris, France – May 23-27, 2011

Retrofit of Intermediate Ozonation for Taste and Odour and Cyanotoxin Removal

Craig Jakubowski1, Paul Thompson1, Mike Brooks2

1. Hunter Water Australia Pty Ltd, PO Box 5007, HRMC, NSW 2310, Australia
2. Armidale Dumaresq Council, PO Box 75A, Armidale, NSW 2350, Australia

Abstract
The 40 ML/d Armidale Water Treatment Plant in the New England area of New South Wales,
Australia, treats surface water from Malpas Dam to provide drinking water to a community of
approximately 25,000 people. Since its commissioning, Malpas Dam has had a history of
cyanobacterial (blue-green algae) events with generally at least one bloom occurring per year.
Liver damage in the community in the 1980s was linked to high levels of Microcystis aeruginosa in
the dam. Powdered activated carbon was employed for mitigation of taste and odours and as a
barrier to cyanotoxins, however the system was considered unreliable, expensive and posed
operational health and safety issues. An extensive process including technology reviews, bench
and pilot testing identified intermediate ozonation and biological filtration as the preferred process
barriers for taste and odours and cyanotoxins from the Malpas Dam supply. The intermediate
ozonation system including air-preparation, ozone generation and ozone contacting was
successfully integrated into the conventional process train, on a site with significant space
restrictions. A design development process was undertaken prior to tendering which enabled key
system features and operating regimes to be selected which removed water quality risk from the
contractor, and gave the principal greater certainty in the technical and operational outcomes and
project cost. The upgraded plant has achieved the intended water quality objectives and also
provided a number of supplementary water quality and operational benefits. This paper describes
the studies undertaken leading to the selection of intermediate ozonation and biological filtration as
the preferred taste and odour and cyanotoxin barrier for Armidale Water Treatment Plant, along
with a description of the final design, and presents performance data and lessons learnt regarding
integration and operation.

Key-words: Ozone, Taste and Odour, Cyanotoxin, Intermediate Ozonation

Introduction
The original Armidale Water Treatment Plant (WTP) was commissioned in 1970, with a major
upgrade commissioned in 1987. The plant has a treated water production capacity of 40 ML/d and
comprises coagulation, flocculation, sedimentation, monosand open gravity filtration, fluoridation
and chlorination. Malpas Dam is the source of raw water for the WTP. Characteristics of Armidale
WTP raw water are provided in Table 1. Cyanobacterial events in Malpas Dam are common, with
generally at least one bloom occurring per year. Generally the events occur between November
and March (the Australian summer season), however high levels of cyanobacteria have been
recorded as early as August and persisted at times until July.

During the algal blooms, contamination of the raw water source with taste and odour compounds
and algal toxins is common. Prior to the 1990s Microcystis aeruginosa was the predominant genus
encountered during blooms. More recently the predominant genus has changed to Anabaena
circinalis, which is capable of releasing the neurotoxin Saxitoxin. Neurotoxins at high levels can
adversely affect human health. Algal counts in the order of 1.75 x 106 cells/mL of Microcystis
aeruginosa and 4 x 106 cells/mL of Anabaena circinalis have been recorded. High levels of
Microcystis aeruginosa in Malpas Dam in the 1980s were linked to liver damage in the Armidale
community [1]. The primary taste and odour compound of concern from Malpas Dam has been
geosmin, with levels as high as 1000 ng/L measured. These levels are extremely high considering
that the odour threshold for geosmin is 4 ng/L. A robust barrier is required to deliver safe and
aesthetically pleasing drinking water to the Armidale community.

IX.5.3 - 1
 Table 1. Armidale WTP raw water quality characteristics.
Raw Water Quality Parameter Units Average Minimum Maximum
Turbidity NTU 1.3 0.2 12.5
True Colour Hazen Units 15.2 1 96
pH 7.8 7.2 8.7
O
Temperature C 16.1 7 25
Alkalinity mg/L as 123.8 90 140
CaCO3
Total Hardness
mg/L as 117.5 115 140
CaCO3
TOC mg/L as 6.33 4.3 15.9
CaCO3
TOC mg/L as 3.6 2.9 4.9
CaCO3
Metals
Iron (total/soluble) µg/L 12/3 0/0 20/10
Manganese (total/soluble) µg/L 55/37 11/1 680/370
Aluminium (total/soluble) µg/L 9/3 5/1 12/9

The conventional treatment process at Armidale WTP was not capable of removing the organic
contaminants to target levels. A powdered activated carbon (PAC) facility was installed and
provided a barrier for the micropollutants, however this approach was expensive, time consuming
for operational staff and was unreliable. PAC dosing also presented a number of health and safety
issues as PAC was purchased in 25 kg paper bags and manually unloaded. Historically, PAC has
been dosed at Armidale WTP for periods of 4 to 6 months per year with doses ranging between 5
mg/L and 60 mg/L. The poor reliability of the PAC system meant customer complaints of
unpalatable, “foul smelling” water were still reported on occasions. Operators worked on a rotating
shift basis to maintain plant operations, which was most challenging during algal bloom events.
Armidale Dumaresq Council (ADC) thus desired a continuous and reliable solution that also
reduced the demands on operator resources.

This paper describes the investigations leading to the selection of ozonation and biological filtration
as the preferred taste and odour and cyanotoxin barrier for the plant. Also presented is a
description of the constructed design, along with performance data including that during a recent
significant cyanobacteria bloom. Lessons learnt regarding the integration of the new units and the
operation of the system are also discussed.

Material and Methods

In September 1998, ADC entered into an agreement with Hunter Water Australia (HWA) to
investigate the suitability of an ozone and biological activated carbon (ozone/BAC) process to
produce safe, palatable, odour free drinking water for Armidale consumers. The project objectives
were to:

§ Critically compare the ozone/BAC process with existing powdered activated carbon
treatment facilities;
§ Consider the financial aspects of both processes; and
§ Recommend the optimum treatment process for Armidale’s water supply.

A technology review was first undertaken to identify suitable alternatives for the removal of taste
and odours and cyanotoxins. Ozonation and biological filtration were found to be used relatively
widely around the world for both dinsfection and organics removal. At this time, the only significant
ozone/BAC process in operation in Australia was the 3 ML/d Edenhope WTP in Victoria [2]. A
desktop feasibility study was then undertaken to determine how ozone could be applied and
integrated at Armidale WTP and the operational and cost implications. The process was deemed
as potentially feasible at this point and worthy of further consideration, with the next step being pilot
testing. Intermediate and tertiary ozonation were considered potential options at this point.

IX.5.3 - 2
There are many variables to consider in the ozone/BAC process. Many of these cannot be
examined in sufficient detail without pilot testing. Hence a pilot testing programme was developed
to investigate the following key issues and criteria:

§ Point of ozone application (intermediate versus tertiary ozonation);

§ Ozone demand and decay;
§ Ozone Ct (mg.min/L) for taste and odour removal;
§ Ozone Ct for saxitoxin removal;
§ Empty Bed Contact Time (EBCT) (minutes) for taste and odour removal and saxitoxin
removal and whether this could be achieved with the depth available in the current filter
cells;
§ Bromate formation potential;
§ Optimum granular activated carbon (GAC) type and filter media size for both organics
removal and filtration performance;
§ Effect of water temperature on the biomass and any subsequent effects on organics
removal performance;
§ Effects on chlorine demand;
§ Cryptosporidium inactivation.

Pilot testing was undertaken at Armidale WTP from December 1998 to April 2000. Settled water
from the WTP sedimentation tanks was used for the intermediate ozone testing, while filtered water
was used for testing of a tertiary ozone/BAC proces. Natural taste and odour levels were increased
at times by spiking with laboratory grade MIB and geosmin.

The complete pilot programme and results are too extensive to be completely detailed in this
paper. The main process performance findings from the pilot plant study included:

§ Ozone/BAC achieved excellent removal of different levels of spiked MIB (230, 550, 1100
ng/L) to less than the treated water target of 10 ng/L.
§ Ozone/BAC achieved excellent removal of high levels of spiked geosmin (1000 ng/L) to
less than the treated water target of 4 ng/L.
§ Natural geosmin levels of 2700 ng/L (geosmin sourced from Malpas Dam surface scum)
were reduced to less than 70 ng/L by ozonation, with subsequent BAC filtration further
reducing geosmin to less than 5 ng/L.
§ The optimum ozonation Ct value for the removal of 25 ug/L of saxitoxin was determined to
be 2 mg.min/L. For the removal of 60 ug/L of saxitoxin the optimum Ct value was 4
mg.min/L. Testing indicated no biologically toxic ozonation by-products were formed after
oxidation of saxitoxin using ozone.
§ Saxitoxin levels of up to 15 ug/L were found to be effectively removed by biological filtration
by the BAC filters independent of ozonation.
§ Pilot plant test work indicated that ozonation was able to achieve complete inactivation of
5000 oocysts/L of Cryptosporidium for an ozone contactor Ct value of 4 mg.min/L.
§ The ozone/BAC process under steady state operating conditions was found to provide an
additional 31% organic carbon removal after conventional treatment.
§ Chlorine reductions of up to 33% were achieved by the ozone/BAC process as compared to
Armidale WTP filtered water.
§ THM formation potentials were reduced by 72% after ozone/BAC as compared to Armidale
WTP filtered water.

IX.5.3 - 3
 § Bromate formation did not exceed The Australian Drinking Water Guidelines (ADWG 2004)
limit for bromate of 0.020 mg/L.

An EBCT of 15 minutes was found to provide greater taste and odour and organic carbon removal
than a 10 minute EBCT. However an EBCT of 15 minutes could not be acommodated by the
current filters at the historical peak day flows. These implications were considered in the upgrade
design and are discussed in latter sections of this paper.

Intermediate ozonation was selected as the most appropriate technique for application of ozone,
as it minimised the amount of additional infrastructure as it makes use of the current filter
structures, providing a significant cost advantage. The is also limited available space at Armidale
WTP (Figure 1) and therefore installation of supplementary filters for biological filtration in addition
to ozone generation and contacting would be a major challenge and required significantly more site
works.

Figure 1. Armidale WTP site showing proximity to residential developments.

A concept design was prepared, followed by a design development process prior to tendering. The
design development approach was used in preference to a design and construct approach, as this
gave Council more certainty in the process and infrastructure that would be provided.

Design Capacity
ADC and HWA examined historical treated water demands in the region and identified that the
average treated water flow was 9 ML/d. The peak raw water usage in the 5 years prior to the
design was 21.9 ML/d, with the peak only exceeding 20 ML/d on five days. Considering the
population of the Armidale community is stable and to limit capital expenditure, an ozonation
design flow of 22.5 ML/d was selected for immediate production requirements. At this flow a BAC
filter EBCT of 10 minutes can be achieved with the current filter structures. The 95-percentile raw
water consumption was 14.9 ML/d, indicating that for most of the time EBCTs of greater than 15
minutes would be achieved. The system was designed however to be enable treatment of up to 29
ML/d if required in the future through addition of a third ozone generation train and modification to
the existing filter structures to provide additional depth and hence EBCT. All piping and contacting
infrastructure were designed and built for the ultimate 29 ML/d capacity.

IX.5.3 - 4
 Upgrade Process
Filter Upgrade
The conversion of the sand filters to GAC/BAC was undertaken first. Specifications were prepared
for both the GAC media and garnet gradings and suppliers selected. The conversion was
undertaken by ADC directly to reduce costs. HWA assisted ADC in planning the works and
prodived supervision of some tasks for the first filters converted.

Table 2 gives the filter media configuration installed in the upgraded filters. The depth of the media
provides an EBCT of 10 minutes at the ozonation design flowrate of 22.5 ML/d. Modifications to the
filter structures would be required in the future if the 29 ML/d treatment capacity was required.

Table 2. GAC/BAC filter design criteria.

Requirement Units Criteria
Area per Filter m2 22.4
No. of Filters - 8
EBCT minutes 10
GAC
Depth mm 880
Effective Size mm 0.8
Uniformity Coefficient - <1.5
Garnet 1
Depth mm 100
Size mm 1.2 – 2.4
Garnet 2
Depth mm 100
Size mm 2.4 – 4.8

The original sand filters were studied in detail prior to the conversion to gauge any indications of
structural issues with the cells or backwashing infrastructure that may have needed rectification.
No major issues were identified and one cell was emptied and cleaned for inspection. No structural
damage was detected. The condition of nozzles was inspected and found to be adequate. Nozzle
heights were also checked by laser level. CCTV inspection of the plenum for each filter was also
undertaken to ensure no material had been deposited in this area due to structural issues.

Installation of the new media commenced in January 2007. The GAC was washed and scraped a
number of times to remove any fine material prior to commissioning. Before returning to service,
the media was disinfected by soaking in around 25 mg/L chlorine. All filters were converted to GAC
in April 2007.

The strong organic adsorption capability of the fresh GAC had immediate benefits on treated water
quality. Once all filters were converted to GAC around a 1 to 2 mg/L reduction in chlorine dosing
requirements was observed. This adsorption capacity diminished over time as the adsorption sites
were exhausted, however removal capability was maintained by colonisation of the media by
bacteria and subsequent biodegradation of organics.

Individual filter turbidimeters were also installed as part of the filter upgrade. The resulting
turbidities from each filter are typically less than 0.2 NTU.

Design Development
During 2006 the design development process was being undertaken to refine the design and select
key process and other design criteria. This included the type of feed gas to be used for ozone
generation, the technique for ozone injection, the configuration of the ozone contactor, how settled
water would be delivered to the ozone contactor and how ozonated water would be returned to the
filter inlet channels, ozone destruction techniques, siting and hydraulic investigations, identification
of power and service water requirements, and development of control and monitoring strategies.
Figure 2 is a schematic of the upgraded Armidale WTP process.

IX.5.3 - 5
 Figure 2. Upgraded Armidale WTP process schematic.

Table 3 presents the key design and operating criteria for the ozonation system for a settled water
flow of 22.5 ML/d over a 24 hour period.

Table 3. Ozonation system design and operating requirements.

Requirement Units Criteria
Maximum settled water flow ML/d 22.5
Ozone dose
Maximum mg/L 6
Typical mg/L 2.5
Ozone generator feed gas - dry air
Generated ozone concentration g.Nm3 30 - 40
Minimum ozone generator production capacity at maximum
kg/hour 5.6
settled water flow and maximum cooling water temperature
No. of installed ozone generators - 2
Ozone generator operation mode - duty/assist
O
Maximum cooling water temperature C 30
Contactor hydraulic residence time at maximum water flow minutes 10
Minimum contactor water depth m 4
Number of contacting compartments - 4
Ozone contactor outlet stilling well residence time minutes 5

One of the key challenges was integrating the plant additions to the site, which neighbours closely
residential dwellings (Figure 1). Aesthetic amenity and noise were critical considerations. In order
to reduce the visual impact and not to block afternoon winter sun to neighbouring properties, the
contactor position, geometry and height required careful consideration. To reduce the height of the
structure a number of features were incorporated:

§ Static mixing was selected over bubble diffusion for ozone injection as the water depth in
the contactor is less critical. With bubble diffusion the water depth needs to be at least 5.5m
to maximise the hydrostatic pressure on the bubbles to minimise their size and maximise

IX.5.3 - 6
 the area for mass transfer. The water depth in Armidale WTP ozone contactor is 4m, saving
1.5m in structural height compared to that of a contactor for bubble diffusion injection.
§ The contactor was constructed at lower than the natural surface level. This meant that the
top level of the water leaving the contactor was below that of the filtered water channel.
§ A pump station was designed to return ozonated water to the filters. Stainless pumps were
required for this due to the corrosivity of the ozonated water.
§ A settled water pit and mild steel concrete lined (MSCL) pipe were constructed at the end of
the settled water channel to feed the ozone contactor using gravity. Constructing the
contactor to allow gravity flow to the filters would have meant the contactor structure would
have been significantly taller than the filters.

A major consideration during the design development process was access and operability. The
objective was to deliver an ozonation facility that met Council’s expectations regarding access to
equipment for operators and maintenance teams. This extended beyond manufacturers guidelines
in some cases. Undertaking the design development process and incorporating these
requirements gave tenderers a clear understanding of Council’s vision and expectations, rather
than in a traditional “design and construct” style contract where these factors are sometimes
focused on less and can leave the client not totally satisfied. The limited available space and the
proximity to residential dwellings meant an innovative design was required incorporating the ozone
contactor into the overall ozone generation system structure.

Procurement Timeline
In June 2007, three contractors were selected through an Expression of Interest process to tender
for the design and construction of the ozonation plant. The tender was issued in August 2007 and
closed in October 2007. Bricon Pty Ltd was awarded the tender in November 2007 for a tender
value of $3.45M.

Preliminary design drawings were issued in January 2008 with construction commencing in April
2008. The commissioning process was completed in February 2009.

Ozone System Description

Ozone is generated from air. Comparisons between oxygen and air sources indicated that for this
scale plant an air fed system was more economical. The ozone system installed features two
ozone generation trains, with one train capable of producing the maximum ozone requirement.
Each train consists of a rotary screw air compressor, desiccant dryer, aftercooler, air receiver and
ITT Wedeco SMA600S ozone generator. Each ozone generator has a capacity of 2.8 kgO3/hour at
40 gO3/m3 at a maximum gas flow of 70 Nm3/hr, and alone can apply an ozone dose of 3.0 mg/L at
the current maximum ozonated water flow requirement of 22.5 ML/d. Space has provided for
addition of both a third air preparation and ozone system in the future if treated water demand
increases. For ozone production requirements above 50% of the maximum production
requirement, both ozone generators operate at equal loads. The air preparation components are
installed in a separate room to the ozone generation equipment components.

Two Statiflo GDS (Gas Dispersion System) injection skids (duty/standby) featuring motive water
pump and side stream static mixers were provided to mix ozone gas with the settled water feed. A
sidestream of settled water is taken off the settled water main and is pumped through the
sidestream static mixers and then re-injected into the main immediately upstream of a Statiflo
DN600 inline static mixer prior to the entry to the ozone contactor.

Ozonated water enters the an inlet chamber in the ozone contactor and then passes through three
over-under contacting compartments providing a combined 10 minute hydraulic detention time at a
flow of 29 ML/d. At the current ozonation design flow of 22.5 ML/d the contactor provides a
hydraulic detention time of 12.9 minutes. The water depth in the contacting compartments is 4 m.
Computational fluid dynamics (CFD) modelling was undertaken during the detail design process to
confirm the adequacy of the contactor baffle design. The water flows over the last compartment

IX.5.3 - 7
outlet weir into a stilling chamber from where it is pumped to the BAC filters. The pumps operate to
maintain a level in the stilling chamber so that the fall from the outlet weir is 0.5m to promote
evolution of residual ozone and de-aeration prior to the BAC filters. The ozone generation system
is typically operated by feedback control to maintain an outlet ozone residual of 0.1 mg/L. Sample
points have been installed in the stilling chamber, but also at the inlet chamber and in each
contacting chamber. Ozone off-gas is treated by an ITT Wedeco COD200 ozone destructor
installed in the ozone generation room.

A number of modifications and additions to the existing plant were required. Cooling water for the
ozone generators is supplied from the elevated filter backwash storage tank. Prior to the upgrade,
chlorine was added for disinfection into the filtered water channel prior to the backwash tank feed
pumps. Suction for the pump had to be sourced prior to the addition of chlorine which required
construction of a new filtered water pit alongside the filtered water channel.

In terms of power supply, the plant transformer capacity was upgraded to accommodate the
increased power requirements for the ozonation process. A new switchroom was provided as part
of the new ozonation plant building to house the new switchboard which supplied the ozone
generators and associated equipment.

In preparation for the incorporation of ozonation, HWA assisted ADC with replacement of the water
treatment plant Square D PLC system with a new Siemens PLC system and Citect SCADA
system. The ozonation plant is controlled by a separate Siemens PLC system which is integrated
with the treatment plant PLC/SCADA system via an Ethernet communication interface.

Results
The commissioning of the intermediate ozone system and biological filtration at Armidale WTP in
April 2009 has enabled the mothballing of the PAC plant and therefore provided an immediate
safety benefit for operational staff. The ozonation process is far less labour intensive for operators,
compared to the task of unloading and monitoring the dosing of PAC and disposing of PAC
packaging and pallets. The PAC has been retained in the short-term for emergency, however it is
intended to be decommissioned. The continuous reliable performance of the ozonation system has
meant this will be possible.

In terms of water treatment, the combination of ozone and BAC has mitigated taste and odour
events since commissioning. Cyanotoxins have not been detected in Malpas Dam since installation
of the ozonation plant and therefore the system has not been challenged in this regard to date.

In early February 2011, a week of very warm weather in the region resulted in a significant bloom
of Anabaena circinalis. On 4th February counts peaked at 660,000 cells/mL equating to a
cyanobacterial biovolume of 165 mm3/L. Operational staff observed that the algae had died on the
5th February and rainfall on 6th February had washed the dead bloom over the dam spillway. Levels
of the algae were down to 42,000 cells/mL (biovolume 11.5 mm3/L) at the dam surface on 7th
February. Despite the flushing of the dead algae from the dam, raw water quality entering the plant
was very poor, with dissolved oxygen around 1 mg/L, DOC 10.6 mg/L and a combined geosmin
and MIB concentration of 299 ng/L, most of which was intracellular. The combined settled water
geosmin and MIB concentration was 41 ng/L, well above the human detection threshold. Table 4
shows the quality of the raw, settled, ozonated and filtered water on the 9 th February. It can be
seen from Table 4 that the ozone process was removing around 83% of the settled water geosmin
and MIB, with the biological filters providing approximately another 10% reduction. DOC removal
by the ozonation and biological filtration process was 33% with ozone contributing 14% and
biological filtration contributing 19%.

IX.5.3 - 8
 th
Table 4. Armidale WTP water quality 9 February 2011 during Anabaena circinalis event.
Parameter Units Raw Settled Ozonated Filtered
Turbidity NTU 4.9 1.2 0.7 0.3
True Colour Hazen Units 40 15 <5 <5
pH 7.7 7.0 7.1 7.1
TOC mg/L 13.2 5.9 5.1 4.2
DOC mg/L 10.6 5.8 5.0 3.9
UVT % 49.8 72.3 87.2 92.3
Geosmin ng/L 282 27 4.7 2.8
MIB ng/L 17 14 2.3 <1
Cyanobacteria cells/mL 28,000 0 0 0

The ozone generators were operating at maximum capacity to apply an ozone dose of 7.8 mg/L to
a flow of 190 L/s. The generators were operating in residual control mode, trying to achieve a 0.1
mg/L residual in the contactor outlet, however this could not be achieved due to the level of DOC.
With the potential for cyanotoxin contamination and variation in ozone demand this was considered
the safest mode of operation at these times. The ozone residual in the contactor inlet compartment
was approximately 2 mg/L and no ozone residual was detectable at the manual sample point at the
outlet of the second contacting chamber. There are opportunities to reduce the ozone demand
through reduction of DOC prior to ozonation. Work is to be undertaken to examine whether
enhanced coagulation (low pH coagulation) can be employed for this purpose. Raw water alkalinity
is relatively high (typically in the range of 120 mg/L as CaCO3) and hence pH depression through
increased alum dosing may be difficult. Alternatives to be examined include acid dosing and the
use of ferric-based coagulants.

Since commissioning in April 2009 ozonation has resulted in a reduction in organic carbon in
filtered water, and a subsequent reduction in chlorine dose for final disinfection along with a more
persistent residual throughout the distribution system.

Figure 3 shows the chlorine dose requirements at Armidale WTP pre and post upgrade. TOC and
DOC data would be the preferred criteria for assessment, however this data is unavailable. This
data will be collected regularly in the future. Chlorine dose provides some indication of the effects
of ozone and biological filtration, however it does not take into account changes in chlorine dose

IX.5.3 - 9
requirements resulting from chlorine decay in the reticulation network due to consumption and
retention time variations and the effects of water temperature. Despite this deficiency, a number of
key points can be observed from

Figure 3:

§ Replacement of filter sand with GAC reduced chlorine demand significantly. The fresh GAC
would have had a high organics adsorptive capacity. Until ozonation was provided the
adsorptive capacity diminished somewhat until the media was colonised by bacteria.
§ Ozonation assisted in reducing chlorine demand even when algae were not recorded in
Malpas Dam. ADC trialled turning off the ozonation system in August and October 2010
and a clear increase in chlorine demand was observed. Turning off of the ozone system
coincided with an increase in raw water and settled water colour, with an inferred increase
in DOC (DOC data not available at this time). Without ozone the chlorine dose
requirements increased markedly, suggesting the DOC in the raw water was more easily
assimilated by the biomass in the filters after reaction with ozone. ADC has since decided
to operate the ozonation system continuously. As discussed above, enhanced coagulation
will be examined to determine if a reduction in DOC prior to ozonation can be achieved
which will further reduce chlorine demand and ozone consumption.
§ Severe cyanobacterial blooms in Malpas Dam not only have the potential to release
metabolites such as taste and odour compounds and cyanotoxins, but also significant
quantities of cellular matter that may have a high oxidant demand. The combination of
ozone and biological filtration may provide a robust barrier to organic micropollutants, but
an increase in chlorine dose may still be required due to increased cellular matter in the raw
water when the algal cells age or die. Optimising dam management practices and aeration
and destratification systems is an important consideration to reduce the performance
demands of water treatment plant barriers.

IX.5.3 - 10
 Figure 3. Armidale WTP chlorine dose requirements pre and post upgrade
to ozonation and biological filtration.

Since commissioning of the new ozone system, a step screen has been installed in the settled
water channel prior to the ozone contactor feed well. The proximity to residential dwellings and
large trees meant that leaves that drop into the sedimentation tank and settled water channel
would flow to the ozone system, where sometimes they would be sucked into the sidestream
injection motive water pumps. The step screen now removes and leaves or debris. The
sedimentation tank scouring cycle has also been modified, with an increased scour frequency and
shorter scour duration selected to minimise the drop in settled water channel and ozone contactor
feed well level. With a long scour duration it was observed that the level in the ozone contactor
feed well would drop too low, causing slugs of air to be entrained with the ozone contactor feed
flow. The entrained air would on occasions cause an airlock in the sidestream injection motive
water pumps and resultant pump failure. Air would then need to be bled from the pump and the
pump reset.

IX.5.3 - 11
 Figure 4. Step screen installed in settled water channel to remove leaf matter and debris
from ozone contactor inflow.

The ozone and biological filtration process at Armidale WTP has provided a continuous and
reliable taste and odour barrier. The onerous and time consuming tasks associated with operating
and maintaining the PAC dosing system and addressing breakdowns and failures are significantly
reduced that plant operators have moved to a 35-hour day-work week from a rotating shift roster.
This is an economic advantage to Council and a work-life balance improvement for operators.

Conclusion

The retrofit of the ozone/BAC process to Armidale WTP has provided a continuously operating
barrier to taste and odours and cyanotoxins, addressing key health and aesthetic concerns for
potable water consumers in the region. The new system has significantly reduced operating labour
requirements for maintaining and taste and odour and cyanotoxin barrier, so much so that the plant
is only attended during daylight hours and operators have moved to a 35-hour week from a rotating
around-the-clock shift basis. Improved chlorine residual stability has also been achieved
throughout the distribution system.

The overall approach taken for the project ultimately resulted in a value for money solution that
addressed the water quality challenges and site specific constraints. A significant contributor to the
outcome was the piloting process, which enabled key design and performance criteria to be
incorporated and considered in the design development process. Undertaking of the design
development prior to tendering removed the water quality objectives risk from the contractor, as
ozone doses and contact times were pre-defined. The design development also allowed Armidale
Dumaresq Council to effectively manage the associated onsite risks and aesthetic aspects of the
plant by pre-selection of the ozone feed-gas and the ozone injection technique and configuration of
the ozone contactor. The delivery of a highly functional and operator friendly system, integrated on
a challenging site, for a tender price of $3.5M represents significant value for a plant of this
capacity.

IX.5.3 - 12
 Acknowledgment

The authors would like to acknowledge the contribution to the project from Armidale Dumaresq
Council’s David Steller, Colin Maciver, Michael Porter and Luke Finnegan along with water
treatment plant operators, led by Chris Keogh.

The authors would also like to acknowledge the contribution of Andrew Boeren and Brad Nixon
from Bricon Pty Ltd.

Finally, we wish to acknowledge the contribution of the following project team members from
Hunter Water Australia: Peter Dennis, Marnie Coates, Jeremy Smith, Jason Sandford, Owen
Driscoll, Kevin Burgess and Darryl Beveridge.

References
1. Falconer I.R., Beresford A.M. and Runnegar M.T.C., Evidence of liver damage by toxin from a bloom of
the blue-green alga, Microcystis aeruginosa. Medical Journal of Australia, 1: 511-514 (1983).
2. Hill, W. and Millican, R., Design and Operation of Ozone/BAC Water Treatment Plant at Edenhope. 61st
Annual Water Industry Engineers and Operators’ Conference. Shepparton, VIC. 2 and 3 September,
1998

IX.5.3 - 13