Figure 1 : General Schematic Trentham WTP

Raw water is pumped from the existing reservoirs to a 25kL tank at the treatment plant.

The microfiltration unit is a US FilterMemcor 16M10C using nominal 0.2 micron pore size polypropylene hollow fibres. Automatic air/water backwashing and clean-in-place systems are provided.

The microfiltered water is delivered under pressure to a pressurised ozone contact tank. Ozone is generated on site from air and injected at the inlet to the ozone contact tank via a service water stream.

The ozonated water is then uniformly applied to an upflow GAC filter vessel with an empty bed contact time of 15 minutes and using coal-based carbon. The GAC filter is backwashed by a combination of ozonated and treated water. GAC washwater is collected in the microfiltration feed tank and is thus retreated within the process.

GAC filtered water is discharged to a 0.7 ML clear water storage. Provision has been made for sodium hypochlorite dosing at the clear water storage inlet. The outflow from the clear water storage is disinfected using the existing chloramination system.

Some unusual features of the plant include:

• full process reversibility. It is possible to alter the process sequence and run ozonation/GAC ahead of microfiltration.
• a stabilisation period is provided for the GAC filter immediately after backwashing to allow its performance to stabilise before it is brought back into service.
• automatic recirculation through the GAC filter occurs if the plant is off-line for more than a preset period of time to prevent the filter from going 'off'.

A summary of the main plant design parameters is presented in Table 2.

Table 2: Design Parameters

Costs
The construction cost of the Trentham plant was around $700,000 including 700 kL clear water storage tank, connecting pipework to and from the existing Reservoir outlet and delivery pumping station.

Operating costs have not yet been fully defined but are estimated to be of the order of $24,000 pa. including labour, parts, maintenance and membranes. Annual treated water production is of the order of 130 ML.

3.0 PERFORMANCE

The plant has been in operation since February 1998. In the sixteen months of operation some key operational knowledge of this type of system has been gained. This can be summarised as follows:

• The plant has operated in the CMF, Ozone and GAC process order. There has been no need to use Ozone/GAC as the first stage.
• Taste and Odour removal has met all Coliban Water's requirements.
• Treated water turbidity has been constantly low (see Figure 1 below).
• True colour removal has been outstanding without the need to add a chemical coagulant ahead of the CMF system. (see Figure 2 below).
• Typical plant performance is shown in Figure 3 below.
• Algae removal rates across the treatment system have been very high (up to 3 log removal) (see Figures 4 and 5 below).
• The plant has been 'user friendly' to operate and requires little fine adjustment to achieve the treated water outcomes.

Figure 1: Turbidity Removal

Figure 2: Colour Removal

Table 3: Typical Plant Performance

4.0 OPERATIONAL ISSUES

In the sixteen month period the following operational issues have been encountered

• A Giardia count was detected in the treated water during commissioning test work. Subsequent test work by Coliban Water established Giardia counts in the Trentham raw water.
• Algae removal across the CMF system in isolation has not been measured at > 4 log as expected.
• Algae counts have been experienced in pipework downstream of the CMF plant.
• Algae counts have increased across the GAC filter indicating some growth in the vessel.
• A membrane failure was experienced after a period of six months resulting in the changeout of a membrane set.
• Backwash water has been above the criteria for return to the reservoir.

4.1 Giardia Issue
The treated water count of Giardia was traced to using raw water to commission the activated carbon vessel. Due to the tight project program the CMF plant was commissioned last so that microfiltered water was not available. GAC needs to be thoroughly backwashed prior to service to remove carbon fines. The Giardia was held in the GAC filter and then subsequently carried through into the treated water during commissioning.

The carbon filter was backwashed for several hours with a maximum dose of ozone. Subsequent testing of the treated water showed no further sign of Giardia. Over the 16 month operation period the treated water has been routinely tested for Giardia with none being detected.

In hindsight the issue could have been averted in a couple of ways,

• Prior knowledge that the raw water contained Giardia.

Until Giardia testing was done as part of the plant's performance criteria, Coliban Water was not aware that the reservoirs contained the cyst species. Awareness of such a factor would have alerted commissioning engineers to not use raw water in backwashing the carbon filter.

• Installation of the CMF unit earlier could have provided microfiltered, and therefore cyst free, water for commissioning ozone and activated carbon filters.

4.2 Algae Issue
As can be seen from Figure 4 below the count of algae in the raw water reservoirs are extremely high over the spring and summer period. This algae has been responsible historically for the taste and odour issues in the Trentham water supply.

Reservoir 2 has had the highest algae counts over the period of the plant's operation. Coliban Water have elected to draw raw water primarily from Reservoir 1 to keep the algae load on the treatment plant as low as possible.

Figure 4 : Algae Counts in Raw and Treated Water

Figure 5 below shows the log removal rating of algae across the treatment system versus the incoming algae count. The higher the incoming algae count the greater is the log removal - approaching 3 log at counts of 22,000 per ml. At lower raw water algae levels the removal rating is much less. This outcome is unexpected with CMF as the first part of the treatment stage.

If we look at the data in Table 4 below an interesting outcome is shown across the treatment process units. The CMF plant is removing algae to very low levels but levels post membrane filtration increase. This is evident in the pipework immediately prior to the activated carbon vessel and in water after the carbon. The only conclusion that can be drawn from this data is that the algae is breeding in the pipework and activated carbon vessel.

It was discovered that the fibreglass vessel containing the carbon was partially transparent to light. It has now been painted to inhibit light. Further results from January to June this year (Table 4) indicate algae levels have fallen in the raw and treated water.

• The apparent ability of algae to breed in pipework and the activated carbon vessel was unexpected. Painting of the carbon vessel has appeared to alleviate the issue to some extent. Algae increases in pipework (presumably dark) post membrane filtration is not expected.
• The actual algae counts are still very low in the treated water to Trentham and no taste and odour issues have been encountered over the spring and summer period.
• The ability of the carbon filter to behave in a biological manner (BOC) is certainly accounting for some of the removal of the naturally occurring organic carbon. This is evidenced by the removal of true colour across the treatment system.

Figure 5: Log removal of Algae across the Treatment System

Table 4 : Algae Counts Across the Treatment System

4.4 Membrane Failure Issue
After some six months of operation, routine integrity testing of the membranes by 'air pressure decay' indicated loss of integrity in the membrane bank.

Modules were removed from the plant and tested in USF Australia's laboratories. The membrane fibres were found to have circular holes penetrating the wall. These holes were some 5 to 40 micron in diameter - much larger than the 0.2 micron average pore size.

The entire membrane bank was replaced under warranty by USF.

The exact cause of the holes in the membrane wall is still not certain. Two possible explanations are:

• The holes may have been caused by a membrane extrusion manufacturing fault.
• The holes may have been caused by some unknown material in the raw water impinging on the membrane surface.

USF's research and development team have concluded that the failure was extremely rare and unlikely to occur again. The current set of membranes have now been in operation for 10 months and show no sign of loss of integrity.

4.5 Backwash Water Issue
Table 5 below shows testwork carried out on the backwash return water from the settling lagoons. The backwash water comes from the CMF plant and Activated Carbon Vessel and returns to the source reservoirs via two simple lagoons in parallel. Some natural settling taking place in these lagoons.

The quality of the backwash water in respect to Colour and Turbidity is higher than normally permitted in a chemically coagulated conventional treatment plant designs. Because the Trentham plant is not dosing chemical coagulants, the backwash water will not settle. There is concern that, over time , this material returning to the raw water reservoirs will impact on the treatment process.

The issue can be resolved quite readily by introducing chemical coagulation to settle the backwash water. This is not seen as a desirable as the plant currently has no need to introduce these chemicals.

Coliban Water and USF Australia propose to address this issue by introducing a more passive 'root zone' wetland approach in the settling lagoons. The proposed reed beds will promote solids capture and biological removal of the organics in the backwash water. This will hopefully continue the 'non chemical' approach that the Trentham WTP is demonstrating.

Table 5 : Backwash Water Return

5.0 CONCLUSION

Experience with no-coagulation operation highlights some significant advantages compared to treatment using chemical coagulation.

• The process does not have to be finely tuned. There is no dosing to be adjusted either for coagulation or pH reduction. Changes in the solids content of the raw water is automatically accommodated by the microfiltration unit and it backwashing frequency. The ozone system can operate under automatic redox control ensuring maintenance of a small ozone residual on the contact tank outlet even under varying raw water conditions. With these two reliable steps ahead of the GAC, the efficiency of its adsorption and biological functions are unaffected.
  
  
• The process is easier to control. As a result of the above features, control of the plant is particularly straight forward. There is no drop off in treated water quality under stop/start conditions or when changes occur in raw water quality.
  
  
• Less chemical handling issues. There is a limited reliance on chemicals so chemical inventory levels do not need careful monitoring, and chemical storage and delivery requirements are minimised.
  
  
• Less sludge protection. No coagulant sludges are produced. The only solids in the wastes generated are those present in the raw water, and pH adjustment is not required prior to return of settled washwater to the source reservoirs.

The operational issues experienced have highlighted the following areas that need further work and understanding at Trentham,

• Algae activity in pipework and Activated Carbon Vessel
• Backwash water handling in the absence of chemical coagulants

The successful implementation and operation of the Trentham WTP utilising a unique three step process of microfiltration, ozonation and GAC/BAC filtration has demonstrated that high quality treated water can be produced without chemical coagulants under appropriate conditions.

The application of this no-coagulant process on other Australian waters may be particularly advantageous under the following water quality conditions:

• presence of giardia/cryptosporidium
• presence of tastes and odours
• presence of algal toxins
• low to moderate algal mass
• low to moderate true colour

6.0 AUTHORS

Michael Muntisov is GHD's National Manager for Water Quality and Treatment based in Melbourne, Jenny Stewart is a Project Manager with Coliban Water and Colin Nash is the Project Sales Manager with USF Australia.

DOWNLOAD (pdf 49K)