The upgrading of a potable water treatment plant

Most people take water for granted. When they turn on a tap they expect the water to come flowing out. They expect it to be safe to drink and they expect it to be pleasant to drink. Supplying this essential commodity to these standards however can be a real challenge to those who are responsible for supplying it.

In 1995, Dunedin's water treatment received an "E" grading – the lowest grade possible. This is a description of the addition of a dissolved air flotation (DAF) process to upgrade an existing Water Treatment Plant so that "A" grade water could be produced. The cold, soft, highly coloured water to be treated was not amenable to clarification processes by conventional settling means. The process chosen represents the leading edge of DAF technology, and incorporates lessons learnt from extensive national and international review. The process design was based on two years of laboratory testing and modelling. Operation of the plant has given results that exceed expectation. Aspects from all seven of the technological areas in the New Zealand technology curriculum can be seen to be embedded in the project.

Conventional water treatment

To remove dirt from water a chemical (coagulant) is added to the water to form a gelatinous floc which enmeshes and sticks to the dirt particles. The initial chemical reaction is called coagulation and the subsequent gentle mixing process is known as flocculation. The floc and the dirt particles are then removed by processes of clarification, and filtration. In clarification, the floc particles are settled out and the clarified water is then filtered to remove any remaining floc and dirt particles.

Various chemicals can be used as coagulants with Aluminium Sulphate ("Alum") being the one most commonly used in New Zealand. Polyelectrolytes (chemicals consisting of long chain molecules) are also used to help with flocculation, clarification and filtration. The long chain molecules trap the dirt particles during flocculation, help settle the floc during clarification by their additional mass, and aid filtration by creating a "sticky" surface on the sand particles.

For optimum performance the water flow through treatment plants needs to be steady.

The Dunedin situation 1995

The most modern treatment plant in Dunedin in 1995 was at Mt Grand. This plant could not produce "A" grade water because it could not cope with the wide range of raw water quality that was delivered to it and because of operational constraints.

The source water for the Mt Grand plant comes from Deep Stream and Deep Creek. These catchments are located 60km inland, to the north west of the city. The waters from both catchments have similar characteristics. They are both cold, soft, acidic, generally of low turbidity, and relatively highly coloured.

As no raw (untreated) water storage is presently available either at the catchments or near the city, the water has to be taken directly from the streams (referred to as "run-of-the-river" supply).

As a result the colour and turbidity in the water reaching Mt Grand vary continually. The changes in these parameters can be dramatic; for example the colour can change from 30¡ Hazen to 80¡ Hazen in two to three hours. Colour peaks decrease more slowly than the turbidity peaks. Colour can trend downwards again over a period of two weeks whereas the turbidity can drop back to its original level in two to three days. In many cases the turbidity drop is much more rapid.

The Mt Grand Treatment Plant

The Mt Grand Water Treatment Plant is situated 300 metres above sea level and it can feed water by gravity to virtually any area of Dunedin. As it can supply approximately 70 percent of Dunedin's water demand it is a major strategic part of the city's water infrastructure.

The existing plant began operating in 1977. It was a direct filtration plant using dual media filters (a layer of anthracite over a layer of sand). Direct filtration means that following chemical flocculation the water is sent directly to filters, without passing through a clarification stage. It did not have a conventional clarification stage in the process because the very small, light floc produced from the raw water is extremely difficult to settle out.

Alum was used as the coagulant and hydrated lime was dosed to adjust the dosed water pH to the optimum valve for most effective flocculation. A very small amount of polyelectrolyte was used to aid filtration.

The continual variations in raw water quality, coupled with the dramatic changes from time to time, made it very difficult to maintain the chemical dosing at the "correct" dose and pH level. Control had always been difficult because of the soft, unbuffered water and the fact that the pH sensing equipment did not operate reliably and consistently. In addition the changes in plant throughput as a result of reticulation system demand changes required time for the process to "settle down" again. For these reasons the treated water quality was adversely affected.

For most of the time the treatment process at Mt Grand is a colour removal process. However, at times of stream "freshes" the turbidity levels peak and essentially two treatment processes are required – colour removal and turbidity removal. The pH for the optimum removal of each is different and the nature of the flocs formed in each process is also different. The colour flocs are very light and do not settle easily; the turbidity flocs are larger and denser and will settle.

The problems to be overcome

To upgrade the treatment process and operational performance of the Mt Grand plant, it was evident that the addition of a clarification process was necessary. This would allow the plant to better cope with the wide range of raw water quality to achieve an "A" grading.

The plant's lack of treated water storage was a severe restraint on the plant in two ways; first, changes in demand in the distribution system resulted in rapid changes in plant flows, and second, the plant was only achieving an average production of 20,000m3/day even though its capacity was 40,000m3/day.

The pipework configuration in the filter block was very basic and if maintenance was required on any one filter then the complete block had to be shut down. This meant the complete loss of treated water production which was a significant problem if work was required during times of high demand.

The sequence of improvements

It was decided that the upgrade would be carried out in the following stages:

• Modification of the pipework at the filter block
  - To allow independent shutdown of the filters for maintenance instead of the former complete shutdown required.

• Provision of treated water storage
  - Two 20,000m3 storage tanks to deal with demand fluctuations from the reticulation system and allow the plant to treat water at a steady throughput. This would improve the treated water quality and allow full plant capacity to be utilised continuously.

• Addition of a clarification stage to the treatment process
  - This would provide another "barrier" to microbiological contaminants in the water. It would also reduce the amount of material removed by the filters to be reduced and allow them to function more efficiently as well.

• Provision of new chemical dosing and storage equipment
  - This would allow all chemicals to be dosed from storage in liquid form. Dosing would therefore be more accurate and more consistent.

Treated water requirements

The aim was to have water which was clear and sparkling to look at, which met public health standards, and which was effectively Giardia and Cryptosporidium free. The treated water was thus to meet the "Drinking Water Standards for New Zealand (1995)", but to be to a higher standard in two areas. One was final colour which was to be less than 5¡ Hazen (standard was 10¡ Hazen) and the other was final turbidity. To ensure removal of Cryptosporidium, the operational goal for turbidity was to be 0.1 nephelometric turbidity units (NTU) or less. This was based on research by Leland [2]. (The standard requirement was 0.5 NTU or less). The most recent Drinking Water Standards have now reduced the turbidity limit to 0.1 NTU or less.

International review

Council engineers inspected water treatment plants in the US and UK that treated water with similar characteristics to Dunedin's. They also took the opportunity to discuss the trends in water treatment with management and operating staff, and with engineering consultants.

Plants that could consistently achieve the treated water standards being aimed for had a number of factors in common:

1. Plenty of Water Storage – Raw and/or Treated
This allowed a consistent water quality to be presented for treatment and also allowed a constant plant throughput.

2. Chemicals were Accurately Dosed
The calibration of dosing pumps was checked regularly throughout the day (plants in the US are manned 24 hours a day). Magnetic flow meters were used on dosing lines at one plant.

3. Plenty of Flocculation Time
This was seen as one of the main factors in producing a good final water.

4. Some Means of Settling/Clarification
This allowed the filters to do a better job and to do it for a longer period of time.

5. Plants Highly Instrumented
Some plants had zeta potential measuring equipment, streaming current meters, turbidity meters, and particle counters. This enabled plant staff to know what was going on in the process.

6. Very Well Equipped on-site Laboratories
Again, this enabled plant staff to check what was going on straight away.

Choice of treatment process

As well as having to meet the required treated water standards the new treatment process would be subject to a number of constraints:

• As much of the existing plant as possible was to be retained.
• The wastewater volume discharged to the sewer was not to be increased, even though the average plant throughput was to nearly double.
• Raw water storage had been advocated for the plant. As algal blooms occurred at the city's other raw water reservoir the treatment process had to be capable of handling algae.
• The use of synthetic polyelectrolytes was to be minimised and preferably eliminated. There was a trend away from the use of such chemicals in some countries and there were indications that it might become a regulated matter in US.

The main stages of a conventional treatment process are flocculation, clarification, and filtration. The first of these assessed was filtration. Possible technologies were evaluated by considering advantages, limitations, capital costs, and operating costs of different filtration systems.

Membrane filtration systems such as microfiltration and nanofiltration were quickly eliminated because of their cost and the fact that the existing filters were still suitable. If an additional barrier to microbes was required after conventional filtration the ozone disinfection was available.

Next, possible methods of clarification were investigated. These included sedimentation, direct filtration, lamella plates, contact absorption, and dissolved air flotation (DAF). The DAF process was a clear winner. It was the only clarification process which could handle algae effectively and it did not require the use of polyelectrolytes. DAF also removed most of the flocculated material from the water and so it reduced the loading on the filters substantially. Thus the wastewater volumes could be kept to the required limit.

DAF had been used as a clarification process in potable water treatment in the UK and Scandinavia for over 20 years. It was an emerging technology in the US. In the process air (which has been dissolved in a carrier stream of water) is released as tiny bubbles at the bottom of a flotation tank. When rising to the surface they collide with the flocculated particles and carry them to the top where a scum layer is formed. The water is thus effectively clarified and the scum layer is removed by either mechanical scrapers or by floating the layer off. It is a robust process and testing in the US has confirmed that it removes cryptosporidium.

With the clarification stage resolved, attention was turned to the water chemistry.

Water chemistry investigations

The key to effective water treatment is a knowledge of the water chemistry involved and the provision of suitable unit processes to allow the chemical reactions to occur fully. In particular it was recognised that effective clarification depended on effective coagulation and flocculation. The determination of optimum conditions for coagulation and flocculation, and the assessment of the effectiveness of DAF as the clarification method were therefore key objectives.

Obtaining more information on the raw water was an additional objective. Plant experience had shown that water being treated on different occasions but with the same colour would react quite differently. Obviously the nature of the organics in the water was different even though the colours were the same.

Two years of testing were carried out to establish the required information.

Plant testing was carried out for the following reasons:

• to see how well the laboratory processes could simulate what actually happened in the plant;
• to try out at full plant scale optimal water treatment as determined at laboratory scale;
• to assess the influence of plant hydraulics on the treatment process at different plant flows;
• to assess whether carbon dioxide could be used to accurately control the dosed water pH.

Laboratory equipment

A specially equipped water testing laboratory was established that was equipped to rival any in New Zealand. The organic carbon analyser and the particle counter were the first such instruments to be purchased in New Zealand by a water providing authority or business.

Conclusions from testing

The following conclusions were drawn from the laboratory and plant testing:

• The DAF process would work successfully at the colours and temperatures likely to be expected.
• Successful DAF depended on effective coagulation and flocculation.
• Alum, lime and carbon dioxide were the chemicals of choice.
• Flocculation energy inputs and detention times were directly related (hence flocculation tank design would be decided by economics).
• Two-stage pH adjusted coagulation/flocculation was preferred for raw water colours above 50¡ Hazen.
• The two-stage process achieved good natural organic matter (NOM) removal.

Specialist process design

To ensure that the best specialist process design advice was obtained for the clarification stage, a partnership was entered into with Purac Ltd (UK). It was felt that a partnership approach would ensure that commercial competitive pressures did not affect the quality of design.

In return, to control costs, Purac only supplied the limited items of equipment that were critical to the integrity of the DAF process. All other equipment was tendered normally, and this allowed most to be sourced from local suppliers.

The DAF process has incorporated the latest international research enabling high loading rates and achieving low recycle rates, especially for low water temperatures. Flexibility is a feature as both contact zone characteristics and recycle rates can be varied.

Chemical dosing - national research

In 1998 the design engineer for the chemical dosing systems and the senior water operator visited the most significant installations around New Zealand to review the current state of art regarding lime and carbon dioxide dosing. The results of this research were dosing systems that produce reliable results in terms of both quantity and quality. In addition the systems have the operator's approval as being practical to operate.

Mechanical and electrical

The Mt Grand Plant operates under automatic control with the operators in attendance for only eight hours a day on week days. The sensors, measuring instruments, and the electrical and electronic control systems therefore play a vital role in the plant's operation.

On-line measurement of parameters such as colour, turbidity, pH, and temperature allows computers to evaluate the data and to make decisions in accordance with pre-programmed algorithms. The electronic and electrical controls are then activated to control all the mechanical equipment throughout the plant so that the required treated water quality is achieved. Measurement of flow, level, pressure, and equipment operating speeds ensure that the process operates safely and efficiently.

Structures and materials

The design and construction of the civil works (buildings, tanks, and large diameter pipework) requires a good working knowledge of structures and mechanisms and of materials.

The DAF building is dominated by the elevated flocculation and DAF tanks. These tanks were elevated so that the new process could operate with the water flowing by gravity from and to the points where it is fitted into the existing Mt Grand facility.

The tanks and their support walls form the major part of the building structure and are formed from in-situ concrete. Concrete was selected for its durability, cost, and ability to act both as tank walls and building structure.

Noise emissions were carefully budgeted for and controlled by using pre-cast concrete exterior walls, plywood lined roof, 6mm thick glass windows, and acoustic louvres on all ventilation openings.

Commissioning and performance

Purac Ltd (UK) sent a specialist to the plant for a month to commission the plant and undertake performance tests. This process went very smoothly even though it was during a period of extremely poor raw water quality.

The results achieved were very promising. Raw water colour was in excess of 80¡ Hazen with turbidity off the scale of the recording turbidimeter (ie >10 NTU). In the past, the plant could not have coped with this water, the sand filters would have rapidly clogged and the plant would have had to have been shut down.

With the addition of the DAF process treated water of 3¡ Hazen colour and less than 0.1 NTU turbidity was produced, and the time between the filter backwashes was as long as that used to be achieved when raw water quality was at its best. These results exceeded expectation. Performance testing of the plant was completed by the end of October 2000.

Note

Roger Oakley is a project engineer for City Consultants, the in-house engineering business unit of the Dunedin City Council. Since 1996 Roger has acted as project manager for much of the DCC's current $64 million water upgrade programme. Murray Petrie is a senior design engineer with City Consultants and was the team leader for the process design and the supervision of the process equipment installation for the upgrade of the Mt. Grand plant. We are grateful to them for making this material available for use by technology teachers. The full paper from which this summary is taken was published in the TENZ 2001 conference proceedings.
For further information contact TENZ@rsnz.govt.nz

	Classroom practice in technology and hangarau | Longer–term programming Professional development | Innovation in technology and hangarau education | Technology and hangarau in the news | Frequently asked questions | Technological practice in the wider community | Assessment in technology and hangarau education| Development in technology and hangarau facilities | Main Index