Total barrier disinfection is an expensive way of providing drinking water.
But at a high risk site in South London it was the only option.
Mark Hansford found out why.
North Orpington Well Station is a key water source in Thames Water's London supply network, with its 9.1Ml/d output - drawn from two boreholes sunk 50m into the chalk aquifer - enough to supply over 60,000 homes and businesses in south London.
It was therefore a major concern when, after a number of raw water quality failures due to high turbidity and coliform levels in the late 1990s, Thames Water was forced to rate the site as a high risk for the Cryptosporidium pathogen.
Although Cryptosporidium was never found at the North Orpington site, coliforms are considered a link to the bug. Like coliforms, it is transmitted through the faeces of animals (see box).
The water quality failures tended to occur after heavy rainfall, and the suspicion was that a foul water sewer was somehow leaking into the aquifer. However, extensive investigations failed to pinpoint the exact cause and source of the contaminant and the operational decision was taken to shut the site down.
'It reached the stage where we had temporary shutdowns after every rainfall, ' exlpains Thames Water engineer's representative Bill Greensmith.
'Although there was no legal pressure to close, it was just not worth the risk.'
By the time the site was finally closed in 1999 Thames Water was already well on the way towards developing a solution.
Outline design for the remodelling of the Well Station began in June 1998 and ultrafiltration membrane plant was selected to provide total barrier disinfection, the first use of the technology for Thames Water and one of the first uses of the technology in the whole of the UK.
UF membranes provide a 'total barrier' by forcing raw water through pores in hollow fibres so small that they are capable of blocking contaminants as small as 0.03microns.
At North Orpington the membranes consist of three 'stacks' with eight modules per stack. In each module there are four elements each 1.5m long containing 10,000 hollow fibre membranes.
Each fibre is 0.8mm in diameter, providing a total porous surface area of 3.360m 2.Raw water is pumped into the membrane stacks via a feed tank at a pressure of 2 bar. The system operates in a 'dead end' mode to minimise the pressure drop and energy costs.
The water flows in through the membranes, passing through the fibres into a central filtrate channel. From there it discharges into a filtered water tank before being pumped to supply at nearly 60m head with a carefully controlled 0.4mg/l dose of sodium hypochlorite to ensure a chlorine residual is maintained in the distribution system.
To stop the membrane fibres from fouling, and to maximise their 3-5 year life span, each element is backwashed on a reverse flow basis.
The process takes place in three stages. A normal backwash every 30 minutes using water taken from the filtered water tank is supplemented by an enhanced wash every eight hours with hypochlorite dosed into the filtered water. With 48 valve sequences in the 90 second process full automation is essential. A weekly chemical clean and integrity test completes the routine.
The £4M contract was let to contractor Morgan Est (formerly Miller Civil Engineering Services) in October 1999 as a design and build using IChemE Green Book conditions of contract. Consultant Binnie Black & Veatch was brought in by Morgan to do the detailed design and PCI Leopold supplied the UF membrane plant.
Work began on site in February 2000 with the target for the new plant to begin operation set for 29 January 2001.
However, the project was not a simple case of ripping out the old and installing the new, explains Morgan Est project manager Dave Collins.
'Although the site was not in regular production, it could still have been called into use in an emergency. We had to build the new around existing force pumps and infrastructure, ' he says.
While new buildings were constructed to house the membrane filters, filtered water tank and chlorine dosing, the existing structure - a listed building housing one of the boreholes and the old force pumps - was converted to contain the new supply pumps and 10m long Motor Control Centre.
To try and help combat the turbidity problems on site, the contract also allowed for the removal of both existing fixed speed borehole pumps and pipework, to be replaced with two new 208m 3/hr 25m head variable speed pumps.
Construction was far enough advanced in October to allow commissioning to begin. Final sequence testing, disinfection and sampling took place during January to allow the project to hit the 29 January deadline - just.
'We went live at about 4pm, ' says Greensmith. 'But it was a period of high demand, so this was vital.'
In all, Thames Water's first foray into total barrier disinfection will have cost the company upwards of £5M. But the technology is now proven, meaning future costs should be reduced.
In North Orpington the Cryptosporidium bug has bitten the dust.
The curse of Crypto Cryptosporidium, transmitted readily by water systems and without a cure, has earned a worldwide reputation as a 'nightmare bug', says the Chartered Institute of Environmental Health. It is difficult to detect, can live for long periods in cold dark reservoirs and pipes, strikes large numbers of people and is potentially fatal for the elderly, young and weak.
Cryptosporidium is a very small protozoan parasite of the Coccidia genus. It is found worldwide in many hosts including birds, fish and mammals. In humans it causes an unpleasant, self-limiting, gastrointestinal illness. Two thirds of people who contract the illness (Cryptosporidiosis) are children.
The parasite is excreted in huge numbers and can be passed on as a secondary infection as it requires a low infective dose of less than 100 organisms. The volume of water sample required to screen the water supply has to be large, often as much as 20 litres. With such a small number of organisms diffused in a large volume of water, it is easy to see how the organisms escape detection.
The biggest concern is Cryptosporidium's resistance to many traditional treatments including chlorination, rendering much of the treatment process irrelevant. How it is able to survive chlorination is not known, but it is believed that its protective 'oocyst' membrane plays a role.
One theory is that the organism might use a pump mechanism to expel toxins from its inside before they cause it harm.
Tests have shown that source protection allied to attention to detail during treatment is the best way to reduce Cryptosporidium numbers in supply water.
However, there is no known safe level for treated water and standards for water quality are not health related.
In Britain, the Water Supply (Water Quality) Regulations 2000, which took effect on 1 January 2001, demand that undertakers assess the risk of Cryptosporidium contamination of their sources of supply.
Where a risk is established, the regulations require undertakers to put in place measures to contain that risk such that the average number of oocysts per 10 litres of water is less than one.
To verify compliance with this treatment standard, water leaving treatment works is subject to a continuous sampling and analysis protocol, and breach of the standard is a criminal offence.