Crypto. removal under scrutiny
Tom Hall, of WRc, reports on Cryptosporidium risk management using membranes and a project designed to unify integrity testing procedures in order to meet DWI requirements
Over the past ten years, concerns over Cryptosporidium in water supplies have led to many developments in water treatment aimed at reducing the risks of waterborne infection. Some of these have been imposed through a regulatory framework for England and Wales established by the DWI. Regulations introduced in 1999 required risk assessments to be carried out at all WTWs, to identify those “at significant risk” from Cryptosporidium. At these sites, a rigorous regime of continuous sampling of treated water was required for Cryptosporidium monitoring, with a regulatory treatment standard of 1 oocyst per 10 litres.
LRV = log10 (feed water
concentration/treated water concentration).
The sizes of pores in membranes are much smaller than Cryptosporidium and intact systems can potentially provide 5 LRV or better for Cryptosporidium-sized particles (i.e. 99.999% removal). DWI did not condone the use of LRV because it was “empirical and based on tests carried out at a range that is outside the requirements of the treatment standard”.
Many water companies have installed membrane plant to remove Cryptosporidium and this trend is continuing. In a survey of 13 of the larger water companies collectively serving more than 90% of the UK population, 49 plants for Cryptosporidium removal were identified as being operational in 2002, or were soon to be brought into service. Most of these will treat groundwater or spring sources, with only three treating solely surface water. The total output from these plants will be 845Ml/d. All but one of these plants and 99% of the output are from three membrane plant suppliers, using hollow fibre design. These are:
- Norit, with X-Flow ultrafiltration (UF) membranes,
- USF Memcor microfiltration (MF),
- Kalsep using Hydranautics UF membranes. (See Table one).
The membrane types used in these plants are on the DWI approved list. However, a clear understanding of the DWI requirements for integrity testing is also needed to gain exemption from regulatory Cryptosporidium monitoring at these sites.
The majority of membrane integrity tests rely on air pressure applied to the wet membrane fibres. The air will diffuse very slowly through the membrane if integrity is not compromised, but much more quickly through holes significantly greater than the pore size of the membrane. The tests measure either the rate of diffusion of air through the membrane at a constant applied pressure (diffusive air flow (DAF)), or the pressure drop (pressure decay test (PDT)). The size of hole that can be identified with these procedures depends upon the test pressure and the tests have been developed primarily to identify the equivalent of a broken membrane fibre, the smallest of which are 0.3mm diameter. Background air flow through intact membranes will mask air flow through an integrity failure, limiting the number of vessels that can be tested concurrently in parallel. Membrane plant suppliers provide information on the test conditions (pressure, duration) and on interpretation of the results.
One membrane plant supplier (Norit) provides an alternative test procedure involving periodic dosing of powdered activated carbon to the feed water and monitoring the permeate with particle counters. This is referred to as the spiked integrity monitoring (SIM) system. Two of the larger plants in the UK (one in Scotland) use this system.
WRc carried out a programme of work for seven large UK water utilities with the primary objectives of providing a unified approach to integrity testing procedures which can be used at a site-specific level to satisfy DWI requirements. Within this, the intention was also to provide a better understanding of the underlying basis to integrity test procedures, to assist in interpretation of the results and facilitate a better dialogue with regulators.
The overall conclusion was that, at the present time, the systems provided by the membrane plant suppliers for carrying out integrity tests represent best available technology for meeting regulatory requirements. These systems cannot provide continuous monitoring of integrity and are incapable of detecting integrity failures for 1µm particles (i.e. a single hole of this size). However, the tests can be carried out relatively easily and can therefore provide frequent assurance the performance of the plant is not significantly compromised. Because the tests are aimed
at identifying broken fibres, they evaluate removal of particles of all sizes.
A mass balance approach used in the work demonstrated that LRV is a fundamental property of the system, rather than an empirical value and can be linked, through the integrity test, with the number of broken membrane fibres in the system.
An approach was recommended for developing and specifying integrity test procedures to meet DWI requirements, which in summary involved:
- establishing LRV targets for the plant,
- establishing the basis of the test and the nature of the integrity failure it aims to identify under the conditions to be implemented,
- identifying the maximum number of integrity failures (e.g. broken fibres) acceptable to avoid compromising the LRV target for the plant,
- defining the test frequency,
- specifying data collection and handling procedures, to demonstrate how these will be used to enhance security of treatment.
This approach has been used successfully to gain DWI approval for integrity testing procedures at a number of sites to date and more will be approved in the near future. As operating experience with membrane plants develops it may be possible to reduce the rigour of the integrity testing regime without compromising the very high degree of
security against waterborne infection that membrane plants can provide