UV Disinfection Workshop – a Guide to Validation
May 2017
Mike Waite, FWR Water Supply Co-ordinator

CIWEM and the International Ultraviolet Association (IUVA) jointly organised this workshop which was, of necessity, very technical. Some of the presentations included too much information for the non-expert to assimilate although this is not a criticism of a very interesting meeting.

The introduction of UV disinfection guidelines by the Drinking Water Inspectorate from the initial 2010 revision has enabled far greater application of UV disinfection technology across the UK. This seminar reviewed the fundamental aspects of this technology and its validated application within the UK.

The meeting began with a presentation from Volker Adam (Heraeus-Noblelight) on the fundamentals of UV technology, including the mode of action of UV and its advantages as a treatment process, either at 257nm which disrupts DNA for disinfection or at 180 nm for the oxidation of organics. He described the three types of lamps used: low pressure, low pressure amalgam and medium pressure. The low pressure amalgam lamp has a much longer useable life (>16,000 hrs). Although UV LEDs are being developed, they are too low-powered and inefficient to be of industrial use.

Ian Mayor-Smith (IUVA & Hanovia) spoke about calculating UV dose/fluence before going on to describe a methodology for measuring fluence. The fluence (UV dose) is obtained by multiplying the fluence rate (or irradiance) by the exposure time in seconds (International Ultraviolet Association). He described how UV intensity reduces with distance from the lamp and that UV energy x sample transmission x time = fluence, and then showed a simple apparatus for directing collimated UV on to a petri dish to determine fluence. Using E.coli T1 bacteriophage, the reduction in phage could be measured and the reduction equivalent fluence (REF) calculated. The log inactivation was proportional to the fluence. There are various sources of possible error such as the need to make serial dilutions when counting the phage and these need to be taken into account by means of a validation factor (VF) when determining the validated dose which will be REF multiplied by VF.

Jutta Eggers (DVGW) then gave a validation overview comparing the three main standards in use: UVDGM (the USEPA approach), DVGW 294 (the main European standard), and ÖNORM (the Austrian standard). The latter two are currently being harmonised for low pressure reactors. UVDGM provides technical information on UV but is not a regulation and has no legal status, whereas DVGW 294 contains requirements which are legally binding at least in Germany, and provides certification which involves regular surveillance of equipment production sites to ensure continued conformity. Certificates are valid for five years.

DVGW itself is an organisation which has many other roles beyond UV validation. Reactor validation by DVGW involves technical testing by the examination of documentation and characterisation of components, and also performance testing both at bench and full scale. All materials coming into contact with the water must have hygienic test certificates of suitability. Lamp UV output and efficiency are measured and properties determined in a standard reactor. The only way to determine the reduction equivalent dose (RED) is by using test microorganisms (biodosimetry). The organisms used are three different phages and Bacillus subtilis. The susceptibilty of the test organisms is determined in the laboratory by producing a calibration curve of reduction against fluence. The reactor is tested by dosing a suspension of the test organism plus lignin sulphonate at a point in the pipework just before the 90º bend at the entrance to the reactor and making counts before and after the reactor. Reactors must achieve an RED of at least 400 J/m² after allowing for a combined ageing and fouling factor of 70% (to become 75% in the next revision). An RED of 400 J/m² is equivalent to a 4 log reduction. Finally, reactors are tested at conditions of high absorbance and emission and low absorbance and emission and must achieve an RED of at least 400 J/m² in both, the lower result being taken for validation.

Karl-Heinz Schön (DVGW) described the critical components of reactor functionality. The control cabinet must trigger shutdown of the reactor in the event of lamp or sleeve failure or overheating, and only permit flow if irradiance, flow and operation of lamps and ballasts are satisfactory. It must also warn if irradiance is getting low for a certain amount of time. In addition, it should compare readings from duty and reference sensors. Disinfection efficiency may decrease due to fouling of sleeves and ageing of components; low pressure lamps are also affected by water temperature. Fouling of sensors can also be important and consequently they must be accessible for cleaning. DVGW requires sensors to have an angular response of ± 80º and be calibrated to a national standard such as NPL (National Physical Laboratory) or the German PTB. A round robin test by three laboratories, checking at least two sensors from each of six manufacturers against the manufacturer's own calibration, showed that while the laboratories were reasonably consistent, the manufacturers' results were often much lower. Manufacturers have since made improvements. DVGW 294 requires checks every two years for certified reactors but not for the individual components.

Christy White (MWH) described validation approaches and challenges with wastewater and storm water in the UK, referring to the results of UKWIR project WW17. Without any disinfection, final effluent after dilution with the receiving water may contain around 104 E.coli/100ml – bathing water targets are 80 E.coli/100ml and shellfish targets are 110/100ml. Treated sewage from 9 million PE (population equivalents), or about 7% of the UK population, requires disinfection before discharge. The UKWIR study looked at company data and concluded that in many cases there was an overachievement of dose and the need for a UV/dose relationship, more effective dose control and site-specific data to set the desired inactivation level. To optimise power use it is necessary to select the right target dose and be able to modify the dose in relation to changes in water quality and flow rate. In the study, site-specific testing determined the log-removal of target organisms required to meet water quality targets, the design envelope (UV transmittance vs flow) and the dose-response relationship for target organisms. Dose-response is found not to be linear. Having done this, a specification for UV equipment could be produced.

Finally, Chris Rockey (South West Water) gave a West Country perspective on UV disinfection systems for public drinking water. South West Water installed its first UV treatment in 1993. The 2000 Water Quality regulations made it a criminal offence not to adequately treat and disinfect drinking water. The company has many good quality groundwater sources which, in a lot of cases, rely solely on UV and have no contact tanks for chlorination. The company is continuing to retrofit additional UV plant. At their Knapp Mill WTW they found that slow sand filtration is not so good after heavy rain, and they have therefore installed UV. The company goals are to have an absolute barrier to Cryptosporidium, minimise disinfection by-products and keep consumer bills down in the long term.