UV Cleaning System Performance Validation

By Mike Oliver
February 2003

The Author is a Research Scientist for Trojan Technologies Inc. in London, Ontario. He has worked on numerous disinfection and advanced oxidation projects since 1996. During the last 2.5 years he has been developing, testing, and validating innovative cleaning systems for use in UV disinfection systems.

All waters and wastewaters foul the quartz sleeves that enclose the lamps used in UV reactors. Quartz sleeve fouling necessitates the need for cleaning. Several cleaning strategies are currently being used in the UV treatment industry. To evaluate these strategies and to validate the performance of any cleaning mechanism, a testing protocol has been proposed. This protocol describes a positive control and criteria to evaluate cleaning performance. This protocol was able to distinguish between performance of offline-manual, automated-mechanical, and automated-chemical/mechanical cleaning methods. Of the strategies tested, the automated-chemical/mechanical cleaning system demonstrated superior cleaning efficacy. Automated-mechanical wiping and un-wiped systems failed rapidly upon onset of fouling.


In UV disinfection systems for wastewater and drinking water, various physical and chemical characteristics of the water being treated result in fouling of the quartz sleeves that enclose UV lamps. Factors such as interfacial temperature, UV intensity, hydrodynamics, and the quartz microstructure and topography allow the attachment of inorganic debris and organic films or greases. These organic and inorganic deposits absorb UV light and decrease the intensity of UV light penetration into the water or wastewater. This fouling decreases the UV dose resulting in reduced disinfection performance. Fouling is a complex process that tends to be site specific and is difficult to predict. The rate of fouling is highly variable from site to site, but eventually all sites accumulate fouling at intervals appropriate to the fouling rate. Fouling necessitates the need for cleaning of the quartz sleeves to maintain optimal system efficiency and performance.

There are several sleeve cleaning options currently used:

  1. offline strategies which require periodic removal of the UV lamp-sleeve structure for soaking in a chemical bath or manual wiping with a chemical cleaner;
  2. automated online strategies that utilize mechanical cleaners (fingers, spirals, seals, rings, etc.) that wipe frequently and require periodic manual chemical cleaning; and
  3. fully automated-chemical/mechanical cleaning systems.

Performance and cost-effectiveness of these options vary. Manual cleaning is labour intensive and expensive. Automated-mechanical cleaning technology typically consists of stainless steel brushes, rubber-type wipers and/or Teflon rings that mechanically remove the foulant from the sleeve. Mechanical cleaning alone is not effective in all water qualities, and in most cases requires periodic offline chemical cleaning. Furthermore, mechanical cleaning requires frequent wiper maintenance. In comparison, automated-chemical/mechanical cleaning systems are effective and require no maintenance for 6 months or longer.

The performance of cleaning systems impacts on equipment sizing to ensure disinfection and is required to support the justification for sizing of equipment. Validation of any cleaning system requires that a positive control and assessment of criteria be established to evaluate performance. Long term testing is an essential part of validation because once the cleaning system is unable to maintain a high degree of sleeve cleanliness performance drops off very rapidly. A pilot study of 3 months or less could not accurately predict long-term performance. In the present study the demonstration of fouling in the absence of cleaning (positive control), monitoring of residuals after cleaning, and the long term performance of the cleaning device in maintaining optical performance of the sleeve were used to evaluate the sleeve cleaning mechanism.

Methodology (Validation Protocol Used)

Due to the complex nature of fouling combined with daily, weekly, monthly, and seasonal variation in water quality, accurate comparison between cleaning strategies can only be achieved by testing the cleaning strategies in parallel using the same effluent over the same time period.

A positive control where the wipers were removed and sleeves left un-wiped was an essential part of the protocol. This was used to assess each site´s fouling potential and quantify fouling rates during the test period. Measurement of fouling rate allows comparison of the relative difficulty of cleaning at different sites and reflects site-specific water quality effects. A site must be shown to foul at an appreciable rate in order to distinguish between cleaning system strategies. This data was also used to determine each site´s off-line chemical cleaning frequency for a non-automated cleaning strategy.

For automated system comparison, new wipers were installed to assess the cleaning strategies being evaluated. This included mechanical wipers and chemical/mechanical wipers. The effectiveness of each cleaning strategy or wiper was tested in duplicate.

Performance was evaluated using UV transmittance (254nm) through the quartz sleeves compared to a new clean reference quartz sleeve. At timed intervals, the sleeves were removed from the channel, rinsed with de-ionized water (to minimize drying spots), and allowed to completely dry. Ten random points were measured along the length of each sleeve within the arc length of the lamp. Double layer sleeve UV transmittance measurements were converted to single layer UVT and reported as relative to a new clean reference sleeve. The mean average of all the measured points of all the sleeves for each cleaning strategy provided the average sleeve UVT for that strategy at that time interval. This non-destructive technique allowed for the sleeves to be replaced in the same location without wiper removal for continued testing.

Several months of controlled testing was necessary to distinguish between cleaning strategies, establish long-term performance, and determine maintenance frequencies. The absence of long term testing could result in incorrect predictions of full system performance; result in system under sizing, and underestimations of maintenance costs.


Four wastewater sites representing different upstream process and water quality were selected to test the validation protocol. Off-line, automated-mechanical, and automated-chemical/mechanical cleaning strategies were evaluated in parallel by equipping all cleaning strategies on a single bank.

Table 1. shows the level of effluent treatment prior to the UV system as well as selected water quality parameters.

Table 1.   Selected Site Treatment and Water Quality Parameters
Site Reference 1 2 3 4
Co-agulantNoneFerric SaltAlumAlum
UVT (%)25 - 4570 - 8065 - 7570 - 80
pH6.5 - 7.57.0 - 7.66.7 - 7.37.0 - 7.6
TSS (mg/L)150 - 2005 - 1015 - 35< 5
TDS (mg/L)600 - 700700 - 800450 - 550600 - 700
Hardness300 - 400200 - 250200 - 250250 - 350
Fe (mg/L)1.5 - 2.00.2 - 0.6<0.02 - 0.050.05 - 0.20

Standard commercial low pressure systems (Trojan UV 3000 Plus) were equipped with positive control sleeves (no wipers), mechanical wipers (two different designs), and chemical/mechanical wipers.

Results from the un-wiped control sleeves indicated the effluent from all sites fouled the sleeves at a quantifiable rate allowing comparison between the automated cleaning strategies. This is shown in Table 2. with fouling rate expressed as the initial percent loss of quartz sleeve UV transmittance per day.

Table 2.   Site Fouling Rates (Initial percent loss of quartz sleeve UVT per day)
Site Reference 1 2 3 4
Co-agulantNoneFerric SaltAlumAlum
Fouling Rate (%)

The loss of quartz sleeve UV transmittance due to fouling indicates that overall system performance was rapidly compromised without cleaning in all cases. Without an automated cleaning system this loss must be compensated with frequent off-line chemical cleaning, increased system size, and higher power input in order to maintain UV dose.

An eighty percent cut-off line for acceptable sleeve transmittance was selected based the NWRI Guidelines for Drinking Water and Water Reuse, 2000 (1). Sleeve transmittance below eighty percent (relative to a new clean sleeve) would compromise disinfection and require module removal for off-line manual chemical sleeve cleaning and wiper maintenance.

Table 3. illustrates the estimated off-line maintenance frequency for the wipers designs testing tested under these conditions.

Table 3.   Estimated Off-Line Maintenance Frequency
Site Reference 1 2 3 4
Co-agulantNoneFerric SaltAlumAlum
No Wipe1 day2 days9 days2 days
Mech Wipe Design 12.5 months40 days6 months3 months
Mech Wipe Design 22 months15 days> 6 months1 month
Chem / Mech Wipe> 4 months> 5 months> 6 months> 5 months

Results showed un-wiped strategies would require cleaning maintenance within days to maintain acceptable sleeve UV transmission. The mechanical wiper designs tested showed varied success at the sites tested. It was not possible to predict mechanical wiper performance based on treatment level or water quality. Several months of testing were required to correctly establish mechanical wiper performance. At all four sites the chemical/mechanical cleaning system was able to maintain sleeve cleanliness above 95% relative to a new clean sleeve over the periods tested. Typically, once the sleeves begin to foul, mechanical wipers are no longer able to keep the sleeves clean and system performance drops off rapidly.

Site specific cleaning performance test results are summarized in Figures 1-4.

Figure 1
Figure 1 - Cleaning Performance Results at Site 1

Figure 2
Figure 2 - Cleaning Performance Results at Site 2

Figure 3
Figure 3 - Cleaning Performance Results at Site 3

Figure 4
Figure 4 - Cleaning Performance Results at Site 4


Field-testing confirmed that fouling is complex, site specific, and difficult to predict. The rate of fouling was highly variable from site to site, but eventually all sites accumulated fouling at intervals appropriate to the fouling rate. A positive control (unwiped sleeves) was used to determine each site´s fouling potential and fouling rate. The upstream treatment processes were not directly linked to fouling rate or cleaning method performance. The water quality data was insufficient to provide a definite link between any water component(s) and fouling rate or cleaning performance.

The criteria established in the protocol to evaluate performance were successful in validating cleaning methods and was able to distinguish between different cleaning strategies (even different mechanical wipers). Mechanical wiper performance at a site was not related to the fouling rate of the site. It would not be possible to predict performance of any cleaning method by simply determining fouling characteristics of a site. The chemical/mechanical wiping systems were able to maintain very high UV sleeve transmittance for extended periods of time at all sites. Mechanical wiper performance was very site specific. At all sites the mechanical wiping systems tested would require periodic removal and manual chemical cleaning of the sleeves.

Long term testing was necessary to clearly distinguish between cleaning strategies and to validate any cleaning method for long-term performance. Cleaning validation must be done using the effluent to be treated and with all the operational parameters expected to be used in a full-scale installation. The effectiveness of the cleaning mechanism should be tested using a minimum of four quartz sleeves (1).

Wiper frequency is a very important variable for all automated cleaning systems. An equilibrium must be made between maintaining acceptable sleeve UV transmittance and accelerated wear on the wiper system parts to provide a cost effective balance between labour and replacement part costs. Mechanical wiping systems generally require high wiping frequencies (on the order of minutes) in order to prevent fouling materials from building up. Chemical/mechanical systems are able to maintain high sleeve UV transmission with much less frequent wiping, as the chemical solution is able to remove all layers of fouling build up.

A maintenance schedule that allowed the sleeves to drop to eighty percent (relative to a new clean sleeve) between off-line chemical cleanings would require 25% more equipment installation to compensate for the loss in UV intensity due to sleeve fouling. Also power input would increase between cleanings in order to maintain UV dose. A wiper frequency is recommended such that the sleeves never fall below ninety five percent sleeve UV transmittance between wiping sequences. For chemical/mechanical systems this translates to a 12-hour frequency at a rapidly fouling site and a 24-hour frequency at a slow fouling site for the low pressure UV systems.

Validation of cleaning methods used in UV disinfection systems require attention to operational parameters, positive control, criteria to evaluate cleaning, and long-term performance. The absence of any one of these factors in any protocol would not allow proper evaluation. The validation protocol applied showed the chemical/mechanical cleaning system to be very effective. The test protocol was able to distinguish between different cleaning methods and is suggested to be used by the industry.



Copyright © 2003, ECO Services International