Evaluation of UV System Performance and Challenges of UV Validation Testing at the Laguna County Sanitation District

Transcription

1 Evaluation of UV System Performance and Challenges of UV Validation Testing at the Laguna County Sanitation District Ufuk G. Erdal 1, *; Jamal Awad 1 ; Marry Vorasis 1 ; Marty Wilder 2 and Mark Moya 2 *Corresponding Author: uerdal@ch2m.com 1 CH2M HILL 2 Laguna County Sanitation District ABSTRACT The Laguna County Sanitation District (LCSD) Water Reclamation Plant, with a design capacity of 3.7 mgd, produces tertiary water that was designed to meet California Title 22 unrestricted re-use criteria. The disinfection in LCSD is provided by an in-line Aquionics Model 7500 medium pressure UV system containing four UV banks. California Department of Health and Services (CDHS) require that all UV systems be validated according to the National Water Research Institute and American Water Works Association Research Foundation UV Disinfection Guidelines (NWRI/AWWARF 2003), before distributing reclaimed water to the end users. A set of laboratory and field testing has been performed to demonstrate the efficacy of the UV equipment. The early test results were, however, not in total compliance with DHS requirements. In two of the five collimated beam tests, the majority of the data (more than 20 percent) did not fall within the "check curves" established by the NWRI/AWWARF guidelines. CDHS directed LCSD to perform two additional collimated beam and full-scale tests to fully satisfy NWRI/AWWARF requirements. With the few changes and adjustments in testing procedure, the UV disinfection system satisfied the requirements for DHS approval. This paper discusses key challenges and presents lesson learned during this validation testing. KEYWORDS Collimated beam tests, DHS, disinfection, reuse, UV, validation INTRODUCTION The Laguna County Sanitation District (LCSD) Water Reclamation Plant has a design capacity of 3.7 mgd which produces filtered-disinfected water to meet California Title 22 unrestricted re-use criteria summarized in Table 1. There are two treatment streams in the plant. One is the 3.2-mgd wastewater treated effluent, which undergoes primary clarification, biofiltration, secondary clarification, and microfiltration. The second flow (0.5 mgd) contains high total dissolved solids (TDS) that is generated as a result of home water softener use within LCSD. The high-tds flow is split from the influent during hours of high TDS concentration and is treated in a separate treatment train consisting of screening, membrane bioreactors (MBR), followed by reverse osmosis (RO) filtration. The effluent from each process train (low and high TDS flows) is combined and conveyed to the UV disinfection system as shown in Figure

2 TABLE 1- Title 22 Unrestricted Reuse Criteria Parameter Turbidity* Pathogens Value 0.2 NTU, 95% of the time within 24-hour period, No higher than 0.5 NTU any time A chlorine disinfection process following filtration that provides a CT (The product of combined chlorine residual and modal contact time measured at the same point) value of not less than 450 milligram- minutes per liter at all times with a modal contact time of at least 90 minutes, based on peak dry weather design flow; or Virus Total Coliform Minimum 5 log (99.999%) inactivation of MS-2 or Poliovirus 2.2 MPN/100, at 7-day period 23 MPN/100, in 30-day period Does not exceed 240 MPN/100 any time FIGURE 1- LCSD Water Reclamation Plant Process Flow Diagram Low TDS Equalization Basin Recycle Raw Sewage Head Works High TDS Flow (0.5 MGD) High TDS EQ Basin Primary Clarifiers (2) Sludge Sludge Biofilter Sludge Secondary Clarifier Membrane Bioreactor Microfiltration Pond A, B, C Separate Flow Equalization Microfiltration Reverse Osmosis UV Disinfection Title 22 Effluent RO Concentrate to Injection Well Anaerobic Digestor No. 1 Secondary Digestor Sludge Drying Beds Supernatant 4414

3 Disinfection of the tertiary treated wastewater is accomplished by the in-line Aquionics Model 7500 medium-pressure UV system that was installed at the LCSD WWTP with a capacity to treat 3.7-mgd of tertiary effluent. A total of 4 UV banks were installed. However, only three will be in operation at one time, with the fourth bank acting as standby. The validation testing on the full-scale, 3.7-mgd Aquionics UV Disinfection System was carried out to satisfy the NWRI/AWWARF guidelines and to confirm that the UV system disinfects the water and aids in producing recycled water suitable for irrigation and meets requirements for unrestricted public access pursuant to Section 60304(a) of the California Code of Regulations, Title 22, Water Recycling Criteria. TEST METHODS AND APPROACH The UV disinfection system as shown in Figure 2 was designed with the influent and performance parameters presented in Table 2. TABLE 2- UV System Design Configuration Parameter Value Design Flow Rate, mgd Transmittance, % Turbidity, NTU UV Dosage, mj/cm 2 Inactivation, log MS-2 removal 65 (minimum) Less than 0.2 NTU 95% of the time, not to exceed 0.5 NTU any time 80 (minimum) 4 (minimum) FIGURE 2. LCSD Aquionics Four-Bank UV System 4415

4 The UV system is designed to achieve a minimum of 4-log (not 5-log) virus removal, since 1-log virus removal credit is given for Zenon Zeeweed 500 Ultrafiltration system by CDHS. To provide a comprehensive validation program for the performance of the Aquionics UV disinfection system, the following tests were performed: Collimated beam tests (for quality assurance/quality control) UV reactor full-scale validation tests UV Lamp aging UV lamp heat tests Collimated Beam Tests Collimated beam tests were carried out under controlled conditions in the laboratory after each UV full-scale test. The wastewater non-disinfected effluent used in the collimated beam tests were the same as their corresponding full-scale tests. Collimated beam studies were performed to establish UV susceptibility of the microorganisms used in the UV reactor validation test. These tests serve as the standards against which the results obtained from full-scale UV tests were compared. The results from the collimated beam tests were used to correlate the microbial inactivation achieved by the full-scale UV unit to the inactivation achieved by an ideal UV system used in the collimated beam tests under standard laboratory conditions (NWRI/AWWARF, 2003). The collimated beam apparatus used in the tests is shown in Figure 3. FIGURE 3- Collimated Beam Test Apparatus Two Aquionics low-pressure UV lamps were used in the collimated beam apparatus. The system operated under 220 V electric supply. The UV lamps are non-ozone generating, 4416

5 germicidal lamps with a primary output of 254 nanometers (nm). These lamps emit 90 percent of its energy at 254 nm. The beam of UV light passes through a 20-inch-long and 2-inch-diameter collimated beam tube to the MS-2 plated wastewater samples, which are continuously stirred on a stir plate. The configuration of the UV lamps ensured a uniform UV intensity field at the surface of the beaker or the Petri dish that contains samples. The UV intensity was determined at the surface of the beaker with an International Light SED240 UV sensor, located in a parallel barrel. The International Light radiometer converts light intensity into units of power-time per area. Temperature and power supply for the UV lamp were monitored, and they remained constant during collimated beam testing. Within 24 hours of each full-scale test, a UV collimated beam test was performed under standard laboratory-controlled conditions defined in NWRI guidelines (2003). The collimated beam tests were performed using the same MS-2 seeded influent in the corresponding full-scale testing. The influent was exposed to known UV intensities at varying UV doses that were achieved by varying the duration of exposures. Sample size of 50 was exposed a series of different exposure times (0-, 1-, 5-, 10-, 15-, 20-, 25-, 35-, 45- and 60-minute exposures) and UV doses ranged from 0 to 120 millijoules per square centimeter (mj/cm 2 ) during each collimated beam test. The exposed sample was plated in triplicate at dilutions appropriate to give 20 to 200 plaque forming unit per plate (pfu/plate). To achieve accurate and reproducible results, the UV intensity was monitored by performing the collimated beam tests under constant conditions of lamp temperature and the voltage of the power supply. The UV intensity at the sample surface was quantified precisely by using a well-calibrated sensor. The UV dose applied to a sample was determined using the following equation. Where: I o t(1-e -kd ) D = f (1) kd D = UV dose measured in mj/cm 2 f = Correction factor (0.975) I o = UV intensity in milliwatts per square centimeter (mw/ cm 2 ) t = Exposure time (sec) k = UV transmittance at 254 nm d = Sample depth (cm) The log inactivation of the test microorganisms at the applied UV dose was determined and then plotted against the applied UV dose to provide a dose-response curve. The intensity probe readings before and after each test did not vary by more than 10 percent and the average of the two readings were used for calculating the UV dose. The collimatedbeam apparatus data were plotted to ensure that the data meet the conditions imposed by the two QA/QC equations (Equations 2 and 3) where at least 80 percent of the data should fall within these two lines in collimated beam tests. 4417

6 Where: -log 10 (N/N o )=0.040*[UV Dose, mj/cm 2 ]+0.64 (2) -log 10 (N/N o )=0.033*[UV Dose, mj/cm 2 ]+0.20 (3) N = Concentration of infective MS-2 after UV exposure No = Concentration of infective MS-2 at dose zero In this report, Check Curves 1 and 2 in graphical illustrations of collimated beam test results were developed using Equation 2 and 3, respectively. Full-Scale UV Validation Tests Full-scale UV validation tests were conducted to determine the efficacy of the Aquionics UV system for disinfection of the LCSD plant effluent to meet the effluent quality criteria (Title 22 criteria) for total and fecal coliform bacteria. Coliphage MS-2 was chosen as the bioassay organism for the UV validation tests. UV Disinfection Guidelines for Drinking Water and Water Reuse also recommended coliphage MS-2 to be used as a bioassay microorganism for validating UV systems for the disinfection of water and wastewater (NWRI/AWWARF, 2003). Five sets of full-scale UV Validation tests were performed. For each test, four flow rates between 1.5 and 3.7 mgd were tested to determine the performance of the system under different system conditions and to cover the design range of flows that could be encountered by the system. The flow rates were changed for about 5 to 10 minutes before collecting the pretreated and post-treated (effluent) samples. The first three banks of the UV system were operational during each test. With the exception of downtimes associated with planned or unplanned maintenance, the UV system was kept operational for 24 hours per day, 7 days per week to simulate actual operating conditions. The flow rates, UV transmittance, water temperature, power levels, UV sensor signals were monitored continuously and were recorded during all the testing. During full-scale testing, different flow conditions were tested to ensure that the treatment goals would be met under varying (actual operational) conditions. The power level was adjusted manually at Power Level 3 and kept constant throughout the validation tests. The transmittance of the wastewater was adjusted as close to 65 percent to meet UV transmittance value enforced by NWRI/AWWARF UV Disinfection Guidelines for UV Validation tests for reuse. Instant coffee was added to the wastewater to adjust UV transmittance of the wastewater. MS-2 and coffee seed were injected from the same location during the first five field testing. This period refers to the old test conditions. Water quality samples including MS-2, coliform and Title 22 quality parameters specified by DHS were collected before and after UV disinfection. The analyses were performed by a number of laboratories. CRG Marine Laboratories were responsible for the MS-2 phage testing, while the LCSD laboratory, and the San Bernardino laboratory located in Lompoc, CA performed the analysis for the coliform and the general chemistry parameters. 4418

7 UV Lamp Heat Tests To determine the heat emitted from the UV lamps tested, the temperatures of the pretreated effluent entering the UV unit and the post-treated effluent leaving the UV unit were measured and recorded during each set of tests. UV Lamp Aging UV lamps lose their output efficiency over time mainly due to solarization. The output they produce in the beginning of their lamp-life usually degrades over time. The lamp input power needs to be adjusted to compensate for the efficiency drop due to lamp aging, as well as factors such as quartz sleeve fouling and lower water UV transmittance. To compensate lamp aging factor, Aquionics UV system offers three power levels which can be adjusted either manually or automatically. The power level was maintained manually at Power Level 3 (highest) for the duration of each field test to impart the desired UV dose under average lamp age of 3,000 hours and low UV transmittance (65%) conditions. RESULTS Old Tests (June-July 2003) Collimated Beam Test Results A total of five collimated beam tests were performed in For each full-scale UV reactor validation test, one collimated beam test was performed using the same non-disinfected filtered effluent via ZeeWeed 500. The dates of tests and collection of effluent samples were June 11, 20, and 24, and July 2 and 8, For each collimated beam test run, a 0 to 90- minute exposure time was conducted to generate the dose-response curve of the system. With the exception of only one collimated beam test, no dose response data fell in QA/QC check curves. NWRI/AWWARF UV guidelines require at least 80% of dose response data between 20 and 150 mj/cm 2 UV dose to fall between the QA/QC check curves. The results, therefore, did not satisfy the minimum requirements outlined by the NWRI/AWWARF guidelines. After analyzing the data some errors were discovered in colimated beam tests. The erros were associated with the accuracy of the UV intensity meter, and the failure to warm up UV lamps prior to start of the test. Therefore, two additional tests were conducted on samples (preserved in the refrigerator) that were collected on June 20 and 24. A correction factor was developed to correct CBT results. Despite the applications of correction factor (2.22), the majority of the data points (80% or more) for the adjusted collimated beam Tests 3 and 4 did not fall in QA/QC check curves thereby not fully satisfying NWRI/AWWARF requirements. DHS, therefore, directed LCSD to perform two additional collimated beam and full-scale tests. The regression lines derived from the three collimated beam tests which satisfied NWRI/AWWARF requirements, were used to calculate the required UV doses to achieve 4- log inactivation of MS-2. The calculated UV doses for 4-log inactivation of MS-2 were 106, 88 and 105 mj/cm 2 for collimated beam Tests 1, 2 and 5, respectively. 4419

8 Full-Scale UV Reactor Validation Test Results Five full-scale UV tests were conducted on June 11, 20, and 24 and on July 2 and 8 of They were completed about a week apart to gauge the impact of water quality changes on the UV system performance. The flows were increased in increments of 50 percent or more, as required by NWRI guidelines, starting from the lowest flow encountered at the WWTP (1.5 mgd). Other flows tested were 2.25 mgd, 3.0 mgd, and 3.7 mgd. Three upstream banks were in operation during the validation testing. The power was set at power level 1 since the lamps were used about 200 hrs before first field test was performed. The results of the full-scale tests and the calculated (assigned) UV doses are summarized in Table 3. Since the collimated beam Tests 3 or 4 did not meet the NWRI/AWWARF requirements, the dose-response curves of these collimated beam tests were not used to calculate the UV doses for the full-scale Tests 3 and 4. The assigned UV doses for these tests were, therefore, marked as NA (not applicable) in Table 3. The calculated UV doses to achieve 4-log removals were 106, 88 and 105 mj/cm 2 for Field Tests 1, 2 and 5, respectively. In other words, the required UV dose for 4-log MS-2 inactivation was in the range between 88 and 106 mj/cm 2. It is expected that the Zeewed 500 ultrafiltration system effectively capture particles and coliform bacteria which are larger than membrane pore size. However, the presence of total and fecal coliform (not shown in Table 4) in UV influent indicated that Zeeweed membrane had leaks and/or breaches during UV field testing. The UV system, however, successfully reduced the total and fecal coliform to non-detactable levels (<2 MPN/100 ). Table 3- Full-Scale Test Results (Old Tests, June July 2003) T UV Dose Field Controller Flow Temperature (C) % mj/cm2 GPM Pre-UV Post-UV Avg Influent MS-2 PFU/100 Avg Effluent MS-2 PFU/100 MS-2 Log Reduction Influent Total Coliform MPN/100 Effluent Total Coliform MPN/100 UV Dose Calculated mj/cm2 Test 1 - June 11, , ND ND , ND ND , ND ND , ND ND Test 2 - June 20, , ,600 < , ,600 < , ,600 < , ,600 < Test 3-June 24, ,076, ,000 < 2 NA , ,000 < 2 NA , ,000 < 2 NA , ,000 < 2 NA Test 4-July 2, , ,000 < 2 NA , ,000 < 2 NA 4420

9 , ,000 < 2 NA ,560, ,000 < 2 NA Test 5, July 8, , ,000 < , ,000 < ,116, ,000 < ,553, ,000 < New Tests (February 2005) Collimated Beam Test Results The following changes to the previous collimated beam test and field test procedures have been made to improve the accuracy of the test results. These improvements were: MS-2 stock solutions and MS-2 pumping rates were carefully adjusted to maintain at least 10 7 PFU/ML in the beginning of each test. MS-2 stock solution was prepared more concentrated to maintain at least 10 7 PFU/ML in UV influent in the beginning of each test Following completion of each UV exposure, a fresh sample was poured into the test beaker to continue the test at another UV exposure. This approach maintained the sample depth at 0.9 cm and eliminated errors associated with imprecise depth measurements. MS-2 and instant coffee were fed separately in each new field testing (Figure 4). It allowed independent adjustment of MS-2 and instant coffee feed rates and the resultant concentrations. Power level was set at 3 to ensure 4-log MS-2 inactivation. FIGURE 4. Schematic of UV System Layout with Injection Ports (Old tests left picture, new tests right Picture) Microfiltration Membranes RO 1 RO 2 Microfiltration Membranes RO 1 RO 2 MS-2 and COFFEE INJECTION PORT RO 3 RO 4 MS-2 INJECTION PORT COFFEE INJECTION PORT RO 3 RO 4 UV System UV System Total length of 12 -inch pipe = 165 ft Total length of 16-inch pipe = 185 ft Total length of 12-inch pipe = 165 ft Total length of 16-inch pipe = 185 ft 4421

10 The new collimated beam tests fully satisfied NWRI/AWWARF UV Disinfection Guidelines for UV Validation tests for reuse as shown in Figures 5 and 6. FIGURE 5- The Relationship between UV Dose and MS-2 Inactivation Collimated Beam Test #1 (February 9) MS-2 Log Inactivation y = x R 2 = Collimated Beam Test #1 Check Curve #1 Check Curve #2 Linear (Collimated Beam Test #1) UV Dose, mj/cm2 FIGURE 6- The Relationship between UV Dose and MS-2 Inactivation Collimated Beam Test #2 (February 16) MS-2 Log Inactivation y = x R 2 = Collimated Beam Test #2 Check Curve #1 Check Curve #2 Linear (Collimated Beam Test #2) UV Dose, mj/cm2 4422

11 Full-Scale Test Results (New Testing) The UV system demonstrated a minimum of 4-log MS-2 removal in 5 out of 8 full-scale testing. This was achieved with the UV assigned dosages between 98 and 110 mj/cm 2 in the and new tests (Figures 5 and 6), respectively. Table 5- Full-Scale Test Results for the New Tests (February 2005) T UV Dose Field Controller Flow Temperature (C) Pre- % mj/cm2 GPM UV Post-UV Avg Influent MS-2 PFU/100 Avg Effluent MS-2 PFU/100 MS-2 Log Reduction Influent Total Coliform MPN/100 Effluent Total Coliform MPN/100 UV Dose Calculated mj/cm2 Test 1 February 9, E ,500 ND E ,200 ND E+07 1, ,000 ND E+07 1, ,500 ND 123 Test 2 February 16, E+07 4, ND E+07 8, ,000 ND E+08 19, ,000 ND E+08 10, ,000 ND 102 This UV dose range is closer to the UV dose that is predicted by the QA/QC lower check curve (check curve #2). Therefore, UV system should be operated close to conservative end to safely meet microbiological water quality criteria of Title 22 reuse regulations. Despite the presence of appreciable amount of coliform in UV influent in all sets of full-scale testing, the total coliform in the disinfected effluent was less then 2.2 MPN/100, the limit set for unrestricted water reuse under the Title 22 requirements. Lamp Heating Tests During the operation of the UV units, the water temperature could rise due to the heat emitted from the lamps. To determine the heat emitted from the UV lamps tested, the temperatures of the pretreated effluent entering the UV unit and the post-treated effluent leaving the UV unit were measured and recorded during each set of tests. As can be seen from table 4, no significant variation of temperature was observed for the pre-uv and post- UV samples taken during the testing. 4423

12 CONCLUSIONS Feeding MS-2 and instant coffee separately allowed independent adjustment of MS-2 and instant coffee feed rates and the resultant concentrations. MS-2 concentration must be at least 10 6 PFU/100 in UV influent to make accurate estimation of UV reactor performance Majority of the collimated beam and full-scale UV validation tests demonstrated 4-log inactivation of MS-2 phage. This was achieved with the UV assigned dosages between 88 and 106 mj/cm 2 and 98 and 110 mj/cm 2 in the old and new tests, respectively. The total and fecal coliform in UV disinfected effluent has consistently met the most probable number (MPN) of less than 2.2 MPN/100 in each field test and daily plant operation satisfying Title 22 reuse water requirements. Neither short-circuting nor back mixing is evident in UV reactor. If it was evident, total and fecal coliform concentrations could be detectable and vary in plant effluent. Inconsitent MS-2 removals in field tests indicate that the UV reactor is not optimized. This is apperant since MS-2 inactivation results were not impacted by varying flowrates and UV transmittance levels. The testing results obtained from the old and new conditions did not fit any of the MLR equations that were developed to estimate UV doses as a function of flowrate, transmittance and power level. Therefore, the use of MLR models to estimate UV doses is not recommended. The calculated UV doses for 4-log MS-2 inactivation were in the range of between 88 and 110 mj/cm 2, with majority of the calculated dose values falling between 100 and 110 mj/cm 2. This UV dose range is closer to the UV dose that is predicted by the QA/QC lower check curve (check curve #2). Therefore, UV system should be operated close to conservative end to safely meet microbiological water quality criteria of Title 22 reuse regulations. UV dose controller was not accurate. The controller over estimated the UV dose by a factor of 1.5 to 2 in new tests and 2 to 4 in old tests. Calibration and accuracy of the online equipment is very essential to assess the UV system performance. Each equipment and instrumentation associated with UV system should be calibrated at least two to three times a year. Finally, the following operating conditions summarized in Table 5 were suggested for operation of the UV system. These conditions were approved by DHS and LCSD started distribution of the recycle water to the end users. 4424

13 TABLE 5- Suggested Operating Conditions for the UV System Operating Parameter Value Action No. of UV banks 3 Will turn on fourth bank if conditions warrant (see below for these special conditions) Flow Rate, gpm 2,500 (maximum) Adjust ZeeWeed to lower flow rate if flows through UV are higher than stated value (with three UV banks running) UV dosage, mj/cm (minimum) If below 100 mj/cm 2 with three banks, divert all flows to the effluent storage ponds located west of the plant and notify end users on reuse water supply quality problem Transmittance, % 65 (minimum) Divert all flows to effluent storage ponds if the transmittance falls below 65% ACKNOWLEDGEMENTS The authors would like to thank Moy Yahya, microbiology laboratory manager at CRG Labs, for his endless help during CBT and field testing. REFERENCES Ultraviolet Disinfection (2003). Guidelines for Drinking Water and Water Reuse 2 nd Edition National Water Research Institute and American Water Works Association Research Foundation. Zar, J. H Biostatistical Analysis. 4th Edition. Prentice-Hall, Inc., Upper Saddle River, NJ. 4425