MIXING UP THE TWINS ACTIVE MIXING AND OTHER WATER QUALITY CONSIDERATIONS FOR STORAGE TANKS

Transcription

1 MIXING UP THE TWINS ACTIVE MIXING AND OTHER WATER QUALITY CONSIDERATIONS FOR STORAGE TANKS Ray Ihlenburg, PE, Senior Technical Director Thomas Waters, EIT, Staff Engineer O Brien & Gere Louisville, Kentucky John Thomas, Superintendent Lebanon Water Works Company Lebanon, Kentucky ABSTRACT The Lebanon Water Works Company (LWWC) serves approximately 8,400 customers in central Kentucky, providing retail service to the City of Lebanon and wholesale water service to the surrounding Marion County via multiple interconnections. As with many water utilities across the United States, the LWWC is faced with the challenge of mapping the optimum course for producing quality water while complying with the Stage 2 Disinfectants / Disinfection By-Products Rule (Stage 2 DBPR). The LWWC is one of only two water utilities in Kentucky that elected to use a computer hydraulic model of their distribution system to develop their Stage 2 DBPR System Specific Study (SSS) plan and Initial Distribution System Evaluation (IDSE) report. The modeling performed for Stage 2 confirmed the potential for very old water in the LWWC s twin 0.94 million gallon (MG) Calvary Road ground-level storage tanks. This water is suspected of contributing to high Total Trihalomethane (TTHM) concentrations. These tanks equally fill and empty through a common main and the tank contents only mix during refill while the main high-service pumps at the water treatment plant operate. Consequently, there are long periods during the day while emptying, and especially at night, when no mixing energy is available because the water plant and high service pumping are not in operation. Hydraulic modeling (MWH Soft s InfoWater) indicated continuous mixing was required during the periods when the high-service pumps were OFF. A mixing system that would operate continuously regardless of the water level in the tanks was preferred. The selected mixing alternative was the PAX Water Technologies active mixing system, which is the first of its kind installed in Kentucky and has been in successful operation since November

2 One mixer was installed in each tank. To assess the effectiveness of the mixers, six thermal sensors / recorders were positioned at 10-ft vertical intervals to monitor for temperature stratification at 30-minute intervals. The recorded temperature data was combined with real-time SCADA level recordings to present a concise picture of the effects of mixing and water temperature in the twin storage tanks. In addition, data obtained from turning one tank mixer off was analyzed to determine the rate at which that tank stratified and then obtained a complete mix upon re-starting the mixer. This was also compared to the other tank with its mixer in continuous operation. 1.0 INTRODUCTION Lebanon, Kentucky is the County Seat for Marion County located in central Kentucky. The Lebanon Water Works Company (LWWC) serves approximately 8,400 customers in the City of Lebanon and the Marion County Water District (MCWD). The water treatment plant is a typical surface water treatment plant with a capacity of 5.2 million gallons per day (MGD). The normal raw water source is the North Fork Rolling Fork River with approximately 300 square miles of drainage area at the raw water intake. The secondary source is the Fagan Branch Reservoir, a pumped storage reservoir with a capacity of 1000 MG. The river water quality is typically very good, however, there is a high TOC potential due to the agricultural and rural nature of the drainage area. Presently, the plant produces an average of 2.6 MGD. The distribution system consists of approximately 65 miles of pipe, the twin 0.94 MG Calvary Road stand pipe tanks (CRTs), the Warehouse Road Booster Pump Station that pumps to the 0.25 MG Springfield Road Elevated Storage Tank (SRT). The SRT forms a pressure zone that is 50 feet above the balance of the system. It is separated by two (2) pressure reducing valves (PRV) that will fully open to provide fire flow back into the lower pressure zone. See Figure 1 for a distribution system schematic. As is the case with many water utilities, competing demands on the system must be met at any given time, and their effects often must be considered together. Concurrent to distribution system evaluations, which were directed at improving water quality in support of Stage 2 Disinfectants and Disinfection Byproducts Rule (Stage 2 DBPR) compliance, the LWWC had to move quickly to make additional system improvements when approached to provide a relatively large daily amount of water to a potential industrial customer. Increased fire protection at the proposed industrial site was also considered. The designed improvements were a new 16-inch water main constructed to serve the proposed industrial park and new development along a newly constructed bypass road around the west side of Lebanon, as well as system modifications intended to increase the capability of the CRTs to provide fire protection while also improving water quality to meet Stage 2 DBPR requirements. Therefore, consideration was taken to dually address these requirements for LWWC s water distribution system. 2

3 3

4 2.0 COMPLIANCE REQUIREMENTS FOR STAGE 2 DBPR The Stage 2 Disinfectants / Disinfection Byproducts Rule was signed by the U.S. Environmental Protection Agency on December 15, 2005 and was published in the Federal Register on January 4, 2006 (71 FR 288) and requires utilities to conduct an Initial Distribution System Evaluation (IDSE) to build on the Stage 1 DBPR by focusing on monitoring for and reducing concentrations of two classes of DBPs total trihalomethanes (TTHM) and five types of haloacetic acids (HAA5) in drinking water (EPA, June 2006). The Stage 2 DBPR has set Maximum Contaminant Levels (MCL) for TTHM and HAA5 to 80 µg/l and 60 µg/l respectively. Maximum Contaminant Level Goals (MCLG) are 60 µg/l for TTHM and 48 µg/l for HAA5. Individual MCLGs have also been established for specific TTHM and HAA5 compounds. Concentrations are determined using a Locational Running Annual Average (LRAA), which is the average of the most recent four quarterly sampling results at each location. Every location must be in compliance. Several options are available to utilities for conducting their IDSE, depending on the availability of TTHM and HAA5 monitoring data and corresponding historical concentrations, and system size. According to the requirements, systems that have ample available TTHM and HAA5 sampling data or choose to develop a hydraulic model can opt to conduct a System Specific Study (SSS) as a means of selecting Stage 2 DBPR compliance monitoring locations. According to the Kentucky Division of Water (KDOW), the Lebanon Water Works Company is one of only two water utilities in the state of Kentucky that has conducted a SSS by means of developing a hydraulic model. The model was developed using MWH Soft Inc. s InfoWater software and was calibrated to fifteen fire-flow tests (static) as well as extended period simulations for pumping and tank level data (time-varied). A Modeling Study Plan was prepared and submitted to EPA in September 2007 to comply with the October 1, 2007 deadline. The Plan identified SSS monitoring locations based on modeling results for high TTHM and HAA5 formation potential, and average water age in the system. LWWC conducted preliminary monitoring at these sites in July 2007 and 2008, since July was determined to be the peak month for DBP formation potential in the system. Laboratory testing on these preliminary samples confirmed that the selected SSS monitoring locations are trouble spots for DBP concentrations. See Figures 2 and 3 at right. TTHM concentrations in water distribution systems are directly proportional to residence time and water temperature. HAA5 concentrations usually peak at about 24 to 48 hours of residence time, and then begin to degrade over time (US Dept of Interior, 2003). For this reason, targeted mitigation strategies for these two regulated DBP classes can differ slightly. This paper focuses on strategies better suited to mitigate TTHMs by reducing residence times and water temperatures, rather than HAA5s, for which mitigation strategies are usually more directed at treatment processes. 4

5 16-inch AWWA DISTRIBUTION SYSTEMS SYMPOSIUM 2010, NATIONAL HARBOR, MD LWWC s IDSE Reports were due to EPA by January 1, 2010 and summarized methodology for selecting the SSS monitoring locations, reported annual SSS and Stage 1 DBPR compliance monitoring results with LRAAs, and proposed Stage 2 DBPR compliance monitoring sites and selection methodology. LWWC must submit their Stage 2 DBPR Compliance Monitoring Plan by October 1, 2013 following approval of their IDSE Report if additional information is required by regulatory agencies, and begin complying with the monitoring requirements. By July 2014, LWWC must begin complying with the Stage 2 DBPR requirements to determine compliance with the operational levels for TTHMs and HAA5s (EPA, June 2006). 3.0 HYDRAULIC MODELING OF ALTERNATIVES The LWWC s water distribution system consists of two pressure zones, one high and one low. High-service pumps deliver water to the system from the water treatment plant to the lower pressure zone. These highservice pumps run from 6 AM to 12 AM midnight, and simultaneously maintain water levels in the two 0.94 MG CRTs that are the main storage facility for the water distribution system. A booster pumping station delivers water from the low zone to the high zone, and is operated according to levels in the high zone elevated SRT. When the high-service pumps shut OFF at night, the CRTs act as the source of water for the system. At this time, the booster station draws water from these tanks to feed the SRT and high zone customers. A comprehensive water system hydraulic model was developed, using MWH Soft s InfoWater software, as part of the LWWCs Modeling Study Plan in the SSS. The model was developed primarily from GIS and water meter data, and was calibrated to fifteen fire-flow tests (static) as well as time-varied tank and pumping levels in extended period simulations (EPS). Modeling runs were intended to address the requirements of the IDSE and involved modeling water age at various locations in the system to determine areas of the system with increased DBP formation potential for monitoring. Modeling confirmed that the CRTs are bypassed while the high-service pumps are ON. This is also known as a side-storage system because the tanks use a single 16-inch main for filling and draining. See Figure 4. When the high-service pumps shut OFF, only 10-20% of the stored water is drained to meet overnight system demands. Thus the CRTs fill/drain fluctuation follows a last-in, first-out (LIFO) pattern, where only the bottom 10-20% is utilized, leading to stratification of the tank, especially during warmer months. This leads to a long residence time problem, which creates an increased potential for formation of TTHMs. Since the CRTs service the system when the high-service pumps are off, modeling showed that the system experiences a slug of high water age water flowing from the tanks and moves though the distribution system throughout the day, once the high-service pumps turn back on at 6 AM. This results in certain areas that receive water from the CRTs, having high TTHM formation potential. West CRT East CRT 1,200 LF 16-inch Figure 4. Twin Calvary Road Side-Storage Tank Configuration Since TTHM formation rate and extent increases with water age and temperature, an obvious mitigation strategy would be to reduce residence time and thermal stratification in the CRTs by mixing. Several alternatives were evaluated and modeled, and are presented below. 5

6 3.1 Baseline To evaluate the efficacy of various mitigation strategies, it was first necessary to determine the baseline operation and resulting water quality in the CRTs. Since filling and draining of the tanks occurs through a common 16-inch main, it was assumed that mixing in the tanks follows a LIFO pattern such that water last in to the tank is the first out of the tank. Modeling conducted under this mixing pattern for the CRTs and under normal high service pumping operation is presented in Figure 5. When simulating water age in the hydraulic model, it is necessary to allow the modeled system enough time to reach equilibrium, such that tanks are responding in a repeating pattern to the input pumping and demand patterns. Equilibrium is typically said to be achieved when plots of water age versus simulation time level out and a repeated pattern can be observed. In the case of the baseline LIFO condition, Figure 5 shows that the system did not reach equilibrium at the end of the simulation period indicating that actual water age in the CRTs could exceed 600 hours (25 days). To limit run model times however, the simulation duration was limited to 600 hours. Figure 5. Water age vs. time: baseline (existing LIFO) conditions 3.2 Industrial Impact CRT Booster Pumping Station Concurrent to planning and modeling conducted to support the IDSE, analyses were also conducted to accommodate a new potential industrial customer. This industrial customer had a process demand of 1,400 gallons per minute (gpm) and a fire flow demand of 2,200 gpm at 70 psi residual pressure. Modeling indicated that LWWC was able to meet the process demand of 1,400 gpm (at 50 psi) with the addition of a new 16-inch water main to the proposed industrial park location, but fire protection would require a booster pump to be installed and maintained by the industrial customer to meet the minimum pressure of 70 psi. Having identified a need for DBP mitigation strategies at the CRTs, LWWC considered adding a booster pump to the CRT system that 6

7 16-inch AWWA DISTRIBUTION SYSTEMS SYMPOSIUM 2010, NATIONAL HARBOR, MD would provide better turnover of the tank during normal operations and make it possible to utilize the entire CRT capacity for fire protection. See Figure 6 for a schematic. The valve on the 16-inch fill/drain pipe was proposed to be an electrically controlled, flow control-type valve such that it is open during filling and closed during pumping. West CRT East CRT PS 16-inch Figure 6. CRT Booster Pumping Station Schematic Table 1 summarizes advantages and disadvantages of constructing a booster pumping station at the CRTs to mitigate water quality issues and provide adequate fire flow pressures to the potential industrial customer. Table 1. CRT Booster Station Pros versus Cons PROS Uses more storage volume for fire protection Increased system operating pressures Allows HSPS to shut down or cycle throughout the day May allow HSPS to pump more water with existing motors Ensures turnover of tank contents CONS Higher capital cost than piping modifications Higher cost of power to re-pump water Additional maintenance Increased SCADA and controls complexity Potential need for emergency generator Reduces importance of emergency generator at WTP The new 16-inch water main was under construction when the industry unexpectedly decided not to expand to Lebanon. It was decided that the CRT booster station option was not necessary at this juncture, as increased fire protection was not needed. Other DBP mitigation strategies, such as mixing, modification of piping to and from the CRTs, and high service pumping operational changes were explored and the booster station shelved for future implementation. 3.3 Adding Mixing to the CRTs Several alternatives were evaluated to add a mixing regime to the CRTs. The intent of adding mixing is to reduce thermal stratification. This maintains chlorine residuals and reduces disinfection by-product s formation potential by reducing temperatures in the upper reaches of the tank by blending with newer, cooler water from the plant. Both passive and active mixing strategies were considered. To quantify their effect in the tanks, all mixing strategies were assumed to result in a completely mixed water volume for modeling purposes. Figure 7 depicts modeling results from changing the CRT mixing designation from LIFO to completely mixed (CM) under 7

8 the same pumping operation (blue plotted line). Hydraulic modeling indicated that adding a mixing mechanism to the tanks would reduce water age from greater than 600 hours to approximately 175 hours. This was a significant improvement in predicted water quality, therefore mixing was pursued as the best alternative. The several types of passive and active mixing strategies evaluated for LWWC are summarized later in this paper. Figure 7. Water age vs. time: baseline (LIFO) and complete mixing comparison 3.4 High-Service Pump Station Operation Operation of the high-service pump station (HSPS) at the WTP also plays a role in the quality of water being stored in the CRTs. The HSPS currently operates daily between the hours of 6 AM and 12AM. When the pumps are off during the night, the CRTs act as the source of water for the system. Overnight demands however, are only enough to drain the tank between 10% and 20%. The tanks then refill throughout the day as the HSPS is operated. As mentioned previously, this generates long residence times in the tanks. Modeling was conducted to determine optimized HSPS controls to reduce residence times in the CRTs. Operating the pumps based on levels in the CRT, rather than time of day, resulted in marked improvements in water age in the tanks. Figure 8 depicts modeling results for two scenarios: 10 feet of drawdown (green plotted line) and 15 feet of drawdown (red plotted line). The modeled controls were such that when water levels in the CRTs dropped to the specified cutoff, the HSPS would turn ON. When the tanks reached their overflow, the HSPS would turn OFF. These operational scenarios maintained system pressures throughout the day while improving water quality in the CRTs to 70 hours/3 days (10 foot operating range) and 60 hours/2.5 days (15 foot operating range), when coupled with complete mixing. Typical recommendations for storage turnover are 1-5 days. Water quality improvements observed in the modeling are due to the combination of complete mixing and increased turnover in the tanks attributable to the increased drawdown frequency during periods of higher demands (during the day). This demonstrates a significant improvement from the baseline conditions. 8

9 Figure 8. Water age vs. time: baseline (LIFO), complete mixing, and complete mixing with modified HSPS operation comparison Table 2 summarizes the advantages and disadvantages associated with changing high-service pump controls at the WTP to improve turnover, which improves water quality in the CRTs. Table 2.HSPS Controls Modification Pros versus Cons PROS Ensures increased turnover of tank contents Allows HSPS to shut down or cycle throughout the day CONS Increased SCADA and controls complexity Changes to operator s daily routines, added training 4.0 EVALUATION OF MIXING SYSTEMS Hydraulic modeling indicated that adding a mixing system to the CRTs was a necessity to begin improving water quality in the tanks and as a first step in making system changes to comply with the Stage 2 DBPR. Both passive and active systems were considered. A passive mixing system relies on the turbulence created by the tank inlet velocity to mix the water, with no additional energy added. Examples of passive mixing systems in storage tanks are duckbill or flap valves, and separate inlet/outlet pipes. An active system is one that directly imparts mechanical energy to the water for mixing, such as a mixer impeller. In each of the approaches considered, the goal was to improve the tanks water quality (reduce DBPs) by reducing thermal stratification, optimizing the tanks turnover rate, and improving the chlorine residual. 9

10 The authors developed the following criteria for evaluating mixing system options for LWWC: Mix tank contents 24/7? Mixing Ability Eliminate thermal stratification? Maximize chlorine residual? Access and lay-down area. Is a crane required to install system in standpipes? Do the tanks have to be drained for installation? Installation & Operational Costs If so, can we coordinate installation with routine maintenance (e.g. painting) to reduce down-time for each tank? Cost of equipment Energy requirements to run system and corresponding cost of power Operational and maintenance requirements / costs Suitability for LWWC s System Can it be adapted to the system without major alterations to facilities? Does it require any major changes in operation and maintenance structure or schedules? Consideration was not only given to the ability of the mixer to improve water quality in the CRTs, but also installation and operational criteria, and how well the mixing system adapted to LWWC s existing distribution system and operation. Options to coordinate mixer installation with scheduled routine tank maintenance (i.e. painting) were also considered to reduce down-time for each tank. Through the evaluation process, it was determined that three classes of mixing systems existed on the market at the time of the study: submersible active mixers, floating draft tube active mixers, and duckbill/flap valve or nozzle-type passive mixers. Table 3 provides a comparison of these three mixing system categories. Each type of mixing system is further discussed below. 10

11 Table 3. Mixing Systems Comparison CRITERIA COMPARISON OF MIXING SYSTEMS SUBMERSIBLE, ACTIVE MIXERS FLOATING, DRAFT TUBE ACTIVE MIXER Applicable to thermally stratified tanks? Yes Yes Applicable to low turnover tanks and standpipes? Good for freeze protection? Engineering design time required per tank Number of hours of mixing per day (Grid Power) DUCKBILL/FLAP VALVE OR NOZZLE PASSIVE MIXERS Yes - If enough tank turnover each day Yes Yes No Yes Perhaps mixing not vigorous enough Yes if enough tank turnover each day Minimal Minimal Extensive 24 hours 24 hours Typically 3-6, depending on fill cycle Solar Option Cost $4,000-$7,000 $8,000-$15,000 N/A Crane onsite for installation No Yes Possibly Installation Cost $2,000-$8,000 $10,000-$16,000 Typically $20,000 + depending on tank size Time required to install 8 hours 24 hours 4-10 days Approximate equipment size 65 lbs. 4 ft. tall lbs ft. tall 200+ lbs ft. tall RPM of device N/A Source: Adapted from PAX Water Technologies data Figure 9. Duckbill valve in storage 4.1 Duckbill/Flap Valve or Nozzle-Type Passive Mixers for the CRTs tank Duckbill/flap valve or nozzle-type passive mixers rely on hydraulic energy of turbulent flow velocity through small openings to mix the tank contents. The valve or nozzle, typically manufactured from some type of rubber or elastomer, is installed at the end of the tank inlet pipe and maintains a flattened duck bill shape until water flows into the tank inlet pipe when it opens slightly to allow the flow to enter the tank at a high velocity. See Figure 9, at right, for a photograph of this type of system in a storage tank. Inlet mixing nozzles were considered based on their many successful installations. When typical LWWC system operation is considered however, it was determined that there is not sufficient flow for high-velocity mixing into the CRTs from the HSPS during the day. As mentioned previously, this is because the CRTs are essentially bypassed when the high-service pumps are on. Modeling indicated that the flow rate to the CRTs during the day was in the range of 500 gpm, or 250 gpm per tank, due to LWWC s typical daily demand patterns and the tanks side-storage configuration. This was not enough to provide suitable mixing velocities. In addition, the inlet mixing nozzle approach does not provide mixing to the water in the CRTs when the HSPS is off (overnight) or if a large 11

12 drawdown condition occurs. Since the tanks refill quickly in the morning when the HSPS comes on and are then bypassed in side storage as flow from the HSPS serves daily demands, little to no flow reaches the tanks during the day when temperatures are highest. Therefore even though the HSPS may be in operation, the CRTs may only experience mixing during the brief refilling period in the morning hours (which are less critical to stratification). Unused and unmixed water can sit in the tanks during the hottest times of day, which leads to thermal stratification and poor water quality. Therefore, a system that mixes tank contents 24 hours a day was preferred. Table 4 summarizes advantages and disadvantages of inlet passive mixing systems with respect to the CRTs and LWWC s system. Table 4. Inlet Nozzle Passive Mixing Systems Pros versus Cons PROS CONS No additional power requirements Installation through existing tank openings No maintenance May not damage existing coatings if self-supported inside the tanks. Supports other system changes such as booster station or parallel main. Requires pipe modification inside each tank. Provides mixing only when HSPS is ON. Not ideal for tanks with low turnover rates (infrequent mixing) Increases potential need for emergency generator at WTP Not effective if water level falls below nozzles. 4.2 Floating Draft Tube Active Mixers for the CRTs Floating draft tube mixers float on the water surface and draw water in through an adjustable draft tube suspended from the underside of the floating mixer. Water exits horizontally out the top of the mixer after passing through an impeller, powered either by the electrical grid or solar panels. Figure 10, at right, depicts a popular floating draft tube-type mixing system. Floating mixers have been utilized successfully in other tanks by the author s company. However, during the comparison of the various mixing options, the client and engineer decided not to pursue this option due to concerns about fluctuating water level in the CRTs, the potential for icing if the mixer turned off (e.g. loss of power), power cabling, anchoring of the mixer, intake draft tube, and size of CRTs existing access openings. Table 5 summarizes advantages and disadvantages of floating draft tube active mixers identified with respect to the CRTs and LWWC s system. Figure 10. Floating Draft Tube Mixing System (Courtesy of SolarBee, Inc.) 12

13 Table 5. Floating Draft Tube Active Mixing Systems Pros versus Cons PROS CONS Mixes CRT water 24 hours a day, independent of HSPS status No pipe modifications Can use solar or grid power for low voltage operation Mixes water regardless of level. Supports other system changes such as booster station or parallel main. Requires power cable out existing hatch or through wall Requires minimum of 24 x 24 opening for installation Requires anchoring or tethering in tank that will allow for variation in water level Concerned about ice formation Utilizes intake hose to draw from bottom and adjust for water level changes 4.3 Submersible Active Mixers the selected system Submersible active mixers are fixed at the bottom of the storage tank and utilize a small motorized impeller to mix water in the tank. The impeller motor draws power either from the electrical grid or solar panels. See Figure 11, at right, for a photograph of one of the submersible active mixers installed in the CRTs for the LWWC. Submersible active mixers had not been used previously by the author s company nor had any been installed in Kentucky as of November However, contact with present users provided the confidence to strongly consider them for this application. Table 6 summarizes advantages and disadvantages of the submersible active mixing systems with respect to the CRTs and LWWC s system. Figure 11. Submersible mixer in storage tank Table 6. Submersible Active Mixing Systems Pros versus Cons PROS Mixes CRT water 24 hours a day, independent of HSPS status CONS Requires power cable out existing hatch or through wall No pipe modifications Can use solar or grid power for low voltage operation Mixes water regardless of level Supports other system changes such as booster station or parallel main Lightweight, tripod-mounted mixer folds to fit through existing access openings 13

14 Ultimately, a PAX Water Technologies active water mixer was selected for the CRTs and installed in November This submersible active mixing system selected for the CRTs included tripod-mounted, motor-driven impellers in each tank, powered from the electrical grid. The impeller motors draw 15 amps with a recommended service life of about 5 years. The 24-volt power supply cable from the control panel to the impeller motor was installed through the tank wall. The installation required a compression fitting to seal the wall penetration and minor welding around the power cable. The power cord and mixer tripod were installed to rest on the bottom of the tank, with the tripod fixed to the tank bottom by magnetic foot pads, which protect the bottom of the tank and anchor the unit in place. Since a wall penetration was made to route power supply cables to the mixer, the tanks had to be dewatered for installation. Therefore installation of the mixers was coordinated with routine tank maintenance (i.e. interior and exterior painting) to minimize down-times for each tank. The following series of figures (12-14) present installation details of Figure 12. Mixer impeller the submersible mixers described above. Figure 13. Tank wall penetration detail Figure 14. Mixer motor control panel 5.0 MIXER EFFICACY To determine the effect the mixers had on stratification in the CRTs and verify the complete mixing assumption made during modeling, temperature probes were installed at 10 foot intervals on a stainless steel cable suspended from the tank access hatch. The probes measured real-time water temperatures at each level at 30 minute intervals over the course of summer Since the CRTs operate as one storage system (i.e. same volume, elevations, and fill/drain patterns), this presented a good opportunity to compare mixed versus unmixed tanks of the same design and operating characteristics. For the months of June and July 2010, the mixer in the East CRT was turned off while the mixer in the West CRT continued to run, with the intent of observing a marked difference in thermal stratification between the tanks. The East CRT mixer was then turned on to observe the amount of time it took for the mixer to de-stratify and blend the tank s contents. The following figures present the temperature data obtained for this study. It is important to note that raw water temperatures observed at the WTP during this study reached into the upper 80 s F. There is potential for greater temperature distribution (wider range of stratification) in the CRTs for conditions where there is greater difference between raw water temperature and ambient air temperature. 14

15 Figure 16. West CRT temperature probe data Figure 15. East CRT temperature probe data AWWA DISTRIBUTION SYSTEMS SYMPOSIUM 2010, NATIONAL HARBOR, MD 15

16 As seen in Figures 15 and 16, temperature probe results followed the authors expectations. Figure 15 shows marked thermal stratification in the East CRT when the mixer is off. An obvious convergence of temperature from each probe elevation can be seen when the East CRT mixer is activated in mid-afternoon July 20 th. The West CRT was the control for the duration of the study (with its mixer on), and exhibited relatively uniform temperature distribution over the height of the tank. This demonstrates a clear improvement to thermal stratification in the tanks. Figure 17 focuses on the time surrounding mixer activation in the East CRT on July 20. After activating the mixer in the East CRT, Figure 17, demonstrates that the mixer achieved completely-mixed, de-stratified conditions in about 8 hours. According to PAX Water Technologies, complete mixing can take up to 24 hours, depending on the size of the tank. Figure 17. East CRT time to achieve complete mixing 6.0 LESSONS CONFIRMED OR LEARNED Planning, design, and installation of the submersible active mixers in the CRTs brought to light several lessons learned and theories confirmed for the authors, and are summarized as follows: It is typically costly to revise tank piping when the initial piping configuration design is LIFO or side-storage Mixing will reduce thermal stratification and there are several methods available. Time should be spent deciding which method accomplishes the desired water quality improvements, but also best fits the storage and distribution system from an installation, operation, and maintenance perspective. 16

17 Installing a tank mixing system is not the only solution for mitigating DBP formation potential in water distribution systems. Utilities should take a comprehensive approach in planning and design for DBP and other water quality improvements for their distribution systems. Install tank wall penetration compression fittings for the mixer power supply cables to be accessible from the outside. This allows for adjustment after installation, should any leakage occur. Have a means of monitoring and recording water temperature throughout the tanks to verify reductions in stratification. Call it commissioning the mixing system. 7.0 CONCLUSIONS Several types of mixing systems were evaluated for the Lebanon Water Works Company s twin Calvary Road Tanks. A collaborative approach was taken by the authors to evaluate systems that address water quality, operation and maintenance, and adaptability to the system. The selected submersible active mixers have demonstrated marked improvements in the CRTs by reducing thermal stratification over a short period of time. The CRTs were good subjects to compare the effect of mixing in tanks of the same design. These tanks have an identical fill/drain profile are subjected to the same environmental conditions. The authors do not consider this tank mixing system improvement to be the final chapter in meeting the compliance requirements for TTHM and HAA5, but are confident that the City of Lebanon and Marion County water quality has been improved. LWWC will continue to evaluate their system to improve their water quality. As mentioned previously, tank mixing should be considered as complementary to other water treatment and distribution system improvements as part of a comprehensive effort to address TTHMs, HAA5s, and other emerging water quality concerns. For example, LWWC is currently evaluating improvements to their WTP to reduce DBP precursors (i.e. organic matter) using granular- or powder-activated carbon (GAC, PAC), new disinfection methods and processes, and evaluating additional HSPS controls directed at increasing tank turnover. 17

18 WORKS CITED United States Environmental Protection Agency (US EPA). Stage 2 Disinfectants and Disinfection Byproducts Rule: A Quick Reference Guide for Schedule 3 Systems. Office of Water. June PAX Water Technologies, San Rafael, California SolarBee, Inc., Dickinson, North Dakota TideFlex Technologies, Carnegie, Pennsylvania United States Department of the Interior, Bureau of Reclamation Technical Service Center. Water Treatment Engineering and Research Group. DBP Trihalomethanes and DBP Haloacetic Acids Fact Sheets. April