Water Quality Improves by Recycling Settled Sludge

by John C. McLane

Many operators wear two hats: one as a water treatment plant operator, the other as a wastewater treatment plant operator. Likewise, many processes used in a wastewater plant can be adopted to benefit drinking water operations. Wastewater operators are responsible for the biological processes of an activated sludge plant, where, at the beginning of the treatment process, settled activated seed sludge is inoculated in incoming raw sewage flow to kick- start the growth of bioreducing organisms in the sewage.

This same process of recirculation works to 1,500–3,000 ntu within a few hours. The Retired chief equally well in the flocculation and coagulation high turbidity levels could persist for up to 20 operator Don processes of drinking water treatment. Recycling days before returning to normal levels. The high Russell (left) fresh settled chemical floc back to the incoming raw water turbidity would carry over through the and Fort Madison raw water flow initiates formation of good- clarifiers and filters, and we’d have periods when superintendent quality chemical floc with exceptionally good finished water turbidity approached or exceeded Norman Dodson solids and impurities removal and settling and the 1.0 ntu maximum contaminant level (MCL). at the upflow filtering characteristics, particularly during cold- In contrast, another seasonal problem we clarifier where water seasons. This practice also provides many faced was low-turbidity cold water. When the recycled labor and cost-savings advantages. temperatures reach freezing or below, surface sludge and the Midwestern Weather Woes water is difficult to treat because lime slurry are ᭤ added through In 1994, I was hired as the superintendent of the cold temperatures slow the necessary the suspended a 3.5-mgd water treatment plant in Fort Madison, chemical reactions, top pipe. Iowa. The plant’s source is 70 percent surface ᭤ floc settling deteriorates because of the water and 30 percent groundwater. The two increased density of water below 4° C, and different water sources have different chemistries and were being treated separately through ᭤ raw water turbidity often drops to coagulation, flocculation, and settling. Lime 2 to 6 ntu. When the surface slurry was (and still is) used as the primary supply source freezes over, the ice John McLane is a water coagulant for solids removal in the surface water cover eliminates the mixing effects treatment process consultant and for softening the groundwater. Each flow of winds, waves, and tributary with his own firm in Fort flows, so the frozen source water Madison, Iowa. A state- required a different pH for accomplishing the certified WTP operator at the individual treatment goal. The flows were then becomes quiescent and clear. highest level in Illinois, combined for recarbonation, filtration, fluoride When this happens, the water is Indiana, and Iowa, he has adjustment, and disinfection. “too clean” to form good-quality served as superintendent, The plant experienced normal, rapid, and floc, which meant that during cold plant engineer, certified weather, our finished water laboratory director, and safety drastic raw surface water fluctuations during engineer at four different spring snow melt and periodic episodes of rain turbidity periodically approached WTPs. He is also an runoff. During these episodes, the surface water or exceeded the 1.0-ntu MCL. instructor at two Illinois community colleges. turbidity would jump from its normal 4–40 ntu continued on page 16

Opflow / January 2004 2003 © American Water Works Association 15 was too labor intensive. The separate treatment Settled Sludge (from page 15) trains made it difficult for the operators to properly control the chemistry of the two source waters. All too frequently, the wrong chemical Lime Slurry Sump Pump Discharge Lime Slurry Drop Pipe feed rates were introduced because operators didn’t have or take the time to monitor both

Flash Mixer Impeller sources. This resulted in decreased process Top Clear Water 4 to 8 ft deep Flash Mixing Zone efficiencies, corresponding plant upsets, and the

Settling Zone rapid deterioration of final effluent quality. By not using the correct or optimum chemical Top Surface of Sludge Blanket Coagulation Zone feed rates, good quality floc would not form and

Skirt Baffle Wall settle out. Clarifier effluent turbidity would increase from the normal 7–10 ntu to 30+ ntu. Excess solids would carry over to overload and Updraft Tube pass through the filters, resulting in filter effluent Floc Zone exceeding the 1.0 ntu MCL. This would also Raw Water Incoming cause rapid filter plugging and reduce filter run times from the usual 30 to 50 hours to short Sump Pump ws) runs of only a few hours. Ultimately, the result Sludge Collection Flights (Plo would be emergency plant shutdowns, public notifications, and boil orders. Waste Sludge Withdrawal Pipe with Intermittent Flow Typical Upflow Clarifier To correct these problems, we combined the groundwater and surface water flows just before they entered the coagulation–flocculation We decided we needed to change our standard Schematic process, producing one combined raw water flow of sludge operating procedures if we wanted to bring the with a consistent chemistry that was less prone recycling finished water turbidity down as low as to rapid changes. This simplified the required lab process. reasonably possible and stay well within required work, minimizing human errors, and decreased standards, providing the very best finished water the operator’s workload. Then, having trained quality that we could. the operators to jar test and simplified the The plant uses ferric sulfate to boost the floc’s laboratory and chemical feeding duties, we were settling characteristics. However, the use of jar ready to take the next step. testing to determine proper chemical dosages of the primary flocculant, ferric sulfate, and Adding Dirt cationic polymers had been discontinued During the cold winter months when sometime earlier. Chemical dosages were turbidity is low, some of my colleagues at other determined by trial-and-error — or just plain treatment plants theorized that we could add guessing — which left much room for error and “dirt” — usually bentonite clay powder — to the caused marginally substandard finished water raw water in an effort to increase turbidity, thus quality. This put the utility on the brink of fostering better floc formation and greater intervention status by the state primacy agency. turbidity removal to keep the finished water Thus, our first steps for improvement were turbidity within specified limits. I had my own ᭤ the reintroduction of proper jar testing experience with bentonite and wasn’t convinced techniques, that was the way to go. A decade earlier, when I was working at the ᭤ the proper and effective interpretation of jar InterState Water Co. in Danville, Ill., we tried test results, and using bentonite, but were not pleased with the ᭤ the proper application of the correct chemical results. The bentonite was too difficult to mix dosages. into solution with the available equipment, and it was a burden on our budget. So we tried to We determined that the process of sampling, recycle the muddy sludge from the bottom of the testing, and controlling chemical feed rates for settling tank, reasoning that because this sludge the two treatment trains — groundwater was already wet and settled, it possessed good softening and surface water solids removal — settling characteristics. And, it was free.

16 2003 © American Water Works Association Opflow / January 2004 However, we didn’t have the Before Sludge Recirculation After Sludge Recirculation right equipment to make the process work Clarifier Effluent 25–70+ ntu 3.0 +/– 0.2 ntu effectively — our (> 90% reduction) boat-mounted sludge pump Filter Effluent > 2.0 ntu < 0.19 ntu was too unwieldy — so we (maximum) (> 90% reduction) abandoned the effort. Maximum Daily 0.42 ntu 0.0347 ntu Later, I read about a Effluent, Annual Average (> 90% reduction) plant-scale pilot study done in 1982 where settled sludge was used to enhance Table 1. Benefits of switching to a recycled sludge process, based on constant raw water turbidity turbidity removal in of 2–4 ntu for raw frozen surface water and 300–3,000 ntu for surface water affected by heavy rainfall or snowmelt. secondary coagulation–flocculation settled the fastest, and it would be at the sedimentation basin was in the tanks; the study confirmed my earlier the bottom of the settling tank. The 30–50 ntu range; now it consistently thoughts about the possibilities of bottom sludge would be the most stays at 3.0 ntu (+/– 0.4 ntu). using recycled sludge. Fort Madison dense and concentrated. Sludge from During our trial phase, we did bring would be a good place to transfer the a higher elevation in the tank would the clarifier effluent down to 1.0 ntu technology to a real-world application. probably be less dense, and its settling by adding extra primary coagulant We determined that the settled characteristics would probably not be and ferric sulfate and various cationic, sludge could be recycled by as good. anionic, and nonionic polymers at a introducing it into the incoming The sump pump discharged its significant cost per mil gal. However, combined raw water flow before the flow into the lime slurry trough, we found that the floc produced coagulation–flocculation process. Our mixed with the lime slurry, and was would readily pass through the filters, first recycling attempt, using the introduced into the incoming raw with the resulting filter effluent existing equipment and plumbing, water flow just prior to entering the registering 0.3–0.5 ntu. provided only an intermittent sludge flash-mixing zone. The results of this We realized that the filter effluent return to the flow entering the flash modification were no less than we were getting with the 3.0 ntu mix and flocculation zone, because the spectacular. clarifier effluent was far superior — a sludge pump was only activated when It has now been more than seven fairly steady 0.022 ntu (+/– 0.005). the sludge pit was full. We added pipes and valves so some years since we began recirculating the The Maximum Day Annual Average of the settled sludge was recyled sludge. The process has proved filter effluent for the past four years directly to the raw water pipes, but the extremely successful, yielding results has been 0.0347 (+/– 0.002) ntu. The sludge return was still intermittent. far better and more numerous than we resultant process efficiencies and However, an increase in solids had hoped for or anticipated. higher-quality finished water also concentration was visible in the High raw water turbidity still means lower disinfectant by-products mixing zone, and we could readily carries over through the clarifier and fewer water quality problems of measure and observe a marked effluent and the filter effluent. But other kinds. reduction of the clarifier effluent those fluctuations were greatly Filter runs originally were 30 to turbidity. From a range of 10–30 ntu, reduced after recirculation of the 50 hours during normal operating the effluent had dropped to 5–7 ntu. recycled floc was implemented. The conditions. Now, they are 150 to We could also see that intermittent undesirable fluctuations in our 200+ hours, with few interruptions. recirculation was not enough. We finished water caused by heavy rainfall The reduction in backwash water needed constant recirculation. or snowmelt and cold water have all usage alone has saved the utility been corrected, and the finished water $46,000 annually. In Iowa, we are not Tapping the Bottom Sludge turbidity is now far below the MCL. allowed to recycle backwash water For our next effort, we installed an Our efforts to improve procedures and because of the concern that recycled inexpensive 3,000-gph submersible processes paid off. backwash might result in a sump pump, with its intake near the Water Quality Benefits concentration of protozoa, such as bottom of the settling zone, out of the Since we began recycling the Giardia and Cryptosporidium, path of the sludge collection sludge, we have routinely experienced reentering the filtration process. equipment. We reasoned the best- low clarifier effluent turbidity. Coliform and heterotrophic plate quality sludge was that which had Originally, the effluent turbidity from continued on page 18

Opflow / January 2004 2003 © American Water Works Association 17 Settled Sludge (from page 17) count (HPC) analyses have been small amount of lime that does attach performed regularly during sludge to the walls is much softer and can be recycling; almost no bacteria at all has removed by a fire hose and backwashing been evident from the top of the filters at flow rates of 15 to 20 gpm/ft2 of and clearwell effluent, and the HPC is filter area, without air scouring — extremely low — far less than 10/mL. another major cost savings. Prior to sludge recycling, HPC results The filter media have shown no were routinely >100,000/mL. Tests signs of deterioration; there has been conducted monthly at more than 60 no increase of lime deposits plating locations in the distribution system the media, and the media no longer are also consistently well below the concentrate in concrete-like layers or USEPA standard of >500 colonies/mL. mudballs. The filter media do not Prior to sludge recirculation, we require replacement at all and may received 30+ taste-and-odor have an indefinite effective, useful life, complaints per week. Now we only which saves us a substantial amount, receive a couple such complaints a because previously the annual budget year, and these are usually traced to for media replacement was some cause other than the drinking $40,000–$80,000. water. Plant operations are less labor Disinfection is more stable and intensive now for our operators. consistent, primarily because of the Operators don’t have to make much lower emergency finished water chemical feed- turbidity and It has now been more than rate adjustments lower organic seven years since we began triggered by loading onto the recirculating the sludge. rapid, large filters. As a result, fluctuations in the total The process has proved raw water trihalomethanes extremely successful, turbidity from and haleoacetic rain runoff or acid levels have yielding results far better snowmelt. Sludge been lowered to and more numerous than recycling one third of the anticipated. significantly new MCLs. buffers these Chlorine demand fluctuations. and usage are also reduced. With this Because the operators are now decreased disinfectant usage, proficient at the practice, we’ve also distribution system chlorine residuals cut back on routine jar testing — are much easier to manage throughout practicing it only when there are the year. sudden changes in the raw surface water quality, when there are major Financial Benefits seasonal temperature changes, or We’ve also reduced the cost of new whenever the routine daily process lime usage by 17 percent. Previously, chemistry tests indicate that the water because of the lack of bonding quality diverges from the optimum between the lime and the floc, lime operating control parameters. would deposit on tank and pipe walls, We have found no disadvantages or mechanical equipment, and filter drawbacks to implementing the sludge media; it took a jackhammer to recycling process modification, which remove it, and frequent media is now employed year-round, not replacement was required. Now, the just during the cold-water season.