Full-scale testing was carried out at two wastewater treatment plants to determine whether residual polymer concentration, measured by filtrate and centrate absorbance at 191 nm, can be used to identify the optimum polymer dose and achieve in-line and real-time dewatering optimization. The first plant uses high speed centrifuges and the second plant uses belt filter presses for dewatering. During the testing, the polymer dose incrementally increased to cover the under-dose, optimum dose and over-dose polymer ranges, and the centrate/filtrate absorbance at 191 nm, turbidity and cake solids were measured. The results showed that absorbance measurements at 191 nm exhibited a parabolic shaped curve with increasing polymer dose, where the minimum absorbance corresponded to the optimum polymer dose. The method can directly measure the residual polymer concentration and determine the optimum polymer dose accordingly, and is planned to be used in the development of a dewatering automation system in the future.

INTRODUCTION

Dewatering plays a key role in reducing the cost of sludge (biosolids) treatment and final disposal. Sludge dewatering is the second highest cost area in a wastewater treatment plant following electrical utility costs, and a small 1 MGD plant can save $200,000/year by optimizing its polymer dose (Hach 2015). In addition, improvements to dewatering can significantly reduce the carbon footprint of a treatment plant. However, optimization of sludge conditioning and dewatering continues to be a challenge due to continuous changes in sludge characteristics and solids concentration during operation. Ideally, polymer dose should be checked and optimized on a real-time basis, but the majority of treatment plants manually adjust the polymer dose by checking cake solids or performing jar tests when they see the necessity. This usually leads to overdosing of polymer, which not only increases the polymer cost for the plant but also decreases cake dryness. A wetter cake, in turn, increases the sludge handling and disposal costs substantially.

Current dewatering optimization systems typically rely on indirect parameters such as solids concentration of incoming sludge or turbidity of the dewatering filtrate, which is not effective in capturing important sludge characteristics (e.g., biopolymers, cations, network strength) and predicting dewatering performance. The most effective method for optimization of sludge dewatering would be to measure the polymer concentration directly. Measuring polymer concentration during conditioning and dewatering would provide information on whether too much or too little polymer is added to sludge, and would also provide a tool to correct the polymer dose as needed. Even though there is a wide range of methods available for measurement and characterization of polymers, none of these methods is suitable for in-line and real-time applications. These methods typically utilize complex analytical methods that rely on chromatography and mass spectrometry, or colorimetric methods that require the addition of various reagents and processing times. The main polymer quantification methods reported in the literature include the starch-triiodide method (Scoggings & Miller 1979), viscosity (Jungreis 1981), calorimetry (Hansen & Eatough 1987), turbidimetry (Clapper et al. 1989), fluorescence spectrometry (Arryanto & Bark 1992), colloid titration (Gehr & Kalluri 1983), radioactive labeling (Nadler et al. 1994), flow injection analysis (Taylor et al. 1998), size exclusion chromatography (Lu et al. 2003), nuclear magnetic resonance spectroscopy (Chang et al. 2002), fluorescence tagging (Becker et al. 2004), and spectrophotometric determination using cationic dyes (Chmilenko et al. 2004).

Over the last few years, a new spectrophotometry based method was developed to measure the residual polymer concentration in filtrate and centrate after dewatering (Gibbons & Örmeci 2013; Al Momani & Örmeci 2014a). It was shown that polyacrylamide polymers strongly absorb light at 190–200 nm and, following the Beer-Lambert Law, there is a linear relationship between concentration and absorbance. The detection limit of the method was shown to range from 0.05–1.35 mg/L depending on the sample characteristics (e.g., water, wastewater, centrate) and polymer type and chemistry (e.g., molecular weight, structure) (Al Momani & Örmeci 2014a). The absorbance of polyacrylamide polymers is typically highest around 190 nm, where the sensitivity of the method is highest, and the detection limit is lowest.

A relationship was also successfully established between the residual polymer concentration and optimum polymer dose, and it was shown that the optimum polymer dose corresponds to the minimum absorbance value in filtrate or centrate during dewatering (Aghamir-Baha & Örmeci 2014; Al Momani & Örmeci 2014b). When the optimum polymer dose is exceeded, and the adsorption sites on sludge solids are saturated, excess polymer passes into sludge effluent and can be detected in centrate or filtrate at 190–200 nm. When the polymer dose is incrementally increased, the filtrate/centrate absorbance exhibits a ‘U’ shaped curve where the dip corresponds to the optimum polymer dose. In the under-dose range, increasing the polymer dose results in a decrease in filtrate/centrate absorbance due to improvement in supernatant quality. The filtrate/centrate absorbance is lowest at the optimum polymer dose. When the optimum dose is exceeded, absorbance values start to increase, corresponding to the increase in the residual polymer concentration in filtrate in the over-dose range.

The objective of this study was to progress from laboratory-scale to full-scale and test the method for polymer dose optimization at two full-scale treatment plants using an in-line UV-vis spectrophotometer (as shown in Figure 1) that can generate real-time data. If successful, results would provide the proof-of-concept for the development of a dewatering optimization and automation system in the future.
Figure 1

The analyzer system consisting of the in-line UV-vis spectrophotometer and the auto dilution system.

Figure 1

The analyzer system consisting of the in-line UV-vis spectrophotometer and the auto dilution system.

MATERIAL AND METHODS

Full-scale dewatering testing was carried out at two wastewater treatment plants. During full-scale testing at these plants, the polymer dose was incrementally increased to cover the under-dose, optimum dose and over-dose polymer ranges, and the centrate/filtrate absorbance at 191 nm, turbidity (Hach turbidimeter) and cake solids (EPA Method 160.3) were measured. The first plant (South Cary, NC, USA) uses high-speed centrifuges, and the second plant (Danbury, CT, USA) uses belt filter presses (BFPs).

The South Cary plant is an advanced activated sludge treatment system designed for nutrient removal. There is no primary treatment system; therefore, only waste activated sludge (WAS) is generated. WAS is blended with WAS imported from two other facilities and aerobically digested in an aerobic sludge holding basin. Following aerobic stabilization, the sludge is dewatered via centrifuge and thermally dried. The aerobically digested sludge has a solids concentration of 2.9%. The plant uses Clarifloc (R) SE-757, which is a high molecular weight cationic polyacrylamide. The facility typically doses polymer between 10 and 15 g/kg dry solids (DS), producing cake solids between 18 and 20%, and consumes approximately 79,500 kg of polymer annually.

The treatment system at the Danbury plant consists of primary settlers, trickling filters and activated sludge. The excess sludge is anaerobically digested (mesophilic). The typical primary sludge to WAS ratio is approximately 70:30, and the solids concentration after anaerobic digestion is 2.1%. The dewatering equipment consists of two 2.5-meter Roediger Pittsburgh BFPs. The plant uses a Mannich polymer (Polydyne Clarifloc C-321) at doses of 5.5–9 g/kg DS, producing cake at 16 to 19% DS.

The analyzer system consisted of the single wavelength (191 nm) spectrophotometer (Real Tech, Inc., ON, Canada), which measured absorbance at 191 nm only, with an auto-sampling/conditioning/dilution system. During these test studies, a full spectrum (191–750 nm) UV-vis spectrophotometer with deuterium light source (Real Tech, Inc., model Real Spectrum Gold Series), which measured absorbance at every wavelength between 191 and 750 nm, was coupled in series with the single wavelength UV-191 analyzer for the comparison of results from two instruments (Figure 1). In order to settle out the larger and heavier suspended solids to avoid internal analyzer tube clogging, a small stilling well with approximately 1–3 minute HRT was employed upstream of the dilution stream. The centrate or filtrate sample is pumped via a mini peristaltic pump to the stilling well. A second mini peristaltic pump continuously conveys supernate from the stilling well into and through the analyzers. A third mini-peristaltic pump provides dilution by pulling potable water from a reservoir at a preset dilution rate. The water is continuously mixed with the sample via the in-line static mixer. The diluted sample then flows through the sample cell in the UV191 analyzer and finally through the full spectrum analyzer to a drain. The proper dilution ratio is predetermined via a screening exercise, where a range of dilution factors are tested over a polymer dose range and the resulting absorbance values are recorded at each polymer dose.

RESULTS AND DISCUSSION

Centrifuge field demonstration test

Full-scale testing at the South Cary plant began in January 2014 and continued over a period of 2 months. After establishing the required dilution ratio (77:1), the sampling frequency was set at one-minute intervals. Centrate sampling and UV absorbance analysis were performed continuously while the full-scale dewatering polymer dose was manually decreased in 4% increments by varying the polymer feed pump speed (e.g. from 36% to 32%) and then increased above the original set point dose likewise in 4% increments. At a time point approximately 45 minutes after each polymer dose change, representative samples of cake were collected and the resulting dry solids concentration was measured.

Figure 2 is a dose response curve at a fixed dilution ratio of 77:1 depicting absorbance measured at 191 nm, with single-wavelength and full-spectrum UV analyzers, and the centrate turbidity. Absorbance and turbidity were recorded approximately 20 minutes after each polymer dose change. The absorbance curve obtained with the single-wavelength (191 nm) analyzer followed a parabolic shaped trend with an absolute minimum at a polymer dose of 11.8 g/kg DS. The minimum absorbance value also corresponded to the minimum turbidity value of the centrate. The plant's operational dose for this day was 17.5 g/kg DS. If operating at the optimal dose as determined by UV-191, the plant could potentially see a 30% reduction in polymer consumption while still maintaining similar final cake solids concentration. The lowest absorbance measurement on the full spectrum analyzer was at a dose of 12.8 g/kg DS, and a similar trend was observed. Some difference in measurements is expected between the analyzers, as the single wavelength analyzer has better sensitivity and its response time is faster and was specifically developed for dewatering applications. The parabola-shaped curve is a result of two conditions: underdosing and overdosing. On the underdosing side of the curve where the slope is negative, the absorbance signal decreases with increasing polymer dose due to the improvement in centrate quality. Around the optimum polymer dose, the absorbance is at a minimum due to the removal of the particles and lack of excess polymer in the centrate. On the overdosing side of the curve where the slope is positive, the absorbance signal again increases due to the increasing excess polymer in the centrate.
Figure 2

In-line and real-time measurement of centrate absorbance during full-scale testing measured at 191 nm with a single wavelength (191 nm) spectrophotometer and a full-spectrum (191–750 nm) spectrophotometer, and their relation to turbidity and optimum polymer dose.

Figure 2

In-line and real-time measurement of centrate absorbance during full-scale testing measured at 191 nm with a single wavelength (191 nm) spectrophotometer and a full-spectrum (191–750 nm) spectrophotometer, and their relation to turbidity and optimum polymer dose.

Figure 3 illustrates the relationship between centrate absorbance at 191 nm and the cake solids obtained from the centrifuge during full-scale testing. The lowest centrate absorbance, which indicates the optimum polymer dose, corresponded to the point where increasing the polymer dose further did not result in a significant increase in cake solids.
Figure 3

In-line and real-time measurement of centrate absorbance during full-scale testing with a single wavelength (191 nm) spectrophotometer, and its relation to cake solids and optimum polymer dose.

Figure 3

In-line and real-time measurement of centrate absorbance during full-scale testing with a single wavelength (191 nm) spectrophotometer, and its relation to cake solids and optimum polymer dose.

Representative centrate samples at each polymer dose were saved, and the absorbance of the samples collected from the full-scale testing (Figure 2) was measured the next day again to confirm the presence of the minimum, and also to determine whether settling out of the particles would affect the results (Figure 4). This experiment was also beneficial to see the impact of reaction time on the detection of the polymer, and whether the polymer would be reacting with other constituents in the centrate matrix and decay. Settling of the particles (less turbid centrate) resulted in flattening the bottom of the curve so it is better, in fact, to have some particles in the centrate to achieve a sharper dip in absorbance, which would make it easier to identify the optimum polymer dose.
Figure 4

The absorbance of centrate samples collected from full-scale testing measured after 24 hours of settling.

Figure 4

The absorbance of centrate samples collected from full-scale testing measured after 24 hours of settling.

Belt filter press field demonstration test

BFP testing began in September 2014 and continued for several weeks. Initial testing consisted of making slight adjustments on the polymer speed pump. The speed of the pump was recorded and then verified with draw-downs for an accuracy check on polymer dose measurement. Initial tests showed that the UV191 analyzer was recording a similar curve trend to that of centrifugation. This particular trend is ideal in that a distinct minimum can be detected and an optimal dose can be selected.

Data collected during the first week of testing are presented in Figures 5 and 6. The initial tests focused on identifying the dilution ratio, since the required dilution would be determined by the filtrate characteristics and would vary from treatment plant to treatment plant. During the full-scale testing, filtrate samples were collected and later analyzed at five different dilution ratios using the in-line spectrophotometer (Figure 5). The filtrate samples were cleaner compared to the previously tested centrate samples and required a lower dilution rate of 14:1, which captured the largest changes in absorbance. The lowest absorbance value of the filtrate samples corresponded to a polymer dose of 8.45 g/kg DS (16.9 lbs/DT), at which dose the filtrate turbidity was also the lowest (Figures 5 and 6). Finding the right dilution rate is important, as this would increase the sensitivity of the measurements and would make it easier to identify the dip in the absorbance that points to the optimum polymer dose. The dip slowly disappeared in over-diluted samples such as 39:1 and 52:1.
Figure 5

Effect of dilution rate on the absorbance of filtrate samples collected during the BFP demonstration testing.

Figure 5

Effect of dilution rate on the absorbance of filtrate samples collected during the BFP demonstration testing.

Figure 6

The absorbance (dilution rate 14:1) and turbidity of filtrate samples collected during the BFP demonstration testing.

Figure 6

The absorbance (dilution rate 14:1) and turbidity of filtrate samples collected during the BFP demonstration testing.

Further investigation was conducted by testing the absorbance signals taken at various sample locations on the press. Modern BFPs typically consist of two belts in a three-zone configuration. The three main zones include a gravity zone, a low-pressure zone, and a high-pressure zone. This particular press employed a rotating flocculation drum as the gravity zone. Filtrate samples from the low and high-pressure zone are typically very high in turbidity (>1,000 NTU) and usually contain a high solids concentration in the filtrate. Both turbidity and absorbance were measured on samples collected in the gravity zone (drum filtrate) and the low-pressure zone. Different dilutions were also tested for absorbance measurements. These data are summarized in Table 1 below.

Table 1

Absorbance and turbidity of filtrate samples collected from gravity and low-pressure zones of a BFP during the full-scale testing

Polymer doseFiltrate absorbance (at 191 nm)
Filtrate turbidity
Cake solids
g/kg DSGravity (14:1)Gravity (30:1)Low press. (30:1)Low press. (50:1)GravityLow press.GravityFinal
5.8 0.907 0.463 0.739 0.443 103 378 7.18 17.26 
6.1 0.899 0.445 0.727 0.428 115 410 18.19 
6.6 0.855 0.425 0.73 0.427 92 310 9.24 18.1 
7.1 0.874 0.44 0.752 0.438 79 318 8.52 17.83 
7.6 0.836 0.4 0.707 0.421 78 300 8.62 18.54 
8.0 0.88 0.439 0.73 0.435 80 317 10.56 17.1 
8.5 0.897 0.451 0.744 0.443 85 315 10.5 19.56 
8.8 0.913 0.481 0.718 0.417 85 284 9.88 18.37 
Polymer doseFiltrate absorbance (at 191 nm)
Filtrate turbidity
Cake solids
g/kg DSGravity (14:1)Gravity (30:1)Low press. (30:1)Low press. (50:1)GravityLow press.GravityFinal
5.8 0.907 0.463 0.739 0.443 103 378 7.18 17.26 
6.1 0.899 0.445 0.727 0.428 115 410 18.19 
6.6 0.855 0.425 0.73 0.427 92 310 9.24 18.1 
7.1 0.874 0.44 0.752 0.438 79 318 8.52 17.83 
7.6 0.836 0.4 0.707 0.421 78 300 8.62 18.54 
8.0 0.88 0.439 0.73 0.435 80 317 10.56 17.1 
8.5 0.897 0.451 0.744 0.443 85 315 10.5 19.56 
8.8 0.913 0.481 0.718 0.417 85 284 9.88 18.37 

Results showed the absorbance trends for samples collected from the gravity zone/drum filtrate were similar to those collected in the low-pressure zone. Both filtrates exhibited the U-shape trend with increasing polymer dose, and the lowest absorbance corresponded to the polymer dose of 7.6 g/kg DS for both filtrate samples at all dilution rates. The turbidity of the filtrates collected from the gravity and low-pressure zones was also lowest at 7.6 g/kg DS. Fewer disturbances and fluctuations were noted on the gravity zone filtrate compared to the low-pressure zone filtrate. The gravity zone had essentially no wash water, which minimized the potential for interference. Sample points downstream (pressure zones) have large amounts of wash-down water and significantly higher suspended solids, so these locations were eliminated as non-ideal. It was concluded that the filtrate collected from the gravity zone was ideal for identifying the optimum dose. Due to the difficulty of taking representative samples from across the belt during the testing, fluctuations in cake solids concentrations were observed.

During the full-scale dewatering testing at the South Cary and Danbury wastewater treatment plants, the Real Tech in-line analyzer was operated for several months without requiring chemical cleaning of sample lines (including periods where centrate turbidity was as high as 1,000 NTU). Following the screening protocol for determination of optimum dilution ratio, absorbance signal dynamic range and signal stability were excellent.

In the future, the unit can be used for in-line and real-time monitoring and adjustment of polymer dose to account for daily changes in sludge characteristics in a full-scale dewatering operation after incorporating a minimum search algorithm and software that keeps the filtrate or centrate absorbance at its local minimum value. The main advantage of the method is that it is quick and simple, does not require any reagents or processing times, and it produces real-time information, which is critical for successful dewatering optimization. Real-time UV-vis spectrophotometers are widely used for water monitoring, and therefore the instrument that is required for this application is available in the market. The data presented herein provides the proof of concept for the method, which allows the development of a dewatering optimization system in the future.

CONCLUSIONS

The full-scale tests carried out at two wastewater treatment plants confirmed the laboratory-scale results that were previously reported by Örmeci and co-workers (Aghamir-Baha & Örmeci 2014; Al Momani & Örmeci 2014a, 2014b). The results showed that the residual polymer concentration in sludge filtrate and centrate could be directly measured at 191 nm with an in-line spectrophotometer, and the optimum polymer dose corresponds to the lowest filtrate or centrate absorbance at 191 nm.

ACKNOWLEDGEMENTS

The authors would like to thank Steven Kosteniuk and Jeff Privott for their assistance and data collection during full-scale testing. The authors would also like to thank Andrew Glover at RealTech, Inc., Andy Russell and Josh Cummings at the S. Cary plant, as well as Walter Royals, Ralph Azzarito and Christian Hoan of Veolia Water North America.

REFERENCES

REFERENCES
Aghamir-Baha
S.
Örmeci
B.
2014
Determination of optimum polymer dose using UV-vis spectrophotometry and its comparison to filtration based tests
.
Residual Science and Technology
11
(
3
),
71
76
.
Al Momani
F.
Örmeci
B.
2014a
Measurement of polyacrylamide polymers in water and wastewater using an in-line UV-vis spectrophotometer
.
Journal of Environmental Chemical Engineering
2
(
2
),
765
772
.
Becker
N. S. C.
Bennett
D. M.
Bolto
B. A.
Dixon
D. R.
Eldridge
R. J.
Le
N. P.
Rye
C. S.
2004
Detection of polyelectrolytes at trace levels in water by fluorescent tagging
.
Reactive & Functional Polymers
60
,
183
193
.
Chang
L. L.
Bruch
M. D.
Griskowitz
N. J.
Dentel
S. K.
2002
NMR spectroscopy for determination of cationic polymer concentrations
.
Water Research
36
,
2255
2264
.
Chmilenko
F. A.
Korobova
I. V.
Nazarenko
S. V.
2004
Spectrophotometric determination of polyacrylamide in aqueous solutions using cationic dyes
.
Journal of Analytical Chemistry
59
(
2
),
124
128
.
Clapper
D. K.
Norfleet
J.
Alford
S. E.
1989
Analyzing PHPA in mud filtrates
.
World Oil
208
(
4
),
77
83
.
Gehr
R.
Kalluri
R.
1983
The effects of short-term storage and elevated temperatures on the flocculation activity of aqueous polymer solutions
.
Water Pollution Research Journal of Canada
18
,
23
43
.
Gibbons
M. K.
Örmeci
B.
2013
Quantification of polymer concentration in water using UV-vis spectroscopy
.
Journal of Water Supply: Research and Technology – AQUA
624
,
205
213
.
Hach
2015
Optimizing Sludge Dewatering: Case Study. DOC043.53.30229
.
Hach Industries, Loveland, CO
,
USA
.
Nadler
A.
Magaritz
M.
Leib
L.
1994
PAM Application techniques and mobility in soil
.
Soil Science
158
,
249
254
.
Scoggings
M. W.
Miller
J. W.
1979
Determination of water soluble polymers containing primary amide groups using the starch-triiodide method
.
Society of Petroleum Engineers Journal
19
(
3
),
151
154
.
Taylor
K. C.
Burke
R. A.
Nasr-El-Din
H. A.
Schramm
L. L.
1998
Development of a flow injection analysis method for the determination of acrylamide copolymers in brines
.
Journal of Petroleum Science and Engineering
21
,
129
139
.