This article was downloaded by: [University of Otago] On: 04 January 2015, At: 21:48 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Environmental Technology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tent20
Variations of morphology of activated sludge flocs studied at full-scale wastewater treatment plants a
b
a
Ewa Liwarska-Bizukojc , Anna Klepacz-Smółka & Olga Andrzejczak a
Institute of Fermentation Technology and Microbiology, Lodz University of Technology, Wolczanska 171/173, Lodz 90-924, Poland b
Department of Bioprocess Engineering, Lodz University of Technology, Wolczanska 213, Lodz 90-924, Poland Accepted author version posted online: 03 Nov 2014.Published online: 26 Nov 2014.
Click for updates To cite this article: Ewa Liwarska-Bizukojc, Anna Klepacz-Smółka & Olga Andrzejczak (2014): Variations of morphology of activated sludge flocs studied at full-scale wastewater treatment plants, Environmental Technology, DOI: 10.1080/09593330.2014.982717 To link to this article: http://dx.doi.org/10.1080/09593330.2014.982717
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Environmental Technology, 2014 http://dx.doi.org/10.1080/09593330.2014.982717
Variations of morphology of activated sludge flocs studied at full-scale wastewater treatment plants Ewa Liwarska-Bizukojca∗ , Anna Klepacz-Smółkab and Olga Andrzejczaka a Institute
of Fermentation Technology and Microbiology, Lodz University of Technology, Wolczanska 171/173, Lodz 90-924, Poland; b Department of Bioprocess Engineering, Lodz University of Technology, Wolczanska 213, Lodz 90-924, Poland
Downloaded by [University of Otago] at 21:48 04 January 2015
(Received 3 October 2013; accepted 31 August 2014 ) Digital image analysis has been intensively developed over the last two decades including its application to describe morphology of activated sludge flocs. However, only few studies concerned the variation of flocs morphology with respect to the operational conditions, particularly oxido-reductive conditions, in a full-scale wastewater treatment plant (WWTP). In this work, morphology of activated sludge flocs was monitored over one year in two different full-scale WWTPs. The main aim of this study was to find the relationship between the operational parameters and morphology of sludge flocs. Simultaneously, the variations in floc size along activated sludge chamber were studied with respect to the oxido-reductive conditions. It was found that the sludge loading rate was one of the most important operational parameters influencing floc size. It was estimated that its values higher than 0.1 kg BOD5 kg TS−1 d−1 contributed to the decrease in floc size. Also, the oxido-reductive conditions influenced the floc size. It was statistically proved that flocs from the anaerobic zone were usually smaller than flocs from the anoxic or aerobic zones. Distribution of floc size in a full-scale WWTP usually could be described by a log-normal model. Keywords: activated sludge; distribution; floc size; image analysis; operational parameters
Introduction Activated sludge systems are commonly applied in wastewater treatment plants (WWTPs) nowadays. After leaving the activated sludge chamber (biological step), the flocculated biomass of activated sludge is usually separated from the treated effluent in settling tanks. The effectiveness of solid–liquid separation and, as a consequence, the quality of the treated effluent depends on the morphology of activated sludge flocs.[1,2] Morphological parameters of the activated sludge flocs can be divided into two categories: (1) parameters describing floc size and (2) parameters describing flocs shape. The following parameters belong, inter alia, to the first category: projected area, equivalent diameter and perimeter. The second category covers circularity, convexity and form factor.[3] Out of the above-mentioned morphological parameters, equivalent diameter (Deq ) is most often used. In accordance with its value, flocs can be classified as small, medium-sized or large. In agreement with the classification elaborated by Eikelboom,[4] flocs of diameter below 25 μm are small, flocs of diameter between 25 and 250 μm are medium-sized and flocs of diameter above 250 μm are large. This classification is still in use irrespective of the fact that whether the traditional (manual) or modern
*Corresponding author. Email: [email protected] © 2014 Taylor & Francis
(computational) microscopic techniques are employed to measure the flocs.[5–7] Techniques of digital image analysis have been intensively developed over the last two decades. They were also adopted to measure morphological parameters of activated sludge flocs. The digital image analysis enables a fast, objective and comprehensive description of floc morphology. It was demonstrated that image analysis was a useful tool to characterize activated sludge biomass in terms of floc size and its relation to the settling properties and concentration of biomass.[8–12] Mesquita et al. [5] proved that the image analysis methodology was adequate for a continuous examination of the morphological changes of aggregates and subsequent prediction of the sludge volume index (SVI). At the same time, image analysis combined with chemometric techniques can be applied to identify activated sludge abnormalities, namely pinpoint floc formation and filamentous and zoogleal bulking.[3] Although image analysis was used several times in the monitoring of activated sludge in full-scale WWTPs,[5,10,12–14] the relationship between morphological parameters of the activated sludge flocs and operational or environmental conditions has not been investigated so far. What is more, activated sludge was usually taken from one sampling
2
E. Liwarska-Bizukojc et al.
Downloaded by [University of Otago] at 21:48 04 January 2015
Figure 1. Scheme of the activated sludge system of the WWTP Zgierz. Variations of mean values of floc diameter along the activated sludge bioreactor in the sampling period.
point located in the aeration zone of an activated sludge system.[5,11,14] It means that the variability of sludge floc morphology along the activated sludge chamber containing zones with different oxido-reductive conditions has not been studied yet. Such operational conditions as sludge age, loading rate, temperature, mixing intensity and dissolved oxygen concentration play a key role in the biological treatment of wastewater in activated sludge systems. Their changes influence the composition, structure and subsequently the morphology of activated sludge flocs. It was found that particularly floc size and its distribution provided useful information on changes of the operational conditions.[15–17] The application of computational image analysis techniques should support the search for correlations between operational conditions of WWTP and morphological parameters of activated sludge flocs. Nevertheless, one should be aware that each real WWTP is an independent stochastic system. In this work, two full-scale WWTPs were monitored over one year. The main aim of this study was to find a relationship between operational parameters and the morphology of sludge flocs. Simultaneously, variations in floc size along the activated sludge chamber were studied with respect to the oxido-reductive conditions.
Materials and methods Description of the WWTP Zgierz Activated sludge was taken from two full-scale WWTPs, i.e. a WWTP in Zgierz and the Combined WWTP in Lodz. Both WWTPs tested treat municipal wastewater from the cities of Zgierz and Lodz, and several communes located in the neighbourhood of these cities as well as industrial wastewater originating mainly from small- and mediumsize plants. A separate sewer system dominates in Zgierz, while in Lodz a combined sewer system is predominant. The average pollutant load to the plant corresponds to approximately 94,000 PE in the WWTP Zgierz, while in
Figure 2. Scheme of the activated sludge system of the Combined WWTP in Lodz. Variations of mean values of floc diameter along the activated sludge bioreactor in the sampling period.
the Combined WWTP Lodz it corresponds to 1 × 106 PE. The contribution of industrial wastewater is usually in the range from 10% to 15% in both WWTPs tested. In the WWTP Zgierz, the biological stage consists of one fivezone bioreactor and secondary clarifier run in the Phoredox process configuration (Figure 1), whereas in the Combined Wastewater Treatment Plant it consists of seven lines of three-zone bioreactor and secondary clarifiers run in the UCT process configuration (Figure 2). In 2012, the average inflow of wastewater was 8720 m3 d−1 in the WWTP Zgierz, while in the Combined WWTP in Lodz it was 174 540 m3 d−1 . The values of other operational parameters of the WWTPs investigated in this study are given in Table 1. Sampling Activated sludge samples were taken from each zone of activated sludge chambers once a month in dry weather conditions. Each time a sample of volume 500 ml was taken from each part of activated sludge chamber and analysed within two hours. In the WWTP Zgierz, the sampling period comprised the whole year 2012, while in the Combined WWTP Lodz it was from March 2012 to February 2013.
3
Environmental Technology
Downloaded by [University of Otago] at 21:48 04 January 2015
Table 1.
Characteristics of operational parameters of the WWTPs studied in the sampling period of time.
Characteristic feature
WWTP Zgierz
Combined WWTP Lodz
Type of activated sludge system Range of average daily flow rate (m3 d−1 ) Average recirculation of activated sludge (m3 d−1 ) COD (mg O2 l−1 ) Average sludge loading rate (g BOD g TS−1 d−1 ) Temperature (°C) SRT (d) TS (g kg−1 ) SVI (ml g TSS−1 )
Phoredox 5510–11,797 9307–12,464 744–4700 0.01–0.132 8.5–22.0 10.2–24.4 4.21–5.88 41–85
UCT (three stages) 149,998–242,311 145,100–237,464 518–985 0.04–0.07 11.9–22.2 7.3–11.5 3.79–4.68 75–135
Image analysis Three vital unstained slides of each activated sludge sample were prepared for image analysis. The activated sludge flocs were observed under a Nikon Eclipse Ni light microscope with 4 × objective. The RGB (red, green, blue model of colour) images were snapped, processed and analysed with the help of NIS ELEMENTS AR software (Nikon). At least 40 images from each sample were processed and analysed with the help of the special automated procedure aimed at measuring the basic morphological
parameters of the flocs, namely projected area, perimeter, equivalent diameter, circularity and convexity. From several to about 60 objects (flocs) were observed in each image. The automated procedure consisted of the following steps: (1) transformation of RGB image to the greyscale image, (2) segmentation (the same level of threshold for all images analysed was set) and (3) measurement of morphological parameters of the activated sludge flocs. The projected area is the basic image analysis parameter found easily by pixel count and its multiplication by
(a)
(b)
Figure 3. Variations of floc diameter in each zone of the activated sludge chamber in the sampling period (a) the Combined WWTP in Lodz and (b) the WWTP Zgierz (the number of objects, i.e. flocs, analysed in each sample of activated sludge varied from 980 to 2450).
4
E. Liwarska-Bizukojc et al.
a scaling factor. The equivalent diameter was measured as the mean value of lengths of the lines between two points on the boundary of the individual object going through its centroid. The perimeter is simply the length of the boundary of the object, while the convex perimeter is the length of the boundary obtained after filling of all concavities of the object. The convexity is the ratio of perimeter to convex perimeter. Circularity is the shape factor that indicates to what extent the measured object is similar to the true circle. If it is equal to one, the object is the true circle. The lower it is, the less circular is the object.
Downloaded by [University of Otago] at 21:48 04 January 2015
Other measurements and analysis Filaments index (range from 0 to 5) was recorded according to Eikelboom.[4] The dominating filaments were identified based on conventional techniques, i.e. morphological and staining (Neisser and Gram) properties. The values of operational parameters as well as the results of physicochemical analysis made in the sampling time were obtained from the WWTPs studied.
Statistical measures and methods The basic statistical elaboration comprising the calculation of mean values, standard deviation (σ ) and goodness of fit for normal distribution of morphological parameters of activated sludge flocs was prepared with the use of MS Excel. Standard deviations (σ ) presented further in Figures 1–3 and 7 express the variability of the values of the morphological parameters shown. The Kolmogorov–Smirnov test (p < 0.05) was used to check the goodness of fit for normal distribution. Several distribution models including log-normal, Gaussian, Lorentz and Voigt were tested to approximate flocs diameter distribution. Additionally, one-way analysis of variance (ANOVA) was applied to estimate if the values of the projected area of activated sludge flocs coming from different parts of the activated sludge chamber were statistically equal. The null hypothesis stating that they were equal was assumed. The measured values of the projected area were log-transformed in order to fulfil the normality assumption.[18] ANOVA implemented in MS Excel (Analysis ToolPak) software was used. The confidence level of 95% was assumed.
Figure 4. Variations of mean projected area of flocs vs. sludge loading rate, SRT and SVI (a) in the Combined WWTP in Lodz and (b) in the WWTP Zgierz (the number of objects, i.e. flocs, analysed in each sample of activated sludge varied from 980 to 2450).
Downloaded by [University of Otago] at 21:48 04 January 2015
Environmental Technology Results and discussion Flocs size The elaborated procedure of image analysis occurred to be successful in fast and automated measurement of morphological parameters of activated sludge flocs. Out of the morphological parameters of flocs measured in this work, those describing floc sizes were particularly important from the practical point of view. Thus, they will be presented and analysed first. Mean equivalent diameters of activated sludge flocs in the WWTP Zgierz varied from 117.1 to 170.9 μm depending on the month and zone of activated sludge chamber, while in the Combined WWTP in Lodz they ranged from 81.1 to 106.7 μm (Figure 3). It meant that the activated sludge flocs could be classified as medium size in both WWTPs studied.[4,7] Mielczarek et al. [7] found that in Danish WWTPs, activated sludge flocs of large size (with a diameter above 250 μm) dominated. However, the diameter of flocs was measured manually, and visual characteristics of flocs were performed in agreement with Eikelboom [4] protocol.[7] At the same time, some previous works [19,20] indicated that activated sludge flocs of equivalent diameter below 100 μm were relatively small and might cause an increase in SVI or even increase in the turbidity of effluent. The latter was not observed because the concentration of suspended solids in the effluent in both WWTPs studied did not exceed 25 mg l−1 . The values of SVI were generally higher in the Combined WWTP in Lodz (Table 1), in which flocs were smaller on average than in the WWTP Zgierz (Figure 4). The values of SVI in the Combined WWTP of Lodz were usually above 100 ml g TS−1 (in 10 out of 12 analysed months); however, they did not exceed 150 ml g TS−1 (Figure 4) which indicated bulking conditions.[21] It meant that the elevated values of SVI in the Combined WWTP in Lodz may have been caused by the disintegration of activated sludge flocs as well as a higher number of filamentous bacteria observed in the activated sludge from this WWTP. No direct correlation between SVI and floc size expressed by the mean projected area was found (Figure 4). At the same time, filaments’ index for sludge from the WWTP Zgierz was equal to 0 or 1, while in the Combined WWTP in Lodz it varied from 2 to 4. The dominating filamentous species were identified on the basis of their morphological characteristics and staining properties. In the Combined WWTP in Lodz Microthrix parvicella and type 0041 were co-dominant, while in the WWTP Zgierz type 0092 dominated.
5
earlier works [15,16] concerning floc size, the log-normal model was selected to describe floc size distribution. In both papers cited above, this model occurred to fit well with the experimental data; however, these were lab-scaleactivated sludge systems. In this work, a log-normal model expressed by the following equation is used: −(ln x/μ)2 A · exp , y = yo + √ 2 · σ2 2·π ·σ ·x
(1)
where x is the floc diameter, μ is the arithmetic mean, σ is the standard deviation, y 0 is the offset frequency and A is the area under the curve within limits ± σ . It occurred that the log-normal distribution fitted well with the data from the Combined WWTP in Lodz. The values of correlation coefficient R2 were in the range from 0.915 to 0.978 depending on the month and zone of the activated sludge chamber. The log-normal distribution is often met in various biological systems. The distribution of particles, chemicals and organisms in the environment is usually log-normal.[22]
Flocs size distribution The distribution of diameters of flocs coming from the WWTPs studied did not correspond to the normal distribution. It was confirmed by the Kolmogorov–Smirnov test and trials of the approximation of the experimental points by the Gaussian curve. Taking into account
Figure 5. Distribution of activated sludge flocs in each zone of the activated sludge chamber in November 2012 in the WWTP Zgierz.
Downloaded by [University of Otago] at 21:48 04 January 2015
6
E. Liwarska-Bizukojc et al.
Figure 6. Distribution of activated sludge flocs in each zone of the activated sludge chamber in November 2012 in the Combined WWTP in Lodz.
The distribution of diameters of flocs coming from the WWTP Zgierz usually fitted to the log-normal distribution. However, the values of R2 were often lower than those determined for the Combined WWTP in Lodz and in some cases they were below 0.900. Thus, attempts of approximation using other distribution curves, such as for example Lorentz or Voigt, were made. However, they were not successful. The values of R2 were at a similar or even lower level as they were obtained for the log-normal distribution. The distribution of activated sludge flocs in both WWTPs in November 2012 is shown in Figures 5 and 6 as an example. It is visible that in the activated sludge coming from the Combined WWTP in Lodz, flocs of diameter up to 60 μm dominated, whereas in the WWTP Zgierz the diversity of activated sludge floc size was significantly wider. This phenomenon was well reflected by the higher values of SD determined for flocs from the WWTP Zgierz (Figure 3). This diversity of floc size is the most probable reason of difficulties in a mathematical description of the distribution of these flocs. In turn, the reason of this
wide diversity of flocs size was the variation of operating conditions in the WWTP Zgierz (Table 1 and Figure 4). In the sampling period (from March 2012 to February 2013), the contribution of activated sludge flocs of a diameter below 60 μm was between 37% and 52% in the Combined WWTP in Lodz, whereas in the WWTP Zgierz it was from 8.5% to 31% in 2012. The values varied depending on the month and zone of the activated sludge chamber.
Effect of operating parameters on flocs morphology Despite the fact that real WWTPs are stochastic systems, the effects of operational parameters on activated sludge have been studied in many works.[7,10,13,14] Analysing the influence of operating parameters on floc morphology in this study, it was observed that sludge loading rate (SLR) and oxido-reductive conditions had a stronger effect on floc size in comparison with other parameters, as for example temperature or sludge concentration (TS). The increase in
Decembers 2012 + +↓ − + November 2012 +↓ +↓ +↓ +↓ Anaerobic vs. anoxic 1 Anaerobic vs. aerobic 1 Anaerobic vs. anoxic 2 Anaerobic vs. aerobic 2
− No statistically significant difference between flocs-projected areas. + The projected areas of flocs are statistically different. + ↓The projected area of flocs from the anaerobic zone is lower than the projected area of flocs from the other zones being compared.
October 2012 − – +↓ − September 2012 − +↓ − − August 2012 +↓ +↓ +↓ +↓ July 2012 +↓ +↓ +↓ +↓ June 2012 − – + − May 2012 − +↓ +↓ +↓ April 2012 − +↓ − +↓ March 2012 +↓ − − − February 2012 − − − − January 2012 − − − −
+↓ +↓ +↓ +↓ − + +↓ +↓ − +↓ + − + + +↓ +↓ +↓ − Anaerobic vs. anoxic Anaerobic vs. aerobic
− +↓ +↓ +↓ WWTP Zgierz
+↓ +↓
November 2012 October 2012 September 2012 August 2012 July 2012 June 2012 May 2012 April 2012 March 2012 Comparison
Combined WWTP Lodz
Table 2. Results of ANOVA test for both WWTPs studied.
Downloaded by [University of Otago] at 21:48 04 January 2015
December 2012
January 2013
February 2013
Environmental Technology
7
sludge loading rate above 0.1 kg BOD5 kg TS−1 d−1 contributed to a decrease in the mean projected area of flocs in the WWTP Zgierz. It is illustrated in Figure 4, in which the value of sludge loading equal to 0.1 kg BOD5 kg TS−1 d−1 was marked by a dashed line. In the Combined WWTP in Lodz sludge loading rate varied in the relatively narrow range and did not exceed 0.07 kg BOD5 kg TS−1 d−1 , thus the decrease in floc size in relation to the increase in sludge loading rate was not found. It was also noticed that a shorter sludge age favoured smaller flocs. In the WWTP Zgierz, smaller flocs were observed in the activated sludge characterized by short sludge retention time (SRT) and exposed to high organic loading (Figure 4). In May 2012, when the sludge loading was 0.133 kg BOD5 kg TS−1 d−1 and SRT was 10.4 d, the smallest flocs of the mean diameter equal to 117.1 μm were found. The occurrence of smaller flocs in the Combined WWTP in Lodz in comparison with the WWTP Zgierz may also be associated with the lower values of SRT used in this WWTP. However, it should be emphasized that no mathematical correlation was found in this WWTP either between floc diameter (or projected area) and SRT or floc diameter and sludge loading rate. Barbusi´nski and Ko´scielniak [15] found that after the substrate overload event, the flocs were more prone to break-up. However, the long-term loading changes caused larger disturbances to the floc size distribution than more rapid but shorter ones.[15] In contrast to the earlier studies, in this work the variations of floc morphology along the activated sludge chambers at different oxido-reductive conditions were investigated. It was observed that activated sludge flocs from the anaerobic chamber were usually smaller than the flocs from the anoxic or aerobic chambers. On the basis of ANOVA, it was revealed in 19 out of 24 compared pairs of data sets that flocs from anaerobic zone in the Combined WWTP of Lodz were statistically different in terms of their size from the flocs of the other two zones. To be more precise, they were smaller than the flocs from the anoxic or aerobic chambers in 16 comparisons of data sets, i.e. in eight cases they were smaller than the flocs from the anoxic zone and in eight cases than the flocs from the aerobic zone (Table 2). In the WWTP Zgierz, activated sludge flocs coming from the anaerobic zone were statistically different in 24 out of 48 compared pairs of data sets (Table 2). What is more important, they were smaller than the flocs from the anoxic or aerobic zones in 21 cases. In Table 2, it is shown in which months and in which zone flocs coming from the anaerobic part were smaller than these from the other four zones of the bioreactor. The results obtained upon statistical analysis for two WWTPs indicated that the anaerobic conditions favoured the decrease in activated sludge flocs. It occurred that neither stronger shear forces nor limitation of carbonaceous substrate, which were more common in the aerobic zones of activated sludge chambers, contributed to the decrease
Downloaded by [University of Otago] at 21:48 04 January 2015
8
E. Liwarska-Bizukojc et al.
in floc size to such extent as oxido-reductive conditions and limitation of dissolved oxygen (DO). In both WWTPs studied, dissolved oxygen concentration was close to zero and redox potential was usually below − 120 mV in the anaerobic zone. DO concentration in the anoxic zone was very low and did not exceed 1 mg l−1 , while the redox potential was usually higher than in the anaerobic part of the bioreactor but varied in a wide range from − 200 to + 50 mV. In the aerobic zone, DO concentration was from 3.5 to 7.2 mg O2 l−1 in the WWTP Zgierz and from 1.9 to 4.6 mg O2 l−1 in the Combined WWTP of Lodz. It indicated that even a small increase in DO concentration or redox potential might have contributed to the increase in floc size. The observations regarding the decrease in floc size in the anaerobic conditions made in this work for two full-scale WWTPs are in agreement with the conclusions formulated by Wilén and Balmér [16] upon the laboratory-scale experiments. They claimed that alternating aerobic/anaerobic conditions, the proportion of smaller flocs (diameter up to 20 μm) increased gradually in the anaerobic period. Moreover, it was found that there was a trend towards larger flocs at a higher concentration of DO.[16] Variations of circularity of activated sludge flocs in both WWTPs studied are shown in Figure 7. The shape
of activated sludge flocs is generally far from the ideal circle, for which it should be equal to one, according to the definitions implemented in NIS Elements software. What is more, despite the significant difference in operational parameters as well as in floc size, the circularity of flocs in both WWTPs tested was at a similar level and varied in a relatively narrow range. In the WWTP Zgierz the mean values of circularity varied from 0.287 to 0.413 depending on the zone and month, while in the Combined WWTP Lodz they were in the range from 0.271 to 0.395 (Figure 7). No correlation between floc circularity and operational parameters was established. It concerns convexity too. Convexity varied from 0.715 to 0.789 in the WWTP Zgierz, whereas in the Combined WWTP Lodz it was from 0.677 to 0.767. In the Combined WWTP of Lodz, the flocs from the anaerobic zone were slightly more circular than those from the anoxic or aerobic ones. However, this trend was not observed in the WWTP Zgierz (Figure 7). Generally, the measurement of floc circularity in a full-scale WWTP is less useful than the measurement of the floc diameter or projected area. However, it may be beneficial in the cases when pollutants such as surface active agents are present in the influent.[20]
(a)
(b)
Figure 7. Variations of floc circularity in each zone of the activated sludge chamber (a) in the Combined WWTP in Lodz and (b) in the WWTP Zgierz (the number of objects, i.e. flocs, analysed in each sample of activated sludge varied from 980 to 2450).
Downloaded by [University of Otago] at 21:48 04 January 2015
Environmental Technology Conclusions (1) Sludge loading rate is one of the most important operational parameters influencing flocs size. There is no direct correlation between SLR and flocs area or diameter. It is estimated that SLR values higher than 0.1 kg BOD5 kg TS−1 d−1 contribute to the decrease in floc size, particularly if lower SLR dominates in everyday operation of a WWTP. The variation of SLR in a wide range (0.0015–0.132 kg BOD5 kg TS−1 d−1 ) induces bigger changes in floc size and as a consequence in floc distribution than the stable values of SLR (0.04–0.07 kg BOD5 kg TS−1 d−1 ). (2) Shorter sludge retention times seem to favour smaller activated sludge flocs (diameter below 100 μm), especially if accompanied by high sludge loading rate. (3) The distribution of flocs size in the full-scale WWTP may vary depending on the WWTP studied. It usually can be described by a log-normal model. (4) Flocs from the anaerobic zone are usually smaller than the flocs from anoxic or aerobic zones. It indicates that oxido-reductive conditions affect the floc size. Acknowledgements This work was made within the project No. 14-0004-10 financed by the National Centre for Research and Development, Republic of Poland.
[6]
[7]
[8] [9] [10]
[11]
[12]
[13]
[14]
[15]
Disclosure statement No potential conflict of interest was reported by the author(s).
[16]
References
[17]
[1] Mohammed RN, Abu-Alhail S, Xi-wu L. Long-term operation of a novel pilot-scale six tanks alternately operating activated sludge process in treating domestic wastewater. Environ Technol. 2013;35:1874–1885. [2] Villain M, Clouzot L, Guibaud G, Marrot B. Impact of oxygen cut off and starvation conditions on biological activity and physico-chemical properties of activated sludge. Environ Technol. 2013;34:901–910. [3] Mesquita DP, Amaral AL, Ferreira EC. Characterization of activated sludge abnormalities by image analysis and chemometric techniques. Anal Chim Acta. 2011;705:235– 242. [4] Eikelboom DH. Process control of activated sludge plants by microscopic investigation. London: IWA Publishing; 2000. [5] Mesquita DP, Dias O, Dias AMA, Amaral AL, Ferreira EC. Correlation between sludge settling ability and image
[18] [19] [20] [21] [22]
9
analysis information using partial least squares. Anal Chim Acta. 2009;642:94–101. Mesquita DP, Ribeiro RR, Amaral AL, Ferreira EC, Coelho MAZ. Image analysis application for the study of activated sludge floc size during the treatment of synthetic and real fishery wastewaters. Environ Sci Pollut Res. 2011;18:1390– 1397. Mielczarek AT, Cragelund C, Erikssen PS, Nielsen PH. Population dynamics of filamentous bacteria in Danish wastewater treatment plant with removal. Water Res. 2012;46:3781–3795. Grijspeerdt K, Verstraete W. Image analysis to estimate the settleability and concentration of activated sludge. Water Res. 1997;31:1126–1134. Da Motta M, Pons MN, Roche N. Automated monitoring of activated sludge in a pilot plant using image analysis. Water Sci Technol. 2001;43:91–96. Da Motta M, Pons MN, Roche N. Monitoring filamentous bulking in activated sludge systems fed by synthetic or municipal wastewater. Bioprocess Biosyst Eng. 2003;25:387–393. Koivuranta E, Keskitalo J, Haapala A, Stoor T, Sarén M, Niinimäki J. Optical monitoring of activated sludge flocs in bulking and non-bulking conditions. Environ Technol. 2013;34:679–686. Koivuranta E, Keskitalo J, Stoor T, Haapala A, Sarén M, Niinimäki J. A comparison between floc morphology and the effluent clarity at a full-scale activated sludge plant using optical monitoring. Environ Technol. 2013;35:1605–1610. Amaral AL, Ferreira EC. Activated sludge monitoring of a wastewater treatment plant using image analysis and partial least squares regression. Anal Chim Acta. 2005;544:246– 253. Mesquita DP, Dias O, Amaral AL, Ferreira EC. Monitoring of activated sludge settling ability through image analysis: validation on full scale wastewater treatment plants. Bioprocess Biosyst Eng. 2009;32:361–367. Barbusi´nski K, Ko´scielniak H. Influence of substrate loading intensity on flocs size in activated sludge process. Water Res. 1995;29:1703–1710. Wilén BM, Balmér P. The effect of dissolved oxygen concentration on the structure, size and size distribution of activated sludge flocs. Water Res. 1999;33:391–400. Govoreanu R, Saveyn H, van der Meeren P, Nopens I, Vanrolleghem PA. A methodological approach for direct quantification of the activated sludge floc size distribution using different techniques. Water Sci Technol. 2009;60:1857– 1897. McDonald JH. Handbook of biological statistics. 2nd ed. Baltimore (MD): Sparky House Publishing; 2009. Eikelboom DH, van Buijsen HJJ. Handbuch für die mikrobiologische Schlammuntersuchung. Munich: F Hirthammer Verlag GmbH; 1992. Liwarska-Bizukojc E, Bizukojc M. Effect of selected anionic surfactants on activated sludge flocs. Enzyme Microb Technol. 2006;39:660–668. Bitton G. Wastewater microbiology. 4th ed. Hoboken (NJ): John Wiley & Sons Inc.; 2010. Limpert E, Stahel WA, Abbt M. Log-normal distribution across the sciences: keys and clues. BioScience. 2001;51:341–352.
Environmental Technology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tent20
Variations of morphology of activated sludge flocs studied at full-scale wastewater treatment plants a
b
a
Ewa Liwarska-Bizukojc , Anna Klepacz-Smółka & Olga Andrzejczak a
Institute of Fermentation Technology and Microbiology, Lodz University of Technology, Wolczanska 171/173, Lodz 90-924, Poland b
Department of Bioprocess Engineering, Lodz University of Technology, Wolczanska 213, Lodz 90-924, Poland Accepted author version posted online: 03 Nov 2014.Published online: 26 Nov 2014.
Click for updates To cite this article: Ewa Liwarska-Bizukojc, Anna Klepacz-Smółka & Olga Andrzejczak (2014): Variations of morphology of activated sludge flocs studied at full-scale wastewater treatment plants, Environmental Technology, DOI: 10.1080/09593330.2014.982717 To link to this article: http://dx.doi.org/10.1080/09593330.2014.982717
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Environmental Technology, 2014 http://dx.doi.org/10.1080/09593330.2014.982717
Variations of morphology of activated sludge flocs studied at full-scale wastewater treatment plants Ewa Liwarska-Bizukojca∗ , Anna Klepacz-Smółkab and Olga Andrzejczaka a Institute
of Fermentation Technology and Microbiology, Lodz University of Technology, Wolczanska 171/173, Lodz 90-924, Poland; b Department of Bioprocess Engineering, Lodz University of Technology, Wolczanska 213, Lodz 90-924, Poland
Downloaded by [University of Otago] at 21:48 04 January 2015
(Received 3 October 2013; accepted 31 August 2014 ) Digital image analysis has been intensively developed over the last two decades including its application to describe morphology of activated sludge flocs. However, only few studies concerned the variation of flocs morphology with respect to the operational conditions, particularly oxido-reductive conditions, in a full-scale wastewater treatment plant (WWTP). In this work, morphology of activated sludge flocs was monitored over one year in two different full-scale WWTPs. The main aim of this study was to find the relationship between the operational parameters and morphology of sludge flocs. Simultaneously, the variations in floc size along activated sludge chamber were studied with respect to the oxido-reductive conditions. It was found that the sludge loading rate was one of the most important operational parameters influencing floc size. It was estimated that its values higher than 0.1 kg BOD5 kg TS−1 d−1 contributed to the decrease in floc size. Also, the oxido-reductive conditions influenced the floc size. It was statistically proved that flocs from the anaerobic zone were usually smaller than flocs from the anoxic or aerobic zones. Distribution of floc size in a full-scale WWTP usually could be described by a log-normal model. Keywords: activated sludge; distribution; floc size; image analysis; operational parameters
Introduction Activated sludge systems are commonly applied in wastewater treatment plants (WWTPs) nowadays. After leaving the activated sludge chamber (biological step), the flocculated biomass of activated sludge is usually separated from the treated effluent in settling tanks. The effectiveness of solid–liquid separation and, as a consequence, the quality of the treated effluent depends on the morphology of activated sludge flocs.[1,2] Morphological parameters of the activated sludge flocs can be divided into two categories: (1) parameters describing floc size and (2) parameters describing flocs shape. The following parameters belong, inter alia, to the first category: projected area, equivalent diameter and perimeter. The second category covers circularity, convexity and form factor.[3] Out of the above-mentioned morphological parameters, equivalent diameter (Deq ) is most often used. In accordance with its value, flocs can be classified as small, medium-sized or large. In agreement with the classification elaborated by Eikelboom,[4] flocs of diameter below 25 μm are small, flocs of diameter between 25 and 250 μm are medium-sized and flocs of diameter above 250 μm are large. This classification is still in use irrespective of the fact that whether the traditional (manual) or modern
*Corresponding author. Email: [email protected] © 2014 Taylor & Francis
(computational) microscopic techniques are employed to measure the flocs.[5–7] Techniques of digital image analysis have been intensively developed over the last two decades. They were also adopted to measure morphological parameters of activated sludge flocs. The digital image analysis enables a fast, objective and comprehensive description of floc morphology. It was demonstrated that image analysis was a useful tool to characterize activated sludge biomass in terms of floc size and its relation to the settling properties and concentration of biomass.[8–12] Mesquita et al. [5] proved that the image analysis methodology was adequate for a continuous examination of the morphological changes of aggregates and subsequent prediction of the sludge volume index (SVI). At the same time, image analysis combined with chemometric techniques can be applied to identify activated sludge abnormalities, namely pinpoint floc formation and filamentous and zoogleal bulking.[3] Although image analysis was used several times in the monitoring of activated sludge in full-scale WWTPs,[5,10,12–14] the relationship between morphological parameters of the activated sludge flocs and operational or environmental conditions has not been investigated so far. What is more, activated sludge was usually taken from one sampling
2
E. Liwarska-Bizukojc et al.
Downloaded by [University of Otago] at 21:48 04 January 2015
Figure 1. Scheme of the activated sludge system of the WWTP Zgierz. Variations of mean values of floc diameter along the activated sludge bioreactor in the sampling period.
point located in the aeration zone of an activated sludge system.[5,11,14] It means that the variability of sludge floc morphology along the activated sludge chamber containing zones with different oxido-reductive conditions has not been studied yet. Such operational conditions as sludge age, loading rate, temperature, mixing intensity and dissolved oxygen concentration play a key role in the biological treatment of wastewater in activated sludge systems. Their changes influence the composition, structure and subsequently the morphology of activated sludge flocs. It was found that particularly floc size and its distribution provided useful information on changes of the operational conditions.[15–17] The application of computational image analysis techniques should support the search for correlations between operational conditions of WWTP and morphological parameters of activated sludge flocs. Nevertheless, one should be aware that each real WWTP is an independent stochastic system. In this work, two full-scale WWTPs were monitored over one year. The main aim of this study was to find a relationship between operational parameters and the morphology of sludge flocs. Simultaneously, variations in floc size along the activated sludge chamber were studied with respect to the oxido-reductive conditions.
Materials and methods Description of the WWTP Zgierz Activated sludge was taken from two full-scale WWTPs, i.e. a WWTP in Zgierz and the Combined WWTP in Lodz. Both WWTPs tested treat municipal wastewater from the cities of Zgierz and Lodz, and several communes located in the neighbourhood of these cities as well as industrial wastewater originating mainly from small- and mediumsize plants. A separate sewer system dominates in Zgierz, while in Lodz a combined sewer system is predominant. The average pollutant load to the plant corresponds to approximately 94,000 PE in the WWTP Zgierz, while in
Figure 2. Scheme of the activated sludge system of the Combined WWTP in Lodz. Variations of mean values of floc diameter along the activated sludge bioreactor in the sampling period.
the Combined WWTP Lodz it corresponds to 1 × 106 PE. The contribution of industrial wastewater is usually in the range from 10% to 15% in both WWTPs tested. In the WWTP Zgierz, the biological stage consists of one fivezone bioreactor and secondary clarifier run in the Phoredox process configuration (Figure 1), whereas in the Combined Wastewater Treatment Plant it consists of seven lines of three-zone bioreactor and secondary clarifiers run in the UCT process configuration (Figure 2). In 2012, the average inflow of wastewater was 8720 m3 d−1 in the WWTP Zgierz, while in the Combined WWTP in Lodz it was 174 540 m3 d−1 . The values of other operational parameters of the WWTPs investigated in this study are given in Table 1. Sampling Activated sludge samples were taken from each zone of activated sludge chambers once a month in dry weather conditions. Each time a sample of volume 500 ml was taken from each part of activated sludge chamber and analysed within two hours. In the WWTP Zgierz, the sampling period comprised the whole year 2012, while in the Combined WWTP Lodz it was from March 2012 to February 2013.
3
Environmental Technology
Downloaded by [University of Otago] at 21:48 04 January 2015
Table 1.
Characteristics of operational parameters of the WWTPs studied in the sampling period of time.
Characteristic feature
WWTP Zgierz
Combined WWTP Lodz
Type of activated sludge system Range of average daily flow rate (m3 d−1 ) Average recirculation of activated sludge (m3 d−1 ) COD (mg O2 l−1 ) Average sludge loading rate (g BOD g TS−1 d−1 ) Temperature (°C) SRT (d) TS (g kg−1 ) SVI (ml g TSS−1 )
Phoredox 5510–11,797 9307–12,464 744–4700 0.01–0.132 8.5–22.0 10.2–24.4 4.21–5.88 41–85
UCT (three stages) 149,998–242,311 145,100–237,464 518–985 0.04–0.07 11.9–22.2 7.3–11.5 3.79–4.68 75–135
Image analysis Three vital unstained slides of each activated sludge sample were prepared for image analysis. The activated sludge flocs were observed under a Nikon Eclipse Ni light microscope with 4 × objective. The RGB (red, green, blue model of colour) images were snapped, processed and analysed with the help of NIS ELEMENTS AR software (Nikon). At least 40 images from each sample were processed and analysed with the help of the special automated procedure aimed at measuring the basic morphological
parameters of the flocs, namely projected area, perimeter, equivalent diameter, circularity and convexity. From several to about 60 objects (flocs) were observed in each image. The automated procedure consisted of the following steps: (1) transformation of RGB image to the greyscale image, (2) segmentation (the same level of threshold for all images analysed was set) and (3) measurement of morphological parameters of the activated sludge flocs. The projected area is the basic image analysis parameter found easily by pixel count and its multiplication by
(a)
(b)
Figure 3. Variations of floc diameter in each zone of the activated sludge chamber in the sampling period (a) the Combined WWTP in Lodz and (b) the WWTP Zgierz (the number of objects, i.e. flocs, analysed in each sample of activated sludge varied from 980 to 2450).
4
E. Liwarska-Bizukojc et al.
a scaling factor. The equivalent diameter was measured as the mean value of lengths of the lines between two points on the boundary of the individual object going through its centroid. The perimeter is simply the length of the boundary of the object, while the convex perimeter is the length of the boundary obtained after filling of all concavities of the object. The convexity is the ratio of perimeter to convex perimeter. Circularity is the shape factor that indicates to what extent the measured object is similar to the true circle. If it is equal to one, the object is the true circle. The lower it is, the less circular is the object.
Downloaded by [University of Otago] at 21:48 04 January 2015
Other measurements and analysis Filaments index (range from 0 to 5) was recorded according to Eikelboom.[4] The dominating filaments were identified based on conventional techniques, i.e. morphological and staining (Neisser and Gram) properties. The values of operational parameters as well as the results of physicochemical analysis made in the sampling time were obtained from the WWTPs studied.
Statistical measures and methods The basic statistical elaboration comprising the calculation of mean values, standard deviation (σ ) and goodness of fit for normal distribution of morphological parameters of activated sludge flocs was prepared with the use of MS Excel. Standard deviations (σ ) presented further in Figures 1–3 and 7 express the variability of the values of the morphological parameters shown. The Kolmogorov–Smirnov test (p < 0.05) was used to check the goodness of fit for normal distribution. Several distribution models including log-normal, Gaussian, Lorentz and Voigt were tested to approximate flocs diameter distribution. Additionally, one-way analysis of variance (ANOVA) was applied to estimate if the values of the projected area of activated sludge flocs coming from different parts of the activated sludge chamber were statistically equal. The null hypothesis stating that they were equal was assumed. The measured values of the projected area were log-transformed in order to fulfil the normality assumption.[18] ANOVA implemented in MS Excel (Analysis ToolPak) software was used. The confidence level of 95% was assumed.
Figure 4. Variations of mean projected area of flocs vs. sludge loading rate, SRT and SVI (a) in the Combined WWTP in Lodz and (b) in the WWTP Zgierz (the number of objects, i.e. flocs, analysed in each sample of activated sludge varied from 980 to 2450).
Downloaded by [University of Otago] at 21:48 04 January 2015
Environmental Technology Results and discussion Flocs size The elaborated procedure of image analysis occurred to be successful in fast and automated measurement of morphological parameters of activated sludge flocs. Out of the morphological parameters of flocs measured in this work, those describing floc sizes were particularly important from the practical point of view. Thus, they will be presented and analysed first. Mean equivalent diameters of activated sludge flocs in the WWTP Zgierz varied from 117.1 to 170.9 μm depending on the month and zone of activated sludge chamber, while in the Combined WWTP in Lodz they ranged from 81.1 to 106.7 μm (Figure 3). It meant that the activated sludge flocs could be classified as medium size in both WWTPs studied.[4,7] Mielczarek et al. [7] found that in Danish WWTPs, activated sludge flocs of large size (with a diameter above 250 μm) dominated. However, the diameter of flocs was measured manually, and visual characteristics of flocs were performed in agreement with Eikelboom [4] protocol.[7] At the same time, some previous works [19,20] indicated that activated sludge flocs of equivalent diameter below 100 μm were relatively small and might cause an increase in SVI or even increase in the turbidity of effluent. The latter was not observed because the concentration of suspended solids in the effluent in both WWTPs studied did not exceed 25 mg l−1 . The values of SVI were generally higher in the Combined WWTP in Lodz (Table 1), in which flocs were smaller on average than in the WWTP Zgierz (Figure 4). The values of SVI in the Combined WWTP of Lodz were usually above 100 ml g TS−1 (in 10 out of 12 analysed months); however, they did not exceed 150 ml g TS−1 (Figure 4) which indicated bulking conditions.[21] It meant that the elevated values of SVI in the Combined WWTP in Lodz may have been caused by the disintegration of activated sludge flocs as well as a higher number of filamentous bacteria observed in the activated sludge from this WWTP. No direct correlation between SVI and floc size expressed by the mean projected area was found (Figure 4). At the same time, filaments’ index for sludge from the WWTP Zgierz was equal to 0 or 1, while in the Combined WWTP in Lodz it varied from 2 to 4. The dominating filamentous species were identified on the basis of their morphological characteristics and staining properties. In the Combined WWTP in Lodz Microthrix parvicella and type 0041 were co-dominant, while in the WWTP Zgierz type 0092 dominated.
5
earlier works [15,16] concerning floc size, the log-normal model was selected to describe floc size distribution. In both papers cited above, this model occurred to fit well with the experimental data; however, these were lab-scaleactivated sludge systems. In this work, a log-normal model expressed by the following equation is used: −(ln x/μ)2 A · exp , y = yo + √ 2 · σ2 2·π ·σ ·x
(1)
where x is the floc diameter, μ is the arithmetic mean, σ is the standard deviation, y 0 is the offset frequency and A is the area under the curve within limits ± σ . It occurred that the log-normal distribution fitted well with the data from the Combined WWTP in Lodz. The values of correlation coefficient R2 were in the range from 0.915 to 0.978 depending on the month and zone of the activated sludge chamber. The log-normal distribution is often met in various biological systems. The distribution of particles, chemicals and organisms in the environment is usually log-normal.[22]
Flocs size distribution The distribution of diameters of flocs coming from the WWTPs studied did not correspond to the normal distribution. It was confirmed by the Kolmogorov–Smirnov test and trials of the approximation of the experimental points by the Gaussian curve. Taking into account
Figure 5. Distribution of activated sludge flocs in each zone of the activated sludge chamber in November 2012 in the WWTP Zgierz.
Downloaded by [University of Otago] at 21:48 04 January 2015
6
E. Liwarska-Bizukojc et al.
Figure 6. Distribution of activated sludge flocs in each zone of the activated sludge chamber in November 2012 in the Combined WWTP in Lodz.
The distribution of diameters of flocs coming from the WWTP Zgierz usually fitted to the log-normal distribution. However, the values of R2 were often lower than those determined for the Combined WWTP in Lodz and in some cases they were below 0.900. Thus, attempts of approximation using other distribution curves, such as for example Lorentz or Voigt, were made. However, they were not successful. The values of R2 were at a similar or even lower level as they were obtained for the log-normal distribution. The distribution of activated sludge flocs in both WWTPs in November 2012 is shown in Figures 5 and 6 as an example. It is visible that in the activated sludge coming from the Combined WWTP in Lodz, flocs of diameter up to 60 μm dominated, whereas in the WWTP Zgierz the diversity of activated sludge floc size was significantly wider. This phenomenon was well reflected by the higher values of SD determined for flocs from the WWTP Zgierz (Figure 3). This diversity of floc size is the most probable reason of difficulties in a mathematical description of the distribution of these flocs. In turn, the reason of this
wide diversity of flocs size was the variation of operating conditions in the WWTP Zgierz (Table 1 and Figure 4). In the sampling period (from March 2012 to February 2013), the contribution of activated sludge flocs of a diameter below 60 μm was between 37% and 52% in the Combined WWTP in Lodz, whereas in the WWTP Zgierz it was from 8.5% to 31% in 2012. The values varied depending on the month and zone of the activated sludge chamber.
Effect of operating parameters on flocs morphology Despite the fact that real WWTPs are stochastic systems, the effects of operational parameters on activated sludge have been studied in many works.[7,10,13,14] Analysing the influence of operating parameters on floc morphology in this study, it was observed that sludge loading rate (SLR) and oxido-reductive conditions had a stronger effect on floc size in comparison with other parameters, as for example temperature or sludge concentration (TS). The increase in
Decembers 2012 + +↓ − + November 2012 +↓ +↓ +↓ +↓ Anaerobic vs. anoxic 1 Anaerobic vs. aerobic 1 Anaerobic vs. anoxic 2 Anaerobic vs. aerobic 2
− No statistically significant difference between flocs-projected areas. + The projected areas of flocs are statistically different. + ↓The projected area of flocs from the anaerobic zone is lower than the projected area of flocs from the other zones being compared.
October 2012 − – +↓ − September 2012 − +↓ − − August 2012 +↓ +↓ +↓ +↓ July 2012 +↓ +↓ +↓ +↓ June 2012 − – + − May 2012 − +↓ +↓ +↓ April 2012 − +↓ − +↓ March 2012 +↓ − − − February 2012 − − − − January 2012 − − − −
+↓ +↓ +↓ +↓ − + +↓ +↓ − +↓ + − + + +↓ +↓ +↓ − Anaerobic vs. anoxic Anaerobic vs. aerobic
− +↓ +↓ +↓ WWTP Zgierz
+↓ +↓
November 2012 October 2012 September 2012 August 2012 July 2012 June 2012 May 2012 April 2012 March 2012 Comparison
Combined WWTP Lodz
Table 2. Results of ANOVA test for both WWTPs studied.
Downloaded by [University of Otago] at 21:48 04 January 2015
December 2012
January 2013
February 2013
Environmental Technology
7
sludge loading rate above 0.1 kg BOD5 kg TS−1 d−1 contributed to a decrease in the mean projected area of flocs in the WWTP Zgierz. It is illustrated in Figure 4, in which the value of sludge loading equal to 0.1 kg BOD5 kg TS−1 d−1 was marked by a dashed line. In the Combined WWTP in Lodz sludge loading rate varied in the relatively narrow range and did not exceed 0.07 kg BOD5 kg TS−1 d−1 , thus the decrease in floc size in relation to the increase in sludge loading rate was not found. It was also noticed that a shorter sludge age favoured smaller flocs. In the WWTP Zgierz, smaller flocs were observed in the activated sludge characterized by short sludge retention time (SRT) and exposed to high organic loading (Figure 4). In May 2012, when the sludge loading was 0.133 kg BOD5 kg TS−1 d−1 and SRT was 10.4 d, the smallest flocs of the mean diameter equal to 117.1 μm were found. The occurrence of smaller flocs in the Combined WWTP in Lodz in comparison with the WWTP Zgierz may also be associated with the lower values of SRT used in this WWTP. However, it should be emphasized that no mathematical correlation was found in this WWTP either between floc diameter (or projected area) and SRT or floc diameter and sludge loading rate. Barbusi´nski and Ko´scielniak [15] found that after the substrate overload event, the flocs were more prone to break-up. However, the long-term loading changes caused larger disturbances to the floc size distribution than more rapid but shorter ones.[15] In contrast to the earlier studies, in this work the variations of floc morphology along the activated sludge chambers at different oxido-reductive conditions were investigated. It was observed that activated sludge flocs from the anaerobic chamber were usually smaller than the flocs from the anoxic or aerobic chambers. On the basis of ANOVA, it was revealed in 19 out of 24 compared pairs of data sets that flocs from anaerobic zone in the Combined WWTP of Lodz were statistically different in terms of their size from the flocs of the other two zones. To be more precise, they were smaller than the flocs from the anoxic or aerobic chambers in 16 comparisons of data sets, i.e. in eight cases they were smaller than the flocs from the anoxic zone and in eight cases than the flocs from the aerobic zone (Table 2). In the WWTP Zgierz, activated sludge flocs coming from the anaerobic zone were statistically different in 24 out of 48 compared pairs of data sets (Table 2). What is more important, they were smaller than the flocs from the anoxic or aerobic zones in 21 cases. In Table 2, it is shown in which months and in which zone flocs coming from the anaerobic part were smaller than these from the other four zones of the bioreactor. The results obtained upon statistical analysis for two WWTPs indicated that the anaerobic conditions favoured the decrease in activated sludge flocs. It occurred that neither stronger shear forces nor limitation of carbonaceous substrate, which were more common in the aerobic zones of activated sludge chambers, contributed to the decrease
Downloaded by [University of Otago] at 21:48 04 January 2015
8
E. Liwarska-Bizukojc et al.
in floc size to such extent as oxido-reductive conditions and limitation of dissolved oxygen (DO). In both WWTPs studied, dissolved oxygen concentration was close to zero and redox potential was usually below − 120 mV in the anaerobic zone. DO concentration in the anoxic zone was very low and did not exceed 1 mg l−1 , while the redox potential was usually higher than in the anaerobic part of the bioreactor but varied in a wide range from − 200 to + 50 mV. In the aerobic zone, DO concentration was from 3.5 to 7.2 mg O2 l−1 in the WWTP Zgierz and from 1.9 to 4.6 mg O2 l−1 in the Combined WWTP of Lodz. It indicated that even a small increase in DO concentration or redox potential might have contributed to the increase in floc size. The observations regarding the decrease in floc size in the anaerobic conditions made in this work for two full-scale WWTPs are in agreement with the conclusions formulated by Wilén and Balmér [16] upon the laboratory-scale experiments. They claimed that alternating aerobic/anaerobic conditions, the proportion of smaller flocs (diameter up to 20 μm) increased gradually in the anaerobic period. Moreover, it was found that there was a trend towards larger flocs at a higher concentration of DO.[16] Variations of circularity of activated sludge flocs in both WWTPs studied are shown in Figure 7. The shape
of activated sludge flocs is generally far from the ideal circle, for which it should be equal to one, according to the definitions implemented in NIS Elements software. What is more, despite the significant difference in operational parameters as well as in floc size, the circularity of flocs in both WWTPs tested was at a similar level and varied in a relatively narrow range. In the WWTP Zgierz the mean values of circularity varied from 0.287 to 0.413 depending on the zone and month, while in the Combined WWTP Lodz they were in the range from 0.271 to 0.395 (Figure 7). No correlation between floc circularity and operational parameters was established. It concerns convexity too. Convexity varied from 0.715 to 0.789 in the WWTP Zgierz, whereas in the Combined WWTP Lodz it was from 0.677 to 0.767. In the Combined WWTP of Lodz, the flocs from the anaerobic zone were slightly more circular than those from the anoxic or aerobic ones. However, this trend was not observed in the WWTP Zgierz (Figure 7). Generally, the measurement of floc circularity in a full-scale WWTP is less useful than the measurement of the floc diameter or projected area. However, it may be beneficial in the cases when pollutants such as surface active agents are present in the influent.[20]
(a)
(b)
Figure 7. Variations of floc circularity in each zone of the activated sludge chamber (a) in the Combined WWTP in Lodz and (b) in the WWTP Zgierz (the number of objects, i.e. flocs, analysed in each sample of activated sludge varied from 980 to 2450).
Downloaded by [University of Otago] at 21:48 04 January 2015
Environmental Technology Conclusions (1) Sludge loading rate is one of the most important operational parameters influencing flocs size. There is no direct correlation between SLR and flocs area or diameter. It is estimated that SLR values higher than 0.1 kg BOD5 kg TS−1 d−1 contribute to the decrease in floc size, particularly if lower SLR dominates in everyday operation of a WWTP. The variation of SLR in a wide range (0.0015–0.132 kg BOD5 kg TS−1 d−1 ) induces bigger changes in floc size and as a consequence in floc distribution than the stable values of SLR (0.04–0.07 kg BOD5 kg TS−1 d−1 ). (2) Shorter sludge retention times seem to favour smaller activated sludge flocs (diameter below 100 μm), especially if accompanied by high sludge loading rate. (3) The distribution of flocs size in the full-scale WWTP may vary depending on the WWTP studied. It usually can be described by a log-normal model. (4) Flocs from the anaerobic zone are usually smaller than the flocs from anoxic or aerobic zones. It indicates that oxido-reductive conditions affect the floc size. Acknowledgements This work was made within the project No. 14-0004-10 financed by the National Centre for Research and Development, Republic of Poland.
[6]
[7]
[8] [9] [10]
[11]
[12]
[13]
[14]
[15]
Disclosure statement No potential conflict of interest was reported by the author(s).
[16]
References
[17]
[1] Mohammed RN, Abu-Alhail S, Xi-wu L. Long-term operation of a novel pilot-scale six tanks alternately operating activated sludge process in treating domestic wastewater. Environ Technol. 2013;35:1874–1885. [2] Villain M, Clouzot L, Guibaud G, Marrot B. Impact of oxygen cut off and starvation conditions on biological activity and physico-chemical properties of activated sludge. Environ Technol. 2013;34:901–910. [3] Mesquita DP, Amaral AL, Ferreira EC. Characterization of activated sludge abnormalities by image analysis and chemometric techniques. Anal Chim Acta. 2011;705:235– 242. [4] Eikelboom DH. Process control of activated sludge plants by microscopic investigation. London: IWA Publishing; 2000. [5] Mesquita DP, Dias O, Dias AMA, Amaral AL, Ferreira EC. Correlation between sludge settling ability and image
[18] [19] [20] [21] [22]
9
analysis information using partial least squares. Anal Chim Acta. 2009;642:94–101. Mesquita DP, Ribeiro RR, Amaral AL, Ferreira EC, Coelho MAZ. Image analysis application for the study of activated sludge floc size during the treatment of synthetic and real fishery wastewaters. Environ Sci Pollut Res. 2011;18:1390– 1397. Mielczarek AT, Cragelund C, Erikssen PS, Nielsen PH. Population dynamics of filamentous bacteria in Danish wastewater treatment plant with removal. Water Res. 2012;46:3781–3795. Grijspeerdt K, Verstraete W. Image analysis to estimate the settleability and concentration of activated sludge. Water Res. 1997;31:1126–1134. Da Motta M, Pons MN, Roche N. Automated monitoring of activated sludge in a pilot plant using image analysis. Water Sci Technol. 2001;43:91–96. Da Motta M, Pons MN, Roche N. Monitoring filamentous bulking in activated sludge systems fed by synthetic or municipal wastewater. Bioprocess Biosyst Eng. 2003;25:387–393. Koivuranta E, Keskitalo J, Haapala A, Stoor T, Sarén M, Niinimäki J. Optical monitoring of activated sludge flocs in bulking and non-bulking conditions. Environ Technol. 2013;34:679–686. Koivuranta E, Keskitalo J, Stoor T, Haapala A, Sarén M, Niinimäki J. A comparison between floc morphology and the effluent clarity at a full-scale activated sludge plant using optical monitoring. Environ Technol. 2013;35:1605–1610. Amaral AL, Ferreira EC. Activated sludge monitoring of a wastewater treatment plant using image analysis and partial least squares regression. Anal Chim Acta. 2005;544:246– 253. Mesquita DP, Dias O, Amaral AL, Ferreira EC. Monitoring of activated sludge settling ability through image analysis: validation on full scale wastewater treatment plants. Bioprocess Biosyst Eng. 2009;32:361–367. Barbusi´nski K, Ko´scielniak H. Influence of substrate loading intensity on flocs size in activated sludge process. Water Res. 1995;29:1703–1710. Wilén BM, Balmér P. The effect of dissolved oxygen concentration on the structure, size and size distribution of activated sludge flocs. Water Res. 1999;33:391–400. Govoreanu R, Saveyn H, van der Meeren P, Nopens I, Vanrolleghem PA. A methodological approach for direct quantification of the activated sludge floc size distribution using different techniques. Water Sci Technol. 2009;60:1857– 1897. McDonald JH. Handbook of biological statistics. 2nd ed. Baltimore (MD): Sparky House Publishing; 2009. Eikelboom DH, van Buijsen HJJ. Handbuch für die mikrobiologische Schlammuntersuchung. Munich: F Hirthammer Verlag GmbH; 1992. Liwarska-Bizukojc E, Bizukojc M. Effect of selected anionic surfactants on activated sludge flocs. Enzyme Microb Technol. 2006;39:660–668. Bitton G. Wastewater microbiology. 4th ed. Hoboken (NJ): John Wiley & Sons Inc.; 2010. Limpert E, Stahel WA, Abbt M. Log-normal distribution across the sciences: keys and clues. BioScience. 2001;51:341–352.
