Abstract

Operation of conventional drinking water treatment plants (DWTPs) is limited by seasonal variations in the characteristics of raw water, such as algae levels and turbidity. In this study, an integrated flotation and sedimentation unit (IFSU) with side flow-inclined plate settlers (SFIPS) was applied to enhance the removal ratio of algae and turbidity in a full-scale field experiment. To accomplish this, the performance of two IFSU reactors (with and without SFIPS) treating algal-rich water, low-turbidity water and high-turbidity water was compared. Extensive experiments were conducted based on a central composite face-centered design, and the results were analyzed using response surface methodology. Polyaluminium chloride dose, NaOH dose, influent turbidity level and influent chlorophyll level were selected as the operating variables to analyze, optimize and model the process. The results showed that SFIPS functioned as a counterflow inducer to enhance the effects of stratified flow patterns and decrease the clarification loading rate in the separation zone during the dissolved air flotation process, leading to considerable increases in algae and turbidity removal in the IFSU reactor. The results of this study will facilitate upgrades of conventional DWTPs to improve their potential.

INTRODUCTION

Conventional drinking water treatment processes, including flocculation, dissolved air flotation (DAF), sedimen­tation, filtration and disinfection, target the removal of impurities such as turbidity, algae and organic matter, etc., from raw wa­ter (Hohner et al. 2016). As freak weather events and anthropogenic activity (including changes in landscape management, inputs of nitrogen and phosphorus amongst others), seasonal variations in the characteristics of raw water have presented more frequently in surface water, resulting in algal-rich water, low-turbidity water and high-turbidity water. These fluctuations can lead to a number of downstream problems such as algae or turbidity breakthrough, filter clogging (Buisine & Oemcke 2003), the presence of toxins (Al-Tebrineh et al. 2010) and the formation of harmful disinfection by-products (Chen et al. 2008). Thus, seasonal variations in the characteristics of raw water have become a great challenge for DWTPs.

When addressing the treatment of algal-rich water in DWTPs, ultrafiltration (UF) and flocculation-DAF have an edge over flocculation-sedimentation. But UF and flocculation-DAF would be ineffective when applied to high-turbidity water during rainy days (Henderson et al. 2010), while flocculation-sedimentation is reportedly more effective for the treatment of high-turbidity water (Li & Gregory 1991). Therefore, a novel water treatment system denoted IFSU (Li et al. 1993) was invented. The main advantage of the IFSU is that it can run flotation or sedimentation processes in a single unit. In these systems, flotation is induced when algal-rich or low-turbidity water is present, while sedimentation is induced when treating high-turbidity water. Since the invention of the IFSU in 1993 (Li et al. 1993), it has been successfully applied in the reservoirs of full-scale water treatment plants with volumes of up to 2 M l/d (Liu et al. 2012; Sun & Liu 2014).

In this study, we used a full-scale in-situ IFSU for the treatment of reservoir water. The main goal of the study was to compare the process performance of two IFSU reactors, with and without SFIPS. Response surface methodology (RSM) was employed to describe and model the variation trends of four significant responses for the treatment of algal-rich water, the removal efficiency of chlorophyll, turbidity, permanganate index (CODMn) and ammonia nitrogen (NH4-N) as a function of three independent variables, polyaluminum chloride (PAC) dose, NaOH dose and influent chlorophyll (Chlin) level. For the treatment of high or low-turbidity water, PAC dose, NaOH dose and influent turbidity (Turin) were selected as the variables, while the removal efficiency of turbidity, CODMn and NH4-N were the responses. The results presented herein will facilitate upgrades of conventional DWTPs to enable the treatment of algal-rich, low-turbidity and high-turbidity water.

METHODS

IFSU reactor and operation

Our experiments were carried out over a few years in a full-scale IFSU plant in the Changkeng Reservoir (N22°24′26.45″, E113°23′48.86″; Zhongshan, Guangdong, China) that was used to generate drinking water. Based on the water quality between 2010 and 2013 (Table 1), the water in the system was made up of algal-rich water, low-turbidity water and low-temperature or high-turbidity water on rainy days.

Table 1

Main characteristics of the Changkeng Reservoir water

Parameters Range Unit 
Chlorophyll 50–10,000 mg/l 
Turbidity 3–300 NTU 
CODMn 0.5–6 mg/l 
Ammonia 0.05–0.6 mg/l 
Parameters Range Unit 
Chlorophyll 50–10,000 mg/l 
Turbidity 3–300 NTU 
CODMn 0.5–6 mg/l 
Ammonia 0.05–0.6 mg/l 

In this study, the 5 M l/d IFSU was divided into two portions, a SFIPS and a non-SFIPS IFSU system. The loading rate was 2.5 M l/d for both systems. Figure 1 shows the schematic diagram of the IFSU with SFIPS, and thumbnail (a) in this figure displays the exterior of the IFSU. The IFSU system consisted of a data monitor section, coagulation section, dissolved air canister (thumbnail (b)), dissolved air releaser (thumbnail (c)), SFIPS (thumbnail (d)) and a scraper and valve. A dissolved air canister maintained a saturator pressure of 0.35–0.40 MPa. The SFIPS inclined at an angle of 60° to the horizontal, and had a perpendicular spacing of 5 cm and a maximum length of 4 m.

Figure 1

Flow diagram of the integrated flotation and sedimentation unit (IFSU). Thumbnail (a) shows the exterior of the IFSU, (b) is a snapshot of the dissolved air canister, (c) is a close-up view of the DAF releaser and (d) displays the side flow-inclined plates and scraper.

Figure 1

Flow diagram of the integrated flotation and sedimentation unit (IFSU). Thumbnail (a) shows the exterior of the IFSU, (b) is a snapshot of the dissolved air canister, (c) is a close-up view of the DAF releaser and (d) displays the side flow-inclined plates and scraper.

Based on the flow process, raw water was first passed into the coagulation section. At this stage, PAC and NaOH were fed as the flocculant and assistant flocculant, respectively. Then, according to the algae and turbidity data acquired from the data monitoring section, the IFSU switched between two options: DAF and sedimentation. For the treatment of algal-rich water or low-turbidity water (0–20 NTU), the recycling ratio of the DAF was set to 8–10%. The DAF process was run by opening scraper A and valve B. The effluent of this process was collected by a conduit set at the bottom of the SFIPS. However, the DAF process would not perform well when the turbidity was greater than 20 NTU. In such cases, the sedimentation process was launched with the opening of scraper B and valve A, as well as the closing of the dissolved air canister, scraper A and valve B.

Experimental design and mathematical model

To describe the interactive effects of the variables studied on the process and eliminate systematic errors while estimating the experimental error and minimizing the number of experiments needed, the central composite design was applied (Kontogiannopoulos et al. 2016). Categorical factors can also be added to this design, which will cause the number of runs generated to be multiplied by the number of combinations of the categorical factor levels. The Design Expert software (Stat-Ease Inc., version 8) was used for the statistical design of experiments and data analysis.

Three independent numerical variables, PAC dose, NaOH dose and Turin, along with SFIPS as a categorical factor, were selected for the treatment of turbidity design. For treatment of algal-rich water, Chlin concentration replaced Turin as the independent numerical variable, while the other independent numerical variables (PAC dose and NaOH dose) and categorical factors were employed in the turbidity experiment design. To fix conditional experiments, the deviation allowance of Turin or Chlin was set to ±5%, while PAC dose and NaOH dose were adjusted according to Turin or Chlin. The design consisted of 2k factorial points augmented by 2k axial points and a center point, where k is the number of numerical variables. Based on the numerical variables, 20 experiments (= 2k + 2k + 6) were designed, with 15 experiments organized in a factorial design (including nine factorial points, six axial points and one center point), and the remaining five involving replication of the central point to obtain a good estimate of the experimental error.

Because a categorical variable (SFIPS) in two levels has also been considered, 40 experiments were conducted. The range and levels of the variables in coded and actual units are given in Table 2, and its deviation allowance was set to ±5%. As in Table 2, Turin and Chlin were the original data of raw water during 2010–2013. Herein, the corresponding responses values (the removal efficiency of chlorophyll, turbidity, CODMn and NH4-N) were documented, only when the levels of Turin and Chlin meeting the permutation of variables in Supplementary Material Tables S1–S3 (available with the online version of this paper). Turbidity, CODMn and NH4-N removal efficiency were measured or calculated as responses to treatment for turbidity, while turbidity, CODMn, NH4-N and chlorophyll removal efficiency were evaluated as responses to treatment of algal-rich water. The experimental conditions and results obtained for treatment of algal-rich water, low-turbidity water and high-turbidity water are shown in the Supplementary Material.

Table 2

Experimental range and levels of independent variables. Chlin indicates influent chlorophyll and Turin indicates influent turbidity

Type of variables Name of variables Range and level
 
− 1 
 Treatment for algal-rich water    
  PAC mg/l 12 
  NaOH mg/l 10 
  Chlin mg/l 1,000 4,000 7,000 
 Treatment for low turbidity water 
Numerical  PAC mg/l 
  NaOH mg/l 1.6 2.6 3.6 
  low-Turin NTU 12 20 
 Treatment for high turbidity water 
  PAC mg/l 22 40 
  NaOH mg/l 8.5 15 
  high-Turin NTU 20 110 200 
Categorical Side flow inclined plate settlers (SFIPS) Two levels (with SFIPS (+) and without SFIPS (−)) 
Type of variables Name of variables Range and level
 
− 1 
 Treatment for algal-rich water    
  PAC mg/l 12 
  NaOH mg/l 10 
  Chlin mg/l 1,000 4,000 7,000 
 Treatment for low turbidity water 
Numerical  PAC mg/l 
  NaOH mg/l 1.6 2.6 3.6 
  low-Turin NTU 12 20 
 Treatment for high turbidity water 
  PAC mg/l 22 40 
  NaOH mg/l 8.5 15 
  high-Turin NTU 20 110 200 
Categorical Side flow inclined plate settlers (SFIPS) Two levels (with SFIPS (+) and without SFIPS (−)) 
After conducting the experiments, the coefficients of the polynomial model were calculated as follows (Khuri & Cornell 1996):  
formula
(1)
where i and j are the linear and quadratic coefficients, respectively, and is the regression coefficient. Analysis of variance (ANOVA) was used for graphical analyses of the data to obtain the interaction between the process variables and the responses. The P value with a 95% confidence level was considered to indicate the effectiveness of the model terms.

Analytical methods

Chlorophyll was determined using the method described by María et al. (2012) and calculated using Liechtenthaler equations (Lichtenthaler 1987). Turbidity was measured using a Hach 2100P turbidity meter (Hach Co., Loveland, Colorado, USA). NH4-N was analyzed by the colorimetric method according to Standard Methods (Nessler Method) (APHA New York 2005). The method USEPA (1983) was applied to determine CODMn.

RESULTS AND DISCUSSION

Results

Model fitting and statistical analysis

Two IFSU reactors, one with and the other without SFIPS, were applied to conduct the aforementioned experiment. The experimental conditions and results for the two IFSU reactors are shown in Supplementary Tables S1–S3, while different degree polynomial models were also presented in Supplementary Table S4 (available with the online version of this paper).

Reactor performance

The ANOVA data for chlorophyll removal responses are shown in Supplementary Table S4. The experimental data were fitted to a reduced quadratic model. Figure 2(a)2(b) shows the 3D surface plots of the response as a function of PAC and NaOH dose at the mean level of Chlin for SFIPS and non-SFIPS IFSU systems. Comparison of the results presented in Figure 2(a) along with those in Figure 2(b) revealed that the SFIPS had an increasing impact on the mean level of Chlin (4,000 mg/l). As shown in Figure 2(c), there was good convergence between the experimental and predicted values of the chlorophyll removal efficiency. The interaction between SFIPS and PAC is presented in Figure 2(d), while the interaction between SFIPS and NaOH is depicted in Figure 2(e)2(f). According to the aforementioned interactional diagrams in Figure 2, higher chlorophyll removal efficiency (in Figure 2(d), 98.3% with SFIPS comparing to 95.1% without SFIPS at condition of 4 mg/l PAC and 6 mg/l NaOH, while 98.1% with SFIPS comparing to 93.6% without SFIPS at condition of 12 mg/l PAC and 6 mg/l NaOH; in Figure 2(e), 98.5% with SFIPS comparing to 94.4% without SFIPS at condition of 8 mg/l PAC and 2 mg/l NaOH, while 96.7% with SFIPS comparing to 93.5% without SFIPS at condition of 8 mg/l PAC and 10 mg/l NaOH) was documented in the presence of SFIPS.

Figure 2

Response surface plots for chlorophyll removal as a function of PAC and NaOH at Chlin concentration of 4,000 mg/l (a) with SFIPS and (b) without SFIPS; (c) plot of predicted vs. actual values for chlorophyll removal; (d) interactive effects of PAC-SFIPS on chlorophyll removal at a NaOH dose and Chlin level of 4 mg/l and 4,000 mg/l, respectively; interactive effects of NaOH-SFIPS on chlorophyll removal at a PAC dose and Chlin level of (e) 8 mg/l and (f) 4,000 mg/l.

Figure 2

Response surface plots for chlorophyll removal as a function of PAC and NaOH at Chlin concentration of 4,000 mg/l (a) with SFIPS and (b) without SFIPS; (c) plot of predicted vs. actual values for chlorophyll removal; (d) interactive effects of PAC-SFIPS on chlorophyll removal at a NaOH dose and Chlin level of 4 mg/l and 4,000 mg/l, respectively; interactive effects of NaOH-SFIPS on chlorophyll removal at a PAC dose and Chlin level of (e) 8 mg/l and (f) 4,000 mg/l.

Instability of effluent turbidity is the main problem associated with DWTPs, especially when they are operating at high loading in the face of seasonal variations in raw water characteristics. Figure 3(a)3(b) shows the 3D surface plots of PAC and NaOH with respect to turbidity for the treatment of low-turbidity water (0–20 NTU) using the DAF process. As shown in Figure 3, the SFIPS had an increasing effect on response at the mean Turin (12 NTU). However, it should be noted that an increasing effect of PAC dose and decreasing effect of NaOH dose were found at Turin of 12 NTU. Figure 3(c)3(d) show similar trends with different values. Figure 3(e) reveals relatively good agreement between the experimental and predicted values of turbidity removal ( in Supplementary Table S4).

Figure 3

Response surface plots for turbidity removal as a function of PAC and NaOH at Turin of 12 NTU (a) with SFIPS and (b) without SFIPS; (c) interactive effects of PAC-SFIPS on turbidity removal at a NaOH dose and Turin of 2.6 mg/l and 12 NTU, respectively; (d) interactive effects of NaOH-SFIPS on turbidity removal at a PAC dose and Turin of 5 mg/l and 12 NTU, respectively; (e) plot of predicted vs. actual values for turbidity removal.

Figure 3

Response surface plots for turbidity removal as a function of PAC and NaOH at Turin of 12 NTU (a) with SFIPS and (b) without SFIPS; (c) interactive effects of PAC-SFIPS on turbidity removal at a NaOH dose and Turin of 2.6 mg/l and 12 NTU, respectively; (d) interactive effects of NaOH-SFIPS on turbidity removal at a PAC dose and Turin of 5 mg/l and 12 NTU, respectively; (e) plot of predicted vs. actual values for turbidity removal.

The results of ANOVA for the treatment of high-turbidity water (20–200 NTU) using the sedimentation process are presented in Supplementary Table S4. The NaOH dose was the minor impact factor and is shown in Figure 4(a)4(b). As shown in Figure 4(c), there was good agreement between the predicted and actual data (R2 = 0.9519 in Supplementary Table S4). In the presence of SFIPS (Figure 4(d)), the higher turbidity removal (94.0% with SFIPS comparing to 88.9% without SFIPS at condition of 40 mg/l PAC and 8.5 mg/l NaOH, while 80.1% with SFIPS comparing to 70.1% without SFIPS at condition of 4 mg/l PAC and 8.5 mg/l NaOH) was recorded, which is in accordance with the results of previous studies (Tarpagkou & Pantokratoras 2014; Ren et al. 2017). Owing to the increased surface area and decreased settling distance, sedimentation processes with SFIPS is able to separate particles more rapidly than conventional settlers (Laskovski et al. 2006; Sarkar et al. 2007). As shown in Figure 4(d), the maximum turbidity removal (94.3%) was found under SFIPS and the highest PAC dose (40 mg/l). High doses of flocculant are always necessary when treating high-turbidity water.

Figure 4

Response surface plots of turbidity removal as a function of PAC and Turin at a mean level of NaOH (8.5 mg/l) (a) with SFIPS and (b) without SFIPS; (c) plot of predicted vs. actual values for turbidity removal; (d) interactive effects of PAC-SFIPS on turbidity removal at a NaOH dose and Turin of 8.5 mg/l and 200 NTU, respectively

Figure 4

Response surface plots of turbidity removal as a function of PAC and Turin at a mean level of NaOH (8.5 mg/l) (a) with SFIPS and (b) without SFIPS; (c) plot of predicted vs. actual values for turbidity removal; (d) interactive effects of PAC-SFIPS on turbidity removal at a NaOH dose and Turin of 8.5 mg/l and 200 NTU, respectively

As shown in Supplementary Tables S1–S3, the ratios of the minimum to maximum CODMn removal were 31.07%: 61.68%, 27.81%:62% and 16.18%:66.33% during the treatment of algal-rich water, low-turbidity water and high-turbidity water, respectively. In the presence of SFIPS, there was an advantage for CODMn removal with the SFIPS over the non-SFIPS (in Figure 5(a), 54.3% with SFIPS compared with 44.2% without SFIPS at condition of 4 mg/l PAC and 6 mg/l NaOH, while 53.8% with SFIPS compared with 42.1% without SFIPS at condition of 12 mg/l PAC and 6 mg/l NaOH; in Figure 5(b), 42.5% with SFIPS compared with 38.8% without SFIPS at condition of 4 mg/l PAC and 2.6 mg/l NaOH, while 38.3% with SFIPS compared with 34.9% without SFIPS at condition of 6 mg/l PAC and 2.6 mg/l NaOH; in Figure 5(c), 66.1% with SFIPS compared with 55.7% without SFIPS at condition of 40 mg/l PAC and 8.5 mg/l NaOH, while 39.7% with SFIPS compared with 32.3% without SFIPS at condition of 4 mg/l PAC and 8.5 mg/l NaOH). Comparison of the results presented in Figure 5(a)5(c) revealed that a high-level PAC dose (12 mg/l PAC in the treatment of algal-rich water, while 6 mg/l PAC in the treatment of low-turbidity water) had a decreasing effect on CODMn removal during the DAF process, while it showed increasing impact during the sedimentation process.

Figure 5

(a) Interactive effects of PAC-SFIPS on CODMn removal during treatment of algal-rich water at a NaOH dose and Chlin of 6 mg/l and 4,000 mg/l, respectively; (b) interactive effects of PAC-SFIPS on CODMn removal during treatment of low-turbidity water at a NaOH dose and Turin of 2.6 mg/l and 12 NTU, respectively; (c) interactive effects of PAC-SFIPS on CODMn removal during treatment of high-turbidity water at a NaOH dose and Turin of 8.5 mg/l and 110 NTU, respectively.

Figure 5

(a) Interactive effects of PAC-SFIPS on CODMn removal during treatment of algal-rich water at a NaOH dose and Chlin of 6 mg/l and 4,000 mg/l, respectively; (b) interactive effects of PAC-SFIPS on CODMn removal during treatment of low-turbidity water at a NaOH dose and Turin of 2.6 mg/l and 12 NTU, respectively; (c) interactive effects of PAC-SFIPS on CODMn removal during treatment of high-turbidity water at a NaOH dose and Turin of 8.5 mg/l and 110 NTU, respectively.

In terms of NH4-N removal, the DAF process was more effective than the sedimentation process. The minimum to maximum ratios were 25.31%:66.59% and 30.23%:72.15% for treatment of algal-rich water and low-turbidity water, respectively (Supplementary Tables S1–S2). However, the NH4-N removal ratio obtained during the treatment of high-turbidity water was only 10.9%:32.47% (Supplementary Table S3). As shown in Figure 6, the effects of SFIPS on NH4-N removal were similar to those observed for CODMn removal.

Figure 6

(a) Interactive effects of PAC-SFIPS on NH4-N removal during treatment of algal-rich water at a NaOH dose and Chlin of 6.0 mg/l and 4,000 mg/l, respectively; (b) interactive effects of PAC-SFIPS on NH4-N removal during treatment of low-turbidity water at a NaOH dose and Turin of 2.6 mg/l and 12 NTU, respectively.

Figure 6

(a) Interactive effects of PAC-SFIPS on NH4-N removal during treatment of algal-rich water at a NaOH dose and Chlin of 6.0 mg/l and 4,000 mg/l, respectively; (b) interactive effects of PAC-SFIPS on NH4-N removal during treatment of low-turbidity water at a NaOH dose and Turin of 2.6 mg/l and 12 NTU, respectively.

DISCUSSION

When SFIPS (Figures 24) was used, IFSU could improve the removal ratio of algae or turbidity separation. The effects of SFIPS occurred via the release of saturated dissolved water into the contact zone during the DAF process, which generated bubbles that caused floc-bubble-aggregates to rise to the surface of the tank (Edzwald 2010). Inside the SFIPS, some small bubbles would aggregate into large bubbles, therefore, small bubbles formed directional upflow along the lower sides of the SFIPS, while large bubbles provided non-directional flow along the upper sides of the SFIPS (Xing et al. 2006). Hence, large bubbles generated more lifting floc-bubble-aggregates than small bubbles. Consequently, the SFIPS investigated in our experiments functioned as a counterflow inducer.

The velocity of the floc-bubble-aggregates in the upper part of the separation zone was also simulated using a waterproof microscope video and computational fluid dynamics during the DAF process (Figure 7) (Li et al. 2017). Prior to the separation zone, we set a baffle to obtain a height of 3.8 m between the surface and the top edge of the baffle. The height produced a cross-flow (above the baffle) water velocity of approximately 200.8 m/h (55.8 mm/s, see Figure 7). These findings were in accordance with those of Lundh et al. (2002), who found that 37 m/h or greater water velocity (above the baffle) favored development of stratified flow in the separation zone. As shown in Figure 7, the highest velocity was found at the most distant point of the separation zone from the contact zone.

Figure 7

Velocity of particles in the upper part of the IFSU separation zone during the DAF process.

Figure 7

Velocity of particles in the upper part of the IFSU separation zone during the DAF process.

This streamlined distribution would conform to the stratified flow pattern of DAF, which was provided by Hazen theory and previous studies by Edzwald et al. (1999; Edzwald 2007). As shown in the conceptual picture of the stratified flow pattern (Figure 8), the horizontal flow near the top of the tank provides a certain clarification area (length × width of the separation zone), which is doubled with the return flow, and then finally tripled with the vertical flow toward the bottom. If one has three flow paths such as this, the clarification hydraulic loading is 1/3 the separation zone loading. In other words, a separation zone loading of 30 m/h reduces to 10 m/h in terms of the Hazen clarification loading (Edzwald 2010). Therefore, we deduced that SFIPS would greatly enhance the effects of the stratified flow pattern and decrease the clarification loading rate in the separation zone during the DAF process, leading to improvement of the algal-rich water and low-turbidity water. With the constant inflow during the DAF process in the IFSU, SFIPS would drastically decrease clarification hydraulic loading, which would increase the retention time of saturated dissolved water and the contact between saturated dissolved water and CODMn or NH4-N in the IFSU. These changes would account for the higher removal ratio of CODMn and NH4-N from IFSU with SFIPS.

Figure 8

Conceptual horizontal stratified flow pattern near the top of the separation zone and vertical plug flow below the bubble blanket. Reprinted from Edzwald (2010).

Figure 8

Conceptual horizontal stratified flow pattern near the top of the separation zone and vertical plug flow below the bubble blanket. Reprinted from Edzwald (2010).

Good coagulation chemistry depends on coagulant dose and pH, with the optimum conditions producing flocs with a charge near zero of zeta potentials and relatively high hydrophobicity. Nevertheless, our study showed that too much NaOH would hinder the removal ratio of algae and low turbidity (see Figures 2(e), 2(f) and 3(d)) during DAF. Since air bubbles is non-polar molecules and exert a negative charge in waters. These negative charges expressed negative zeta potentials, which is attributed to smaller anions that reside at the bubble–water interface at a greater concentration than larger hydrated cations (Edzwald 2010). Therefore, we hypothesized too much NaOH would release abundant hydrated cations, which squeezing out anions and changing the bubble–water interface from negative zeta potentials to positive zeta potentials. Consequently, the positively charged bubble–water interface would affect the interaction between bubble and PAC particle floc, which lead to the lower removal ratio obtained from flocs with high NaOH dose during DAF. However, more micro-data need to be analyzed to elucidate the above experiment result.

In terms of treating high-turbidity water via IFSU, the performance of the sedimentation process was equivalent to that of other settling plants, and it overcame the ineffectiveness of the DAF process for treatment of high-turbidity water. The IFSU integrates the flotation and sedimentation process in a single unit, therefore, it has a smaller plant footprint, which is a significant advantage for facilities in large cities. Moreover, the need for only one structure for DAF and settling reduces costs when compared with those of conventional plants, which have a horizontal layout of separate units.

CONCLUSION

This study presents new findings regarding the application of IFSU to the treatment of algal-rich, low-turbidity and high-turbidity water. The results revealed a significant positive effect of SFIPS on the IFSU performance. SFIPS functioned as a counterflow inducer, which enhanced the effects of stratified flow and decreased the clarification-loading rate in the separation zone during the DAF process. The IFSU with SFIPS improved the removal ratio of algae, turbidity and CODMn. Moreover, the RSM model derived in this study could describe the variations in the process as a function of the variable studied. However, application of the RSM models was limited to the range of the variables studied. Therefore, to predict the IFSU performance under conditions outside of the range studied or under different operating conditions, additional experiments are needed to verify the model under the desired conditions. This study also showed that the use of IFSU with SFIPS is a new window to introduce a promising and economical method to enable the upgrade of conventional DWTPs.

CONFLICT OF INTEREST

The authors declare that they have no conflict of interest.

ACKNOWLEDGEMENTS

The authors would like to acknowledge National Natural Science Foundation of China (Grant No. 51608140, Grant No. 51778164 and Grant No. 41401355), The Academician Workstation for Urban Environment Governance in Guangzhou Municipal Engineering Design & Research Institute (Grant No. 2013B090400006) and the Guangdong Provincial Department of Education Project of production, teaching, research cooperation (Grant No. 2012B091000055) for financial support of this work.

REFERENCES

REFERENCES
Al-Tebrineh
,
J.
,
Mihali
,
T. K.
,
Pomati
,
F.
&
Neilan
,
B. A.
2010
Detection of saxitoxin-producing cyanobacteria and Anabaena circinalis in environmental water blooms by quantitative PCR
.
Applied and Environmental Microbiology
76
,
7836
7842
.
American Public Health Association (APHA)
2005
Standard Methods for the Examination of Water and Wastewater
.
21st edn
,
American Public Health Association
,
New York
.
Buisine
,
F.
&
Oemcke
,
D.
2003
Seasonal influence of waste stabilisation pond effluent on DAF/F (dissolved air flotation/filtration) process operation
.
Water Science and Technology
48
(
2
),
357
364
.
Chen
,
C.
,
Zhang
,
X.
,
Zhu
,
L.
,
Liu
,
J.
,
He
,
W.
&
Han
,
H.
2008
Disinfection by-products and their precursors in a water treatment plant in North China: seasonal changes and fraction analysis
.
Science of the Total Environment
397
,
140
147
.
Edzwald
,
J. K.
2007
Developments of high rate dissolved air flotation for drinking water treatment
.
Journal of Water Supply: Research and Technology-AQUA
56
,
399
409
.
Edzwald
,
J. K.
2010
Dissolved air flotation and me
.
Water Research
44
,
2077
2106
.
Edzwald
,
J. K.
,
Tobiason
,
J. E.
,
Amato
,
T.
&
Maggi
,
L. J.
1999
Integrating high-rate DAF technology into plant design
.
American Water Works Association Journal
91
,
41
.
Hohner
,
A. K.
,
Cawley
,
K.
,
Oropeza
,
J.
,
Summers
,
R. S.
&
Rosario-Ortiz
,
F. L.
2016
Drinking water treatment response following a Colorado wildfire
.
Water Research
105
,
187
198
.
Khuri
,
A. I.
&
Cornell
,
J. A.
1996
Response Surfaces: Designs and Analyses
.
CRC Press
,
Boca Raton, FL
.
Kontogiannopoulos
,
K. N.
,
Patsios
,
S. I.
&
Karabelas
,
A. J.
2016
Tartaric acid recovery from winery lees using cation exchange resin: optimization by response surface methodology
.
Separation & Purification Technology
165
,
32
41
.
Laskovski
,
D.
,
Duncan
,
P.
,
Stevenson
,
P.
,
Zhou
,
J.
&
Galvin
,
K. P.
2006
Segregation of hydraulically suspended particles in inclined channels
.
Chemical Engineering Science
61
,
7269
7278
.
Li
,
G.
&
Gregory
,
J.
1991
Flocculation and sedimentation of high-turbidity waters
.
Water Research
25
,
1137
1143
.
Li
,
Z.
,
Sun
,
Z.
&
Liu
,
W.
1993
Lateral flow flotation-sedimentation tank with inclined planes
.
Water Supply and Sewage Disposal
2
,
6
9
(
In Chinese
).
Li
,
L.
,
Sun
,
Z.
&
Zhang
,
R.
2017
Numerical simulation of sedimentation processes in a novel air flotation-sedimentation tank
.
Journal of Water Process Engineering
18
,
41
46
.
Lichtenthaler
,
H. K.
1987
Chlorophylls and carotenoids: pigments of photosynthetic biomembranes
.
Methods in Enzymology
148
,
350
382
.
Liu
,
G.
,
Sun
,
Z.
&
Liu
,
W.
2012
New flotation/sedimentation process for reconstruction of treatment project of algae-laden water from south reservoir
.
China Water and Wastewater
28
,
55
57
(
In Chinese
).
María
,
C. F.
,
Buffing
,
M. F.
,
Marcel
,
J.
,
Carlos
,
V. L.
&
Wijffels
,
R. H.
2012
Performance of Chlorella sorokiniana under simulated extreme winter conditions
.
Journal of Applied Phycology
24
,
693
699
.
Ren
,
P.
,
Nan
,
J.
,
Zhang
,
X.
&
Zheng
,
K.
2017
Analysis of floc morphology in a continuous-flow flocculation and sedimentation reactor
.
Journal of Environmental Sciences
52
,
268
275
.
Sun
,
Z.
&
Liu
,
G.
2014
Manganese removal by new flotation/sedimentation and V-type filter
.
China Water and Wastewater
30
,
34
41
(
In Chinese
).
USEPA
1983
Methods for Chemical Analysis of Water and Wastes
.
USEPA
,
Washington, DC
.
Xing
,
C. H.
,
Yamamoto
,
K.
&
Fukushi
,
K.
2006
Performance of an inclined-plate membrane bioreactor at zero excess sludge discharge
.
Journal of Membrane Science
275
,
175
186
.

Supplementary data