Abstract

A pilot-scale CFIC® (continuous flow intermittent cleaning) reactor was run in anoxic conditions to study denitrification of wastewater. The CFIC process has already proven its capabilities for biological oxygen demand removal with a small footprint, less energy consumption and low cost. The present study focused on the applicability for denitrification. Both pre-denitrification (pre-DN) and post-denitrification (post-DN) were tested. A mixture of primary treated wastewater and nitrified wastewater was used for pre-DN and nitrified wastewater with ethanol as a carbon source was used for post-DN. The pre-DN process was carbon limited and removal rates of only 0.16 to 0.74 g NOx-N/m²-d were obtained. With post-DN and an external carbon source, 0.68 to 2.2 g NO3-Neq/m²-d removal rates were obtained. The carrier bed functioned as a good filter for both the larger particles coming with influent water and the bio-solids produced in the reactor. Total suspended solids removal in the reactor varied from 20% to 78% (average 45%) during post-DN testing period and 9% to 70% (average 29%) for pre-DN. The results showed that the forward flow washing improves both the DN function and filtration ability of the reactor.

INTRODUCTION

Low cost, small footprint, less energy consumption and product reuse are some of the main considerations for new system developments in water and wastewater treatment. The MBBR (moving bed biofilm reactor) process, developed about 25 years ago, could address some of these issues and has been very successful compared to the conventional biological wastewater treatment processes (Ødegaard et al. 1999; Rusten & Paulsrud 2009). The CFIC® (continuous flow intermittent cleaning) reactor, which has been developed and patented by Biowater Technology with the help of external R&D institutions, is considered as the next-generation biofilm process of the well-known MBBR (Rusten et al. 2011).

The CFIC biofilm reactor concept and process have been explained in detail by Rusten et al. (2011) and Stang et al. (2013). The CFIC reactor contains highly packed biofilm carriers, which hinders the movement of the carriers in the reactor during normal operation. This carrier bed reduces the footprint of the reactor by increasing the specific protected surface area (m²/m³ reactor volume) for biofilm growth. Treatment efficiency of this system is increased by facilitating good air distribution and good contact between water, substrates and the biofilm grown on the carrier surface. The CFIC reactor has its own ability to separate particles from the water. The compacted carrier bed functions as a ‘filter’ for the particles that come with influent water as well as the bio-solids produced in the reactor. Intermittent reactor washing removes the retained particles and the bio-solids, facilitating separate bio-sludge treatment (Rusten et al. 2011). Studies of biological aerated filters and use of biofilm carriers as a filtration media show that removal increases with the height of the filter (Mann & Stephenson 1997). Liao & Ødegaard (2002) showed that total suspended solids (TSS) removal decreased, but not linearly, with increasing filtration rate (m/h).

Treatment efficiency of the CFIC reactor for removal of organic matter has been tested and verified in bench-scale, pilot-scale and full-scale treatment plants (Stang et al. 2013; Siljudalen et al. 2014). The purpose of this research was to demonstrate the applicability of CFIC for nitrogen removal and the results from a pilot plant for denitrification (DN) are presented.

MATERIALS AND METHODS

Pilot-scale testing was carried out at the Nedre Romerike (NRA) wastewater treatment plant (WWTP) in Lillestrøm, Norway. NRA is one of Norway's largest WWTPs, having a full-scale MBBR process for nitrogen removal. The full-scale treatment process consists of mechanical pre-treatment, primary sedimentation, biological treatment with MBBRs and a chemical coagulation process followed by secondary sedimentation. Presently the plant is running with a load of about 110,000 population equivalents and a flow of around 50,000 m3/day.

The CFIC pilot was a cylindrical reactor, made of transparent PVC with an inner diameter of 500 mm and a height of 1,800 mm. Table 1 presents the details of the pilot bioreactor.

Table 1

Bioreactor data for the pilot-scale testing at the NRA WWTP

CFIC reactor
Total wet volume of the reactor 0.29 m³ 
Water depth 1.5 m 
Type of biofilm carriers BWTSTM 
Biofilm carrier fill 66% 
Protected biofilm surface area 123.4 m² 
CFIC reactor
Total wet volume of the reactor 0.29 m³ 
Water depth 1.5 m 
Type of biofilm carriers BWTSTM 
Biofilm carrier fill 66% 
Protected biofilm surface area 123.4 m² 

BWTS™ carriers from Biowater Technology were used to fill approximately 66% of the empty bed volume. The carrier shape provides a large void volume (typically an 85% void volume in a 100% fill situation) for growth and accumulation of biomass. Table 2 gives more details with a drawing of one of the biofilm carriers.

Table 2

Biofilm carriers used for pilot-scale testing of CFIC process at the NRA WWTP

Type of biofilm carrier BWTSTM  
Protected biofilm surface area 650 m2/m3 
Length 14.5 mm 
Height 18.5 mm 
Width 7.3 mm 
Number of cells per carrier element 
Type of biofilm carrier BWTSTM  
Protected biofilm surface area 650 m2/m3 
Length 14.5 mm 
Height 18.5 mm 
Width 7.3 mm 
Number of cells per carrier element 

The pilot reactor for the present studies had some modifications from the original CFIC reactor design explained in Rusten et al. (2011) and Stang et al. (2013). Unlike the original CFIC reactor design, the CFIC-DN reactor did not run with two different water levels during normal mode and forward flow washing mode, see Figure 1.

Figure 1

The CFIC DN pilot (a) during normal operation, and (b) during the forward flow washing cycle.

Figure 1

The CFIC DN pilot (a) during normal operation, and (b) during the forward flow washing cycle.

Influent water was fed from the bottom, making upward movement in the reactor. Because the density of carrier (BWTS) material is slightly less than the density of water, all the carriers were lifted up with the upward movement of water and remained packed at the upper part of the reactor during the normal operation. The filter height (carrier bed) of the CFIC pilot reactor was constant (105 cm) (Figure 1(a)). A mixer was used with low rpm (maximum of 45 rpm) to keep the influent water mixed before it reached the biofilm carriers. The propeller of the mixer was placed between the water inlet (bottom of the reactor) and lower level of the carrier bed when it was in normal mode. Anoxic conditions were maintained during normal operation.

The washing cycle was started with coarse bubble aeration at the bottom of the reactor. The turbulence produced by the coarse air bubbles facilitated good mixing and removal of biomass via collision of carriers, and the biomass was washed out from the reactor with the wash-water flow (Figure 1(b)).

Unlike the earlier CFIC design, outlets for both normal effluent and wash-water were at the top water surface level of the reactor. During normal operation, the effluent outlet was kept open while the wash-water outlet was kept closed. The normal effluent outlet was kept closed and the wash-water outlet was kept open during the washing mode. Normal effluent and wash-water were collected in separate containers.

Two different treatment concepts, pre-denitrification (pre-DN) and post-denitrification (post-DN) were tested in pilot scale. Pre-DN pilot testing was started in December 2015 and full sampling was run during January and February 2016. Pre-DN is a process where denitrifying bacteria use biodegradable carbon from the influent wastewater. Primary effluent after pre-sedimentation at the NRA WWTP was used to provide the carbon source. Removing a part of the influent particulate organic material by primary treatment does not have a significant effect on DN rates (Newcombe et al. 2011; Razafimanantsoa et al. 2014). Rusten et al. (2016) studied the DN rates with filtered wastewater as carbon source and concluded that the removal of organic matter with a filter cloth with openings as small as 33 μm had no negative effect on the DN process. In order to provide adequate nitrate for the DN process, effluent water from bioreactor 3 (nitrified wastewater) of the NRA WWTP was mixed with the above-mentioned primary treated wastewater before it reached the pilot. A mixture with 50% primary treated wastewater and 50% nitrified wastewater was used for pre-DN.

Post-DN is the DN process where an external carbon source is used to provide biodegradable chemical oxygen demand (COD). Nitrified wastewater from the full-scale bioreactors (bioreactor 3) was used as influent to the pilot. The post-DN process in the pilot was initiated in August 2015 and sampling was started from 15th September. The carbon source is the controlling factor of the structure and function of the denitrifying bacterial community (Wang et al. 2017; Yang et al. 2018; Xu et al. 2018). When the carbon source is specified, the DN process in different reactors often possesses similar dominant populations (Lu et al. 2014). Ethanol was used as the external carbon source and different N loading rates and carbon-to-nitrogen ratios, expressed as g COD/g NO3-Neq (COD/N), were tested during post-DN tests.

Automatic samplers were used for flow-proportional samples of the influent and the normal mode effluent from the reactor. A continuous flow sampler was used for wash-water sampling during reactor cleaning. Manual grab samples were taken before and after the cleaning cycle. TSS was measured by an external laboratory. Whatman GF/C glass fiber filters were used for filtration of samples and analysis of TSS. TSS was analyzed according to Standard Methods SM 2540 D and E (APHA/AWWA/WEF 2005). All the other parameters were analyzed on-site, using Dr Lange test cuvette kits, a thermostat LT200 and a DR 2800 spectrophotometer (Hach Lange, Dusseldorf, Germany). Biomass on the biofilm carriers was also measured routinely as total solids (TS), following the method explained in Rusten et al. (2016). In addition wastewater influent flow, NO3-N (inlet and outlet), pH (inlet and outlet), dissolved oxygen (DO) (influent and inside the reactor), temperature (influent and inside the reactor) and TSS (effluent and wash-water) were measured using online sensors (Hach Lange, Dusseldorf, Germany).

RESULTS AND DISCUSSION

Pilot-plant influent concentrations in the pre-DN and post-DN periods, respectively, are given in Table 3. All the data are based on flow proportional samples collected during the test periods. Pre-DN samples were taken after mixing primary and nitrified wastewater in the inlet tank of the pilot. Post-DN samples were taken from the inlet tank after adding the external carbon source.

Table 3

Characterization of pilot-plant influent during the pre-DN and post-DN tests

Pre-DN
Post-DN
AverageMaxMinPercentile
nbAverageMaxMinPercentile
n
20802080
pH 7.22 7.44 7.05 7.10 7.33 16 7.10 7.35 6.81 6.98 7.23 19 
Temperature (°C) 10.1 11.2 9.3 9.5 10.5 16 13.7 15.5 10.5 12.5 15.1 16 
TSS (mg/l) 185 300 56 100 230 16 96 170 54 72 120 19 
Total COD (mg/l) 333 470 120 254 430 19 225 328 132 182 264 19 
FCOD (mg/l)a 100 180 41 77 116 19 117 184 67 82 147 19 
NH4-N (mg/l) 14.5 24.0 3.7 6.8 21.5 19 0.2 0.6 <0.1 0.1 0.3 19 
NOx-N (mg/l) 9.2 17.1 4.9 6.3 12.5 19 15.9 21.8 9.1 13.1 18.3 19 
NO3-N (mg/l) 6.1 8.8 2.2 5.0 7.5 19 14.3 21.6 6.2 10.8 17.6 19 
NO2-N (mg/l) 3.1 10.9 0.6 0.8 5.1 19 1.6 6.7 0.1 0.4 2.4 19 
PO4-P (mg/l) 0.54 0.93 0.22 0.27 0.65 19 0.26 0.63 0.01 0.03 0.47 17 
DO (mg/l) 2.07 2.40 1.47 1.62 2.18 15 6.10 7.90 5.00 5.44 6.48 18 
Alkalinity (mg CaCO3/l) 114 176 61 81 150 16 64 93 46 55 71 19 
Pre-DN
Post-DN
AverageMaxMinPercentile
nbAverageMaxMinPercentile
n
20802080
pH 7.22 7.44 7.05 7.10 7.33 16 7.10 7.35 6.81 6.98 7.23 19 
Temperature (°C) 10.1 11.2 9.3 9.5 10.5 16 13.7 15.5 10.5 12.5 15.1 16 
TSS (mg/l) 185 300 56 100 230 16 96 170 54 72 120 19 
Total COD (mg/l) 333 470 120 254 430 19 225 328 132 182 264 19 
FCOD (mg/l)a 100 180 41 77 116 19 117 184 67 82 147 19 
NH4-N (mg/l) 14.5 24.0 3.7 6.8 21.5 19 0.2 0.6 <0.1 0.1 0.3 19 
NOx-N (mg/l) 9.2 17.1 4.9 6.3 12.5 19 15.9 21.8 9.1 13.1 18.3 19 
NO3-N (mg/l) 6.1 8.8 2.2 5.0 7.5 19 14.3 21.6 6.2 10.8 17.6 19 
NO2-N (mg/l) 3.1 10.9 0.6 0.8 5.1 19 1.6 6.7 0.1 0.4 2.4 19 
PO4-P (mg/l) 0.54 0.93 0.22 0.27 0.65 19 0.26 0.63 0.01 0.03 0.47 17 
DO (mg/l) 2.07 2.40 1.47 1.62 2.18 15 6.10 7.90 5.00 5.44 6.48 18 
Alkalinity (mg CaCO3/l) 114 176 61 81 150 16 64 93 46 55 71 19 

Data are based on 24-hour flow proportional samples. Pre-DN samples were analyzed after mixing the two water types.

aCOD, measured on filtered samples.

bNumber of samples.

As reported by Ekama et al. (1986) the biodegradable soluble organic matter concentration reduces to approximately zero for wastewater biologically treated at 20 °C and with more than 3 days solids retention time. This is applicable to the nitrified effluent taken from bioreactor 3 at the full-scale WWTP. Only the amount of COD in filtered sample (FCOD) provided by external carbon was considered as biodegradable soluble COD (BSCOD) for post-DN tests. For the pre-DN tests the BSCOD was typically 55–60% of the FCOD in the primary effluent wastewater.

In order to change the NO3-N loading during the tests, an external NO3-N supply was used. The commercial product called Nutriox, having 113 g NO3-N/l was dosed when necessary, during the post-DN tests. Influent samples were taken downstream of where Nutriox was mixed into the wastewater.

Since the wastewater was not subjected to pre-precipitation, PO4-P concentrations in both pre-DN and post-DN, which were 0.54 and 0.26 mg/l (average) respectively, were sufficiently high for the DN process. NOx-N is the sum of NO2-N and NO3-N. The reactor was run in anoxic conditions. DO of influent water is shown in Table 3 and average DO measured inside the reactor was 0.82 mg/l and 0.21 mg/l for pre-DN and post-DN respectively.

Pre-denitrification

Figure 2 shows DN rates versus loading rates during the tests. Loading and removal rates were calculated as grams NOx-N per m² available biofilm area in the reactor, per day. DN rate during this study varied between 0.16 and 0.74 g NOx-N/m²-d, with average of 0.35 g NOx-N/m²-d. It can be observed that the DN rates are increasing with increasing loads. Opposite to that, the percentage removal was very high with low loading rates and it went down when the NOx-N loading was increased. Close to 93% NOx-N removal was achieved with low loading rates. This shows that the low DN rates achieved were mainly due to lack of easily BSCOD in the raw wastewater. In addition the low temperature during the testing also influenced the results.

Figure 2

DN rates versus NOx-N loading for pre-DN tests. All the samples were 24 h flow proportional samples. Hydraulic retention times are based on empty bed volume and the influent flow. Temperature is the water temperature inside the reactor.

Figure 2

DN rates versus NOx-N loading for pre-DN tests. All the samples were 24 h flow proportional samples. Hydraulic retention times are based on empty bed volume and the influent flow. Temperature is the water temperature inside the reactor.

Maximum DN rate achieved was 0.74 g NOx-N/m²-d with 1.4 g NOx-N/m²-d loading, which equals 50% removal. When compared with the earlier experiences of MBBR, these results can be considered as good DN rates for pre-DN at lower temperatures. Rusten et al. (1995) reported maximum pre-DN rates of 0.68 g NOx-N/m²-d at temperatures from 7 to 16.4 °C (load of 3.3 g/m²-d), in MBBR tests at Nordre Follo WWTP, Norway. In another study at Frevar WWTP, Norway, the maximum removal rate that could be achieved was 0.54 g NOx-N/m²-d at 11.4–12.7 °C with a high influent BSCOD of 50 mg/l (Rusten et al. 2000).

During the DN process in the CFIC reactor, bacteria in the biofilm attached to the carriers consume nitrate and carbon from wastewater to produce nitrogen gas. This process also adds more and more bio-solids to the reactor. Three graphs in Figure 3 show FCOD, TCOD and TSS removal versus loading. Percentage FCOD removal during the testing period varied between 0.5% (0.02 g FCOD/m²-d) and 39% (5.6 g FCOD/m²-d). This low removal is shown with low FCOD loading rates.

Figure 3

FCOD, TCOD and TSS removal during the pre-DN tests.

Figure 3

FCOD, TCOD and TSS removal during the pre-DN tests.

TCOD removal varied between 7.5% (1 g TCOD/m²-d) and 62% (18.4 g TCOD/m²-d). This removal is considerably larger than the FCOD removal. TSS removal in the third graph varied from 9% (1.8 g TSS/m²-d) to 69% (4.2 g TSS/m²-d). The CFIC-DN pilot ran anoxic and was designed to improve the NOx-N removal but not for optimizing the biological oxygen demand removal. TCOD removal is a result of both the bacteria consumption during the DN process and particle filtration by the carrier bed. TSS removal was achieved mainly due to the filter effect of the carrier bed.

Post-denitrification

High DN rates were obtained with increasing NOx-N load and addition of adequate amounts of ethanol as an external carbon source. Data cover a temperature range from 11 to 16 °C and empty bed residence times from 22 to 40 minutes. We have seen that the influent water contains considerable amounts of NO2-N and DO. Both NO2-N and DO will consume organic substrate inside the reactor. Therefore, we used the conversion described by Koopman et al. (1990) to calculate an equivalent concentration of NO3-N (NO3-Neq) by converting influent NO2-N and DO to equivalent amounts of NO3-N. For this conversion, 1.0 mg NO2-N is considered as 0.60 mg NO3-Neq and 1.0 mg DO is considered as 0.35 mg NO3-Neq.

DN rates versus loading is shown in Figure 4. The influent NO3-Neq concentrations varied from 11.2 to 23.8 mg/l (average 14.5). BSCOD varied from 61 mg/l to 140 mg/l, with corresponding COD/N ratios from 3.3 to 12.5 g COD/g NO3-Neq.

Figure 4

DN rates versus NO3-Neq loading for post-DN tests. All the samples were 24 h flow proportional samples. Hydraulic retention times are based on empty bed volume and the influent flow. Temperature is the water temperature inside the reactor. COD/N ratio is based on g BSCOD/g NO3-Neq.

Figure 4

DN rates versus NO3-Neq loading for post-DN tests. All the samples were 24 h flow proportional samples. Hydraulic retention times are based on empty bed volume and the influent flow. Temperature is the water temperature inside the reactor. COD/N ratio is based on g BSCOD/g NO3-Neq.

Higher DN rates were achieved with increasing NO3-Neq loads. The maximum DN rate achieved was 2.2 g NO3-Neq/m²-d with 2.8 g NO3-Neq/m²-d loading, which equals 79% removal. DN rates higher than 1.5 g NO3-Neq/m²-d were obtained when the loading was higher than 2 g NO3-Neq/m²-d. Temperature range and varying hydraulic retention time (HRT) during the testing may also have an effect on the DN rate.

The COD/N ratio is one of the key parameters for biological nitrogen removal. The minimum theoretical requirement for anoxic DN is 2.86 g COD/g NO3-Neq. Due to other natural processes, such as biomass growth, this value will always be higher than the theoretical value (Rusten et al. 1996; Fu et al. 2009). For the calculations in this paper, we have used the amount of FCOD added as the external carbon source and the amount of NO3-Neq. Using COD/N ratios expressed as g COD/g NO3-Neq is especially useful when comparing substrate consumption ratios in systems run with different influent DO concentrations.

Figure 5 shows the average removal rates with different ranges of COD/N ratios. Here the NO3-N, NO3-Neq and FCOD removals are directly related to the biological consumption. TSS and TCOD removal are due to both bio-consumption and filtration by the carrier bed. It is clearly shown in the figure that the COD/N range 4–6 gave the best NO3-N, NO3-Neq and FCOD removal rates with average removals of 1.4, 1.7 and 5.7 g/m²-d respectively, corresponding to a consumption ratio of 3.4 g FCOD/g NO3-Neq.

Figure 5

Removal rates with increasing BSCOD/N ratios. Average removal for each BSCOD/N range is shown.

Figure 5

Removal rates with increasing BSCOD/N ratios. Average removal for each BSCOD/N range is shown.

We can also in Figure 4 see that the two best results with the COD/N range 4–6 was with high NOx-N loads. Most of the tests were run with COD/N ratios between 6 and 8. As Figure 5 shows, the range 6 to 8 also had good removal rates with 1.2, 1.4 and 5 g/m²-d for NO3-N, NO3-Neq and FCOD respectively, corresponding to a consumption ratio of 3.6 g FCOD/g NO3-Neq. These consumption ratios are in full agreement with previous results from testing of ethanol as external carbon source in MBBRs at 10–15 °C (Rusten et al. 1996). TCOD removal also followed the same pattern, but it shows a sudden increase at the highest COD/N ratio. Removal of TSS and TCOD mainly vary with the filter effect of the CFIC reactor, which is discussed in a later section.

Figure 6 shows three graphs of FCOD, TCOD and TSS removal versus loading. Percentage FCOD removal during the testing period varied between 11% (1.1 g FCOD/m²-d) and 61% (10.5 g FCOD/m²-d). TCOD removal varied between 15% (4 g TCOD/m²-d) and 59% (15.7 g TCOD/m²-d). TSS removal achieved a maximum of 78% (6.1 g TSS/m²-d). Minimum removal was 20% (2.4 g TSS/m²-d). It is important to note that the reactor also removes a large portion of the biomass created in the reactor, and this is not captured in the removal efficiencies that are calculated based on influent and effluent TSS concentrations.

Figure 6

FCOD, TCOD and TSS removal during the post-DN tests.

Figure 6

FCOD, TCOD and TSS removal during the post-DN tests.

Filter effect and intermittent washing

The density of the biofilm carrier material is slightly less than the density of influent wastewater. This makes all the carriers slightly buoyant and packed at the upper part of the reactor, during the normal mode with calm upwards flow. These highly packed carriers limit the free movements inside the reactor and serve as a ‘filter’, to keep the particles and bio-solids inside the filter bed, and reduce solids concentrations in the effluent from the CFIC reactor. An intermittent cleaning process removes the loose and loosely attached sludge outside and inside the carriers. Figure 7 visualizes the filter effect of the CFIC reactor. Figure 7(a) shows the biofilm carrier bed after 24 hours of normal mode operation. The picture was taken just before starting the forward flow washing process during post-DN. The picture in Figure 7(b) was taken after washing on the same day.

Figure 7

CFIC reactor before (a) and after (b) washing cycle. Pictures were taken immediately prior to and immediately after the forward flow washing cycle. Graph (c) is an example of TSS concentrations in regular effluent and wash-water for CFIC tests. Data collected online 5th–8th November 2015. Forward flow washing daily from 09:30 to 10:30.

Figure 7

CFIC reactor before (a) and after (b) washing cycle. Pictures were taken immediately prior to and immediately after the forward flow washing cycle. Graph (c) is an example of TSS concentrations in regular effluent and wash-water for CFIC tests. Data collected online 5th–8th November 2015. Forward flow washing daily from 09:30 to 10:30.

Basic reactor washing was started at 09:30 and continued for 1 hour each day. The volume of washing water varied from 0.26 to 1.2 m³, depending on the influent flow rate. The average value of 0.64 m³ was approximately double the empty bed volume of the reactor. Figure 7(c) shows an example of variation of TSS concentrations during a normal run and forward flow washing cycle with post-DN tests. TSS was lower than 100 mg/l during normal operation and increased to up to maximum 3,000 mg/l during the washing cycle.

Wash-water was sampled, starting with every 5 minutes for the first half hour and every 10 minutes thereafter. Figure 8 shows TSS variations in wash-water during washing, together with the normal effluent TSS immediately before the washing cycle started. When comparing the two graphs, it is clear that the wash-water TSS production during post-DN was always higher than during pre-DN. When the influent flow during post-DN was high (0.77 and 0.79 m³/h), wash-water had the highest TSS concentrations.

Figure 8

TSS concentrations in wash-water at pilot-scale CFIC tests. Washing was done from 09:30 to 10:30 (1 hour) every day. Figure to the left represents the pre-DN tests and figure to the right is from post-DN tests. Corresponding flow rates are shown in brackets.

Figure 8

TSS concentrations in wash-water at pilot-scale CFIC tests. Washing was done from 09:30 to 10:30 (1 hour) every day. Figure to the left represents the pre-DN tests and figure to the right is from post-DN tests. Corresponding flow rates are shown in brackets.

Table 4 compares the removal rates of the CFIC reactor just before and after the washing cycle. Grab samples were taken directly from the influent tank and effluent tank, before and after the washing cycle. Table 4 shows the averages of 28 samples (14 before and 14 after) from pre-DN and 32 samples (16 before and 16 after) from post-DN tests.

Table 4

Removal rates before and after forward flow washing cycle

Pre-DN
Post-DN
(g/m2-d)BeforeAfterBeforeAfter
TCOD 4.6 4.7 9.8 12.8 
FCOD 1.7 2.3 7.8 10.8 
NOx-N 0.34 0.37 1.1 1.3 
TSS 2.6 2.5 5.1 5.6 
Pre-DN
Post-DN
(g/m2-d)BeforeAfterBeforeAfter
TCOD 4.6 4.7 9.8 12.8 
FCOD 1.7 2.3 7.8 10.8 
NOx-N 0.34 0.37 1.1 1.3 
TSS 2.6 2.5 5.1 5.6 

Grab samples were taken just before and after the washing cycle. Sample numbers were 14 for pre-DN and 16 for post-DN.

Data from post-DN tests show an improvement of the removal rates after the washing cycle. As shown in Figure 7, after 24 hrs the reactor was full of bio-solids and particles. These bio-solids and particles make it harder for the nutrients to reach the live bacteria cells on the biofilm and thus the removal rate will be reduced. The washing cycle removes the excess solids from the reactor and facilitates nutrients to reach live bacteria. This improves the nutrient removal process.

In addition to that, when the filter bed is filled and blocked by debris, filter break-through can happen and this facilitates particle movement through the filter. This makes poor TSS and TCOD removal before washing the filter. Washing improves the filter conditions again and results in better particle removal after the filter wash. Graphs in Figures 7(c) and 8 show that effluent TSS concentrations can also be very low prior to washing. We have observed that effluent TSS was at its highest 6–12 h after washing.

NOx-N removal rates were much lower during pre-DN and it did not show a clear difference in N removal before and after the washing cycle. Since the bio-solids production was low, the filter was not saturated within 24 hours filtration time and therefore the TSS and TCOD removal rates were also not increased by filter washing.

Gram sludge TSS removed per m² washing cycle per biofilm varied from 3.0 to 7.3, with an average of 4.6 (g TSS/biofilm m²) for pre-DN. The same was 3.5 to 13.5 with an average of 6.0 (g TSS/biofilm m²) for post-DN tests.

Biofilm carriers were sampled immediately prior to and immediately after the washing cycle, to measure biomass on the biofilm carriers. Table 5 shows the average measurements and biomass removal by washing cycle.

Table 5

Biomass growth on the carriers in pre-DN and post-DN tests

Pre-DN
Post-DN
Before washAfter washBefore washAfter wash
TS on carriers, mg TS/carrier Max 47.3 46.8 44.9 40.6 
Min 39.3 37.9 39.3 38.0 
Average 43.6 42.4 43.3 39.1 
g TS/m² biofilm surface area Average 26.8 26.1 26.4 24.1 
TS removed during wash, % Max 14 
Min 
Average 10 
Pre-DN
Post-DN
Before washAfter washBefore washAfter wash
TS on carriers, mg TS/carrier Max 47.3 46.8 44.9 40.6 
Min 39.3 37.9 39.3 38.0 
Average 43.6 42.4 43.3 39.1 
g TS/m² biofilm surface area Average 26.8 26.1 26.4 24.1 
TS removed during wash, % Max 14 
Min 
Average 10 

Average biomass on the carriers in both pre- and post-DN tests was approximately 43 mg TS/carrier, equivalent to 26.5 g TS/m² biofilm surface area. Table 5 shows that 10% (equivalent to 2.3 g TS/m² biofilm surface area) of the biomass attached to the carriers was removed by reactor washing during post-DN and it was only 3% (equivalent to 0.7 g TS/m² biofilm surface area) in pre-DN tests. The rest of the particles in the wash-water were influent TSS and eroded biomass that had been trapped inside or between the biofilm carriers during CFIC mode, without being attached to the biofilm carriers.

CONCLUSIONS

The CFIC process, which has earlier been demonstrated to have several advantages for removal of organic matter, was evaluated for its applicability for nitrogen removal. Results confirmed satisfactory removal rates both for pre-DN and post-DN in pilot scale.

Good denitrification rates, with a maximum of 0.74 g NOx-N/m²-d, were achieved in pre-DN tests. These rates compares favorably with other biofilm reactors using primary treated wastewater as carbon source in cold temperatures. In the post-DN process, with addition of ethanol as external carbon source, DN rates up to 2.2 g NO3-Neq/m²-d were achieved. The optimum COD/N ratio was found to be between 4 and 6 g COD/g NO3-Neq.

Biofilm carriers in the CFIC reactor function as a ‘filter’ for the particles that come with influent water as well as the bio-solids produced in the reactor. Filtration in the present study of an anoxic denitrifying CFIC reactor was satisfactory (up to 60% TCOD removal). The results showed that the forward washing improves both the DN function and filtration ability of the reactor. The reactor forward flow washing process successfully removed particles and bio-solids from the reactor. The volume of wash-water varied from 0.26 to 1.2 m³ per wash, equivalent to 0.9 to 4 empty bed volumes. Indicating the higher denitrification rate and bio-solids production with post-DN, removal of attached solids was a lot higher per carrier for post-DN than for pre-DN.

The pilot testing demonstrated that an anoxic CFIC reactor for DN can successfully operate with the same water volume and water surface level during both normal mode operation and during the forward flow washing cycle. This is significant since it will enable much simpler construction and operation of these anoxic reactors, compared to a reactor with a higher water level during washing. The benefit of an anoxic pre-DN CFIC reactor with only one water level will be very obvious when it is coupled with downstream nitrifying CFIC reactors in a full-scale plant.

It is anticipated that even better results can be achieved in full-scale operations with further optimizations, enabling the application of CFIC with the many advantages the concept represents.

ACKNOWLEDGEMENT

The authors acknowledge the technical assistance from Mr C. L. Otis, and valuable help from the management and operators at the Nedre Romerike WWTP, Aquateam COWI AS and Biowater Technology AS, Norway. The CFIC project was financed by grant no. 238995 from the Regional Research Council of the Capital of Norway, NRA WWTP and Biowater Technology AS, Norway.

REFERENCES

REFERENCES
APHA/AWWA/WEF
2005
Standard Methods for the Examination of Water and Wastewater
, 21st edn.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
Koopman
B.
,
Stevens
C. M.
&
Wonderlick
C. A.
1990
Denitrification in a moving bed upflow sand filter
.
Research Journal of the Water Pollution Control Federation
62
(
3
),
239
245
.
Liao
Z.
&
Ødegaard
H.
2002
Coarse media filtration for enhanced primary treatment of municipal wastewater
.
Water Science and Technology
46
(
4–5
),
19
26
.
Mann
A. T.
&
Stephenson
T.
1997
Modelling biological aerated filters for wastewater treatment
.
Water Research
31
(
10
),
2443
2448
.
Newcombe
R.
,
Lopp
M.
,
Higby
T.
,
Martini
R.
&
Rusten
B.
2011
Particle size separation implications on COD removal before BNR: a case study
.
WEF Proceedings, Nutrient Recovery and Management
2011
,
477
484
.
Ødegaard
H.
,
Rusten
B.
&
Siljudalen
J. G.
1999
The development of the moving bed biofilm process – from idea to commercial product
.
EWPCA Journal European Water Management
2
(
3
),
36
43
.
Razafimanantsoa
V. A.
,
Ydstebo
L.
,
Bilstad
T.
,
Sahu
A. K.
&
Rusten
B.
2014
Effect of selective organic fractions on denitrification rates using Salsnes Filter as primary treatment
.
Water Science and Technology
69
(
9
),
1942
1948
.
Rusten
B.
&
Paulsrud
B.
2009
Environmental Technology Verification of a biofilm process for high efficiency nitrogen removal from wastewater
. In:
CD-proceedings of WEFTEC 2009
,
October 10–14
,
Orlando, Florida
, pp.
4378
4391
.
Rusten
B.
,
Hem
L.
&
Ødegaard
H.
1995
Nitrogen removal from dilute wastewater in cold climate using moving bed biofilm reactors
.
Water Environment Research
67
(
1
),
65
74
.
Rusten
B.
,
Wien
A.
&
Skjefstad
J.
1996
Spent aircraft deicing fluid as external carbon source for denitrification of municipal wastewater – from waste problem to beneficial use
. In:
51st Purdue Industrial Waste Conference Proceedings
,
May 6–8, 1996
, pp.
359
369
.
Rusten
B.
,
Hellström
B. G.
,
Hellström
F.
,
Sehested
O.
,
Skjelfoss
E.
&
Svendsen
B.
2000
Pilot testing and preliminary design of moving bed biofilm reactors for nitrogen removal at the FREVAR wastewater treatment plant
.
Water Science and Technology
41
(
4–5
),
13
20
.
Rusten
B.
,
Stang
P.
,
Rogne
E.
,
Siljudalen
J.
&
Marcolini
L.
2011
Development of a compact, cost effective, and energy efficient, biofilm reactor for wastewater treatment and effluent reuse
. In:
CD-proceedings of WEFTEC 2011
, pp.
5222
5235
.
Rusten
B.
,
Razafimanantsoa
V. A.
,
Andriamiarinjaka
M. A.
,
Otis
C. L.
,
Sahu
A. K.
&
Bilstad
T.
2016
Impact of fine mesh sieve primary treatment on nitrogen removal in moving bed biofilm reactors
.
Water Science and Technology
73
(
2
),
337
344
.
Siljudalen
J. G.
,
Marcolini
L.
,
Xin
G.
,
Stang
P.
&
Rusten
B.
2014
Sustainable wastewater treatment with the innovative CFIC® biofilm reactor process
.
Paper Presented at Singapore International Water Week (SIWW)
.
Stang
P.
,
Siljudalen
J.
,
Marcolini
L.
&
Rusten
B.
2013
Performance of the CFIC reactor – a novel cost effective and energy efficient biofilm reactor for wastewater treatment
. In:
Invited Key Note Paper Presented at IWA 9th International Conference on Biofilm Reactors
,
May 28–31
,
Paris, France
.
Wang
D.
,
Fu
Q.
,
Xu
Q.
,
Yiwen
L.
,
Ngo
H. H.
,
Yang
Q.
,
Zeng
G.
,
Li
X.
&
Ni
B.
2017
Free nitrous acid-based nitrifying sludge treatment in a two-sludge system enhances nutrient removal from low-carbon wastewater
.
Bioresource Technology
244
,
920
928
.
Yang
G.
,
Wang
D.
,
Yang
Q.
,
Zhao
J.
,
Liu
Y.
,
Wang
Q.
,
Zeng
G.
,
Li
X.
&
Li
H.
2018
Effect of acetate to glycerol ratio on enhanced biological phosphorus removal
.
Chemosphere
196
,
78
86
.