Investigating nitrogen removal using simultaneous nitrification-denitrification in transferring wastewater through collection networks with small-diameter pipes

Wastewater collection networks are among the oldest major urban infrastructure. Building networks to collect and transfer wastewater from residential, industrial, and commercial regions to treatment plants is an optimal method, and one of its goals is improving public health by reducing epidemics (Ebrahimi raviz & Amini rad 2016).

Engineers concentrate mainly on hydraulic parameters when designing wastewater collection systems, and relatively little attention is paid to biological and chemical conditions. However, chemical and biological processes cannot be ignored because, if unsuitable designs are used, the wastewater can become septic, leading eventually to microbiologically induced corrosion. These changes first affect the network itself, then the performance of the treatment plant and its facilities, and, later, the environment and living organisms (Hvitved-Jacobsen 2013; Swamee & Sharma 2013).

Various chemical and biological wastewater collection network characteristics have been studied, including the reduction in infectious loads arising from micro-organic activities (Raunkjær et al. 1995; Hvitved-Jacobsen et al. 1998; Hvitved-Jacobsen 2013; Liu et al. 2015a, 2015b).

Sulfate-reducing bacteria produce hydrogen sulfide, which comes into contact with pipe surfaces in collection networks and is oxidized to sulfuric acid. This penetrates the concrete surface, forming the corrosive compound calcium sulfate (Ganigue et al. 2011). The amount of oxygen dissolved in the water, its temperature, the concentration of sulfate ions, and its flow rate and retention time are among the factors influencing corrosion. H₂S production indicates that biological reactions are occurring in the network, and shows that such reactions are possible (Nielsen & Hvitved-Jacobsen 1988).

Many people pay attention to contaminants such as taste and smell, color, turbidity, suspended solids, hardness, etc., but ignore contaminants like nitrate that are more hazardous and have effects that become apparent some time later (Terblanche 1991; Qaderi et al. 2011). Excessive nitrate concentrations in urban wastewater can cause environmental and public health problems. The release of urban wastewater to the environment without prior treatment, could enable it to enter water sources used by others for drinking water. Nitrate at high concentrations in drinking water reacts with hemoglobin to produce methemoglobin, and can cause blue baby syndrome in infants under six months of age. Nitrate is also carcinogenic at high concentrations (EPA 1994, 2007; Sungur et al. 2013).

The most important contaminating nitrogen compounds in wastewater include the ammonium, nitrite, and nitrate ions. The ammonia present in wastewaters can be converted and/or removed using a variety of physical, chemical, and/or biological methods. The biological methods are certainly the most important, however (Benefield & Randal 1980; Grandclément et al. 2017; Park et al. 2017).

Microorganisms play a major role in the nitrogen cycle. Among others, Ganesh et al. (2014), and Mathioudakis & Aivasidis (2009) studied how nitrogen sources were removed from wastewater, and investigated nitrifying and denitrifying conditions. Simultaneous nitrification and denitrification (SND) is a newly identified biological removal method for nitrogen compounds, and a number of researchers have shown that some processes are capable of SND in reactors (Ding et al. 2011; Vijayalayan et al. 2014). Nitrification takes place in the outer layers and denitrification in the inner layers of biofilms (Ganesh et al. 2014).

Understanding and identifying physical, chemical, and biological reactions in networks enables the qualitative aspects of wastewater to be considered in their design and use. Thus, collection networks in sustainable urban wastewater management can both transfer wastewater and provide partial wastewater treatment during transfer – i.e., they can be used for the simultaneous removal carbon and nitrogen. Modern treatment plant processes can remove nitrogen very efficiently but their high operating costs led to research on more economic methods that can be used for nitrate removal. The aim of the work discussed here was to use the design of gravity-driven sewage collection networks with small diameter pipes to study the potential for nitrogen removal, at laboratory scale, by considering the biofilm thickness on the pipe walls.

The PVC pipes used were of 10 centimeter diameter with a total volume of 120 liters. Gravity flow could be generated by changing the heights of the two forward supports – left-hand end in Figure 1. Plastic netting was fixed inside the pipes to provide a suitable substrate for biofilm growth and proliferation. Two glass pipe lengths were built in and used to observe how the microorganisms were attached to the pipe walls.

The pilot-plant incorporated two tanks (numbers 1 and 2) – see Figure 1 – with volumes of 140 and 40 liters, respectively. The network inlet was placed as low as possible in tank 2 to minimize wastewater accumulation. A pump with a 75 mm (3 inch) diameter outlet and maximum discharge of 40 l/min was selected. As one of the most important parameters in obtaining a stable system is the flow or discharge rate, a flow meter and control valve were also built into the system, to enable adjustments to the flow rate whenever necessary.

For biofilm thickness observation, Rectangular (8 × 12 cm), PVC sheets with netting stuck on one side were mounted respectively at the beginning and end of the system. These two PVC sheets were at flow direction, and the biofilm on them was observed daily (Figure 1).

Microorganisms from the aeration pond in the wastewater treatment plant at Shahrak Yasreb, Ghaemshahr, were used to start biofilm growth in the pilot-plant. The main components of the synthetic wastewater used are listed in Table 1. It is very similar to domestic wastewater.

Attempts were made to keep the chemical oxygen demand (COD) concentration in the network constant and close to that of the treatment plant's urban wastewater influent, to obtain stable conditions similar to those occurring naturally in existing urban wastewater collection networks. The COD concentration in the system was thus controlled daily. If it fell, a solution containing between 500 and 600 mg-COD/L was added in the appropriate volume, through tank 1. The circulation system operated for 45 minutes every hour.

The standard titration method (American Public Health Association et al. 2014) was used to determine the COD concentration. The concentration of dissolved oxygen (DO) was measured using a Multi-Function Meter (model WA-2017SD) manufactured by Lutron Company, Inc, while a Multiparameter Photometer (HI-83200 model, made by HANNA Company, Germany) was employed to measure the concentrations of ammonia, ammoniacal nitrogen, nitrate and nitrate-nitrogen.

During the first 10-day period, the estimated mean biofilm thickness in the glass-tube samplers was less than 1.0 mm. Over the 3 ten-day periods, thickness increased to 0.5 to 1.5 mm on the surface of the samplers. Previous research (Daigger et al. 1999; Pochana et al. 1999; Moya et al. 2012; Ganesh et al. 2014) has shown that SND can occur under these conditions, but biofilms of greater thickness are required for stability. During the third 10-day period, the biofilm became almost uniform across the sampler surfaces, with thicknesses of between 2.5 and 5.5 mm.

The SND process is affected by the thickness of the biofilm formed. Usually the thickness is about 0.08 to 0.1 mm but SND only occurs when it is at least 0.125 mm, because it depends on the limited penetration of DO (Wilen & Balmer 1999).

SND can offer advantages over conventional nitrification and denitrification, which requires two separate tanks. SND reduces the space required and simplifies the process. It also uses between 20 and 40% less carbon and, hence, produces up to 30% less sludge (Seifi & Fazaelipoor 2012).

An attempt was made to fit the linear function to the points obtained from the results but this could not be used to recognize the trend type. The fluctuations of nitrite concentration shown in 4-b, 5-b and 6-b were caused by changes in nitrogen content as SND occurred. All three diagrams – Figures 4–6 – show fluctuating nitrate concentrations, accompanied by nitrate removal after between 1 and 2 consecutive hours of nitrate production. Since the ammonia content (ammoniacal nitrogen) in the system always decreased, it appears that the reduction in nitrate simultaneously with its production indicates simultaneity of the nitrogen removal stages.

When the pilot-plant was first operated using wastewater containing 340 mg-COD/L, the initial concentration of ammonia was 5.57 mg/L because of the C/N/P ratio, which yielded an initial nitrogen concentration of 4.58 mg/L. The curves in Figure 4 indicate that nitrification began when the ammonia and DO concentrations began to fall, while the nitrate concentration rose from its initial value of zero. The fall and sudden decline in DO concentration, the lowered pH and the decrease in alkalinity, all indicated that nitrification had started. During the first two hours, this increasing trend in nitrification continued, but there was a sudden decline in the third hour. This pattern in time could be explained by the biofilm thickness on the pipe wall and the distance covered by the nitrate through the biofilm, which would indicate the use of nitrate resources and the removal of nitrate-nitrogen from the system.

After ten hours, the total nitrogen removal reached 51% with residual ammoniacal nitrogen representing 61% of the initial quantity (Figure 4). 24 hours after the start, 21% of the ammoniacal nitrogen remained and nitrogen removal had reached 78%. Figure 4 shows that the COD concentration also declined during the first three hours to 20% of its original level – i.e., removal was 80%. It can be inferred from this that nitrogen removal began when the COD concentration decreased sharply. The ammonia removal trend over the first 10 hours was such that a linear function, with slope and y-intercept– 0.25 and 4.3, respectively, gave the best fit, and the estimated regression coefficient was 0.94.

Figure 5 shows the results from the second test, when the initial COD and ammonia concentration was 570 and 7.69 mg/L respectively and an estimated ammoniacal nitrogen concentration 6.32 mg/L. As before, in the 340 mg/L trial outlined above, the extent of changes in carbon and DO concentration, in biofilm thickness, and in nitrogen concentration all indicate that SND occurred. The changes observed in DO concentration suggest that the system was aerobic at all times. The concentration of DO and alkalinity both declined considerably at the start of nitrification.

The trend of the changes in ammonia concentration in this test differed from that in the previous one, but the regression coefficient of the best-fit linear equation was 0.92, with slope −0.6 and y-intercept 5.7. In other words, the rate of reduction in this test was better than before. The total nitrogen content in this test fell by 75% after 10 hours, and 84% after 24 hours. However, the COD content declined considerably in the early stages, with almost 80% removal in the first three hours, after which the proportional removal of nitrogen appeared to increase gradually, causing decreasing nitrogen content after the first three hours. The ammoniacal nitrogen concentration ten hours after the start had fallen to 24% of its initial value, and declined to 16% after 24 hours.

Figure 6 shows the results of the third test. The initial COD and ammonia concentration 860 and 11.53 mg/L respectively (ammoniacal nitrogen 9.48 mg/L). In this test, as before, falls in the concentrations of DO, alkalinity, and ammonia could be observed within an hour of start-up, and nitrate production – the start nitrification – began in the second hour.

The rate of ammonia removal in this test was higher than in the previous two, and the slope of the line and its y-intercept were – 0.71 and 8.9, respectively, with a regression coefficient of 0.95. After ten hours, 67% of nitrogen had been removed and the ammoniacal nitrogen concentration was 33% of its initial value. Nitrogen removal had risen to 78% at the end of 24 hours and only 23% of the original ammonia content remained. The COD concentration was almost constant after five hours, some 88% of the initial content having been removed in the first four hours.

Comparisons of the three different test runs, show that total nitrogen removal efficiency was at its best in the second run – 570 mg-COD/L. The trends observed and the biofilm formation in the pipes, both indicate a high probability that SND occurs in such systems. Given the system's aerobic operation and the apparent absence of anaerobic conditions, it seems that denitrification must have occurred simultaneously with nitrification. The results suggest that although high initial nitrogen concentrations helped to increase the removal efficiency of SND, there was a maximum initial nitrogen concentration beyond which removal efficiency would not increase. This optimum concentration had to be determined, taking into account the factors influencing it.

The results indicate that wastewater collection networks can remove a considerable amount of the contaminants present in wastewater before it enters a treatment plant. This protects surface- and ground- water resources, and raises the level of health in the community, especially in remote areas.

The technical results of the work show that the average thickness of the biofilm formed on the plastic netting on the inner pipe walls, after 25 days and during the tests, was between 2.5 and 5.5 mm. Such a thickness is suitable to support SND.

The concentrations of ammonia and nitrate-nitrogen observed show that the initial carbon and nitrogen concentrations in the wastewater are a very important factor in nitrogen source removal from wastewater by SND. Although the rate of COD removal improved with increasing initial concentration, nitrogen removal did not exceed a maximum level. A removal rate of 76% was achieved for ammoniacal nitrogen with an Hydraulic Retention Time (HRT) of 10 hours, when the initial COD concentration was 570 mg/L.