Side stream treatment nitrogen removal: alternatives for New York City

New York City Environmental Protection (NYCEP) owns and operates fourteen waste water treatment plants (WWTPs). However, only eight of the fourteen plants have dewatering facilities some of which handle digested sludge which is either barged or piped from multiple WWTPs. As part of their long term nitrogen control program (LTCP) to address hypoxia issues in the receiving water bodies of Long Island Sound (LIS) and Jamaica Bay, the City is in the process of upgrading their plants for biological nitrogen removal (BNR). The existing step-feed activated sludge process was modified to include a series of anoxic/oxic zones to achieve nitrogen removal. However, the reject water from the dewatering of the digested sludge, called centrate, can contribute up to 40% of the total nitrogen load in some of the plants. Therefore, a more cost effective approach in these cases was to treat centrate for nitrogen removal separately (Janus & van der Roest 1997). In addition, significant savings can be realized in separate centrate treatment (SCT) if in lieu of the traditional nitrification/denitrification route, newer process alternatives such as using the nitrogen shunt wherein ammonia oxidation is restricted to nitrite only and then reduced to nitrogen gas, are incorporated (Henze et al. 2008).

Several treatment approaches are being considered or are in use by the NYCEP for SCT that include:

• Conventional nitrogen removal in a dedicated tank (SCT), segmented in anoxic/oxic zones with alkalinity and carbon addition. Such SCTs are located at the Hunts Point, Bowery Bay, and 26^th Ward WWTPs.

• A full-scale facility of the SHARON^® process at the Wards Island WWTP.

• A pilot of a single stage partial nitritation/anammox (PNA) MBBR process. The pilot study was conducted by the City College of New York (CCNY) at the 26th Ward WWTP.

Table 1 shows the average concentration of the different parameters along with the standard deviation (Std. Dev.) and maximum values for a three-year period. The large variations in the values of the Std. Dev. reflects the big fluctuations experienced by the facility as a consequence of it being a centralized dewatering facility accepting digested sludge from different WWTPs.

Centrate is fairly basic with a relatively stable pH. However, available alkalinity in centrate is limiting for anammox process based on stoichiometry. Removal of nitrogen by anammox requires a ratio of nitrite to ammonia of 1.32:1. Hence, any additional removal will require supplemental alkalinity addition.

Conventional BNR includes oxidation of ammonia to nitrate (nitrification) and then reduction of nitrate to nitrogen gas (denitrification). Nitrification requires high aeration and supplemental alkalinity addition for pH control and an external carbon source for denitrification (Ahn 2006).

SHARON^® (Single reactor High activity Ammonia Removal over Nitrite) is an alternative process for nitrogen removal from ammonia enriched streams (Hellinga et al. 1998). In this process, controlling solid retention time and taking advantage of high temperature, ammonia oxidizers dominate, while the nitrite oxidizing bacteria get washed out (Ganigué et al. 2012). Therefore, ammonia is oxidized to nitrite and nitrite is reduced to nitrogen gas by an external carbon source. The inherent ability to achieve nitritation predominantly results in 25% less aeration demand, 40% less external carbon source addition, 30% lower sludge production and 20% less CO₂ emissions compared to conventional nitrogen removal (Sri Shalini & Joseph 2012).

As part of the Nitrogen Removal Applied Research Program, NYCEP undertook installation of the world's largest full scale SHARON^® reactor at the Wards Island WWTP, a site of one of the main centralized sludge dewatering facilities. The SHARON^® process is currently being operated by NYCEP and being evaluated for nitrogen removal using methanol/glycerol as the carbon source.

The SHARON^® system was designed to treat an average flow of 1.85 MGD (7,000 m³/day) with a nitrogen load of 5,000 Kg N/day and to achieve 85% of total inorganic nitrogen (TIN) removal. The process is operated at a temperature of 95°F (35°C), slightly above the temperature of centrate from the dewatering of sludge from mesophilic anaerobic digesters. The biochemical reactions occurring within the SHARON^® reactors are exothermic, generating sufficient heat that necessitates cooling provisions during parts of the year to sustain a constant temperature profile.

In the PNA process, only 50% oxidation of ammonia to nitrite by ammonia oxidizing bacteria (partial nitritation) is required and subsequently, the anammox bacteria oxidizes the remaining ammonia using the nitrite produced in the previous step to nitrogen gas but produces a small amount of nitrate. Hence, the maximum theoretical nitrogen removal by this process is 89% with the remaining 11% being nitrate produced.

Anammox has a long duplication time of 0.003 h⁻¹ at 32–33°C (Strous et al. 1998). Slow growth rate and low biomass yields of these bacteria lead to long start up periods. In order to make the full scale implementation of anammox feasible, improving biomass retention by minimizing wash-out is critical. However, the tendency of these bacteria to form biofilms on abiotic surfaces or as compact suspended aggregates in the form of granules allows for design of processes with highly efficient biomass retention (Fernández et al. 2008).

The institution of the PNA process for side stream treatment in municipal wastewater treatment plants reduces biomass production, eliminates the need for external carbon sources and decreases the greenhouse gas emissions by approximately 90% since CO₂ is consumed and N₂O emissions are non-existent, thus reducing the overall operating expenses by approximately 60%. (Henze et al. 2008; Malamis et al. 2013).

Currently, there are only a handful of small full-scale implementation sites using PNA for nitrogen removal from centrate in the United States. The MBBR pilot study at the 26^th Ward WWTP showed the enormous potential of this process for New York City when 80–90% of nitrogen removal was achieved with minimal alkalinity addition (Mehrdad et al. 2013).

Table 2 provides a comparison of the volumetric loading and removal rates of the three technologies and the obvious advantage from a foot print stand point is clearly evident from the table.

The MBBR pilot at 26^th Ward WWTP received the highest nitrogen loading rate of 0.46 kg N/m³.day. With this loading, high nitrogen removal rate of 0.28 kg N.m³.day was achieved by PNA process.

All three processes described above are viable alternatives that are also being used at various installations around the world. Two of the three are currently in use in the New York City system and have been critical components for the overall nitrogen removal program. However, as WWTPs are viewed increasingly as water resource recovery facilities, process selection and operations must consider energy usage and possibly aim to be a net zero or positive energy facility. PNA is a nitrogen removal process that fits this low energy, low carbon footprint framework exceedingly well with its low aeration for nitritation and no external carbon requirements for denitritation.

With this objective, the existing AT-3 SCT is compared to SHARON and the proposed PNA process for treating the same amount of flow and at similar levels of removal.

The annual cost was estimated for three processes based on the centrate characteristic at the 26^th Ward WWTP with a flow of 1.3 MGD (4920 m³/day) and an average ammonia concentration of 400 mg N/L as shown in Table 1. All calculations are based on the stoichiometry of each of the processes. The annual cost is calculated based on the actual expenses for caustic and glycerol consumption in NYC WWTPs by NYCEP in 2012.

Table 3 delineates the associated costs of the AT-3 SCT process at 80% removal versus SHARON and the anammox process at the same level of removal. In order to achieve 80% TIN removal from the centrate at the 26^th Ward WWTP by the PNA process, alkalinity addition is required because the PNA process does not get any credit of alkalinity as the other two processes do. Based on the ammonia nitritation stoichiometry, the inherently available alkalinity in centrate which is shown in Table 1 is sufficient to oxidize only 48% of the ammonia content. This amount of alkalinity would restrict ammonia removal to between 60 and 70% of nitrogen removal compared to the 89% achievable based on stoichiometry.

To treat 1.3 MGD (0.057 m³/s) of centrate for 80% of TIN removal efficiency, AT-3 SCT using conventional treatment and the SHARON process using nitritation and carbon denitritation cost approximately $2.5 and $1.8 million annually, respectively. However, the PNA process costs $0.57 million per year for the same level of treatment, as shown in Table 3. The substantial part of the savings is derived from the external carbon source cost. Since there is no credit for denitrification in the PNA process, the caustic consumption is marginally higher than what is consumed by the SHARON^® process.

Table 4 compares the three processes on a cost per pound of nitrogen removed and on all counts, the PNA process comes out as the winner. For an 80% removal efficiency, the carbon costs for the SHARON^® process is approximately half of the conventional process while the cost of caustic is highest for the PNA process because of zero credits for denitrification. In spite of this, the total cost per pound of N removed is only 65% of the SHARON^® and 35% of the conventional SCT process.

In comparing the three SCT processes for nitrogen removal for the City of New York, it has become abundantly clear that the PNA process can deliver the same or higher degree of removal at a substantially lower operating costs which translates into savings in the millions of dollars. As an example, by implementing PNA at the 26^th Ward WWTP, even at the 70% removal level, potential savings compared to conventional removal at a flow of 1.4 MGD (0.061 m³/s) will be of 1,100 megawatt hours of electricity per year, 2,000 metric tons of glycerol per year, 2,600 metric tons of CO₂ emissions per year resulting in an overall saving of US $ 2.2 Mil per year

As the City moves forward to implement the PNA process, it will continue to accrue substantial savings for the next several decades which include 60% reduction in aeration requirements, 90% reduction in sludge production, 50% reduction in alkalinity and no external carbon requirement.