The role of innovative technologies in achieving energy-neutral wastewater treatment

The energy content of municipal wastewater is two to four times greater than the energy required to treat it (Tchobanoglous 2009). However current technologies and practices do not exploit this to the full extent. As a result, the energy demand for wastewater treatment remains significant: in the United Kingdom it represents 15% of operating costs for the water industry (UK KTN 2008); in the United States, it represents 3% of the total electricity demand (US EPA 2006). It is anticipated that this demand will continue to grow globally as emerging countries improve their level of sanitation and developed countries pursue higher levels of treatment for reuse and discharge to the environment.

Meeting the objective of energy-neutral wastewater treatment can be achieved by following two parallel paths: (i) minimizing the energy required for the removal of solids, organics and nutrients, and (ii) maximizing the conversion of organics to usable energy. Both of these paths require innovative technologies to achieve the goal.

A new energy-neutral wastewater treatment flowsheet is presented in this paper. The flowsheet meets the following objectives:

• 1. It is ‘electricity-neutral’. The electricity produced meets the electricity demand for treatment and useable heat is not accounted for in the energy balance. If useable heat were taken into account, the solution would be largely energy positive.

• 2. It is capable of nitrogen removal using the proven nitrification-denitrification metabolic pathway.

• 3. It does not rely on co-digestion (e.g., of food wastes) to increase energy production.

• 4. It has the potential to be cost-competitive with conventional activated sludge (CAS) treatment.

• 5. It is applicable for new plants and retrofitting existing activated sludge plants.

This paper compares the new energy-neutral flowsheet to a reference CAS flowsheet using simulation software. Experimental results on various aspects of the new flowsheet are presented in companion papers (Adams et al. 2014; Peeters et al. 2014).

Recent efforts in municipal wastewater treatment have been focused on the reduction of energy consumption and on the diversion of organic carbon to anaerobic digestion to produce energy through the combustion of biogas in a combined heat and power (CHP) system (US EPA 2008; WEF 2009). Energy-neutrality has been achieved at the Strass plant in Austria, but with the addition of external organic wastes to the anaerobic digesters (Wett et al. 2007). Several plants in North America are pursuing an energy-neutrality goal (Bunce et al. 2013; Stinson et al. 2013).

The pursuit of energy-neutrality should not be achieved at the expense of effluent quality; nitrogen removal has proven particularly challenging for the technical solutions that have been proposed to-date. In general, maximizing the diversion of organic carbon to energy production does not leave enough soluble carbon for conventional nitrification-denitrification. Therefore, recent research has investigated alternate pathways for the removal of nitrogen in order to reduce the energy consumption for nitrogen removal and overcome the challenge of carbon limitation. The alternate pathways include nitrite shunt (i.e., nitritation – denitritation) and de-ammonification (i.e., partial nitritation and anaerobic ammonia oxidation through anammox bacteria). These approaches face several challenges, including:

• 1. Process conditions. De-ammonification is proven for side-stream treatment of ammonia-rich liquors (e.g., anaerobic digestate), but does not appear adapted for mainstream implementation. There are many difficulties, including the low temperature of wastewater, and the suppression of nitrite oxidizing bacteria (NOB) which compete with heterotrophs and anammox bacteria for nitrite. Stinson et al. (2013) listed 10 potential NOB suppression/inhibition mechanisms under investigation.

• 2. Process control. Conventional dissolved oxygen (DO) control is not sufficient to maintain conditions necessary for alternate nitrogen removal pathways. Online inorganic nitrogen instruments are required together with aeration cycling to control biological conditions and minimize effluent total nitrogen (Bunce et al. 2013).

• 3. Effluent and air emissions. These pathways are not conducive to low ammonia concentration and may leave significant amounts of nitrite in the effluent (Bunce et al. 2013). Furthermore, higher nitrite leads to increased emission of N₂O, a greenhouse gas with a global warming potential 300 times greater than CO₂ (De Clippeleir et al. 2012).

• 4. Cost. Maximization of carbon diversion can be expensive, especially if it is done with two suspended sludge systems (i.e., the A/B process). Furthermore, Shiskowski (2013) has demonstrated that side-stream de-ammonification is not cost effective strictly based on energy savings.

• 5. Operation and Maintenance. Finally, the impact of these new processes on operation and maintenance has not been evaluated and may only be feasible for large, sophisticated facilities.

The first component of the new flowsheet is an enhanced primary treatment step that shunts a portion of the organic matter in raw sewage to sludge treatment. It is based on rotating belt sieving. The objective is not necessarily to maximize removal as sufficient organic matter must be allowed to flow to the biological treatment step for nutrient removal. Process control of the technology (e.g., modulation of belt speed) allows for tuning of the organics removal, which is not possible with conventional primary clarification.

The second component of the new flowsheet is a hybrid membrane-aerated biofilm reactor (MABR) process. This concept was first proposed by Côté et al. (2003) and refined by Downing & Nerenberg (2007). An innovative MABR product is described in a companion paper (Adams et al. 2014). Hollow fiber membranes are arranged in modules and cassettes that are deployed in a way similar to immersed hollow fiber filtration membranes used for MBR (e.g., the ZeeWeed² 500 product). Atmospheric air is fed down the lumen of hollow fibers and oxygen is selectively transferred over nitrogen to the biofilm growing on the outer surface of the membrane without the formation of bubbles.

The key feature of this process is a fully oxygenated nitrifying biofilm immersed into an otherwise anoxic suspended biomass. The hybrid MABR enables nitrification by the biofilm, while denitrification and COD removal are performed in the low sludge retention time (SRT) suspended-growth portion of the reactor. Organic solids that are not removed in enhanced primary treatment can flow through the biological process and end up in sludge treatment essentially un-oxidized. A secondary clarifier is used for mixed liquor retention; it can be replaced with a membrane filtration system when reuse quality water is required.

The metabolic reaction pathway for nitrogen removal proposed for the new flowsheet is proven. It is based on autotrophic nitrification and heterotrophic denitrification. This pathway is well understood as it is the result of 100 years of activated sludge evolution and has been implemented in thousands of plants around the world. The process can be controlled to achieve the most stringent total nitrogen effluent concentrations. However, this pathway and, in particular its application in a conventional suspended growth process has had major limitations that are addressed by the new hybrid MABR process proposed in this flowsheet:

• 1. It is energy intensive to fully oxidize ammonia to nitrate. The MABR gas transfer membrane can transfer the oxygen required for full nitrification very efficiently, i.e., at an aeration efficiency >6.0 kg O₂/kWh (Adams et al. 2014) compared to 1 to 2 kg O₂/kWh for fine bubble aeration (Metcalf & Eddy 2003). The result is a 4X reduction in the energy for process aeration.

• 2. Nitrification by slow-growing autotrophic bacteria requires a long SRT, which translates into large tanks and expensive facilities. The nitrifying biomass is attached to the MABR membranes, which allows designing the suspended biomass portion of the process at low SRT, and therefore reduced tank volumes.

• 3. When using bubble aeration, there is a fundamental competition for DO; complete nitrification also results in aerobic metabolization of the organic matter (COD). In the suspended growth portion of the hybrid MABR reactor, the mixed liquor is maintained under anoxic conditions to promote co-denitrification and limit the oxidation of organic matter.

• 4. If there is a stringent limit on total nitrogen, a carbon source may need to be added for denitrification. In the hybrid MABR process, the primary effluent COD is available for denitrification, thus eliminating or reducing supplemental carbon needs compared to conventional nitrification-denitrification processes.

The third component of the new flowsheet involves using anaerobic digestion for the stabilization of primary and secondary sludges and the production of biogas. There is a strong synergy between the rotating belt sieve used for enhanced primary treatment and sludge treatment processes when the former is used for waste activated sludge (WAS) co-thickening. This concept is explored in a second companion paper (Peeters et al. 2014).

The fourth component of the new flowsheet is a CHP system that is used to convert the biogas into usable energy (electricity and heat).

The new energy-neutral flowsheet was compared to a CAS flowsheet to evaluate the energy balance for both. The model plant had an average daily flow of 18,925 m³/d (5 MGD) with the following influent characteristics: BOD/SS/TN/TP = 220/220/40/7 mg/L. It was assumed that the peak hourly flow was two times the average daily flow, the monthly loading peak was 1.35 times average and the winter design temperature was 15 °C. The target effluent characteristics were BOD/SS/TN/TP = 15/15/10/1.0 mg/L; both plants were designed to fully nitrify.

The unit operations of the two flowsheets are described in Table 1.

The process design of the two plants was simulated with GPS-X (Hydromantis Inc.) using default parameters to obtain treatment efficiency, sludge production and the sizing of reactors. HRT and SRT of the biological processes were adjusted to meet the target effluent characteristics stated above. Unless noted otherwise, the same model parameters in GPS-X were used to simulate the two flowsheets.

The plant design philosophy and sizing of equipment were consistent with previous work (Young et al. 2013). Electricity requirements were estimated using the methodology described in Côté et al. (2013).

Selected process modeling results for the CAS and new energy-neutral flowsheets are presented in Table 2. The overall plant electricity balance is presented in Table 3. The electricity consumption is for treatment, and does not include provision for head-works pumping. Design conditions and results are presented in this section.

Both flowsheets included identical 6 mm coarse screens with trash compactors, and grit removal chambers with grit washer/classifiers, with energy consumption of 160 and 170 kWh/d, respectively, for a total of 330 kWh/d.

The CAS flowsheet included two primary clarifiers. The removal of total suspended solids (TSS) was 60%; this corresponded to removal of BOD/COD of 37% based on partitioning of the organic matter in GPS-X. Energy consumption for the clarifier drives, the primary scum pumps and the primary sludge pumps was estimated to be 200 kWh/d.

The new energy-neutral flowsheet was designed with rotating belt sieves, specifically four LEAPprimary³ LP60 units available from GE Water & Process Technologies. Equivalent TSS and COD/BOD removal as for primary clarification was assumed. Energy consumption for the belt drives, the cake removal blower and the sludge compactors was estimated to be 530 kWh/d.

For both flowsheets, phosphorus was removed chemically by alum addition, at a dose of 45 kg Al/d; dosing pumps and rapid mixers had an energy consumption of 110 kWh/d. The additional chemical dosing energy for the CAS flowsheet was for methanol addition.

The CAS flowsheet biological reactors were designed as four lines in parallel, each consisting of 2 anoxic zones and 4 aerobic zones in series. A total HRT of 9.6 h and SRT of 12 d, with a MLSS concentration of 2.7 g/L were established to achieve the effluent total nitrogen target of 10 mg/L; the addition of 450 L/d of methanol was required for denitrification. Two recycle lines were included, for nitrate (4.0Q) and return activated sludge (1.0Q). Aeration blowers were sized to deliver 5,770 Nm³/h, which allowed meeting an oxygen transfer rate (OTR) of 4,300 kg/d using fine bubble diffusers in a 5 m deep tank, using an alpha factor of 0.60. The CAS flowsheet included two secondary clarifiers.

The new flowsheet biological reactors were designed as four lines in parallel, each consisting of four hybrid MABR reactors in series. A total HRT of 7.7 hours and SRT of 5 days, with a MLSS concentration of 1.8 g/L for the suspended biomass portion were able to meet the effluent total nitrogen target of 10 mg/L without methanol addition. Each of the four MABR trains contained immersed MABR cassettes as described by Adams et al. (2014). A single recycle line was included for return activated sludge (0.5Q). Aeration blowers were sized to deliver 1,100 Nm³/h to meet the OTR of 2,900 kg/d. It should be noted that the oxygen demand is 33% lower than that for the CAS flowsheet due the lower SRT of the suspended biomass. The exhaust air from the MABR cassettes is used to renew mixed liquor and provide mixing within the cassette volume (Adams et al. 2014). However, since the cassettes only occupy about 20% of the volume of the tanks, additional mixing was sized to keep the mixed liquor in suspension. For both the MABR tanks and the CAS flowsheet anoxic tanks, it was assumed that mixing could be accomplished with a power input of 4 W/m³ (Badr et al. 2012). The new flowsheet included two secondary clarifiers.

For both flowsheets, the supernatant from solids dewatering was returned to the head of the biological reactors. The COD and ammonia content of these streams represented 2% and 15% of the influent load, respectively.

The hybrid MABR plant has a total reactor volume of 6,000 m³, also in 5 m deep tanks. Aerobic degradation (nitrification and soluble organic matter oxidation) takes place in the biofilm while denitrification happens in the bulk. The MABR membranes, once covered with a biofilm, are not efficient at delivering oxygen to the bulk mixed liquor since the biofilm represents a relatively thick diffusion layer. Therefore, the entire reactor volume is anoxic or slightly aerobic in the downstream tanks. The footprint of the MABR biological system is approximately 25% smaller than that of the CAS system.

The blower energy consumption for the CAS biological system (2,440 kWh/d) represented typical conditions; the aeration efficiency calculated with the OTR in Table 2, 4,300 kg O₂/d / 2,440 kWh/d = 1.8 kg O₂/kWh, is in the middle of the range of what can be achieved with fine bubble aeration (Rosso et al. 2005). The blower energy consumption of the hybrid MABR biological system (490 kWh/d) is 80% lower than that of the CAS system and corresponds to an aeration efficiency of 6.0 kg O₂/kWh. Energy for mechanical mixing of anoxic zones of the two systems is approximately the same. Total energy for biological treatment with the hybrid MABR system is 60% lower than the CAS system.

As stated above, it was assumed that the same amount of primary sludge was produced in the two flowsheets (2,500 kg/d) in order to facilitate the comparison of the biological treatment steps. GPS-X predicted that the CAS and hybrid MABR systems would produce 1,580 kg/d and 1,930 kg/d of WAS, respectively. The higher WAS for the new flowsheet is due to operating at lower SRT and under bulk anoxic conditions, in spite of the fact that external carbon was not added for denitrification. Overall, the new flowsheet produced 8% more bio-solids as compared to the CAS flowsheet.

Sludge thickening for the two flowsheets was handled differently. With the CAS flowsheet, the primary sludge was thickened with gravity thickeners to 8% and the WAS was thickened with rotary drums to 6%. The blended sludge had TSS of 7.1%. With the new flowsheet, the WAS was co-thickened with the primary sludge using the RBS to a blended sludge value of 10.0% (Peeters et al. 2014).

For both flowsheets, the mixed sludges were anaerobically digested with a hydraulic retention time of 25 days. The new flowsheet has a digester 28% smaller than the CAS flowsheet because the mixed sludge has a higher solids concentration. The volatile solids reduction in the new flowsheet was 67% and biogas production was 1,900 m³/d, as compared to 59% and 1,500 m³/d for the CAS flowsheet. The higher production of biogas produced in the new flowsheet (+26%) was due to the higher amount of sludge and higher volatile solids content.

The energy consumption for sludge treatment of the CAS flowsheet had 5 components: 150 kWh/d for the primary sludge gravity thickener, 50 kWh/d for the WAS rotary drum thickener, 140 kWh/d for sludge blending and holding tank mixing, 460 kWh/d for anaerobic digester gas mixing and pumping, and 200 kWh/d for the dewatering centrifuge, for a total of 1,000 kWh/d.

The energy consumption for sludge treatment of the new flowsheet had 3 components since thickening and blending were achieved by the RBS: 60 kWh/d for mixing the holding tank, 360 kWh/d for anaerobic digester gas mixing and pumping, and 200 kWh/d for the dewatering centrifuge, for a total of 620 kWh/d.

The conversion efficiency of biogas into useable energy through CHP systems is well established (US EPA 2008). For the size range of interest in this project, reciprocating engines have an electrical power efficiency of 22–40% and an overall efficiency of 70–80%. In the electricity balance presented in Table 3, a power efficiency of 35% was used to convert biogas energy into useable electricity. Usable heat was not considered in the energy balance.

The CAS flowsheet taken as reference in this project had a specific energy consumption of 0.29 kWh/m³ (Table 3). This is low when compared to values reported by Monteith et al. (2007), ranging between 0.35 and 0.65 kWh/m³, for two reasons. First, our analysis considered energy for treatment only while literature numbers often include head-works pumping. Second, our analysis was based on a plant utilization factor (PUF) of 100% while many plants surveyed are not running at full capacity; conservation of the energy benefits as the PUF decreases depends on the fraction of the power input that can be turned down with the flow (variable power, e.g., aeration) versus the fraction that always runs at full capacity (base power, e.g., clarifier mechanism).

In the new energy-neutral flowsheet, electricity neutrality was achieved through a combination of significant savings in biological treatment (−60%) and enhanced biogas production (+26%). A large portion of the savings are due to more efficient oxygen transfer (−1,950 kWh/d), but also to elimination of nitrate recycling (−400 kWh/d) and running at a lower return activated sludge rate (−160 kWh/d). The gains with biogas production are attributed to diverting more organics to anaerobic digestion (+350 kg/d).

The MABR had very high oxygen transfer efficiency (>60%) and aeration efficiency (6.0 kg O₂/kWh), while working with a low SRT suspended biomass that allowed shunting more sludge to energy production. In a CAS system, oxygen transfer efficiency and shunting biomass are incompatible objectives. Rosso et al. (2005) showed the alpha factor and the standard oxygen transfer efficiency dramatically dropped at SRT less than 5 days.

A detailed cost analysis of flowsheets containing MABR in comparison with CAS was performed by Aybar et al. (2012) with CapdetWorks (Hydromantis, Inc.). This analysis showed that the MABR-containing flowsheets have much lower energy consumption as compared to CAS flowsheets and that their cost effectiveness is sensitive to membrane-related factors, including membrane costs, mixing energy requirements and lifetime of membranes.

While the cost of the new energy-neutral flowsheet introduced in this paper has not been fully evaluated in comparison to a CAS flowsheet, it is recognized that deployment of MABR membranes will represent an additional cost. However, the modelling and design work reported above identifies several opportunities to reduce the capital cost of the new hybrid MABR as compared to a CAS process:

• Reduction of the size of the biological reactors (≈25%)

• Elimination of fine bubble diffusers

• Reduction of the size of blowers (≈75%)

• Elimination of the nitrate recycle stream

• Elimination of the external carbon dosing system

• Reduction of the plant footprint

Furthermore, the new flowsheet would reduce O&M costs by making the plant electricity-neutral and eliminating the need for external carbon addition for denitrification.

In the proposed new flowsheet, additional capital and O&M cost benefits are provided by the replacement of primary clarifiers with rotating belts sieves; these benefits are discussed in a companion paper (Peeters et al. 2014).

Energy-neutral wastewater treatment is an important goal, but it should not be achieved at the expense of effluent quality or plant operability. A new flowsheet is proposed based on a hybrid MABR process. This new flowsheet achieves energy-neutrality (even better, electricity-neutrality) while removing nitrogen using the proven nitrification-denitrification metabolic pathway. Furthermore, it is compatible with solid-liquid separation by conventional clarification or membrane filtration.

The hybrid MABR biological process is based on two sludges, one fixed and one suspended. Since oxygen is not transferred through bubbles, oxidation reactions for nitrification and BOD removal can take place in an otherwise anoxic reactor which allows suspended and colloidal solids to flow through the MABR without being oxidized. The suspended biomass is managed at a low SRT and the organics can be sent to anaerobic digestion through the WAS. Furthermore, denitrification using the influent COD is enhanced.

The new energy-neutral flowsheet was compared to a CAS flowsheet using a wastewater treatment simulator (GPS-X from Hydromantis Inc.). Both flowsheets included complete wastewater and sludge treatment with anaerobic digestion and CHP production. The CAS flowsheet had a specific electricity consumption of 0.29 kWh/m³ while the new flowsheet had 0.17 kWh/m³, a reduction of about 40%. Electricity produced through the CHP system was 0.18 kWh/m³ for the CAS flowsheet and 0.21 kWh/m³ for the new flowsheet, an increase of 18%. Overall, the new flowsheet was electricity-neutral.

Experimental validation of the new flowsheet is in progress. Initial piloting results support the oxygen transfer efficiency and nitrification capacity of the new MABR membrane (Adams et al. 2014), and the synergistic use of rotating belt sieves for primary treatment (Peeters et al. 2014).