Improved nitrogen removal efficiency by implementation of intermittent aeration

The intermittent aeration system is an activated sludge system in which nitrification and denitrification take place in one reactor. Thus, the activation tank is not structurally divided into a denitrification section and a nitrification section, but both processes take place in one activation tank and are separated from each other only in time (Hanhan et al. 2011). Similar to a carousel system, there is no need of implementation of internal recycle. It is necessary to alternate oxic and anoxic conditions in the tank so that both processes can take place to a sufficient extent. Due to its arrangement, the system allows to extend the aerobic age of the sludge arbitrarily. It is advantageous, especially during treatment plant commissioning or during winter operation, and furthermore it allows easy optimization of the nitrogen removal process (Fiala 2005).

Intermittent aeration control for increasing the efficiency of total nitrogen removal focuses on optimal distribution of aerated and non-aerated phases in the activation process. The aim is to achieve the highest possible quality of effluent of total nitrogen varying between 4 and 10 mg/L on a long-term average basis (Henze 1991). In addition, high efficiencies (90–97%) in removing organic compounds can be expected (Dan et al. 2021). To identify the optimal distribution of aerated and non-aerated phases, several approaches to determine the required length of each phase for both ammoniacal nitrogen removal and nitrate removal in real-time can be used, as discussed in the following subsections.

In the Czech Republic, the most commonly used method of intermittent aeration control is setting fixed periods for nitrification and denitrification. The main advantage of this setting is its simplicity and low requirements for instrumentation and control system. However, the main limitation of this system is its inability of flexible response to changes in the quantity and composition of the influent wastewater. In an ideal system the aerated period length would be reduced during night hours due to lower ammoniacal nitrogen loading. Conversely, during peak loading, the aeration time would be extended to ensure sufficient nitrification. In fixed period systems adjustments of settings are only possible on the scale of seasonal changes and they have to be evaluated by a qualified process engineer based on results of the operational laboratory monitoring. The procedure for the correct setting of fixed aeration times and their adjustments according to variations in the treatment plant load is described in this article.

These systems were originally developed for SBR reactors making their application in intermittent aeration activation difficult. Their basic feature is that the termination of the nitrification or denitrification phase is not prompted by the concentration of the relevant nitrogen form but by changes in other parameters in the controlled tank. These are points at which typically 90–100% removal of the relevant nitrogen form is achieved and which are detected by time-dependent parameters such as pH, oxidation-reduction potential (ORP) or respiration rate (OUR) measured in real time. An example of typical application is a system based on the termination of nitrification/denitrification phase after maximum/minimum ORP value is reached or alternatively the termination of nitrification phase after the OUR decreases as detected by continuous derivation. These systems can contribute to energy savings of about 5% of the WWTP consumption without affecting the nitrogen removal efficiency (Caulet et al. 1998; Corominas et al. 2006).

The ability to control aeration based on the monitored concentrations of oxidised and reduced nitrogen forms is naturally used in many systems. One of the simplest is a system that turns aeration off when a minimum defined concentration of ammoniacal nitrogen is reached and on when a maximum defined concentration of ammoniacal nitrogen is reached (Kalker et al. 1999). In a system where total nitrogen removal is required, use of nitrate nitrogen instead of ammoniacal nitrogen is preferable. The implementation of such a system to intermittently control the aeration has a positive effect on both power consumption and the effluent concentration of total nitrogen (Hauser 2006). These systems need complex real time analytical instrumentation measuring ammoniacal nitrogen and nitrates and can save 5–15% of WWTPs energy or reduce nitrogen concentrations in the outflow by about 30–60% (Lukasse et al. 1999).

The principles of fuzzy logic are commonly used in industrial automation today. For the purpose of control of intermittent aeration, a system that determines the ratio between the duration of the aerated and non-aerated phases based on the magnitude of the ammoniacal nitrogen concentration has been developed and its derivative with respect to time. Such a system can lead to a reduction of the effluent nitrogen concentration of 1% with a simultaneous saving of 10% of WWTP electrical energy consumption. (Kalker et al. 1999).

An example of a system using the mathematical model is given by Lukasse et al. (1998). A simplification of the ASM1 model (see Henze et al. 2000) is used as a process model, where the switching of oxygen functions is replaced by a discrete model. Here, ammoniacal and nitrate nitrogen concentrations are measured not by real-time probes but by an automatic analyser with a measurement interval of 20 minutes. The model therefore always predicts values within this time frame and then compares them with reality. This comparison is the basis for the application of the optimisation algorithm, which then ensures the adjustment of the model for further prediction. Using this system a saving of 6% of WWTP energy consumption was achieved without any negative impact on the nitrogen effluent concentrations.

However, even with the use of aerated and non-aerated time automatic control systems, some limitations in the settings of these times cannot be avoided. It is necessary to set the minimum duration of each phase so that frequent switching on and off does not lead to faster wear of the aeration system, especially the blowers. Therefore, the minimum phase duration is usually limited to a minimum of 15 minutes (Holenda et al. 2007). This minimum value is also needed to allow the bacteria to consume dissolved oxygen in the tank after switching the aeration off and to change their metabolic pathways from oxic to anoxic. The maximum duration of the non-aerated phase must be limited if the tank is not equipped with mixing so that sedimentation of the activated sludge does not occur (Holenda et al. 2007). Therefore, it is recommended to leave aeration off for a maximum of 120 minutes.

Design of an intermittent aeration system must consider that, unlike a continuously aerated activation system, the intermittent aeration blowers run for shorter time and the required amount of oxygen in kg must therefore be delivered to the system in a shorter time interval. For example, if intermittent aeration will operate in a 1:1 ratio (non-aerated vs. aerated period), it is necessary to dimension the aeration system for a double air supply compared to continuous aeration (Novák et al. 2012). The intermittent aeration system is highly efficient and can be used very effectively to intensify existing WWTPs. Provided that the system is not loaded near to its designed capacity, the conversion of operation to intermittent aeration is usually successful (Novák et al. 2012).

The aim of this work is to verify the possibility of using intermittent aeration to decrease outflow total nitrogen concentrations in municipal WWTPs of different capacities (15,780, 23,000 and 806,250 PE). A simple intermittent aeration control option based on time intervals was used in the experimental work, and the results of WWTPs are compared with each other and with literature sources.

Intermittent aeration was applied to three WWTPs with the following parameters (Table 1):

The WWTP Sedlčany is a mechanical-biological wastewater treatment plant designed for 23,000 PE₆₀. Wastewater collected by the combined sewer system flows through the grit trap and then is pumped via a pumping station to mechanical treatment which consists of automatic fine screens and a sand trap. Subsequently wastewater is divided into two identical lines consisting of a primary settling tank and biological stage arranged as the D-D/N-N system with internal recirculation. Each nitrification tank is equipped with an oxygen probe, the blower is controlled by a frequency converter based on the measured oxygen concentration. Chemical precipitation of phosphorus is achieved by coagulant (ferric sulphate) dosing in front of the settling tanks. Primary and excess sludge are managed in sludge management, which consists of mechanical thickening, sanitation equipment (pasteurization), anaerobic mesophilic digestion tank, storage tank, gas management and mechanical sludge dewatering.

The WWTP Uhříněves is located in the southeast of Prague and treats wastewater from the Uhříněves and Dubeč districts. It is a mechanical-biological treatment plant with a design capacity of 15,780 PE₆₀ and with gravitational flow, equipped with storage and dewatering facilities for excess sludge. Mechanically pre-treated wastewater flows into a two-line biological stage arranged as the D-D/N-N-postD-postN system. The separation of treated water and sludge takes place in a pair of circular settling tanks. Phosphorus is precipitated by dosing ferric or ferric-aluminium coagulant. The possibility of external substrate dosing for improved denitrification is not currently used due to the favourable ratio of organic substrate to nitrogen.

After screening, the wastewater is mechanically pre-treated in lamella sedimentation tanks Densadeg 4D with integrated sand and grease removal. Biological treatment of wastewater is carried out in four identical activated sludge process lines. Each line consists of a three-stage cascade activated sludge process – ALPHA, which is an activated sludge technology with three anoxic/oxic zones in series. The oxic zones are constructed as carousel tanks. The inflow is distributed to individual anoxic zones in a ratio of 39:33:28. External substrate can be dosed to denitrification sections. The regeneration zone is common for two lines. Mixed sludge from the activated sludge process tanks is led to 40 rectangular clarifiers. Return sludge is subjected to oxic regeneration (Kos et al. 1992). The regeneration zone is also supplied with reject water (centrate) for in-situ bioaugmentation of nitrification. Chemical phosphorus removal is carried out as post-precipitation at the third stage of treatment in lamella sedimentation tanks Densadeg 2D. A mixture of primary and tertiary sludge is thickened in gravity thickeners, excess biological sludge is pre-thickened in gravity thickener Drainis Turbo and thickened by centrifuges. All thickened sludge, together with separated grease, is pumped to the sludge management unit for further processing.

The operation of the WWTP Sedlčany was affected by a significant change in operating conditions. In 2017, the load of the WWTP was 19,395 PE₆₀, which represented 84% of the total design capacity. In 2018, gradual reduction in operation of a very important dairy producer of industrial wastewater began. At the end of the year, the dairy factory shut down completely. In the beginning of 2019 the dairy's technologies were fully terminated and only wastewater from remediation work was produced. Since May 2019, the dairy factory has not discharged any wastewater into the sewerage network of the town of Sedlčany.

In 2020, the first year after the closure of the dairy factory, the load of WWTP was 8544 PE₆₀, which represented only 37% of the total design capacity. A significant decrease in dry-weather flow, BOD₅, and COD_Cr was observed. A comparison of the individual parameters is given in Table 2. As the load and the share of easily biodegradable substrate decreased, problems with high oxygen concentrations in the nitrification tanks began to occur. In addition, excessive foaming of activation tanks was observed, which was probably also related to the high oxygen input.

The average outflow total nitrogen concentration obtained before the introduction of intermittent aeration was 10.2 mg/L. Taking into account the high values from the spring 2019, the concentration even reached 14.2 mg/L. After the introduction and optimization of the intermittent aeration setting, values of around 6.2 mg/L were achieved and remained stable for the rest of the year. These results are comparable with fuzzy logic control systems but are still lower than systems based on rule-based systems for N-NH₄⁺ and N-NO₃⁻ concentrations (Srb 2011). As it is a smaller WWTP, it is not equipped with separate metering of electricity consumption for aeration. The aim of introducing intermittent aeration in this case was not to save electricity but to improve nitrogen removal under conditions of reduced load below design values. The overall energy consumption decreased after the introduction of intermittent aeration went from 1.6 to 1.5 kWh/kg BOD₅.

Table 3 shows that the treatment plant in Uhříněves was loaded to approximately 55% of its design capacity until 2021 when the connection of additional producers to the treatment plant was approved.

The biological line was operated in the regime of maximum use of denitrification, meaning that sections with the alternative D/N and the post-denitrification section were operated as anoxic all the time, even during winter periods. Biological line is equipped by blowers in the 2 + 1 assembly, where one blower is designed to supply air to one biological line. Due to the low loading of the plant, air was supplied to both biological lines by one blower. The blower power was operatively adjusted based on oxygen concentration measured by online probes at the end of nitrification sections. Even though only one blower was in operation, the total nitrogen in the effluent was just below the required emission limit of 15 mg/L. During low loading periods, oxygen concentrations in nitrification sections increased above 3 mg/L, which in turn led to the introduction of excessive amounts of oxygen into the denitrification sections. Therefore, in 2018, intermittent aeration was implemented.

In Uhříněves, intermittent aeration was controlled simply by selecting the running time and idle time of the blowers. In the mode of intermittent aeration, the supply of air to each line was ensured by a separate blower. During the aeration time, the blower output was controlled by the oxygen probe in accordance with the defined maximum and minimum oxygen concentration. If the oxygen concentration surpassed the default maximum during the aeration phase, the blower switched off. If the minimum oxygen concentration value was reached, the blower switched on. During the rest period, the blower did not switch on to provide the desirable length of anoxic period. The set running time of the blowers changed during the year, ranging from 40 to 90 minutes, the selected rest time was most often 20 minutes.

The differences in the average outflow total nitrogen concentrations in the periods with permanent and intermittent aeration are summarized in Table 4.

Due to the low load of the WWTP and the sufficient power, blowers are able to reach the required oxygen concentration (currently 1.6 mg/L) in the nitrification section in a matter of minutes. Thus, intermittent aeration at this WWTP is functioning very effectively. Average outflow total nitrogen concentration was reduced from the initial 14.4 to 7.5 mg/L after implementation of intermittent aeration, which represents a 48% reduction. These results are comparable with systems based on rule-based systems for N-NH₄⁺ and N-NO₃⁻ concentrations (Srb 2011). As it is a smaller WWTP, it is not equipped with separate metering of electricity consumption for aeration. The total energy consumption decreased after the implementation of intermittent aeration from 2.9 to 2.6 kWh/kg BOD₅ removed. The aim of the implementation of intermittent aeration in this case was not to save electricity, but to improve nitrogen removal without the need to inject external substrate.

The NWL was designed with full aeration of all nitrification zones with the possibility of setting the required oxygen concentration individually for each nitrification section. NWL was put in operation without sludge inoculation and when its commission began all aeration basins were fully filled with river water. After two months of gradual increase of loading, full load was achieved with an average daily BOD load of 48.4 t/d and total nitrogen load of 11.0 t/d. Full nitrification process was reliably achieved in 43 days from the start of operation. While the effluent N-NH₄ concentrations were close to zero, the denitrification capacity was not sufficient and the outflow total nitrogen limit was not met due to high concentrations of nitrates.

One set of blowers is shared for all first nitrification sections and both regeneration tanks. The second set of blowers is shared for all second and third nitrification sections and for degassing tanks of all four activation tanks. The two sets of blowers for the biological part of the treatment are located in two separate blower rooms that are not connected by any piping. The simple solution of reducing the number of blowers in operation to ensure intermittent aeration in all four activation tanks was not possible. Another complication was the minimum capacity of each blower (11,450 Nm³/h) that had to be met to avoid frequent blowers’ shutdowns. The NWL commissioning was monitored very carefully, both by investor and by contractor, particularly in order to meet the targets set by the FIDIC Yellow Book procedures. Any changes to the original vision, including intermittent aeration, had to be implemented very carefully. Therefore, three individual sequential steps of implementation were chosen.

The second step was the automation of a proven method of intermittent aeration. After a short time of operation, it became apparent that intermittent aeration with fixed phase lengths was not optimal because it did not respond to the variable daily loads. Peak of N-NH₄⁺ concentrations at the influent caused overloading of the nitrification sections resulting in higher N-NH₄⁺ concentrations of around 5–7 mg/L at the outlet of the activation tanks. To prevent nitrification overload, adjustable ox:anox ratio conditions during the 4-hour cycle were introduced. The duration of the anoxic conditions was variable from 25 to 80 minutes and different settings were allowed for individual nitrification sections N1, N2 and N3. However, this solution still did not provide full adaptation to the variable daily load.

In the third step, it was necessary to ensure automatic adaptation of the activation conditions to the inlet load variability (N-NH₄⁺ concentration 30–70 mg/L).

For historical reasons, the original design did not include automatic aeration control based on the N-NH₄⁺ or N-NO₃⁻ concentration which are measured by online probes in each section of the activation. Automatic control of oxygen concentration based on N-NH₄⁺ concentration is planned after the trial phase will be completed.

Therefore, to ensure at least a partial response of the intermittent aeration to the inlet load, 24-hour zones were created. Each hourly zone contains information on the length of the anoxic phases. The length is manually determined according to the typical daily N-NH₄⁺ inlet load unevenness. Inlet N-NH₄⁺ concentration is measured by an online probe. The setting correction is made according to the N-NO_x concentration in the denitrification sections. At the outlet of each activation both of these forms of nitrogen are measured. The setting must be checked regularly by process engineers to ensure that it corresponds to seasonal fluctuations. However, a daily check is not necessary.

The length of anoxic phases is adjustable and varies from 25 to 108 minutes. However, the start of the anoxic conditions phase in each of the four activations must be delayed in time regardless of the inlet load in order to meet the technical requirements of the blowers. For this reason, an anoxic phase longer than 108 to a maximum of 120 minutes in the extreme is not possible.

For a faster transition from oxic to anoxic conditions the required dissolved oxygen concentration was reduced from 3.0, 2.5 and 2.0 mg/L to 1.5, 1.5 and 2 mg/L in oxic sections N1, N2, N3 respectively and subsequently to 1.5 mg/L in all oxic sections.

The introduction of automatic switching of the oxic and anoxic conditions led to a further reduction of the N-NO₃⁻ concentration in the WWTP effluent and, in particular, contributed significantly to the sustainability of effluent quality. During the summer periods in 2020 and 2021, the total nitrogen concentration in the effluent was below 6 mg/L.

Implementation of intermittent aeration led to a significant improvement. The required effluent quality of 10 mg/L of total nitrogen as annual average was met without a problem. At the same time, electricity consumption has also been reduced, as only 1–2 blowers are needed for most of the time of day, instead of the original 4–6. Due to the rapid application of intermittent aeration, it was not possible to accurately evaluate the energy savings at NWL. However, savings of up to 40% can be expected (Huang et al. 2021). Blower consumption decreased from 11,091 MWh/year in 2019 to 5490 MWh/year in 2020, although the decrease in influent load between these two years was less than 10%. Thus, the 50.5% decrease in consumption corresponds to the literature data presented. Even though the reduction in energy consumption could not have been calculated due to the unavailability of energy consumption data during the commissioning period, the estimation based on the number of blowers in operation shows a significant reduction of energy consumption comparable with systems based on rule-based systems for N-NH₄⁺ and N-NO₃⁻ concentrations (Srb 2011).

The impact on outflow total nitrogen concentrations is summarized in Table 5. The achieved results show a possible decrease of outflow total nitrogen concentration by up to 57%. When comparing the results from individual WWTPs there are differences in the objectives and results of intermittent aeration and in the size of the WWTPs. The largest implementation is NWL Prague where the largest range of values for the aeration on and off times exist. This is due to the large range of load changes and the possibility of more dynamic adjustments on a large WWTP equipped with on-line analytical techniques. As a result of these dynamics we can see a very stable outflow total nitrogen concentration. At both smaller WWTPs (Uhříněvěs and Sedlčany) it can be seen that the times are set much more tightly than at NWL, while at Sedlčany WWTP there was basically only one adjustment (extension) of the times. This is consequently reflected in a greater variability of the outflow total nitrogen concentrations.

Intermittent aeration at WWTP with capacity reserve allows a decrease of outflow total nitrogen concentrations without increasing investment and operating costs. In addition, this system saves electricity and allows a wide range and variability of process control options. Although the time-based control is very simple and inexpensive to implement, it is highly efficient and can be used very effectively to intensify existing WWTPs. Provided that the system is not loaded near to its designed capacity, the conversion of operation to intermittent aeration is usually successful. The system enables up to 57% reduction of outflow total nitrogen concentrations and/or energy consumption comparable to more complex systems, such as systems based on rule-based systems for N-NH₄⁺ and N-NO₃⁻ concentrations or fuzzy logic or mathematical model systems. As the intermittent aeration proved its potential to adapt existing WWTPs to different load scenarios, the authors highly recommend considering the possibility of using intermittent aeration in the design stage of blowers and aeration systems for new WWTPs.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.