Nutrient Removal Project
Simultaneous Nitrification and Denitrification in a Sequencing Batch Reactor
Will Lambert
Robert Nwaokoro
Stephen Russo
CEE 453
12/8/04
Abstract
Nitrogen removal is important part of wastewater treatment, especially in coastal regions where nitrogen loading of oceans and estuaries can cause algal blooms. The most common method of removal uses microbes. Unfortunately, there are several steps to achieve denitrification with an organic nitrogen feed. Many bacteria convert organic nitrogen to ammonia, but only certain anaerobes, namely Nitrobacter and Nitrosomonas, are responsible for the conversion of ammonia to nitrite and nitrate, or nitrification. Furthermore, it is only from the nitrite or nitrate form that anaerobic denitrifying bacteria can produce nitrogen gas. Therefore, traditional nitrogen removal in waste treatment plants occurs in first and aerobic stage and then an anaerobic stage. This can either occur in separate tanks or in one tank that is sequenced to go through each state. Recently, further research has been done on simultaneous nitrification and denitrification. It seems that at very low levels of dissolved oxygen, anaerobic zones can be produced, allowing both processes to occur concurrently.
The purpose of this research is to explore the possibilities of simultaneous nitrification and denitrification in a sequencing batch reactor at a laboratory scale. We used aeration cycles of 8, 18, and 24 hours and a dissolved oxygen level of 0.5 mg/L. Removal efficiencies were on the range of 10-20%, and increased with longer aeration times. Due to unexpected experimental difficulties, we were not able to take as many samples as we would have liked, so there are many options for further research.
Introduction
The objective of our research is to evaluate the efficiency of simultaneous nitrification and denitrification (SND) in a sequencing batch reactor (SBR). Nitrogen removal is an important part of wastewater treatment. Nitrogen is a limiting nutrient for algal growth in oceans and estuaries. Furthermore, organic nitrogen in wastewater is converted to nitrates, which can cause Methemoglobinemia (“blue-baby” syndrome) if it is present in drinking water. Biological nitrogen removal has been highly favored by wastewater treatment plants due to the increasingly stringent effluent discharge requirements on nitrogen. Biological nitrogen removal involves ammonification (converting organic nitrogen to ammonia), nitrification (biological oxidation of ammonium nitrogen to the nitrite), and denitrification (reduction of nitrate to free nitrogen gas via a nitrite intermediate). Traditional nitrogen removal in wastewater treatment plants occurs in separate aerobic and anaerobic stages. SND systems tend to be a cheaper means of achieving these requirements as it uses less bio-reactor volume, consumes less time, produces less sludge and has lower energy costs than current sequential nitrification and denitrification systems.
A potential problem with the SND setup is the anoxic requirement of the denitrification process which we intend to synchronize with the aerobic nitrification process. An aerobic condition (a measurable bulk liquid dissolved oxygen (DO)) inhibits the nitrifying bacteria Nitrosomonasand Nitrobacter, while an anaerobic situation (bulk liquid DO 0.5 – 2 mg/L) inhibits denitrifying bacteria such as Thiobacillus denitrificansandPseudomonas denitrificans. Denitrification occurs in a series of steps where nitrate is converted to nitrite and ultimately to gaseous nitrogen. A general schematic of the nitrogen cycle as it occurs in activated sludge treatment is given below:
Figure 1: Schematic of the nitrogen cycle in wastewater treatment plants.
A compromise situation at which both nitrification and denitrification can occur simultaneously requires the bioreactor to operate at the minimum DO concentration necessary to support nitrification ( 0.5 mg/L) (Munch, et al 1996).
The major challenge is creating the optimal conditions necessary for an efficient SND process: appropriate hydraulic/solids retention times, optimal aeration rates, and suitable feed introduction strategy. These conditions must be met while minimizing odor generation and the accumulation of free ammonia gas. Another potential difficulty is the selection of filamentous bacteria at low DO that interfere with settling and can cause sludge bulking in a reactor. However, these same bacterial flocs (sludge bulking) may provide sites of anoxic zones where denitrification can take place.
Ideally, dissolved oxygen control should be based on levels of nitrate and nitrite in the system, but access to reliable nitrogen probes is limited. Therefore, airflow rates will be controlled by the DO level in the system using the proportional integral derivative (PID) control. Nitrate and nitrite levels will be measured using the Hach Cadmium Test with the aid of a UV Spectrophotometer. Ammonia levels in the system are monitored using the micro-phenate test, also with the aid of a UV Spectrophotometer. Assuming that all organic nitrogen is converted to ammonia, these tests allow us to measure total nitrogen in an effluent sample. The nitrogen input of the waste is known, so nitrogen removal is easily calculated. Reported total nitrogen removal efficiencies vary between 73% and 91% (Sverdlikov, et al 1999). The Sverdlikov Group also found that complete nitrification, 90% denitrification, and 91% total nitrogen removal efficiency is attainable at high hydraulic and volumetric loading rates (1999).
Plant Setup and Characteristics
To implement Simultaneous Nitrification-Denitrification, we used an automated Sequencing Batch Reactor. The reactor system was built using these main components:
4 L reactor vessel
magnetic stirrer
peristaltic pump
pipe-thread solenoid control valves
tubing, bulk head fittings, clamps and clamp holders.
dissolved oxygen probe and pressure sensors
diffuser stone
data acquisition and process controller
The reactor vessel was mounted on the magnetic stirrer base, set to a stir speed of 5. The synthetic waste solution stored in the shelf refrigerator was input to the reactor using the peristaltic pump controlled by the process controller. The solenoid control valves were an integral part of the plumbing system by controlling:
the flow of compressed air
the drainage of effluents
the regulation of waste and tap water influent
The dissolved oxygen probe was fastened by the clamps and situated such that contact with aeration bubbles via the diffuser stone was limited. The pressure sensor provided the process controller with reactor fluid volume. The process control software allowed us to automate the SBR using desired set points.
Plant Operation
Figure 2: Schematic of the reactor system used. Air delivery system shown in detail.
To operate the plant, the above components were configured as illustrated in Figure 2 above. A test run of the system identified and led to the repair of any leaks in the plumbing connections. The plant was then configured to execute five states involved in wastewater treatment cycle in a SBR:
Fill with waste
Fill with water
Aerate
Settle
Drain
The above setup was designed to automatically cycle through the five states in sequence, repeating at the completion of each cycle. The process controller automatically defaults to an off-status if any error occurs in any of the steps. However, we were unable to implement a configuration for a system shutdown in case of a sensor read failure.
Fill with Waste
A 100x stock synthetic waste was diluted to a 20x feed waste and both were kept in the shelf refrigerator. Table 1 below shows the recipe for the desired solution in our plants.
Table 1: 100x stock synthetic waste contents and their concentrations.
Compound / Chemical Formula / Concentration (mg/L)Starch / 84.40
Casein / 125.00
Sodium acetate / C2H3O2Na3H20 / 31.90
Capric acid / C10H20O2 / 11.60
Ammonium chloride / NH4Cl / 75.33
Potassium phosphate / K2HPO4 / 6.90
Sodium hydroxide / NaOH / 175.00
Glycerol / C3H8O3 / 12.00
Refridgerated storage prevented microbial degradation of the synthetic waste until its use in the reactor. The synthetic waste contained 40.9 mg/L of nitrogen and a chemical oxygen demand (COD) of 325 mg/L. The waste flow was configured to enter the reactor via a Y-piping connection that also fed the tap water. During the Fill with Wastestep, the peristaltic pump was configured by the process controller to:
flow waste feed for 18.3 seconds
deliver 140 mL of waste
These configurations were based on the measured peristaltic pump flow rate of 459 mL/min.
Fill with Water
The tap water delivered consisted of “modified” tap water and was made by adding the contents in Table 1 below to tap water.
Table 2: "Modified" tap water contents and their concentrations.
Compound / Chemical Formula / Concentration (mg/L)Magnesium sulfate / MgSO47H2O / 69.60
Sodium molybdate / NaMoO42H2O / 0.15
Manganese sulfate / MnSO4H2O / 0.13
Cupric sulfate / CuSO44H2O / 0.08
Zinc suflate / ZnSO47H2O / 0.48
Calcium chloride / CaCl22H2O / 22.50
Iron chloride / FeCl36H2O / 18.33
Cobalt chloride / CoCl26H2O / 0.42
This modified tap water was fed from a central supply that was passed through the same piping as the synthetic waste. This periodic flushing minimized the bacterial build-up on the waste delivery line. The Fill with Water state was configured by the process controller to deliver the “modified” tap water until two conditions were met:
- 350 seconds elapsed, which corresponded to the flow rate in the Fill with Waste state.
- the reactors volume reach 4 L, averaged over 5 seconds.
The second condition acted as a fail-safe if the peristaltic pump failed to deliver the appropriate amount of fluid. The pressure sensor attached at the bottom of the reactor provided the necessary data to monitor the volume.
Aerate
The aeration state was configured to flow air into the reactor and stir the reactor’s contents. The aeration time was varied (8, 18, and 24 hours) in order to identify its impact on the extent of the SND process. The aeration rate was controlled by implementing a Proportional, Integral and Derivative (PID) control system. The target DO level was set to 0.5 mg/L by the PID control. PID sets the value of the control parameter (airflow) based on the sum, the integral and the derivate of the error. Equation 1 below is the general PID function:
,equation 1
where, Kc is controller gain (tuning parameter), TI is the integral time (tuning parameter), TD is the derivative time (tuning parameter), is the difference between measured value and set point (measured oxygen concentration minus desired oxygen concentration), /t is the error rate of change (note that this is the same as the dissolved oxygen concentration rate of change), is the area under the curve of the error as a function of time, and u(t) is the airflow rate that the controller sets. The difference between the process variables and the user-defined set-point is the error (which is reduced to zero to find an output, in this case the airflow rate.
Equation 1 is simplified when brought to the user-interface of LabView, as shown below in Equation 2 :
P I D
.equation 2
For a response to be shown in airflow rate, the values of P, I, and D must be changed accordingly from their default zero values. LabView then calculates the respective values of the PID parameters found in Equation 1.
To identify appropriate values for the values of P, I, and D, trial and error method was utilized to observe the changes in the DO of the reactor. When the DO approached 0.5 mg/L, effective control was established.
The aeration state relied heavily on the performance of the diffuser stone and the DO probe. The DO probe was recalibrated and its membrane was changed on a weekly basis, which ensured its accurate acquisition ability. The diffuser stone was changed once it became clogged and ceased to deliver air.
Settle
The settle state lasted for 1 hour after aeration in order to settle the sludge and biomass sufficiently, to prevent evacuation from the reactor. The allocated settling time was observed to be more than ample for settling the bio-solids in the reactor, but was left unchanged.
Drain
The drain state was the final state in the operational cycle for the SBR, and discharged water from the reactor. The state ran until a residual volume of 1.2 L in the reactor remained, which ensured a continual microbial population in the reactor.
Plant Activation
The plant was activated by the introduction of 4 L of waste-mixed liquor from the City of Ithaca wastewater treatment plant.
Experimental Methods
Testing Methods
The extent of SND occurring in the reactor was assessed by measuring the ammonia and total nitrate concentrations, as ammonia-nitrogen and nitrate-nitrogen. These tests were used as indicators of reactor performance, as two different microbial populations were attempted to be maintained in the reactor. This method was used instead of the Total Suspended Solids (TSS) method because the dominating bacterial population was not known. Satisfactory levels of nitrification and denitrification were hoped to be established via these methods.
Essential Parameters
The essential parameters in the tests were the concentration of nitrogen as ammonia and the concentration of nitrogen as nitrite/nitrate. The total nitrogen in the synthetic waste is known, so assuming all organic nitrogen is converted to ammonia, this information allows us to determine nitrogen removal.
Ammonia Standards via the Micro-Phenate Method
The micro-phenate method was used to determine the total ammonia concentration in the reactor as ammonia-nitrogen. Standards for the micro phenate method were prepared with the following species:
Phenol solution: prepared by introducing 1.11 mL of liquefied phenol (>89%) into a 1.5 mL cuvet with a micro-pipette. Dilution to 10 mL by the addition of 95% v/v ethyl alcohol under the fume hood.
Sodium nitroprusside: prepared by dissolving 50 mg sodium nitroprusside in 10 mL E-pure (deionized) water. Storage in a plastic bottle which was protected from light.
Alkaline citrate: prepared by dissolving 200 g trisodium citrate and 10 g sodium hydroxide in deionized water diluted to 1 L.
Sodium hypochlorite: Clorox®.
Oxidizing Solution: prepared for each test by mixing 1 mL of the alkaline citrate solution with 0.25 mL of Clorox in a 4.5 mL plastic cuvete.
Ammonia standards were prepared by dissolving 1.389 mg of ammonium chloride in 1 mL of water, diluted to 1 L, making a solution containing 1 mg/L of nitrogen. The stock solution was diluted with deionized water to concentrations of 0, 0.2, 0.4, and 0.6 mg N-NH4/L.
1 mL of each of the ammonium standards was pipetted to a separate 1.5-mL cuvete in a fume hood, with the following additions:
40 mL phenol solution
40 mL sodium nitroprusside solution
100 mL oxidizing solution
The cuvete samples were covered and left to stand for a minimum of 1 hour inside a bench drawer. Analysis then followed at an absorbance of 640 nm using a UV-spectrophotometer. The measured absorbances were used to obtain a calibration curve.
Nitrate-nitrogen Standards
The nitrate concentration in the reactor was determined as total nitrate-nitrogen, using the cadmium gravimetric method with Nitriver® and Nitraver® reagent packets produced by the Hach Corporation. Nitrate-nitrogen standards were prepared by creating a solution of sodium nitrate in distilled water at a concentration of 100 mg/L. This stock solution was diluted into concentrations of 0, 2, 6 and 10 mg/L of nitrate-nitrogen. 0.5 mL of each of the standards was pipetted into a cuvete and a Nitraver® packet was added under the fume hood. The cuvet was then covered, shaken for 3 minutes, and allowed to stand for 30 seconds. Nitriver® packets were added and the new solutions were shaken for 30 seconds and allowed to stand for 10 minutes. The absorbance of these standards was also measured using the UV-spectrophotometer at a wavelength of 542 nm, which generated a calibration curve.
Sampling
Samples were taken from the reactor at the end of the aeration state, filtered using a Millipore filter and stored in the shelf refrigerator. For the ammonia samples, no smell of hydrogen sulfide was detected, so no further additional treatment was required. These samples were diluted by 10x and 100x and were treated using the micro-phenate procedure described above. The dilution was performed because we were not aware of the residual ammonia-nitrogen concentration in reactor at the end of the respective aeration phases. It was discovered from the standards preparation that a highly concentrated solution would give inaccurate absorbance readings (the absorbance readings would exceed the accuracy limits of the spectrophotometer). The 10x dilution fell in the range of the standard calibration curve and its use was continued.
The nitrate-nitrogen samples were acquired from the reactor in a similar way outlined above. These samples were treated using the nitrate-nitrogen procedure previously described. The UV-spectrophotometer was used to determine the nitrate-nitrogen concentration in the samples.
Results and Discussions
PID Control
PID control was set to maintain a DO level of 0.5 mg/L in the reactor during the aeration state. To assess the performance of the control mechanism two concurrent DO data sets, during the aeration state, were analyzed against time. The data was collected on November 30, 2004 and December 1, 2004 from 4:42:54 PM to 9:42:54PM and from 10:55:20 PM to 3:55:20 AM. For each data set, the average DO and its standard deviation was:
4:42:54 PM to 9:42:54PM, values taken from4:51:54PM on, when DO dropped below target value of 0.5 mg/L
- Average DO: 0.505 mg/L
- Standard Deviation: 0.015
10:55:20 PM to 3:55:20 AM, values taken from11:04:20PM on, when DO dropped below target value of 0.5 mg/L
- Average DO: 0.538 mg/L
- Standard Deviation: 0.042
The graphs of DO response versus time are shown below in Figure 3 and 4. PID control accurately maintained an approximate DO of 0.5 mg/L in the reactor throughout the aeration state, as the samplefigures below clearly show. We were originally planning on changing the values for P and I, trying to maximize control, but the program worked much better than expected and there was no reason to change the values.
Figure 3: Dissolved Oxygen (mg/L) vs. Time during aeration state on 11/30/04 from 4:42:54 PM to 9:42:54 PM.