Development and Implementation of a Real-Time ... - ACS Publications

This paper presents the development and implementation of a real-time control strategy based on end-point detection of biological reactions responsibl...
0 downloads 4 Views 608KB Size
Ind. Eng. Chem. Res. 2005, 44, 3367-3373

3367

RESEARCH NOTES Development and Implementation of a Real-Time Control System for Nitrogen Removal Using OUR and ORP as End Points Sebastia` Puig,† Lluı´s Corominas,‡ M. Teresa Vives,§ M. Dolors Balaguer,| and Jesu ´ s Colprim* Laboratory of Chemical and Environmental Engineering (LEQUIA), EQATA Department, Campus Montilivi s/n, Facultat de Cie` ncies, University of Girona, E-17071 Girona, Spain

Joan Colomer⊥ Control Engineering and Intelligent Systems Group (eXIT), Campus Montilivi s/n, Escola Polite` cnica Superior, University of Girona, E-17071 Girona, Spain

This paper presents the development and implementation of a real-time control strategy based on end-point detection of biological reactions responsible for carbon and nitrogen removal in order to optimize the sequencing batch reactor (SBR) process. The control system is an algorithm that automatically adjusts the cycle length to the influent wastewater characteristics according to the end points. The algorithm acts in the reaction phases of the SBR cycle using the oxygen uptake rate (OUR) and oxidation-reduction potential (ORP) values as the aerobic and anoxic phase end points, respectively. Real-time control was employed in the 1-m3 pilot-plant SBR treating urban wastewater with some industrial components. Despite the influent variability, the effluent levels, 57 mg of COD‚L-1 and 4.7 mg of N‚L-1, were lower than those of the European Directive 91/217/CEE. These results demonstrate that the real-time control, using a conventional online ORP probe and the respirometer measurement, OUR, can be applied in the SBR to wastewater treatment. Introduction Treatment of contaminated wastewater by means of biological processes has been widely implemented from classical urban wastewater to industrial wastewaters.1 Thus, from economical and operational points of view, biological treatment has proved to be a robust and more energy-efficient way of treating biodegradable wastewaters if good process control could be ensured.2,3 The classical configurations of a bioreactor, continuous stirred tank reactors (CSTRs) or plug-flow reactors (CPFRs), for carbon and nutrient removal are based on the activated sludge process. The microorganisms responsible for biodegradation are kept under suspension. These bioreactors have the operational conditions needed for biological reactions. After the bioreactor, a separation unit (i.e., settler) is commonly used for internal recycling of the biomass to the bioreactor in order to increase the process efficiency and final effluent quality.2 Another option for classical CSTR or CPFR is to use sequencing batch reactors (SBRs), where the biological * To whom correspondence should be addressed. Tel.: +34-972-418-281. Fax: +34-972-418-150. E-mail: [email protected]. † E-mail: [email protected]. ‡ E-mail: [email protected]. § E-mail: [email protected]. | E-mail: [email protected]. ⊥ E-mail: [email protected]

Figure 1. End points of biological nitrogen removal during the aerobic (A-D) and anoxic (E and F) phases.

processes are conducted in a complete mix reactor following a sequence of different operational conditions.4 In contrast to conventional systems, the SBR technology uses the same reactor as a final settler to get a clarified effluent without suspended matter. The common practice used in SBRs is based on the execution of a predefined cycle over time.5,6 However, it is possible to take more advantage of the flexibility of the SBR technology by finding the correct duration of aerobic and anoxic phases to achieve complete nitrification and suitable denitrification, respectively.7 In SBRs operating with a fixed cycle scheme, the variations in the influent composition are considered by assuming a reaction phase that is able to deal with the worst operational

10.1021/ie0488851 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/12/2005

3368

Ind. Eng. Chem. Res., Vol. 44, No. 9, 2005

Figure 2. Schematic overview of the pilot-plant SBR.

conditions. Otherwise, in some cases different predefined cycles could be used under different influent flow or concentration conditions. Instead of using a fixed SBR cycle, abstraction of knowledge from simple online probes (dissolved oxygen, DO; pH and oxidation-reduction potential, ORP) can be used as an indicator of the SBR cycle phase status. Determining the degree of nitrogen removal could be characterized by observing end points in the probes’ signal profiles. Figure 1 presents the nitrogen removal end points that are observed during a cycle. There are many end points of the nitrification process in the aerobic phase: ammonia valley (Figure 1A), residual carbon manipulation point (RCMP point;8 Figure 1B), RO2 point9,10 (Figure 1C), and an oxygen uptake rate (OUR) decrease,11,12 the so-called ROUR or average viability3 (Figure 1D). In the anoxic phase, two end points can indicate the end of denitrification: nitrate knee10 (Figure 1E) and nitrate apex (Figure 1F). Many researchers have used these end points,13-16 but any combination of “in situ” respirometry measurements, an online OUR, and the ORP probe detects the end of the nitrification and denitrification processes, respectively. Furthermore, this strategy has been applied to treatment of urban wastewater without the addition of any carbon source. It is important to notice that the respiration rate is directly linked to two important biochemical processes that must be controlled in a wastewater treatment plant (WWTP): biomass growth and substrate consumption.17 Many respirometry-based control strategies have been proposed in the literature, but very few real implementations are reported.17 However, the ORP signal is commonly used to establish the end of the denitrification process.18-20 The aim of the research presented in this paper is to develop a real-time control system for the optimal treatment of real urban wastewater for carbon and nitrogen removal. This control system allows each aerobic and anoxic reaction phase to be automatically adjusted to their optimal duration in order to achieve complete nitrification and denitrification. The control strategy is based on online monitoring and analysis of two characteristic parameters, the online OUR and the ORP, to detect the end points of the aerobic and anoxic reaction phases, respectively. Materials and Methods Pilot-Plant SBR. The pilot-plant SBR (Figure 2) was composed of a 1-m3 stainless steel square reactor. The SBR treated 600 L of wastewater per day from Cassa`-

Figure 3. Operational periods of the pilot-plant SBR with filling strategy.

WWTP. The cycle characteristics to achieve complete nitrification and denitrification were defined by previous laboratory-scale studies.21 Figure 3 presents the fixed cycle used during the experimental period. An 8-h cycle with six feeding steps was implemented, introducing the wastewater in the anoxic phases and then alternating the anoxic and aerobic phases for improving carbon and nitrogen removal in WWTPs.21,22 The cycle was divided into a reaction phase (395 min; 46.2% under aerobic conditions), settling (60 min), and draw (25 min). The pilot-plant SBR was DO-controlled at a 2.0 mg of DO‚L-1 set point by an air on/off strategy in order to achieve complete nitrification and avoid a high DO concentration at the start of the anoxic phase. The wastage in the SBR was performed under mixing and aeration conditions in the last aerobic phase for controlling the sludge or solids retention time of the system, assuming equal concentrations of solids in the wastage and in the reactor.23 Furthermore, the plant was equipped with a monitoring and control system consisting of three parts: probes (from Endress-Hauser), data acquisition and switch on/ off cards (from National Instruments), and interfaces developed in LabWindows (from National Instruments). The SBR was equipped with DO-temperature (OXIMAX-W COS 41), pH (CPF 81), and ORP (CPF 82) probes. Their signals, filtered in the transmitter, were captured by a data acquisition card (PCI-6025E), and control was conducted using a power relay output board (SC-2062), which allowed optimal functioning of the equipment. The software, consisting of user-friendly interfaces, was able to repeat over time a previously defined operation cycle by controlling the switch on/off process of filling, wastaging, and drawing of peristaltic pumps, the mixing device, and the air supply. Online mean values of pH, ORP, DO, and temperature were obtained every 5 s (software sampling time, 0.05 s) and stored in a simple ASCII file for further processing. Analytical Methods. Total suspended solids, volatile suspended solids (2540E), total and soluble chemical oxygen demand (COD and CODs), ammonium (NNH4+), total Kjeldahl nitrogen (TKN), nitrites (N-

Ind. Eng. Chem. Res., Vol. 44, No. 9, 2005 3369

Figure 4. pH, ORP, and DO profiles during the 8-h total cycle length employed in the pilot-plant SBR.

NO2-), and nitrates (N-NO3-) were all measured according to standard methods.24 The total nitrogen (TN) was calculated as the sum of TKN, N-NO2-, and N-NO3- concentrations as mg of N-TN‚L-1. “In Situ” Online OUR Measurements. The functioning of aerobic processes, such as activated sludge, depends on the availability of DO. This is due to the biomass consuming oxygen related to their activity when oxidizing organic matter and ammonia as well as to endogenous respiration. The respiration rate usually is measured with a respirometer.3,17 In this case, the reactor itself can serve as a respirometer. The mass balance for oxygen in a well-mixed region of the reactor is presented in eq 1, though it must be taken into account that no influent or effluent flow occurs in an SBR during the reaction phase:

dSo/dt ) KLA(Sosat - So) - OUR

(1)

where So ) DO concentration in the reactor (mg‚L-1), Sosat ) saturation DO concentration (mg‚L-1), and KLA ) oxygen mass transfer coefficient (h-1). Nevertheless, if no aeration is conducted, or during air off periods, the mass balance can be reduced to eq 2:

dSo/dt ) -OUR

(2)

Hence, to obtain the respiration rate, only the DO derivative has to be determined. This can be done by measuring the decrease in DO as a function of time due to respiration, which is equivalent to approximating the differential terms with a finite difference term.17 The dynamics of the sensor must be taken into account, so the first measurements (50 s) after deactivation of the aeration system were not used. Next DO values were acquired until the valve was opened again, and finally the linear regression was obtained. Results Background. The pilot-plant SBR located at Cassa`WWTP (Girona, NE Spain) ran for 200 days at a fixed cycle. The results obtained22 demonstrated that these conditions may be useful in treating real wastewater (532 ( 220 mg of COD‚L-1 and 53.6 ( 25.0 mg of N‚L-1 on average), keeping effluent levels (54 ( 25 mg of

COD‚L-1 and 4.7 ( 5.6 mg of N‚L-1 on average) lower than those of the European Directive 91/217/CEE (125 mg of COD‚L-1 and 15 mg of N‚L-1). Figure 4 presents pH, ORP, and OD profiles of a typical 8-h cycle of the pilot-plant SBR when nitrification/denitrification was completed. In the aerobic phases, the ammonia valley (Figure 4A) cannot be easily observed because of the stripping effect of the on/off aeration control. During anoxic phases, the appearance of nitrate knee (in the ORP profile, point C) and nitrate apex (in the pH profile, point B) indicates the end of the denitrification process. It may be possible to optimize the aerobic and anoxic stages using these end points to adapt the cycle length to the influent characteristics without affecting the effluent quality. On the basis of both the monitoring and control systems and the background knowledge acquired, a module to control the alternation of aerobic and anoxic phases was developed. This control was based on the detection of end points. Because of the fluctuations in the pH profile during the aerobic phase, which could complicate the detection of the ammonia valley, ROUR was proposed as the parameter for controlling the end of the aerobic phase. The nitrate knee is a robust parameter, and therefore it is used as the end point for the anoxic phase. Control System Module. The control system acts on the aerobic and anoxic phases of the SBR cycle. The aerobic phase length is controlled by the OUR measurements, while the ORP profile is used during the anoxic phase. To identify the end of the aerobic phase, different cycles were analyzed using OUR data from previous cycles. Figure 5 presents the evolution over time of online OUR measurements in aerobic phases of different cycles all together. At the beginning of the phase, an increasing trend for OUR values can be seen, which is caused by the changing conditions (from the anoxic phase to the aerobic phase). This transient response of the activated sludge is most likely from the sequence of intracellular reactions involved in substrate degradation by the activated sludge.25 For this reason, the control system waits for a minimum time, tmin, of 5 min. After tmin, the OUR profile decreases slowly until ROUR. After that, a minimum OUR value, OURmin, near 35 mg of O2‚L-1‚h-1 was achieved for most cases; therefore,

3370

Ind. Eng. Chem. Res., Vol. 44, No. 9, 2005

Figure 5. Analysis of previous OUR data corresponding to aerobic phases of different cycles.

Figure 6. Analysis of previous ORP data corresponding to the anoxic phase of different cycles.

this was the end-point reference. At this time, referred to by Watts3 as the “time to endogenous”, the system was assumed to be under endogenous conditions when at least 95% of the organic materials in the waste have been treated. This rate is also proportional to the biomass activity. Many factors, such as the following, have an impact on the biomass activity and thus activated sludge plant treatment: total load, biomass viability, and influent treatability or inhibition due to toxicity or temperature.3,17 After the minimum OUR was achieved, a stabilization time, twait, of 2 min in the control strategy was applied to ensure that the system was under endogenous conditions. Furthermore, the function of the device was also to guarantee the process performance by adopting a maximum time length (tmax) of 30 min for the aerobic phase, equal to the fixed cycle.

Figure 7. Real-time control strategy flow diagram.

In the anoxic phase (Figure 6), the parameter used to control the phase duration was the ORP. To determine a relationship between the end of the anoxic phase and the ORP, data from different profiles corresponding to anoxic phases were analyzed. The nitrate knee end point occurs depending on the characteristics of the influent wastewater and on the performance of the SBR. When the ORP values from different profiles obtained in the SBR (Figure 6) are compared, the nitrate knee always took place above an ORP value of -120 mV, ORPmin. The ORP minimum value has been used by other authors14,18 and depends on the influent characteristics and the aim of the study. As in the aerobic phase, tmin and twait (2 and 5 min, respectively) were also applied. Control Strategy Developed. The software control program developed using LabWindows was based on three synchronized modules that communicate with each other: the signal acquisition and monitoring device, the online OUR measurement module, and the control system module. The control system module was responsible for detecting the aerobic and anoxic phase end points using minimum OUR and ORP values (OURmin and ORPmin, respectively). The control strategy scheme, presented in Figure 7, was implemented in the control program. First, the control system detected whether the phase was aerobic, anoxic, or other (i.e., filling, settling, wastage, or extraction). If the phase was anoxic or aerobic, after a minimum time (tmin), it started to check the measured values with the fixed minimum values to find the end point. When the measured value was lower than the minimum, a waiting time (twait) for stabilization was added. The function of the device was also to guarantee the process performance by adopting a maximum time length (tmax) for each phase. If the minimum value was not achieved before tmax, the control system changed to the next phase. Implementation of the Control Strategy. After development of the control system, the next step was to test it. For this reason, the control system was employed in the pilot-plant SBR for 3 months in treating real urban wastewater. Figure 8 shows ORP (A) and OUR (B) profiles of a typical cycle from the pilot-plant SBR with the developed control system. In comparison with the 480 min of the fixed cycle, the cycle length applying the control system was around 390 min. The anoxic periods were reduced by around 56%. The largest reductions always took place during the first steps

Ind. Eng. Chem. Res., Vol. 44, No. 9, 2005 3371

Figure 8. ORP (A) and OUR (B) profiles of an optimized cycle using the control system.

Figure 9. Evolution of the influent (black triangles) and effluent (white triangles) for the total COD during the pilot-plant SBR operation.

Figure 10. Evolution of the influent (black circles) and effluent (white circles) for nitrogen removal during the pilot-plant SBR operation at Cassa`-WWTP.

because of the rapid decrease of the ORP profile (Figure 8A) because the nitrate concentration in the reactor was the same as the effluent concentration of the previous cycle. At the end of each anoxic phase, the nitrate knee was clearly observed. After the nitrate knee appeared,

the control system detected that the ORP values were lower than the ORPmin values, and it changed to the aerobic phase. Figure 8B also shows the evolution of the OUR profile in the aerobic phases of the cycle. In this phase, the

3372

Ind. Eng. Chem. Res., Vol. 44, No. 9, 2005

reduction time was around 12%. The largest reduction always occurred in the last step, when the substrate was depleted and OUR values reached the aerobic end-point value (OURmin). Despite the high cycle length reduction, which depends on the cycle used and the influent characteristics, the aim of the control system was to adapt the cycle length to the influent conditions. If a high effluent quality could be achieved, this would demonstrate that the control system was operating correctly. Figure 9 shows the evolution of the organic matter, both influent and effluent, during the two experimental periods: operation with the fixed cycle and with the control system (after day 220). Despite the variability of the organic matter (black triangles) influent (527 ( 220 mg of COD‚L-1 on average), the organic matter effluent (57 mg of COD‚L-1 on average) was always lower than the standard requirements. There were no major differences between the two periods in the effluent profile. Figure 10 presents the TN evolution, influent and effluent, during the experimental period. In a way similar to that of the organic matter, the high nitrogen influent concentration variability (51.1 ( 17.1 mg of N‚L-1 on average) did not affect the effluent quality, 4.7 mg of NTOT‚L-1; it was always lower than the European Directive values. At days 260 and 290, the pilot-plant SBR suffered some faults/incidents that concluded with a malfunction of the DO control system (i.e., air flow to the reactor was maintained active for a whole cycle, concluding with a high nitrification and a lower denitrification) and a change in the settling properties of the sludge (i.e., some colloidal organic matter was observed in the effluent) because of a strong variation of the influent wastewater quality (i.e., influent COD and nitrogen highly decreased because of a continuous rainy period). Nevertheless, after such incidents, the SBR performance rapidly recovered the normal efficiencies without application of extra actions. Conclusions From the DO monitoring during the aerobic phases of the SBR, using an air on/off control has permitted “in situ” measurement of an online OUR. Thus, with the OUR calculation and the ORP values, it is possible to estimate the status of the biological processes and to control real-time aerobic and anoxic phases of the SBR operational cycle. From the analysis of the aerobic and anoxic phases, a control module was designed, adjusted, and implemented for real-time control of a SBR treating urban wastewater, reaching effluent levels lower than the legal requirements of the European Directive 91/ 217/CEE (i.e., 57 mg of COD‚L-1 and 4.7 mg of N‚L-1). Finally, the obtained results demonstrated that a realtime control based on the use of in situ measurement of OUR and ORP monitoring is useful for identifying the end of the aerobic and anoxic phases of a step-feeddefined SBR cycle, respectively. Acknowledgment The authors thank CDTI-Spanish Government and INIMA Servicios del Medio Ambiente (Grupo OHL), Spanish Government (MCYT-DPI-2002-04579-C02), and University of Girona for their financial support in this study.

Literature Cited (1) Metcalf & Eddy, Inc. Wastewater engineering: treatment and reuse; McGraw-Hill Higher Education: New York, 2003. (2) Grady, J.; Daigger, G.; Lim H. Biological wastewater treatment; Marcel Dekker: New York, 1999. (3) Watts, J.; Garber, W. F. Respirometric control of the activated sludge process. Proceedings of the IAWQ Specialized Conference on Sensors in Wastewater Technology, Copenhagen, Denmark, 1995. (4) Mace, S.; Mata-Alvarez, J. R. Utilization of SBR Technology for Wastewater Treatment: An Overview. Ind. Eng. Chem. Res. 2002, 41 (23), 5539. (5) Keller, J.; Watts, S.; Battye, W.; Chong, R. Full-scale demonstration of biological nutrient removal in a single tank SBR process. Water Sci. Technol. 2001, 43 (3), 355. (6) Ayesa, E.; Goya, B.; Larrea, A.; Larrea, L.; Rivas, A. Selection of operational strategies in activated sludge processes based on optimization algorithms. Water Sci. Technol. 1998, 37 (12), 327. (7) Andreottola, G.; Foladori, P.; Ragazzi, M. On-line control of a SBR system for nitrogen removal from industrial wastewater. Water Sci. Technol. 2001, 43 (3), 93. (8) Ra, C. S.; Lo, K. V.; Maivinic, D. S. Real-time control of twostage sequencing batch reactor system for the treatment of animal wastewater. Environ. Technol. 1998, 39, 343. (9) Mauret, M.; Ferrand, F.; Boisdon, V.; Sperandio, M.; Paul, E. Process using DO and ORP signals for biological nitrification and denitrification validation od a food-processing industry wastewater treatment plant on boosting with pure oxygen. Water Sci. Technol. 2001, 44 (2-3), 163. (10) Plisson-Saune, S.; Capdeville, B.; Mauret, M.; Deguin, A.; Baptiste, P. Real-Time control of nitrogen removal using three ORP bending-points: signification control strategy and results. Water Sci. Technol. 1996, 33 (1), 275-280. (11) Demuynck, C.; Vanrolleghem, P.; Mingueau, C.; Liessens, J.; Verstraete, W. NDBEPR process optimization in SBRs: reduction of external carbon-source and oxygen supply. Water Sci. Technol. 1994, 30 (4), 169. (12) Vanrolleghem, P.; Coen, F. Optimal design of in-sensorexperiments for on-line modelling. Water Sci. Technol. 1995, 31 (2), 149. (13) Al-Ghusain, I.; Hao, O. J. Use of pH as control parameter for aerobic/anoxic sludge digestion. J. Environ. Eng. 1995, 121 (3), 225. (14) Battistoni, P.; De Angelis, A.; Boccadoro, R.; Bolzonella, D. An automatically controlled alternate oxic-anoxic process for small municipal wastewater treatment plants. Ind. Eng. Chem. Res. 2003, 42, 509. (15) Klapwijk, A.; Brouwer, H.; Vralijk, L.; Kujawa, K. Control of intermittently aerated nitrogen removal plants by detection endpoints of nitrification and denitrification using respirometry only. Water Res. 1998, 32 (5), 1700. (16) Yu, R. F.; Liaw, S. L.; Cho, B. C.; Yang, S. J. Dynamic control of a continuous-inflow SBR with time-varying influent loading. Water Sci. Technol. 2001, 43 (3), 107. (17) Spanjers, H.; Vanrolleghem, P.; Olsson, G.; Dold, P. Respirometry in control of the activated sludge process: Principles; Scientific and Technical Report No. 7; IAWQ: Bristol, U.K., 1998. (18) Tomlins, Z.; Thomas, M.; Keller, J.; Andic, J. M.; Urbain, V. Nitrogen removal in a SBR using the OGAR process control system. Water Sci. Technol. 2002, 46 (4-5), 125. (19) Kishida, N.; Kim, J. H.; Chen, M.; Sasaki, H.; Sudo, R. Effectiveness of oxidation-reduction potential and pH as monitoring and control parameters for nitrogen removal in swine wastewater treatment by sequencing batch reactors. J. Biosci. Bioeng. 2003, 96 (3), 285. (20) Chang, C. H.; Hao, O. J. Sequencing Batch Reactor System for Nutrient Removal: ORP and pH Profiles. J. Chem. Technol. Biotechnol. 1996, 67, 27-38. (21) Vives, M. T. SBR Technology for wastewater treatment: suitable operational conditions for nutrient removal. Ph.D. Thesis, University of Girona, Girona, Spain, 2004. (22) Puig, S.; Vives, M. T.; Corominas, L.; Balaguer, M. D.; Colprim, J. Wastewater nitrogen removal in SBRs, applying a step-

Ind. Eng. Chem. Res., Vol. 44, No. 9, 2005 3373 feed strategy: from lab-scale to pilot-plant operation. Water Sci. Technol. 2004, 50 (10), 89. (23) Irvine, R. L.; Busch, A. W. Sequencing batch biological reactorssan overview. J.sWater Pollut. Control Fed. 1979, 51 (2), 235. (24) APHA. Standard Methods for the Examination of Water and Wastewater, 19th ed.; American Public Health Association/ American Water Works Association/Water Environment Federation: Washington, DC, 1995.

(25) Vanrolleghem, P. A.; Sin, G.; Gernaey, K. V. Transient response of aerobic and anoxic activated sludge activities to sudden substrate concentration changes. Biotechnol. Bioeng. 2004, 88 (3), 277.

Received for review November 18, 2004 Revised manuscript received February 25, 2005 Accepted March 4, 2005 IE0488851