Acta Chini. Slov. 2001, 48, 625-636. 625 ASMI LIMITATIONS FOR MODELING OF WASTEWATER TREATMENT PROCESSES* Igor Plazl, Tine Koloini Department of Chemical Engineering, Universty of Ljubljana, Aškerčeva 5, 1001 Ljubljana, Slovenia Meta Levstek, Olga Burica, Marjetka Stražar Domžale-Kamnik Wastewater Treatment Plant, Študijanska 91, 1230 Domžale, Slovenia #This paper is dedicated to Professor Roman Modic on the occasion of his 90* celebration. Received 01-06-2001 Abstract In modern wastewater reactor technology dynamic modelling proved to be a valuable tool for the simulation of wastewater treatment performance, especially in cases of problematic inflow characteristics and no optimal process conditions. Due to stringent EU legislation the Domžale-Kamnik WWTP of 200,000 PE is confronted with the need to upgrade the plant in the sense of nutrient removal. As the full-scale plant is also subject to changing process conditions, a CSTR pilot plant of 15 liters was installed. The system was operated in the single sludge mode at the specified process conditions (temperature, dissolved oxygen, flow, recycle ratio), and was fed with artificial wastewater. The effluent composition (organic nitrogen, ammonium, nitrate, COD), some stoichiometric and kinetic parameters, and operation parameters of the pilot plant were determined in the laboratory by standard laboratory methods. These results served as input parameters for the calibration of the steady-state model based on real conditions according toIAWQmodelNo.l. The dynamic response to a step disturbance of the input flow was measured and comparison between dynamic response of the calibrated ASMI model and experimental data is represented. The disagreement between the experimental and by model predicted results can be explained with ASMI limitations. Introduction Understanding, describing and predicting the dynamic behavior of the activated sludge process for removing carbonaceous and nitrogenous compounds from wastewater needs to account for a large number of reactions between a considerable number of components. In the past many different models were introduced and presented in the literature using different process kinetics and model components. To create a uniform model structure in 1987 the IAWQ Task Group on Mathematical Modelling introduced ASMI , which became a major reference for many scientific and practical projects, and describes all the necessary reactions and components. Many authors have reported I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... 626 Acta Chini. Slov. 2001, 48, 625-636. results of practical modelling according to ASMI (simulations, determination of kinetic and stoichiometric constants) ' ' . Our aim was to set up a dynamic model based on ASMI for a laboratory scale pilot plant and to obtain answer to two questions: 1. What is the reliability of the laboratory determined kinetic and stoichiometric constants compared to process parameters? In other words, which parameters should be chosen as calibration parameters? 2. Can the dynamic model based on ASMI, successfully calibrated at steady-state conditions, correctly predict the dynamic behavior of the activated sludge process at nonsteady-state conditions? Experimental The CSTR pilot plant of 15 liters including nitri-denitrification was operated at the following volumetric relations: anoxic stage Vanx = 5 L, oxic stage Vaer = 10 L, settler Vs = 5 L (Figure 1). It was runing for a few months under constant operating conditions (flow, recycle ratio, oxygen concentration, temperature,...). F0, SSO, SNO0, SNHO, SNDO XBH0,=XBA0=0 F1,SS1,SN01, SNH1.SND1 XBH1.XBA1 F2, SS2, SN02, SNH2, SND2 XBH2, XBA2 SOanx Vanx FIR, SS2, SN02, SNH2, SND2 XBH2, XBA2 SOaer Vaer F3, SS3, SN03, SNH3, SND3 XBH3=XA3=0 FR, SS2, SN02, SNH2, SND2 XBHR, XBAR FER, SS2, SN02, SNH2, SND2 XBHR, XBAR FW, SS2, SN02, SNH2, SND2 XBHR, XBAR Figure 1: Scheme of the pilot plant I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... Acta Chini. Slov. 2001, 48, 625-636. 627 The pilot plant was fed with artificial wastewater of known chemical composition (yeast extract, meat extract, casein peptone, sodium acetate, ammonium chloride, potassium hydrogen phosphate, some inorganic compounds such us magnesium carbonate, sodium chloride, etc., and up to 20 vol % of settled municipal wastewater) ' . The chemical composition of the artificial wastewater is presented in Table 1. Table 1: Artificial wastewater composition in the first and second experiment Component Unit First experiment (17.6.-29.9.1999) Second experiment (10.5.-17.5.2000) COD mg02/L 650 450 N-Kjel mgN/L 65 65 N-NH4 mgN/L 20 45 N-org mgN/L 45 20 Total phosphorus mgP/L 6 6 At steady state conditions we analyzed the influent, effluent and the water of the anoxic and oxic reactor for the following compounds: COD (SS), ammonium nitrogen (SNH), Kjeldahl nitrogen (SNH + SND) and nitrate nitrogen (SNO) according to ISO standards. Influent and effluent were analyzed on daily average samples, anoxic, oxic stage and settler were analyzed on spot samples. Spot samples were also analyzed for MLSS, total COD and total N-Kjeldahl which served as an estimate of active heterotrophic (XBH) and autotrophic biomass (XBA). At the same time, we also determined the kinetic and stoichiometric constants using respirometry and batch I S 7 S 0 tests ' ' ' ' . All these data were used in the model calibration. After first stationary experiment (17.6.-29.9.1999) we performed second experiment (10.5.-17.5.2000) under different initial conditions. We measured and analyzed the same parameters and compounds as in the first experiment. These data served us for model verification. All experimental results are presented in Table 2. I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... 628 Acta Chini. Slov. 2001, 48, 625-636. Table 2: Experimental results First experiment Second experiment (17.6-29.9.1999)____________________(10.5.-17.5.2000) Symbol UNIT I. STATIONARY STATE -Average value Day:l 2 3 5 7 IL STAT. STATE IL STAT. STATE l.DAY AFTER STEP FLOW CHANGE 3. DAY AFTER STEP FLOW CHANGE 5. DAY AFTER STEP FLOW CHANGE F0 L/h 1.5 + 0.1 1.47 1.44 2.75 2.5 2.5 FIR L/h 3.9 + 0.1 3.64 3.53 3.54 3.54 3.54 FR L/h 1.5 + 0.1 3.13 3.2 3.2 3.1 3.1 FW L/h 0.017 + 0.002 0.0125 0.0125 0.0125 0.021 0.021 XB1 mg CODc/L 5400 3038 3160 1626 2222 2716 XB2 mg CODc/L 5033+1180 2994 3130 1724 2250 2736 XBR mg CODc/L 6869+1728 4622 3980 6986 5490 5210 sso+xso mg COD/L 689 ± 63 451 388 439 514 366 SSO mg COD/L SSI mg COD/L 48.6+13.6 19.2 40.3 49.6 74 SS2 = SS3 mg COD/L 33.4 ±5.8 19.7 22.4 25.6 32.1 33.2 SNO0 mg N-NO3/L 3.2 + 0.6 0 0 0 0 0 SNOl mg N-NO3/L 0.3+0.3 0.027 0.08 0 0 0 SN02 = SN03 mg N-NO3/L 6.3+2.1 10.2 10.7 9.15 8.42 9.6 SNH0 mg N-NH4/L 20.7 + 4.0 55.1 48.3 44.7 68.8 53.5 SNH1 mg N-NH4/L 6.1+0.9 9.83 7.64 20 29.6 35.5 SNH2 = SNH3 mg N-NH4/L 0.6 + 0.3 0.17 0.17 7.03 14.1 16.4 SND0+X ND0 mg N-org/L 46.5 ±7.7 9.4 14.6 21.0 22.3 SNDO mg N-org/L SND1 mg N-org/L 3.2 ± 1.9 1.2 2.4 2.3 4.2 SND2 = SND3 mg N-org/L 2.6 ±2.0 0.9 0.7 4.3 6.5 Simulation model A mathematical model was built up according to ASMI for activated sludge, including nitrogen removal. For each reactor we wrote mass balance Equations (Equation 1 to 19) based on the known concentrations (SS, SNO, SNH, SND, XA, XH). I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... Acta Chini. Slov. 2001, 48, 625-636. 629 Anoxic reactor: Vanx--^1- = F0-SS0 + (FIR + FER)-SS2-Fl-SSl—— • RHanx • Vanx dt YH (1) + PSI • (bH • XBH1 + bA • XBA1) • Vanx Vanx- dSN01= FO SNOOf (FIR+ FER) SN02-F1- SNO1- 1-YH • RHanx Vanx (2) dt 2,86- YH Vanx • SNH1 - FO • SNHO + (FIR + FER) • SNH2 -Fl- SNH1 - iNx • RHanx • Vanx dt V ' (3) + ka-SNDl-XBHl-Vanx Vanx—^^1=F0-SND0+(FIR+FER>SND2-Fl-SNDl-ka-SNDl-XBHl- Vanx dt (4) +PSI-iNx-(bH-XBHl+bA-XBAl)Vanx dXRHI Vanx---------= FIR XBH1+FER XBHR-F1XBH1+RHanx Vanx-bH-XBHl Vanx (5) dt dXBA 1 Vanx------------- FIR • XBA1 + FER • XBAR - Fl • XBA1 - bA • XBA1 • Vanx (6) dt Oxic (aerobic) reactor dSS2 1 Vaer-^^ = Fl-SSl-(FIR + F2)-SS2--------RHaer-Vaer dt YH (7) + PSI • (bH • XBH2 + bA • XBA2) • Vaer Vaer • SN°2 -Fl- SNO 1 - (FIR + F2) • SN02----— • RAaer • Vaer (8) dt YA Vaer-^5=F 1- SNH 3-(FIR+F2) SNH2-iNx RHaerVaer-dt +ka-SND2XBH2Vaer f. 1 ^ lNxf— •RAaerVaei (9) Vaer----^_ = F \. SND \ _ (FIR + F2). SND2 - ka • SND2 • XBH2 • Vaer dt (10) + PSI • iNx • (bH • XBH2 + bA • XBA2) • Vaer dXRH2 Vaer------------ Fl • XBH1 - (FIR + F2)-XBH2+RHaer- Vaer -bHXBH2- Vaer (11) dt HXB A9 Vaer------------ Fl • XBA1- (FIR + F2)-XBA2+RAaer- Vaer -bAXBA2- Vaer (12) dt Settler Vs.^^ = F2-SS2-F3-SS2-FR-SS3 (13) dt I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... 630 Acta Chini. Slov. 2001, 48, 625-636. dSN03 = F2. SN02 - F3 • SN02 - FR • SN03 F2-SS2-F3-SS2-FR-SS3 : F2 • SNH2 - F3 • SNH2 - FR • SNH3 - F2 • SND2 - F3 • SND2 - FR • SND3 = F2-XBH2-FR-XBHR F2-XBA2-FR-XBAR dt Vs dSS3_ dt Vs dSNH3 dt Vs dSND3 dt Vs dXBH3 dt Vr. dXBA3 dt (14) (15) (16) (17) (18) (19) Process Rates (anoxic and aerobic growth of heterotrophs and autotrophs) RHanx=MaxH- SSI V KS+SS1 KOH V KOH+SOanx SNOl V KNO+SNOl f RHaer = MaxH • SS2 RAaer = MaxA ¦ V V KS+SS2 SOaer V KOH + SOaer XBH2 etag-XBHl (20) (21) SNH2 KNH + SNH2 f SOaer A vKOA +SOaer Flow rates XBA2 Fl =F0 + FIR + FR-FW F2 = F1-FIR F3 = F2 - FR FER = FR-FW (22) (23) (24) (25) (26) Active biomass concentration determined from total biomass concentration XBH = fa(l-deNI)XB (27) XBA = fa deNI XB (28) I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... Acta Chini. Slov. 2001, 48, 625-636. 631 Table 3: Experimental and modeled values for model parameters determined by the calibration of the first stationary condition at 20 °C. Symbol of parameter unit Experimental value Vanx L 5 Vaer L 10 Vs L 5 SOanx mg02/L 0.0 SOaer mg02/L 5.5 F0 L/h 1.5 FIR L/h 3.9 FR L/h 1.5 FW L/h 0.02 XB1 mg MLSS/L 4300 XB2 mg MLSS/L 4900 XBR mg MLSS/L 6800 XBH1 mg COD/L 2846.2 XBA1 mg COD/L 425.3* XBH2 mg COD/L 3243.3 XBA2 mg COD/L 484.6* XBHR mg COD/L 4500.9 XBAR mg COD/L 672.5* SSO mg COD/L 689.4 SSI mg COD/L 48.6* SS2 = SS3 mg COD/L 33.4 SNO0 mg N-NO3/L 3.2 SNOl mg N-NO3/L 0.3 SN02 = SN03 mg N-NO3/L 6.3 SNHO mg N-NH4/L 20.7 SNH1 mg N-NH4/L 6.1 SNH2 = SNH3 mg N-NH4/L 0.6 SNDO mg N-org/L 46.5 SND1 mg N-org/L 3.2 SND2 = SND3 mg N-org/L 2.6 The model, rebuilt for steady-state operation, can now be used for calibration using appropriate numerical methods . Results and discussion The sludge process variables (Table 3) and some laboratory determined kinetic and stoichiometric constants (Table 4) served as the input calibration parameters. Calibration was not possible even in the case when all stoichiometric and kinetic constants were free for calibration. For our experimental pilot plant activated sludge process with nitrogen removal included it was shown, that calibration based on the ASMI model is successful only when some experimental data, namely the I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... 632 Acta Chini. Slov. 2001, 48, 625-636. concentrations of active biomass and readily biodegradable substrate, together with some kinetic and stoichiometric constants were used as calibrated parameters. Those data are shown in Table 3, marked with an asterisk for process variables and in Table 4 for kinetic and stoichiometric constants. This opens the question of the reliability of some of the analytical methods and the results for activated sludge process variables, especially the active concentration of biomass. On the other hand, the definition of kinetic and stoichiometric constants given by the ASMI model cannot properly describe the activated sludge process - for example, the heterotrophic yield cannot be fixed as a constant for different reactors. The new concept of ASM3 confirmed our opinion ' . Table 4: Comparison of stoichiometric and kinetic constants determined experimentally, by calibration and presented by ASMI at 20°C. Symbol Unit experimentally determined modeled ASMI1 STOICHIOMETRIC CONSTA NTS YH mg CODc/mg COD 0,68 0,67 0,67 (0,46 - 0,69) YA mg CODc/mg N 0,27 0,24 (0,07 - 0,28) iNx mg N/mg CODc 0,04 - 0,08 0,1228 * 0,086 fa 0,634 0,634 deNI 0,13 0,008 KINETIC CONSTANTS KS mg COD/L 10-65 161,7 20(10-180) KOH mg 02/L 0,2 0,2(0,01-0,15) KOA mg 02/L 0,4 0,4(0,5-2) KNO mgN/L 0,5 0,5(0,1-0,2) KNH mgN/L 1,0 1,0 1,0 MaxH 1/h 0,22 - 0,38 0,0986 * 0,25(0,125-0,55) MaxA 1/h 0,01 0,0377 0,032 bH 1/h 0,007 0,00927 0,0258 (0,002 - 0,067) bA 1/h 0,00616 0,0021-0,0063 ka L COD/mg.h 0,0003 * 0,0033 etag 0,07* 0,8(0,6-1) Legend: * - the calibrated value is very different from the value given in the literature On the basis of the calibration described, the calibrated parameters for steady-state operation of the activated sludge process were used for process simulation based on the dynamic model equations (Equations 1-18). I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... Acta Chini. Slov. 2001, 48, 625-636. 633 -SS1 (model) 3 4 5 day of II. experiment — SS2 (model) • SS2 (experiment) ¦FO Figure 2: Dynamic model prediction of COD concentration in the anoxic (SSI) and oxic reactors (SS2) compared with experimental data ,__. 30 U) ?s E *w ts X 20 z CO ,_" 15 X z CO 10 A 35,5 ------- ------------A-29-,6--------------- - -A 29 - - - - - • 16,4 9,83 1 • 14,1 i 7,64 • 7,03 - i 0,17 0,17----------- •------ i >---------------------- ---------------1-------------------------------------------1— ---------------------------1--------------------------------1— ------------------1---------------------1- 1.5 a O riav of II. exneriment -------SNH1 (model) • SNH2 (experiment) A SNH1 (experiment) --------F0 -SNH2 (model) Figure 3: Dynamic model prediction of ammonia concentration in the anoxic (SNH1) and oxic reactors (SNH2) compared with experimental data I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... 634 Acta Chini. Slov. 2001, 48, 625-636. The comparison between experimental and dynamic model response on the step change of the influent flow are presented in Figures 2 and 3. It can be seen that the dynamic model predictions are not in good agreements with the experimental data. Reasons for that could be part of the ASMI limitations. Although the calibrated ASMI model can be proved for the steady state operations, the dynamic simulations based on IAWQ model No.l are questionable. The model, already most parametric sensitive, anticipates that kinetic and stoichiometric parameters are constants for the aerobic nitrification and anoxic denitrification step. Setting up a new model configuration according to the ASM3 model can solve some of these problems. On the other hand, we have to take into consideration the problems of reliability of experimental results. We should emphasize the difficulties in determing the concentrations of active heterotrophic and autotrophic biomass and the correctness of some kinetic and stoichiometric constants used in the model simulation, which are determined in the laboratory under o different process conditions. Conclusions A mathematical model was built up according to ASMI for wastewater treatment process taking place in the pilot wastewater treatment plant including nitrogen removal. The model was successfully calibrated for the steady-state operational conditions and the calculations of the dynamic response on the step change of the influent flow were compared with the experimental data. The deviating between predicted and experimental results was explained with ASMI limitations and with possible problems of reliability of experimental values. Parameters, the quantities that go alongside the problem, cannot but be lumped, because, were they not, they would have to be treated as variables. Acknowledgment The research was supported by Grant J2-7508-0103 of the Ministry of Science and Technology of Slovenia. I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... Acta Chini. Slov. 2001, 48, 625-636. 635 Symbols Symbol Unit bA 1/h bH 1/h deNI - etag - etah - F L/h fa - iNx mg N/mg CODc ka L COD/mg.h KNH mgN/L KNO mgN/L KOA mg02/L KOH mg02/L KS mg COD/L MaxA 1/h MaxH 1/h SND mgN/L SNH mgN/L SNO mgN/L SO mg02/L SS mg COD/L Vaer L Vanx L Vs L XB mg CODc/L XBA mg CODc/L XBH mg CODc/L YA mg CODc/mg N YH mg CODc/mg COD Name Decay coeff. for autotrophic biomass Decay coeff. for heterotrophic biomass Fraction of autotrophic biomass Correction factor for MaxH under anoxic conditions Correction factor for hydrolysis under anoxic conditions Volumetric flow Active fraction of the biomass Mass N/mass COD in biomass Ammonification rate Ammonia half-saturation coeff. for autotrophic biomass Nitrate half-saturation coeff. for heterotrophic biomass Oxygen half-saturation coeff. for autotrophic biomass Oxygen half-saturation coeff. for heterotrophic biomass Half-saturation coeff. for heterotrophic biomass Max. specific growth rate for autotrophic biomass Max. specific growth rate for heterotrophic biomass Soluble organic-N concentration Ammonia-N concentration Nitrate-N concentration Dissolved oxygen concentration Readily biodegradable (soluble) substrate concentration Volume of aerobic - oxic reactor Volume of anoxic reactor Volume of settler Total biomass concentration in COD units Active autotrophic biomass concentration in COD units Active heterotrophic biomass concentration in COD units Autotrophic yield Heterotrophic yield References 1. W. Gujer, M. Henze, T. Mino, MVan Loosdrecht, Activated sludge models ASMI, ASM2, ASM2D and ASMS., Published by IWA Publishing, London, England, 2000. 2. H. Spanjers, P.A. Vanrolleghem, G. Olsson, P. Dold, Respirometry in control of the activated sludge process: principles, IAWQ task group on respirometry, England, 1998. 3. M. Henze, P. Harremoes, J.C. Jansen, E. Arvin, Wastewater Treatment- Biological and Chemical Processes, Springer, Berlin Heidelberg, 1995. 4. D. Orhon, N. Artan, Modelling of the activated sludge systems, Technomic, USA, 1994. 5. L.C.L. Grady, G.T. Daigger, H.C. Lim, Biological Wastewater Treatment, 2nd edition, Marcel Dekker, USA, 1999. 6. J. Bever, A. Stein, H. Teichmann, Weitergehende Abwasserreinigung, 3rd edition, R.Oldenbourg, Germany, 1995. 7. M. Roš, Respirometry of Activated Sludge, Technomic, USA, 1993. 8. H. Spanjers, Respirometry in Activated Sludge, Wageningen Agricultural University, Netherlands, 1993. 9. A.F. Rozich, A.F. Gaudy, Design and Operation of Activated Sludge Processes Using Respirometry, Lewis Publishers, USA, 1992. 10. T.E. Cloete, N.Y.O. Muyima, Microbial community analysis: The key to the design of biological wastewater treatment systems, IAWQ Scientific and Technical Report No. 5, England, 1997. I. Plazl, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater... 636 Acta Chini. Slov. 2001, 48, 625-636. 11. D.H. Eikelboom, Process control of activated sludge by microscopic investigation, IWA Publishing, England, 2000. 12. W. Gujer, M. Henze, T. Mino, M.Van Loosdrecht, Activated sludge model ASM3, Water Sci. Technol. 1999, 39(1), 183-193. 13. G. Koch, M. Kuhni, W. Gujer, H. Siegrist, Calibration and validation of activated sludge model No. 3 for Swiss municipal wastewater, Wat. Res. 2000, 34(14), 3580-3590. 14. I.Plazl, G.Pipus, M.Drolka, T.Koloini, Parametric Sensitivity and Evaluation of a Dynamic Model for Single-Stage Wastewater Treatment Plant, Acta Chim. Slov. 1999, 46(2), 289-300. 15. M.N. Pons, Modeling and simulation of a municipal waste watertreatment plant by activated sludge, Proceedings of the IMACS Symposium on mathematical modelling, Vienna, 1994, 3, 431-434. 16. M. Maurer, W. Gujer, Dynamic modelling of enhanced biological phosphorus and nitrogen removal in activated sludge systems, Proceedings of the IAWQ 19th Biennial Conf. Water Quality International, 1998, 1, 192-199. Povzetek Dinamično modeliranje procesov čiščenja odpadnih vod seje uveljavilo kot zelo uporabno urodje pri načrtovanju in optimiranju čistilnih naprav ter pri razumevanju kompleksnih procesov čiščenja. Še posebno se uporabnost modeliranja iskaže pri odpadnih vodah zahtevnejših karakteristik in pri ne-optimalnih procesnih pogojih. Stroga zakonodaja Evropske skupnosti in velika vsebnost dušika v odpadni vodi narekujeta nadgradnjo čistilne naprave Domžale-Kamnik z 200,000 PE. V ta namen potekajo študije nitrifikacije in denitrifikacije na pilotni dvostopenjski čistilni napravi, ki vsebuje 5 L anoksični in 10 L aerobni standardni mešalni reaktor ter 5 L usedalnik. Pilotno napravo, ki omogoča spremljanje in nadzor procesnih pogojev (temperatura, raztopljeni kisik, pretok, recikel), smo v našem primeru napajali z umetno odpadno vodo. Vrednosti procesnih spremenljivk med procesom čiščenja (organski dušik, amonijak, nitrat, KPK) in nekaterih stohiometričnih in kinetičnih parametrov smo določili s standardnimi laboratorijskimi metodami. Te rezultate smo uporabili pri kalibraciji s stacionarnim matematičnim modelom, razvitim po ASMI. Uspešno kalibriran in potrjen model za stacionarno obratovanje smo potem uporabili za napoved dinamičnega odziva pilotne čistilne naprave na stopenjsko motnjo vstopnega toka. Primerjava med eksperimentalnimi rezultati in matematično napovedjo je pokazala na omejitve ASMI modela. I. Plazi, T. Koloini, M. Levstek, O. Burica, M. Stražar :ASM1 limitation for modeling of wastewater...