721

IWA Publishing 2012 Water Science & Technology

65.4

2012

Industrial wastewater platform: upgrading of the biological process and operative congurations for best performance
Anna Laura Eusebi, Alessandro Massi, Emiliano Sablone, Martina Santinelli and Paolo Battistoni

ABSTRACT
The treatment of industrial liquid wastes is placed in a wide context of technologies and is related to the high variability of the inuent physicalchemical characteristics. In this condition, the achievement of satisfactory biological unit efciency could be complicated. An alternate process (AC) with aerobic and anoxic phases fed in a continuous way was evaluated as an operative solution to optimize the performance of the biological reactor in a platform for the treatment of industrial liquid wastes. The process application has determined a stable quality efuent with an average concentration of 25 mg TN L1, according to the law limits. The use of discharged wastewaters as rapid carbon sources to support the anoxic phase of the alternate cycle, realizes a reduction of TN of 95% without impact on the total operative costs. The evaluation of the micro-pollutants behaviour has highlighted a bio-adsorption phenomenon in the rst reactor. The implementation of the process dened 31% of energy saving during period 1 and 19% for the periods 2, 3 and 4.
Key words
Anna Laura Eusebi (corresponding author) Martina Santinelli Paolo Battistoni Department SIMAU, Universit Politecnica delle Marche, Via Brecce Bianche, 60100 Ancona, Italy E-mail: a.l.eusebi@univpm.it Alessandro Massi SEA Ambiente S.r.l., Localit Saline, 60020 Camerata Picena, Ancona, Italy Emiliano Sablone Ingegneria Ambiente S.r.l., Via del Consorzio 39, 60015 Falconara Marittima, Ancona, Italy

| energy savings, industrial liquid wastes, micro-pollutants bio-adsorption, nutrient removal,

operating congurations, plant upgrading

INTRODUCTION
The treatment of industrial liquid wastes is placed in a wide context of technologies, related to the high variability of the physicalchemical characteristics of the inuent ow. The complexity of the ow schemes is associated with the necessity to respect the discharge limits in a public sewer, in terms of macro- and micro-pollutants. In this scenario, the removal of organic compounds and nutrients through biological processes is highly conditioned by the presence in the inuent of inhibitory factors and non-biodegradable substances, which compromise the denitrication phase (ygard et al. ). In this condition, to maintain the efciency of the biological unit and to reduce phenomena of kinetic inhibition, the raising of the hydraulic retention time and the external carbon dosage are adopted. As a consequence, operating costs related to the aeration and the use of reagents increase. Hence, despite its simple implementation and high efciency, the use of the BNR (biological nutrient removal) process to treat industrial wastewaters can be problematic (Battistoni et al. ). The treatment difculties also occur
doi: 10.2166/wst.2012.889

in the application of process schemes characterized by separated sections of nitrication and denitrication (DN) or by intermittent aeration in a single reactor (Sequential Batch Reactor or Carousel process) (Wang et al. ). This paper concerns the results achieved by the research project, developed by the Universit Politecnica delle Marche and co-nanced by Marche Region, aimed at optimizing the performances of an industrial wastewater treatment platform (SEA Ambiente S.r.l.).

MATERIALS AND METHODS

The treatment platform The industrial wastewater treatment platform (Figure 1) has a design capacity equal to 1,050 m3 week1 and a legal authorization of daily discharge of 400 m3 d1. The water line of the plant is provided with the storage unit, the

722

A. L. Eusebi et al.

Upgrading biological processes and operative congurations for the best performance

Water Science & Technology

65.4

2012

Figure 1

Platform ow scheme.

chemicalphysical and biological treatments and the ltration part. The sludge and the solids treatment line consists of the dewatering unit and the stabilization section. During the upgrading period, the biological unit, characterized by separate biomass in two stages (OX1 and OX2 with volume equal to 380 and 400 m3 respectively), performed the total oxidation process (TO). The equalized (volume 130 m3) inuent to the biological tank consists of water from the chemicalphysical units, dewatered ow, rainwater and other liquids which do not require primary treatments. The aeration is provided with two compressors (30 kW of power installed for each) and a reserve one (19 kW). The authorized discharge of the nal efuent to the urban sewage system allows its further treatment in the municipal wastewater treatment plant of 66,000 PE. The biological upgrading and the experimentation periods The AC process is a continuous technology based on the automatic alternation of oxic and anoxic phases into the same reactor. The aerobic period enables the complete ammonia oxidation whereas the anoxic one allows the nitrate denitrication. The automation of the process derives from a simplied mathematical model (Battistoni et al. ), which considers a complete mixed system, to determine the nitrates and ammonia variation as in the reactor and as in the efuent ow. According to the AC process, the durations of the alternating oxicanoxic phases were automatically determined by a patented control device which operates an on-line data processing of DO and ORP signals outputting from the probes installed in the biological basins. The percentages of nitrogen removal for the different forms are related to the length of the aerobic and anoxic phases. The control optimization of the process is guaranteed by an automatic device: the system elaborates

statistically the performed cycles, giving information on the phase changing (due to setpoint values, optimal conditions (OCs), time basis) and on the time lengths (average, minimum, maximum) (Figure 2). The percentages of nitrogen removal for the different forms are monitored with the performance evaluation and the mass balances and are related to the length of the aerobic and anoxic phases. The ORP and DO proles could be not reliable when the plant was overloaded or overaerated. In these cases two other controllers enter in operation to determine the phase length: the set-points basis and the time length basis. The process exibility permits its adaptation to reactors of different shape (rectangular, circular, annular), by reclaiming the existing structures. In fact, to perform the AC technology, only the optimization of the air supply is required, together with the installation of mixers to maintain the biomass suspension and of probes to control the process. After the start-up of the process, according to the operating steps summarized in Table 1, the experimental phases commenced. The logical approach was the maintenance of one reactor in total oxidation conditions to cope with variable and overloading inuent characteristics, trying to optimize the

Figure 2

Level AC control automation.

723

A. L. Eusebi et al.

Upgrading biological processes and operative congurations for the best performance

Water Science & Technology

65.4

2012

Table 1

Experimental phases and implemented processes

Period

Days

OX1

OX2

External carbon

Dosage

Recycle

1 2 3 4

23 32 24 15

AC TO TO TO

TO AC AC AC

No Yes No No

OX2

No No No Yes

measured according to the Kristensen method (Kristensen et al. ). The heavy metal concentrations were dened with the Varian mod. AA 240-FS spectrometer equipped with a vapour generating accessory by cold vapour atomic absorption.

RESULTS AND DISCUSSION

performances with the application of the AC process either in the OX1 (period 1) or in the OX2 (period 3). Moreover, to evaluate the effect of an added external carbon source period 2 was developed. Finally, the activation of the recycle of a quote (40%) of the efuent ow from the second secondary clarier to the OX1 reactor was studied during period 4 to determine the results obtainable with a higher HRT (hydraulic retention time). Process control and monitoring The chemical and physical characterization is performed, according to Standard Methods (APHA ), through the analysis of grab (ve samples every week for the equalization basin and one sample every week for the efuent from OX1) and average (one sample every week for the efuent from OX2) samples. Hence, during the different experimental phases, the macro-pollutant removal efciencies of the entire process and of each single step were evaluated. Moreover, kinetic nitrication rates are evaluated in OX1 and OX2, before process upgrading. The specic uptake rate of ammonia of the biological process was
|
Biological inuent characterization COD (mg L1) NH4-N (mg L1) NO3-N (mg L1) NO2-N (mg L1) TN (mg L1) Cl (mg L1) PO4-P (mg L1)

Inuent and efuent characterization and process evaluation The main characterization is developed during one year of monitoring phase before the experimental work (Table 2). The biological inuent ow rate, evaluated in the equalization tank was calculated equal to 50 m3 d1 on annual scale. The COD (chemical oxygen demand) concentration (average value of 4,000 mg L1) reected the typical variability of industrial wastewater, in the range from 1,900 to 7,500 mg L1. Similarly, the changeability of the ammonia concentration, between 26 and 77 mg L1, determined high COD/TN ratios. This condition is only apparently favourable for a possible denitrication phase considering the elevated percentages of non-biodegradable carbon in the inuent. The nitrates concentration rose up to 20 mg NO3-N L1 during the monitoring period whilst the nitrites one was constant and equal to 1 mg NO2-N L1. Moreover, the characterization highlighted an average salinity equal to 1,740 mg Cl L1; inhibitory conditions for the following nitrication phase (Beccari et al. ). In fact, the kinetic tests at 20 C determined a nitrication rate on a volatile
W

Table 2

pH

COD/TN

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

8.1 7.9 7.9 8.1 7.9 7.6 7.7 7.9 8.0 7.7 7.9 8.1

3,342 6,018 2,111 2,332 2,278 1,975 3,980 3,708 7,588 4,715 3,886 4,290

25.7 42.1 26.9 38.0 39.6 59.0 52.3 27.0 51.3 77.6 72.3 55.6

8.40 2.50 5.78 7.71 4.00 3.60 2.81 1.44 2.55 2.95 11.05 19.15

1.20 1.00 1.00 1.00 1.00 1.00 1.00 1.22 1.00 1.18 1.30 1.30

35 46 34 47 45 64 56 30 55 82 85 76

1,880 1,687 1,155 1,682 1,964 2,500 1,938 1,583 1,422 1,768 1,240 2,085

3.25 3.00 3.00 3.00 3.00 3.00 3.24 3.00 4.67 3.55 3.21 3.00

114 281 76 67 57 38 112 210 103 89 30 81

724

A. L. Eusebi et al.

Upgrading biological processes and operative congurations for the best performance

Water Science & Technology

65.4

2012

suspended solids basis (VSS) equal to 0.01 kg NH4-N kg VSS1 d1 in OX1 and 0.04 kg NH4-N kg VSS1 d1 in OX2, remarking the evident impact of the inuent in the rst reactor. This complexity of the treated ow rate needs to adopt high hydraulic retention times in the biological sections (7.8 and 8.2 d in OX1 and OX2 respectively). The results in terms of optimization of nutrient removal, achieved from the experimental periods, are summarized in Tables 3 and 4. In order to prevent possible inhibitory phenomena in the kinetic reactions, it is important to highlight that the MLSS (mixed liquor suspended solids) concentration was maintained constant and equal to 10,470 1,800 and 2,867 360 mg L1 in OX1 and OX2, with percentages of volatile solids of 61 5% and 56 3% respectively. The nitrogen load (LTN) rate values (NLR), calculated on the volumes (NLR between 0.006 and 0.010 kg m3 d1 in OX1 and between 0.003 and 0.007 kg m3 d1 in OX2) and on the volatile biomass (LVSS) basis (LTN LVSS1 from 0.0007 to 0.0017 kg TN kg VSS1 in OX1 and from 0.0012 to 0.0030 kg TN kg VSS1 in OX2), excluded conditions of overloads in the two biological reactors. The rst and third periods were characterized by the application of the AC process in either OX1 or OX2 (Table 1), maintaining TO in the other reactor without adding external carbon. The evident decrease of NOx-N

concentrations (NO2-N and NO3-N equal to 1 mg L1) in the efuent was due to intermittent anoxic phases during period 1, whilst the ammonia concentrations settled to an average equal to 32 mg NH4-N L1. The following phase of TO in OX2, to completely remove the residual ammonia, permitted us to obtain an efuent with concentrations of 3.8 mg NH4-N L1 and 44 NO3-N L1 (Table 3). Although a higher possibility to exploit the degradability of the organic inuent load for the denitrication phase in OX1, the ow scheme adopted in period 1 was compromised by a low nitrication velocity, determining an increase of NOx-N in the OX2 efuent. Contrarily in the third period, when the conguration of the process was opposite to period 1, the ammonia removal in OX1 guaranteed an average efuent concentration from OX2 of 1 mg NH4-N L1 (Table 3) and of 24 mg NO3-N L1 (Table 4). In this condition, the denitrication phase of the alternate cycles in OX2 was strongly limited by the low availability of carbon (RBCOD and SBCOD) after the TO step. During period 2, in order to compensate the shortage of biodegradable carbon, the denitrication phase of AC was provided with the provision of external carbon in OX2. The carbon source was added through beverages and/or dairy wastewater conferred to the platform, with an average carbon concentration around 100,000 mg L1. The removal

Table 3

Inuent and efuent of OX1 OX1 Inuent COD NH4-N

1

Efuent Cl ) (mg L
1

COD/TN )

NLR (kg m
3

LN/VSS d
1

MLSS
1

VSS/SS ) (%)

COD (mg L1)

NH4-N (mg L1)

NO2-N (mg L1)

NO3-N (mg L1)

Period

(mg L

(mg L

(kg Ntot kg VSS

(mg L

1 2 3 4

3,459 2,541 4,009 3,358

69 55 73 33

2,112 1,769 2,750 2,333

42 34 50 119

0.008 0.009 0.010 0.006

0.0017 0.0017 0.0021 0.0007

10,367 9,743 8,505 13,265

64 65 63 53

706 623 435 1,585

32 9 1 3

1 20 9 1

1 15 14 9

Table 4

Inuent and efuent of OX2 OX2 Inuent COD NH4-N

1

Efuent Cl ) (mg L
1

COD/TN )

NLR (kg m
3

LN/VSS d
1

MLSS
1

VSS/SS ) (%)

COD (mg L1)

NH4-N (mg L1)

NO2-N (mg L1)

NO3-N (mg L1)

Period

(mg L

(mg L

(kg TN kg VSS

(mg L

1 2 3 4

706 623 435 1,585

32 9 1 3

1,452 2,750 2,125 2,250

160 17 25 542

0.007 0.005 0.003 0.004

0.0026 0.0030 0.0019 0.0012

2,967 3,180 2,670 2,650

55 57 59 53

651 516 389 1,035

3.8 1.4 1.0 1.0

0.6 0.9 0.5 0.5

44 2 24 7

725

A. L. Eusebi et al.

Upgrading biological processes and operative congurations for the best performance

Water Science & Technology

65.4

2012

efciency in this period highlighted efuent characteristics equal to 1.4, 0.9 and 2 mg1, as ammonia, nitrites and nitrates respectively (Table 4). Finally, during the fourth period of experimentation, the AC process was maintained in the OX2 reactor, without providing the dosage of external carbon and by activating the recycle ow rate of the efuent from the secondary clarier to OX1, in order to increase the hydraulic retention times and to optimize the nitrate removal. In fact the lack of biodegradable carbon in OX2 determined a reduction of the specic denitrication velocity in the reactor. The nitrogen concentrations in the efuent were equal to 1 mg NH4-N L1, 0.5 mg NO2-N L1 and 7 mg NO3-N L1 (Table 4). Moreover, a characterization in terms of micro-pollutants in the efuent of OX1 and OX2 was made in the liquid phase (Table 5) and in the solid one for the 137 days included in the experimental period. The values highlighted high average range for the metals analysed, which were similar for Cu (0.10 mg L1 in OX1 and 0.13 mg L1 in OX2), Ni (0.59 mg L1 in OX1 and 0.62 mg L1 in

OX2) and Zn (1.03 mg L1 in OX1 and 1.27 mg L1 in OX2) in the two reactors. A decrement of the efuent amount could be evaluated for chrome and iron between OX1 and OX2. Although, the standard deviation and the statistical parameters reported in Table 5 show the high variability of the micro-pollutant amounts for iron, nickel and zinc in both the reactors and for chrome in the rst one. Although this disadvantaged conditions for the growth and kinetic behavior of the biomass, the ammonia uptake rates measured in the biological units were on average during the fourth periods equal to 0.01 kg NH4-N kg VSS1 d1 in OX1 and 0.04 kg NH4-N kg VSS1 d1 in OX2. Moreover, the MLSS concentrations in the two reactors are particularly different, in the range from 13,000 to 8,500 mg L1 for the rst (Table 3) and almost constant to 2,800 360 mg L1 in the second one (Table 4). A strict correlation between the biomass concentration and some heavy metals removals was evaluated (Figure 3). In fact, during the period, when there is a net growth of the MLSS concentration, the metals in the solution decrease. The

Table 5

Efuent heavy metals concentrations from biological reactor Efuent OX1 Cr (mg L1) Cu (mg L1) Fe (mg L1) Ni (mg L1) Zn (mg L1) Efuent OX2 Cr (mg L1) Cu (mg L1) Fe (mg L1) Ni (mg L1) Zn (mg L1)

Average Min. Max. St. Dev.

0.13 0.04 0.50 0.18

0.10 0.01 0.25 0.06

0.76 0.04 2.70 0.55

0.59 0.34 1.11 0.19

1.03 0.33 2.65 0.54

0.04 0.04 0.04 0.00

0.13 0.01 0.24 0.06

0.41 0.04 1.61 0.28

0.62 0.32 1.04 0.16

1.27 0.25 2.84 0.80

Figure 3

Biomass, chrome and copper concentrations in OX1.

726

A. L. Eusebi et al.

Upgrading biological processes and operative congurations for the best performance

Water Science & Technology

65.4

2012

bioabsorption conducted by the biomass, through the chemicalphysical interactions between micro-pollutants and the functional groups of the biological cell, permitted the accumulation for the two heavy metals (Cr, Cu). The

higher concentration of the biomass in OX1 (Table 3) compared with that in OX2 (Table 4), determined the main role of the rst reactor as promoter of micro-pollutants removal. Nitrogen removal efciency The results obtained from the experimental periods can be evaluated by dening the removal percentages (E%) in terms of total nitrogen (E%TN) and ammonia (E%NH4N), in OX1, OX2 and in the whole biological section (Figures 4 and 5). In the rst period, the percentage of total nitrogen removal in OX1 was equal to 61%, penalized from the TO phase in OX2, with a total removal efciency of 45%. The global ammonia removal equal to 94% was mainly reached in OX2 (88%), whilst it was lower for the inuent overloads in OX1 (54%). Similarly, during period 3, the shortage of biodegradable carbon in OX2 penalized the TN global removal, equal to 71%. During period 2, by supplying biodegradable carbon through liquid wastes conferred to the platform, the elasticity of the AC process enabled us to obtain global removal of total nitrogen equal to 95% (90% in OX2) and of ammonia of 97% (84% in OX2). The periodic feed of material rich in RBCOD and the necessity to contain operating costs, motivated the activation of the efuent recirculation to OX1 in the fourth period, to improve the HRT and the removal effect of the rst reactor. The global removal efciencies of TN and NH4-N reached percentages equal to 56% (63% in OX1 and 34% in OX2) and 97% respectively. By considering the period prior to the experimentation, the introduction of the denitrication phase and the

Figure 4

ETN% Global, OX1, OX2.

Figure 5

ENH4-N% Global, OX1, OX2.

Figure 6

TN inuent and efuent biological unit.

727

A. L. Eusebi et al.

Upgrading biological processes and operative congurations for the best performance

Water Science & Technology

65.4

2012

Table 6

Energy consumption and saving

Period

2-3-4

OX1 OX2 Total consumption with AC Total consumption PreAC Energy saving kWh d1 kWh d1 %

AC TO 901 1,298 31

TO AC 1,047 1,298 19

exibility of the process enhanced the nitrogen removal, as reported in Figure 6. Nevertheless, with the increase of total nitrogen in the inuent, from an average value of 57 mg L1 (before the upgrading) to 78 mg L1 (after the AC process application), the TN concentration in the efuent during the experimental period was on average equal to 25 mg L1, independently of the applied conguration. Energy savings An evident reduction, from 10 to 25% (Balku ), of the energy consumption is determined by applying the AC process. This condition, in comparison with a traditional DN process, is related with the alternation of oxic and anoxic phases, with the use of the combined oxygen and with the absence of the internal recycle. The energy saving, referred to the biological sections (OX1 and OX2) and to the installed electro-mechanics (compressors and mixers), was evaluated for OX1 in period 1 and for OX2 in the other periods. Energy savings equal to 31 and 19% were respectively obtained (Table 6).

inhibitory effect of inuent or the low percentages of biodegradable carbon. The different congurations enhanced the reduction of nitrates and nitrites, maintaining the ammonia removal over 90% and substantially higher ammonia uptake rate, despite the heavy metals in the solution. The supply of uids rich in biodegradable carbon to support the denitrication during period 2, permitted us to obtain a total nitrogen removal equal to 95%. The choice to recycle the efuent to OX1 could determine a modest improvement of the nitrogen removal and cover only temporarily the deciency of the carbon source. The bioadsorption phenomenon was evaluated for chrome and copper in the rst reactor. Finally, the AC application enhanced during period 1 and periods 2, 3, 4 respectively the reduction of energy consumption; savings of 31 and 19% were obtained.

REFERENCES
APHA Standards Methods for the Examination of Water and Wastewater, 21st edition. American Public Health Association, Washington, DC. Balku, S. Comparison between alternating aerobic-anoxic and conventional activated sludge systems. Water Research 41, 22202228. Battistoni, P., Morini, C., Pavan, P. & Latini, F. The retrotting of an extended aeration process to optimise biological nitrogen removal in liquid industrial wastes. Environmental Technology 20, 563573. Battistoni, P., Boccadoro, R., Bolzonella, D. & Marinelli, M. An alternate oxic-anoxic process automatically controlled. Theory and practice in a real treatment plant network. Water Science and Technology 48 (1112), 337344. Beccari, M., Marani, D. & Ramadori, R. A critical analysis of nitrication alternatives. Water Research 13 (2), 185192. Kristensen, G. H., Jorgensen, P. E. & Henze, M. Characterization of functional microorganism groups and substrate in activated sludge and wastewater by AUR, NUR and OUR. Water Science and Technology 25, 4357. ygard, J. K., Gjengedal, E. & Ryset, O. . Size charge fractionation of metals in municipal solid waste landll leachate. Water Research 41 (1), 4754. Wang, B., He, S., Wang, L. & Shuo, L. Simultaneous nitrication and de-nitrication in MBR. Water Science and Technology 52 (1011), 435442.

CONCLUSION
The optimization of the performances of the biological unit of the industrial wastewaters treatment platform was investigated through different congurations and AC process application. The strategic choice to perform a biological double stage with separate biomass assured an elevated efciency in terms of macro-pollutant removal, notwithstanding the variability of mass loading, the potential

First received 15 July 2011; accepted in revised form 30 September 2011