-Combined biological and physico-chemical treatment of ~

~filtered pig manure wastewater: pilot investigations ~

cn()CD

s. Kalyuzhnyi*, v. Sklyar', A. Epov'l. Arkhipchenko", I. Barboulina", o. Orlova" and 6A. Klapwijk'" ~". Dept. of Chemical Enzymology, Chemistry Faculty, Moscow State University, 119899 Moscow, Russia ~(E-mail: svk@enz.chem.msu.ru) ~

,,-.. Institute of Agricultural Microbiology, Podbelsky shosse 3, 189620 St-Petersburg-Pushkin 8, Russia Q.

0... Sub-dept. of Environmental Technology, Wageningen University, 6700 EV Wageningen, The ~Netherlands 2:

~U1

Abstract Combined biological and physico-chemical treatment of filtered pig manure wastewater has been z0investigated on the pilot installation operated under ambient temperatures (15-20.C) and included: i) UASB- ~N

reactor for elimination of major part of COD from the filtrate; (ii) stripper of CO2 + tluidised bed crystallisator "C"Cfor phosphate (and partially ammonia) removal from the anaerobic effluents in the form of insoluble minerals - ~

struvite (MgNH4PO 4) and hydroxyapatite (Ca5(PO 4)aOH); (iii) aerobic-anoxic biofilter for polishing the final dI--I

effluent (elimination of remaining BOD and nutrients). Under overall hydraulic retention time (HRT) for the @

system of 7.8 days, the total COD, inorganic nitrogen and total phosphorous removals were 88, 65 and ~74%, respectively. A decrease of the overall HRTto 4.25 days led to 91,37 and 82% removals for total ~COD, inorganic nitrogen and total phosphorus removals, respectively. The approaches for further §:improvement of effluent quality are discussed. ~Keywords Integrated system; nutrient removal; phosphate precipitation; pig manure wastewater; UASB ~

~reactor 0

0~

Introduction

More than 30 million tonnes of pig manure wastewater containing 2-4% total solids isproduced in Russia at big complexes and medium-scale farms due to flushing technologyused for cleaning (Arkhipchenko, 2000). A possible solution for sustainable utilisation andtreatment of diluted manure streams is the preliminary mechanical separation of solidand liquid fractions followed by separate biological and physico-chemical treatment ofboth fractions (Kalyuzhnyi et al., 1999). Various treatment steps involved in this approachwere investigated on the laboratory level (Kalyuzhnyi et al., 1999, 2000) during theexecution of the joint Russian-Dutch project "The development of integrated anaerobic-aerobic treatment of liquid manure streams with maximisation of production of valuableby-products (fertilisers, biogas) and re-utilisation of water" (1999-2001). They served as abasis to design a pilot installation (Figure 1) for treatment of filtered pig manure waste-water. This paper discusses the results obtained during the experimental evaluation of thisinstallation: i) COD elimination from filtered pig manure wastewater using a UASB reac-tor; ii) optimisation of phosphate precipitation from anaerobic effluents; iii) performanceof a biofilter for the removal of remaining BOD and nutrients.

Materials and methodsManure wastewater and pilot installation

The raw manure wastewater (RMW) was taken directly from a pig farm, using a flushingtechnology for cleaning and located on the territory of municipal solid waste treatmentplant in St. Petersburg. The RMW was decanted and filtered through a tissue filter (in full 79

"""~, ~".,.,.u.,,~, r-..o""U"U"" ,YU""VV1V1V/5Y \-".. C""'1"UU1/5) Wll"I" I' Willi IUl:illt:U.

- VASE reactor. The UASB reactor was made from transparent plastics and had the follow-~ ing size: cross-section (rectangular) - 22.6 cm2, height - 206 cm, working volume - 44.61.~ It was seeded with 10 1 of anaerobic sludge originating from an anaerobic digester treatingf sewage sludge (Moscow). During the start-up period (1 month), the reactor was operated in:r~. semi-continuous mode to adapt the sludge to new feeding substrate. Then it was switched; on a continuous regime and a gradual increase of organic loading rate (OLR) was applied,... by decreasing hydraulic retention time (HRT).

Phosphate precipitation block. This block consisted of air stripper (diameter - 20 cm,height - 20 cm, working volume - 6 1) for CO2 removal to increase pH and fluidised bed

crystallisator (FBC, diameter- 7.8 cm, height 105 cm, total volume-51) forcrystal1isationof phosphate minerals such as struvite (MgNH4PO4) and hydroxyapatite (Ca5(PO4)30H).Both reactors were made from transparent plastics. The FBC was initially filled by 1 kg ofwashed sand (0.25-0.5 rom fraction) as a source of nuclei to promote phosphate crystallisa-tion from supersaturated effluents of the stripper. The fluidisation was performed using anairlift loop.

Biofilter. The biofilter was made from transparent plastics and packed by road metal (0.5-2cm fraction). It had the following size: diameter - 13 cm, height - 145 cm, working volume- 19 1) and functioned in alternating aerobic/anoxic regime for treatment of the FBC efflu-

ent. During aerobic phase (duration - 30 or 20 min), the feeding was stopped, while air at a

Table 1 Range of variation of some characteristics of FMW, gll (average values are given in brackets)

CODIot CODss COD,., COD,., pH N.H3 Plot PPQ4

3.7-12.4 0.2-4.9 0.3-3.8 2.6-9.9 5.2-8.7 0.37-1.45 0.08-0.24 0.04-0.14(8.1) (2.1) (1.2) (4.9) (6.8) (0.75) (0.15) (0.09)

~~

, , ., , ., , ., .

i i ! "1 f , :

I I i

I I i :~t

~) Figure 1 Flow sheet of the pilot installation for treatment of liquid fraction of pig manure wastewater

ensure an aaequa!e mlxIllg III !ne OlOllllcr. rrogrammaolC IllUlll-l:ll£lllllCllllllcr "Willi lUl£l1

time cycle of 1 hour) controlled all 3 pumps used. Secondary (nitrifying) sludge fromwastewater treatment plant of pig complex "Vostochnyi" (Leningrad province) was used as -seed sludge for formation of the attached biofilm. The excess of sludge was periodically .cnwithdrawn from the top ofbiofilter. ~

'<c:N,,-

Analyses. COD was analysed spectrophotometric ally using Hach tubes. Raw samples of ~.influents or effluents were used for determination of total COD (CODtoJ, 4.4 ~m folded ;paper-filtered (Schleicher & Schue1l5951/2' Germany) samples for determination offiltrat- ,..

ed COD (CODfilJ and 0.45 ~m membrane-filtered (Schleicher & Schuell ME 25,Germany) samples for determination of soluble COD (CODsoJ. The suspended solids COD(CODss) and colloidal COD (CODcoJ were calculated by the differences between CODtotand CODfiltr' CODfilt and CODsol' respectively. All other analyses were performed 2-3times per week by standard methods (APHA, 1992) or as described previously (Kalyuzhnyiet al., 1999, 2000). Due to technical problems, the measurements of total nitrogen were notmade. All gas measurements are recalculated to standard conditions (1 atm, DOC).Statistical analysis of data was performed using Microsoft Excel.

Results and discussionCOD elimination using UASB reactor

The results of the UASB treatment of the FMW are presented in Figure 3 and they are out-lined in Table 2. It can be seen that during Period I (days 0-32, Figure 2a), the HRT was onthe average 3.5 days resulting in the average OLR of 1.7 g COD/l/d (Table 2). The totalCOD removal was 45% while removals of suspended solids (SS), colloidal and solubleCOD fractions were 69, 59 and 38% (on the average), respectively (Table 2). In spite of bigfluctuation of influent pH, the effluent pH was rather stable - around 7.5 (Figure 2c). The

specific methane production was also a subject of some variations and accounted (on theaverage) for 0.23 nl/l/d. This value is somehow below the theoretically expected one (0.27nl/l/d) taking into account the observed COD removal. The discrepancy can be mainlyattributed to entrapment of some part of the undigested SS by the reactor sludge bed. Asexpected, the ammonia concentrations slightly increased due to anaerobic hydrolysisof proteinaceous substances in the FMW (Figure 2d). On the contrary, the concentrations oftotal phosphorus and phosphate substantially dropped during the anaerobic treatmentof FMW (Figures 2e-f and Table 2). As in laboratory experiments (Kalyuzhnyi et al.,2000), this was attributed to partial precipitation of phosphate minerals (presumably:hydroxyapatite and struvite) inside ofUASB reactor.

After a decrease ofHRT during Period II (days 33-75, Figure 2a) to on the average 2days resulting in an increase of OLR to 5 g COD/l/d (on the average, Table 2), the totalCOD removal step-wise increased to around 70% (Figure 3b, 50-75). This was due toincreased removal of colloidal and soluble COD fractions (on the average - 74 and 63%,

respectively) compared to Period I (Table 2). On the contrary, a slight decrease of SSremoval was detected due to increased wash-out of sludge and other entrapped particulatematter clearly observed throughout Period II. The specific methane production rate (Figure2c, days 33-75) followed a tendency of increasing total COD removal (Figure 2b, days33-75) though some discrepancies with the theoretically expected one were observed. Inspite of acidic influents fed to the reactor, especially in the end of Period II (Figure 2c), theeffluent pH was stabilised around 7 due to volatile fatty acids (VFA) consumption (data 81

HRT, days 3.2-4.3 (3.5) 1.4-3.0 (2.0)OLR,gCOD/l/d 1.3-2.9(1.7) 3.0-7.4(5.0)- InfluentCODlol,g/l 3.7-10.1(6.0) 6.1-12.4(9.3)

~ Effluent CODlo" g/l 2.2-4.4 (2.9) 2.7-6.6 (3.7);0; Total COD removal, % 21-56 (45) 20-77 (60)~

'f Influent CODss' g/l 0.2-1.9 (1.2) 0.1-4.9 (2.2)~ EffluentCODss,g/l 0.1-1.1(0.5) 0.4-2.7(1.1)~. Suspended solids COD removal, % 55-96 (69) 17-94 (56)~ Influent COD I' g/l 0.3-2.3 (0.9) 0.5-2.3 (1.3)~ co=- Effluent CODcol' g/l 0.1-0.6 (0.3) 0.1-1.3 (0.5)

Colloidal COD removal, % 33-96 (59) 59-91 (74)Influent CODsol' g/l 2.6-4.5 (3.4) 3.0-10.0 (5.8)

EffluentCODsol,g/l 1.1-3.1 (2.1) 0.9-3.2(2.1)Soluble COD removal, % 20-57 (38) 38-84 (63)Influent pH 6.7-8.7(7.7) 5.2-6.9(6.1)Effluent pH 7.2-7.9 (7.5) 6.7-7.7 (7.1)CH4 production, nlll reactorld 0.14-0.34 (0.23) 0.34-1.41 (0.8)Influent N-NH3, g/l 0.37-1.45 (0.78) 0.52-1.1 (0.74)Effluent N-NH3, g/l 0.55-1.26 (0.90) 0.56-1.45 (0.84)Influent total phosphorus, g/l 0.08-0.19 (0.15) 0.12-0.24 (0.15)Effluent total phosphorus, g/l 0.07-0.14 (0.11) 0.09-0.13 (0.11)Total phosphorus removal, % 16-47 (29) 8-53 (27)Influent P-PO4' g/l 0.04-0.08 (0.07) 0.06-0.14 (0.10)Effluent P-PO4, g/l 0.03-0.07 (0.05) 0.04-0.08 (0.05)P-PO4removal, % 5-67(31) 32-81(50)

- Influent4' ~ \- MJ 8~ ~12 00 RemOVal EfflU t 10 ~

~3 ':V'V' 6~ Q 9 80 "i~ - HRT 0 0 b 60 >

2 -OLR 4 ~ ':; 6 ~ .. 40 ~~ . ~ ,...":I: 1 2~ ~ 3 . ' ,~ 20 ~

a 00 0 0 0

0 20 40 60 80 0 20 40 60 80Time, days Time, days

-Info pH . - - . Eff. pH - CH4 -Influent - - . - Effluent9 1.5 1.5~~~1~' ~ ~ .'8 ~ 0» 1.2 ,f.. ,.

1"d ..; . 0 ':I: . :1:0.9 ',. ,','.'~7 = ~O,6 "'"0""

6 0.5 u Z 0,3 d

5 0 0.00 0 20 40 60 80

Time, days Time, days

-Influent - - . -Effluent - - - - Effluent -Influent~ 0'3 1~!~i~JRemOVal 60~. ~ 0.15~~~JRemOVal : ~ c: 0.2 40 ~ .0.10 ~3 . 0 2 - 40 000.1 . .,' "".. - 20 ~ ~0.05 ,. '00.' ~".." 20 ~~ e ~, f ~

0.0 0 0.00 00 20 40 60 80 0 20 40 60 80

Time, days Time, days

Figure 2 Performance of UASB reactor treating the FMW: a - HRT and OLR; b - total influent and effluent

COD and total COD removal; c - influent and effluent pH and methane production per litre of reactor volume

per day; d - influent and effluent ammonia; e - total influent and effluent phosphorus and total phosphorus

82 removal; f - influent and effluent phosphate and phosphate removal

il :;nUUIU oe nUleU llliilllle \.-VU iiIlU pnU:;pniile reIIlUViil:; UOliillleu un PUUl U/'I.03D rCiiL;lUl

under sub-mesophilic temperatures (15-20°C) were comparable to the results obtained inthe lab scale trials under mesophilic regime (30°C) (Kalyuzhnyi et al., 2000). However, -such exhaustion of easily degradable COD (e.g. VF A) in the anaerobic effluents might cre- ~ate problems for biological post-treatment (e.g. for biological N and P removal). ~

'<c:N:r

Phosphate removal from anaerobic effluents via precipitation ~.

In spite of substantial precipitation of phosphate minerals inside of UASB reactor, this ;process was incomplete and can be continued by adjusting the pH to the optimal supersatu- =-

rating value, which is above 9 (Kalyuzhnyi et al., 2000). The results of continuous pilot-scale phosphate precipitation promoted by air stripping of CO2 to increase pH andcrystallisation in the FBC are shown in Figure 3 and Table 3. The total HRT in the phos-phate precipitation block was initially set as - 1 day (-0.6 days in the stripper and - 0.4 days

in the FBC). During days 0-32 (Figure 3a), this block demonstrated very good efficiencywith regard to phosphate removal- 84% (on the average, Table 3) ensuring the effluentphosphate concentration below 10 mg/!. Some drop in ammonia concentrations was alsodetected (Figure 3b, days 0-32). Besides suspected struvite formation, some losses ofammonia probably occurred due to its stripping into the gas phase at pH values higher than8 which were usually observed in the precipitation block. In addition, biological nitrifica-tion of ammonia was gradually developed in the FBC as the effluents contained 0.2-0.3 gN-NO3 during days 35-45 (data not shown). Since occurrence of ammonia nitrification,which became almost complete during days 38-45 (Figure 3b) and led to pH drop below 8,had a deteriorating impact on the phosphate removal (Figure 3a, days 38-45), the total HRTfor precipitation block was reduced to - 0.25 days at day 47. This resulted in a gradual

increase of phosphate removal (Figure 3a, days 47-75) to the average value of73% with theaverage effluent phosphate concentration of 15 mg/l for Period 11-2 (Table 3). Also ammo-nia nitrification almost stopped as only negligible concentrations of nitrate and nitritewere observed in the effluent during this period (data not shown). The overall removal ofammonia (presumably due to struvite precipitation and stripping) was accounted for 32%

(Table 3).

Biofilter performance

A successful start-up of biofilter in the nitrifying mode was achieved in 1 month. Then itwas switched on alternating (aerobic-anoxic) operation and the corresponding results arepresented in Figure 4 and Table 4. During Period I (days 0-32), when duration of anoxicand aerobic phases was 30 min each, the average HRT was 3.3 days while the average OLRwas 0.86 g COD/l/d (Table 4). The average total COD removal accounted for 74%, thoughthe removal of individual COD fractions was not uniform- 84, 40, 79% (on the average) for

- Influent - - - Effluent - Influent - - - Effluent0.09 - Removal 100 1.5~RemOVai 10~ 80~ ~ b 80~

~ 0.06 60 -i .; 1 .. 60 ...:; ~ ==' .. ~

~ 40 e Z 05 . 40 e, 0.03 .. ..-,1\ "z . _. -. 20 ~

~ . of. 20 ~ '" ., ,---"'~ 0 ~00.00 0

0 20 40 60 80 I) 20.40 60 80T. da TlIne, days

lIne, ys

Figure 3 Dynamics of phosphate (a) and ammonia (b) concentrations in the phosphate precipitation block 83

HK I , cays 1 U.:lO

Influent P-PO4' gll 0.027-0.071 (0.050) 0.046-0.081 (0.057)Effluent P-PO4, gll 0.001-0.010 (0.007) 0.011-0.027 (0.015)

- P-PO4 removal, % 64-98 (84) 56-83 (73)

r;> Influent N-NHa, gll 0.548-1.260 (0.896) 0.670-1.450 (0.916)~ Effluent N-NHa' gll 0.343-0.845 (0.560) 0.530-0.980 (0.747)'f N-NHa removal, %1 13-53 (37) 23-46 (32)N:T~

~.

; SS, colloidal and soluble matter, respectively (Table 4). In spite of significant variations in=- influent concentrations, the total COD effluent concentrations were fairly stable, slightly

oscillating around 0.72 g COD/l (Figure 4b). The efficiencies of nitrification and denitrifi-cation (Figures 4d and e) were 75 and 65% (on the average) resulting in the average inor-ganic nitrogen removal of 49% (Table 4). A more than double increase of phosphateconcentrations in the biofilter effluents compared to those of phosphate precipitation block(see Tables 3 and 4) was likely due to the dissolution of the small phosphate precipitatesentering into the biofilter with the influents (these small precipitates were not accountedduring soluble phosphate analysis because the samples were centrifuged before analysis).

Due to occurrence of ammonia nitrification in the phosphate precipitation block result-ing in the low influent ammonia concentrations entering to the biofilter (Figure 4d, days33-46), the duration of anoxic phase of biofilter operation was increased to 40 min and thatof aerobic phase was decreased to 20 min keeping the average HRT on the same level of

5 -HRT-OLR 20 -Influent - - -~ . ~ 5 -Removal ,10

~ 4 1.5 § ~ 4 t,~:/:~~f:~:::: 80 ~~. 3 1.0 U Q 3 60 ~~2 ~ 82 40 0

:I: 0.5 ~ b 51 a a ~ 11~:~~~r20 ~0 0.0 0 j -_.~- ~---~ , ___~b_- ~ 0

0 15 30 45 60 75 90 0 15 30 45 60 75 90

Time, days Time, days

- Inf. pH - - - - Eff. pH - Influent - - - Effluent9 -ALR 05 "" -Removal~~i~l' ~ 0'9 i~~k3jl00 0.4:f ~ 80~ :I: 8 . '. '. - 0.3 ~ ~ 0.6 . .', 60 ~

Q. 02 Z , ~ 40 07' bO Z 03 .~ e

c 0.1 ~. bO. .', d 20's1,-, ...6 0 0.0.. 0

0 15 30 45 60 75 90 0 15 30 45 60 75 90

Time, days Time, days

- - - - Eff. N-NO3 - Rem. N-NO3 - P-P04 - Total P- Rem. (NH3+NO3) ~ 0'04

1~:~~~ ~0.3 ~~X=:;=. l00~ ~ ~£~::~\J~~I . .\! . 80 . bO 0.03'" . '. I -= ..r00.2 .'.' e 60 ~ 5 0.02Z., .' 40 e =Z 0.1. '.. . " e 0.01.'. 20~ ~0.0 ~.~ 0 0.00

0 15 30 45 60 75 90 0 15 30 45 60 75 90

Time, days Time, days

Figure 4 Operational performance of biofilter treating effluents from nutrient precipitation block: a - HRT

and OLR; b - total influent and effluent COD and total COD removal; c - influent and effluent pH and ammo-

nia loading rate (ALR); d - influent and effluent ammonia and ammonia removal; e - effluent nitrate, its

84 removal and inorganic nitrogen removal; f - effluent phosphate and total phosphorus

.. .. -~ r:c~ Aerobic phase, min : 30~, 20 ji;{ ~: 30

Anoxic phase, min ' 30 ;,~.. 40 ': 30HRT, days 2.7-4.6 (3.3) 3.0-3.4 (3.3) 1.6-2.6 (2.0) -

OLR, 9 COD/l/d 0.28-1.56 (0.86) 0.69-1.20 (1.04) 1.06-1.61 (1.39) .cn

$ALR, 9 N-NH4/1/d 0.05-0.299 (0.168) 0.006-0.066 (0.028) 0.268-0.427 (0.365) ~Influent CODtot' g/l 2.20-4.42 (2.93) 2.90-3.84 (3.55) 2.60-4.50 (3. 15) ~Effluent CODtot' 9/1 0.55-0.99 (0.72) 0.38-0.50 (0.45) 0.65-0.94 (0.82) ~Total COD removal, % 60-85(74) 61-90(85) 64-79(71) ~.

InfluentCODss,g/l 0.19-1.17(0.52) 0.40-3.17(1.74) 0.10-2.15(0.54) ;EffluentCODss,g/l 0.01-0.06(0.05) 0.00-0.03(0.02) 0.03-0.24(0.12) =-

CODss removal, % 50-95 (84) 98-100 (99) 65-90 (79)Influent COD co!' 9/1 0.10-0.60 (0.31) 0.14-1.31 (0.45) 0.08-1.30 (0.70)Effluent CODcol' 9/1 0.08-0.50 (0.23) 0.01-0.07 (0.04) 0.05-0.26 (0.15)CODcol removal, % 6-73 (40) 75-97 (86) 43-86 (63)Influent CODooi' g/l 1.14-3.13(2.10) 2.14-2.80(2.36) 0.93-3.23(1.89)Effluent CODsoi' g/l 0.08-0.62 (0.43) 0.35-0.43 (0.39) 0.42-0.67 (0.56)CODool removal, % 69-93 (79) 81-88 (83) 44-81 (68)Influent pH 7.6-9.0 (8.5) 7.2-8.7 (7.8) 7.6-8.1 (7.9)Effluent pH 7.0-8.7(8.2) 7.8-8.5(8.2) 7.7-8.1(7.9)Influent N-NH3' g/l 0.343-0.845 (0.560) 0.060-0.210 (0.094) 0.503-0.835 (0.738)Effluent N-NH3' g/l 0.055-0.230 (0.134) 0.004-0.045 (0.014) 0.280-0.563 (0.448)N-NH3 removal, % 67-87 (75) 64-94 (84) 21-49 (39)Effluent N-NO3, g/l 0.025-0.268 (0.140) 0.046-0.325 (0.195) 0.003-0.075 (0.017)'N-NO3 removal, % 25-89 (65) 14-87 (45) 77-99 (94)#Inorganic nitrogen removal, % 22-70 (49) 14-78 (42) 20-49 (37)Effluent total P, 9/1 0.020-0.044 (0.031) 0.022-0.049 (0.031) 0.011-0.034 (0.021)Effluent P-PO4, g/l 0.016-0.020 (0.018) 0.016-0.022 (0.020) 0.010-0.024 (0.018)

$ALR - ammonia loading rate'Calculated as: (1-[N-NO3Je/([N-NO~in + [N-NH3Jin - [N-NH3JeJ}'100#Calculated as: {1-([N-NO3Jef + [N-NH~eJ/([N-NO3Jin + [N-NH~in)}'100

3.3 days during Period 11-1 (days 33-46, Table 4). In spite of increase of the average OLR

till 1.04 g CODIl/d, the total COD removal as well as COD removals of individual COD

fractions also increased (compared to Period I) resulting in the average effluent concentra-

tion of total and soluble COD of 0.45 and 0.39 g/l (Table 4). It is likely that the latter con-

centration represents a hardly biodegradable (neither under anaerobic nor under

aerobic/anoxic conditions) fraction of COD in the pig manure wastewater. Though theeffluent ammonia concentrations were relatively low - around 0.014 g Nil (Figure 4d, days

33-46), the inorganic nitrogen removal was on the average 42% (Table 4) due to insuffi-

cient development of denitrification process resulting in high effluent nitrate concentra-tions - 0.195 g Nil (on the average). It is likely that the concentration of easily

biodegradable COD was insufficient to fulfil the denitrification requirements of the

system.In order to improve the nitrate removal, the HR T was decreased to around 2 days (Figure

4a, Table 4, Period 11-2) while duration of anoxic and aerobic phases was equalised to 30

min each. As a result, the average OLR increased to 1.39 g CODNd but the total CODremoval as well as COD removals of individual COD fractions decreased with respect to

Period 11-1, being comparable with those for Period 1 (Table 4). The better availability of

easily biodegradable COD immediately resulted in an almost complete nitrate removal(Figure 4, days 47-87) and the average effluent nitrate concentration accounted forO.017 g

Nil during this Period. However, this operation regime had a detrimental effect on nitrifica-tion - the average ammonia concentration accounted for 0.448 g Nil. Besides higher 85

c ~ j ll,_I",~," ~ ~~OV~ II,Il,",~, 8 40 ~i 8 40 .~ 20 .~ 20

.so .! 0- ~ FMW UASB PPB BF ~ FMW UASB PPB BF

~~ Figure 5 Relative decrease in total COD, soluble N and total P concentrations after each treatment step: a

~ -Period I; b- Period 11-2 (FMW -filtered manure wastewater, UASB- UASB reactor; PPB-phosphateN"5 precipitation block; BF - biofilter)':$.~.,=- ammonia loading rate applied (less ammonia stripping in the PPB), this was probably due

to decrease of concentration of autotrophic nitrifying bacteria in the biofilter being out-competed by heterotrophic microflora under the elevated OLR imposed on the systemduring Period 11-2.

Figure 5 summarises the average data obtained in the pilot trials of combined biologicaland physico-chemical treatment of FMW (Figure 1) during Periods I and 11-2 (the datafor period II-I are not shown due to occurrence of nitrification in the PBB, which shouldbe avoided by decreasing HRT in this block). It is seen that under the overall HRT forthe system of7.8 days (3.5 days in the UASB + 1 day in the PPB + 3.3 days in the biofilter),the total COD, inorganic nitrogen and total phosphorus removals were 88, 65 and 74%,respectively (Figure 5a). A decrease of the overall HRT to 4.25 days (2 days in both theUASB and biofilter and 0.25 days in the PPB) led to 91, 37 and 82% removals for totalCOD, inorganic nitrogen and total phosphorus removals, respectively (Figure 5b).

Conclusions1. The UASB reactor was quite efficient for removal of bulk COD presented in the FMW

even during operation under sub-mesophilic conditions (15-20°C).2. The PPB was able to decrease the concentration of soluble phosphate in the anaerobic

effluents up to 7-15 mg P/l, but the measure should be taken to prevent an entrance ofsmall phosphate precipitates into the biofilter where they can dissolve giving a rise insoluble phosphate concentrations of the final effluents. The formed in the PPB phos-phate minerals (presumably, struvite and hydroxyapatite) have a perspective to be soldas fertilisers or as raw material for this industry (e.g. the price of magnesium-ammoniaphosphate in Russia is 100-150$/ton). However, stripping of ammonia in the FBBshould be minimised, since ammonia release to the atmosphere causes acid rains. Thelatter can also be prevented by installation of acid tramp (e.g. with concentrated nitricacid) before discharge of stripped air into the environment. The concentrated ammonianitrate formed in the tramp can be used as raw material for fertiliser/chemical industryor directly as a liquid nitrogen fertiliser.

3. The application of aerobic/anoxic biofilter as a sole polishing step was acceptable fromaesthetic point of view (the effluents were transparent and almost colourless and odour-less) and BOD elimination (the resting COD was hardly biodegradable). But the efflu-ent nutrient concentrations (especially nitrogen) were far from the current standards fordirect discharge of treated wastewater. The possible actions to improve an overallnitrogen removal in this system can include further playing with HRT and increase ofduration of aerobic phase to achieve at least a complete nitrification. We are currentlyinvestigating these possibilities.

4. If the nitrogen removal will be further optimised, the possibility of re-use of treated86 wastewater for flushing should be investigated with regard to pathogen limits and

r r ' ---,

Acknowlegements -The financial supportofNWO (grant No 047-07-18) is gratefully acknowledged. ~

A~

-<

References ~:T

APHA (1992). Standard Methods for Water and Wastewater Examination, 17th ed. Amer. Public Health ~.Assoc., Washington, DC. ~

~Arkhipchenko, I.A. (2000). The livestock waste in Russia: current situation. In: Proc. Internat. Con! :-

.. Microbial ecotechnology in processing of organic and agricultural wastes" (22-23 August, 2000, St.

Petersburg, Russia), Archipchenko I.A. and Kalyuzhnyi S. Y. (eds.), pp. 49-53.

Kalyuzhnyi, S., Sklyar, Y., Fedorovich, Y., Kovalev, A., Nozhevnikova, A. and Klapwijk, A. (1999). The

development of biotechnological methods for utilization and treatment of diluted manure streams. Wat.

Sci. Technol., 40(1), 223-229.

Kalyuzhnyi, S., Sklyar, Y., Rodriguez, J., Archipchenko, I., Barboulina, I., Orlova, 0., Epov, A.,Nekrasova, Y., Nozhevnikova, A., Kovalev, A., Derikx, P. and Klapwijk, A. (2000). Integrated mechani-

cal, biological and physico-chemical treatment of liquid manure streams. Wat. Sci. Technol., 41(12),

175-182.

87